# Adaptive cycling training plans: what actually works when life gets in the way URL: https://www.adaptcycling.com/guides/adaptive-cycling-training-plans Updated: 2026-09-03 Author: Jim Camut I built AdaptCycling because the plans I trained off as a pro broke the moment I became a dad. Every training app in 2026 claims to be "adaptive." The word has been diluted to near meaninglessness. An adaptive cycling training plan, operationally, has to do three things: read your real training data, adjust the plan — not just the workout — when life disrupts a week, and explain why each session exists. Most apps stop at the first; only a handful attempt the second; almost nobody does all three. ## What 'adaptive' actually means in 2026 An adaptive plan adjusts in response to what you actually did, what you couldn't do, and what's coming. There are three operational tiers: intensity-adaptive (workouts get harder or easier inside a fixed plan), load-adaptive (the training-load target shifts based on your current fitness), and life-adaptive (the plan itself restructures when life intrudes). The third is what most amateurs actually need. The term got co-opted because nobody enforces a definition. TrainerRoad's Adaptive Training is the cleanest example of intensity-adaptive — Progression Levels move workout difficulty inside a Plan Builder structure, and the product is excellent at what it does. Xert sits closer to load-adaptive: its proprietary Training Load model continuously fits curves to every effort and recommends the next session against a fitness signature. Both are real engineering. Neither, on its own, restructures your week when a sinus infection knocks out Wednesday's VO2 session and you wake up Sunday wondering whether to pretend the missed work didn't happen, double up Saturday, or give up on the plan entirely. Life-adaptive is the hardest because it requires four systems working together: a memory of your goal and its constraints, a live connection to what you actually rode, a plan generator that respects periodization invariants the model cannot simply violate, and a solver that can reshape weeks without abandoning the macro arc. The classical periodization frameworks — Friel's Base/Build/Peak/Race/Transition macrocycle [Friel 2018], Issurin's block-periodization model [Issurin 2010], and Coggan and Allen's training-load framework [Allen et al. 2019] — all assume an athlete capable of executing the prescribed week. None of them prescribes what the plan should do when the athlete cannot. That's the gap life-adaptive software has to fill. ## The three failure modes of static plans Static plans fail in three predictable ways: the missed workout (the plan assumes you rode), the unplanned effort (an unscheduled group ride doesn't fit the script), and the extended disruption (illness, travel, work crisis). Each is normal life. Good plans handle all three without requiring you to manually redesign your week. Failure mode one — the missed workout. The plan says 4×8 sweet spot at 92% of FTP; you got home at 9pm, fed kids, and the trainer never spun. A static plan does not care. The next page in the binder still says Saturday's 3-hour endurance ride and Sunday's hill repeats. You're left to triage the rest of the week with no framework: skip, double up tomorrow, redo the whole structure? Foster's training-monotony research [Foster 1998] showed across 25 athletes that the most reliable behavioral predictor of illness and overtraining was high load combined with low daily-load variance — exactly what "making up" missed sessions tends to produce. The right answer is almost never to compress the missed work into the next available slot. Failure mode two — the unplanned effort. You planned a 60-minute zone 2 recovery ride. A friend called; you went on a 90-minute group ride averaging 230 normalized power instead. A static plan pretends it didn't happen and tells you to do tomorrow's threshold session anyway. A connected plan reads the ride, recognizes the intensity stimulus is already booked for the week, and rebuilds tomorrow as recovery rather than work. Seiler's foundational research on training-intensity distribution [Seiler 2010] and follow-up descriptive studies on well-trained athletes [Stöggl & Sperlich 2015] both show that getting the weekly intensity ratio right matters more than any individual session — but only if your plan can read what you actually did. Failure mode three — extended disruption. Sick for a week. Travel with no bike. Work crunch. A static plan tells you to "pick up where you left off" — which is the worst advice in amateur training. Mujika and Padilla's foundational detraining work [Mujika & Padilla 2000] documents that VO2max and maximal aerobic power begin meaningful decline within 10–14 days of insufficient training stimulus, with metabolic and neuromuscular changes inside two weeks. The return ramp should start at roughly 60–70% of pre-break load and progress gradually under a controlled progression rate. The Meeusen ECSS/ACSM consensus statement on overtraining [Meeusen 2013] is explicit that resumed training without a graded reintroduction is a leading cause of non-functional overreaching. A life-adaptive plan handles this automatically; a static plan asks you to be your own coach at the worst possible time. ## Why your training plan shouldn't make you feel guilty Guilt is a performance-killing emotion. Athletes who feel guilty about missed workouts tend either to overtrain — trying to "make up" for missed sessions — or to disengage from the plan entirely. The research on motivation in exercise is unambiguous: autonomy-supportive framing produces durable adherence; controlling, compliance-shaming framing produces dropout. The Self-Determination Theory literature on physical activity is one of the most replicated bodies of work in behavioral science. Teixeira et al.'s 2012 systematic review of SDT in exercise [Teixeira et al. 2012] synthesized 66 empirical studies and found that intrinsic and identified (autonomous) motivation reliably predict longer-term exercise adherence, while introjected motivation — the kind driven by guilt and self-imposed pressure — predicts short-term effort but long-term burnout. Ng and colleagues' meta-analysis across 184 health-behavior data sets [Ng et al. 2012] reached the same conclusion across the broader health domain. Athletes who train because they feel they should are statistically the athletes most likely to quit. Look at the language in most training apps. "Below target." "Compliance score." Red indicators for missed sessions. Streak counters that reset. The design treats the athlete as the problem and the plan as the source of authority. But a missed workout is almost always information about the rider's life — a sick kid, a delayed flight, a deadline — not a character flaw. Apps that frame missed sessions as failures generate exactly the controlling, externally regulated motivation pattern the SDT literature [Teixeira et al. 2012, Ng et al. 2012] identifies as the strongest behavioral predictor of dropout. AdaptCycling's restructure flow is built around the opposite assumption. The athlete is doing their best, life is complicated, and the coach's job is to meet them where they are. One tap for "Not today," one for "I'm sick," one for "Travel — no bike for four days," and the plan rebuilds — no judgment, no compliance score, no shame. The point is not to be soft. The point is that an adherence-destroying app is, by definition, not adaptive — because the rider it most needs to adapt for is the one who has stopped opening it. ## The architecture of a plan that survives real life A life-adaptive plan needs four components: a data layer reading actual rides, a memory system holding goals and constraints across months, a plan generator that respects periodization invariants, and a restructure engine that can reshape weeks without breaking the macro arc. Each maps to a specific technical decision. Data layer. A ride account connects on sign-in — Wahoo, Hammerhead, Polar or intervals.icu, with Strava seats capped and Strava riders joining a short waitlist — the coach reads the ride history already in it, and every new ride lands on upload. FTP is estimated continuously from the power curve rather than enforced through a forced ramp test. TSS, IF, normalized power, CTL, ATL, and TSB are computed from every activity using the standard frameworks [Allen et al. 2019]. Whatever the prescribed workout, the plan knows what the athlete actually did within seconds of the ride landing. Without this layer, no downstream component can adapt to anything real. Memory layer. Typed durable facts — permanent (geography, equipment, injuries), episodic (last week's bonk, this month's race travel), goal (the August 16 crit), preference (early-morning rides only) — with explicit lifecycle. Episodic facts decay on a half-life so a bad week two months ago doesn't dominate the prompt; goals expire on their event date; permanent facts persist. The result is that the coach always has the right amount of athlete history in front of it without drowning in noise. Plan generation. A two-stage LLM pipeline: stage one produces a coaching assessment; stage two writes the plan. Both are constrained by deterministic solvers the model cannot violate. Weekly TSS may not increase by more than the safe ramp range that mirrors Coggan and Allen's chronic-load progression [Allen et al. 2019]. Recovery weeks are mandatory every third or fourth week. Tapers drop volume by 41–60% over the final two weeks while preserving intensity, mirroring Bosquet et al.'s 27-study tapering meta-analysis [Bosquet et al. 2007]. Block-periodization principles [Issurin 2010] govern the relationship between consecutive build blocks. The model proposes; the solver enforces. Restructure engine. When the athlete taps "Not today" or "I'm sick" or logs an out-of-script group ride, the engine reads the remaining week, the affected dates, the current fitness profile, and the goal-event distance, then rebuilds. The macro arc — total time-in-zone allocations across the mesocycle, the relative weighting of base and specificity — is preserved. The micro adjustments — which day holds the threshold work, whether tomorrow becomes recovery, whether next week absorbs an extra easy day — are recomputed. This is the part most static plans cannot do because they treat the week as a fixed sequence rather than a flexible budget. ## The periodization invariants any adaptive plan must respect An adaptive plan that ignores the periodization rules is not adaptive — it is just chaotic. Four invariants matter: a sane progression rate, built-in recovery weeks, a real taper before the goal, and an intensity distribution that respects the 80/20 rule. Adaptation that breaks any of these costs fitness, not adds to it. Progression rate. Foster's training-monotony work [Foster 1998] formalized the relationship between weekly load, daily-load variance, and overtraining risk: high load combined with high monotony is the strongest behavioral signal that an athlete is heading for non-functional overreaching. Coggan and Allen's CTL framework [Allen et al. 2019] operationalized the same principle as a chronic-load ramp rate — typical guidance keeps weekly CTL increases inside roughly 4–7 TSS/day per week during a build phase, with deload weeks cutting load substantially before resuming the ramp. An adaptive plan that lets a missed week "compensate" by stuffing the next week with extra intensity violates this invariant directly. The rebuilt week must respect the ramp the original week assumed. Recovery and taper. The Meeusen consensus [Meeusen 2013] is unambiguous: scheduled recovery weeks every third or fourth week are not a nicety, they are the mechanism through which supercompensation actually happens. Bosquet et al.'s 2007 taper meta-analysis [Bosquet et al. 2007] across 27 studies converged on the numbers most coaches already use — a two-week taper reducing volume by 41–60% while preserving intensity produced the largest effect on competition performance. An adaptive plan that lets life disruption push the goal-event-week taper out of compliance with these numbers is not adapting; it is reverting to a static schedule that happens to start later. Intensity distribution. Seiler's polarized model [Seiler 2010] — roughly 80% of training time at low intensity below the first lactate threshold, 20% at high intensity above the second, very little tempo in between — is the most replicated finding in endurance training. Stöggl and Sperlich's descriptive analysis of well-trained athletes [Stöggl & Sperlich 2015] confirmed that elites cluster around polarized or pyramidal distributions; Filipas et al.'s 2024 controlled trial in recreational male cyclists [Filipas 2024] showed pyramidal distribution produced significant lactate-threshold and body-composition gains over 16 weeks. An adaptive plan that compensates for a missed easy ride by inserting more threshold work is moving the rider exactly the wrong direction. The reshuffle must preserve the weekly intensity ratio, not just the weekly TSS total. ## What to look for when choosing an adaptive plan in 2026 Five questions separate genuinely adaptive plans from intensity-only adapters. Does it read your actual ride data directly? Does it correctly handle an unplanned group ride? Can it restructure after a missed week without making you restart? Does every workout explain why? And what's the failure mode when you break the script three weeks in a row? Most apps pass one or two of these. Static-plan apps fail the first question outright. Intensity-adapters like TrainerRoad pass questions one and two with caveats — Adaptive Training adjusts workout difficulty inside the structure but does not reshape the week when work falls apart. JOIN passes three to four for moderate disruption but tends to struggle when schedule variance gets high week-over-week. Xert handles question two well through its Training Load model, but the question of restructuring after a missed week sits closer to a recommendation engine than a planner. AdaptCycling is engineered to pass all five — and the trade-off is real software complexity, not a marketing tagline. The five-question test is one of the few ways to evaluate an adaptive claim without running a 12-week trial. The fifth question — what happens when you break the script — is the real tell. If the answer is "we flag you as overtraining" or "the plan gets more conservative until you comply," the app is punishing the athlete for having a life. The Meeusen consensus [Meeusen 2013] is clear that overtraining is a load problem, not a compliance problem; the SDT exercise literature [Teixeira et al. 2012] is clear that controlling motivation produces dropout. Genuinely adaptive software has to absorb life and reshape around it, not interpret life as the rider's failure to perform. ## Common questions **What's the difference between adaptive training and a dynamic training plan?** They overlap in marketing but differ technically. "Dynamic" usually means the plan UI updates live as new data arrives — your CTL chart reads today, the workout hides if completed. "Adaptive" means the plan content itself changes — workouts swap, weeks restructure, the macro arc adjusts to disruption. AdaptCycling is both; many apps marketed as "dynamic" are actually static plans with a reactive UI. **Does AdaptCycling work if I don't have a goal race?** Yes. If you select "I just want to get faster," the plan runs a rolling 12-week structured improvement cycle anchored to your ride history — base, build, peak, short taper, repeat — with no event date driving the taper. The periodization invariants [Allen et al. 2019, Friel 2018] still apply; the difference is that the peak-and-recover cycle just keeps cycling instead of pointing at a date. **What happens if I keep missing workouts for weeks?** The plan progressively reduces load expectations rather than scolding you. After 10–14 days of substantially reduced training, Mujika and Padilla's detraining work [Mujika & Padilla 2000] shows VO2max and maximal aerobic power begin meaningful decline, so the rebuilt plan starts the return ramp at roughly 60–70% of pre-break load and progresses gradually rather than pretending the gap didn't happen. Marking a disruption explicitly helps the restructure pick the right return profile. **Can I still use it if I prefer unstructured outdoor riding?** Absolutely. Every outdoor ride is read from your connected account, classified, and folded into the week's accounting. If your Saturday group ride covered the prescribed intensity work — say, an hour with normalized power above threshold — the coach recognizes the stimulus is already booked and rebuilds Sunday as recovery. The Filipas 2024 pyramidal-distribution trial in recreational male cyclists [Filipas 2024] is one of several lines of evidence that the weekly intensity ratio matters more than which sessions deliver it. **How do adaptive plans compare to a human coach?** For the top 5% of amateurs chasing specific results, a great human coach still wins on judgment and outside-perspective accountability. For everyone else, the gap has narrowed dramatically in 2026. A well-designed AI coach — Strava-connected, periodization-respecting, life-adaptive — costs roughly 1/13th the price of 1-on-1 coaching and responds in seconds. Most self-coached amateurs are the target audience for AI coaching, not personal coaching. **How adaptive should an adaptive plan actually be?** Adaptation that violates the periodization invariants — Foster's monotony rule [Foster 1998], Bosquet's taper data [Bosquet et al. 2007], the polarized 80/20 distribution [Seiler 2010] — is not adaptive, it is chaotic. The solver should restructure the week to absorb disruption, not lengthen the build phase past the point of return, and not "make up" missed work by stacking intensity. The right answer is constrained adaptation: maximal flexibility inside the rules, rigid enforcement of the rules themselves. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. 3. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 4. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. 5. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. 6. **Mujika & Padilla 2000.** [Detraining: Loss of Training-Induced Physiological and Performance Adaptations. Part I: Short Term Insufficient Training Stimulus](https://pubmed.ncbi.nlm.nih.gov/10966148/). Sports Medicine. 7. **Meeusen 2013.** [Prevention, diagnosis and treatment of the overtraining syndrome: Joint consensus statement of the European College of Sport Science and the American College of Sports Medicine](https://pubmed.ncbi.nlm.nih.gov/23247672/). European Journal of Sport Science. 8. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 9. **Stöggl & Sperlich 2015.** [The training intensity distribution among well-trained and elite endurance athletes](https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2015.00295/full). Frontiers in Physiology. 10. **Filipas 2024.** [Effects of a 16-Week Training Program with a Pyramidal Intensity Distribution on Recreational Male Cyclists](https://pubmed.ncbi.nlm.nih.gov/38251291/). Sports (MDPI). 11. **Teixeira et al. 2012.** [Exercise, physical activity, and self-determination theory: A systematic review](https://pubmed.ncbi.nlm.nih.gov/22726453/). International Journal of Behavioral Nutrition and Physical Activity. 12. **Ng et al. 2012.** [Self-Determination Theory Applied to Health Contexts: A Meta-Analysis](https://pubmed.ncbi.nlm.nih.gov/26168470/). Perspectives on Psychological Science. --- # The self-coached cyclist: a research-backed playbook for training without a coach URL: https://www.adaptcycling.com/guides/self-coached-cyclist Updated: 2026-06-16 Author: Jim Camut I went from racing professionally in Europe to training around two daughters and a full-time job. The gap between those two athletes is exactly what 'self-coached' means in 2026. The vast majority of cyclists who train with structure don't have a coach, and the research on what separates the riders who keep improving from the ones who plateau is unambiguous. This is the practical, evidence-backed guide to coaching yourself well — what to copy from the elites, what to ignore, and what to never get wrong. ## What 'self-coached' actually means in 2026 Self-coached doesn't mean unstructured. It means you make the periodization, intensity, and recovery decisions a coach would otherwise make — using your own data, books, and tools. The category has grown because cycling tech has democratized what used to require a professional. The ceiling is no longer access to information. It's the discipline to apply it. The numbers are blunt. USA Cycling has roughly 60,000 licensed members [USAC 2024]; the broader U.S. amateur cycling population is in the tens of millions [PeopleForBikes 2024]. Most of those riders train without a paid coach because real coaching costs $250–$600 per month for a written-plan tier and $400–$800+ for true 1-on-1 [CTS 2024]. For amateurs whose goals are local races, gran fondos, or just being faster on Saturday, the math doesn't pencil out. What's different in 2026 versus 2010 is infrastructure. Used power meters are sub-$300. Strava reads every ride and computes summary metrics for free. Open tools like Intervals.icu give you CTL, ATL, and TSB charts that used to require WKO+ and a trained eye. AI coaches like AdaptCycling build adaptive plans for the price of a Strava subscription. The technology that historically lived inside a coach's office now lives on your phone. The trap is that access to data is not the same as coaching. A self-coached cyclist with five seasons of Strava data and a $1,200 power meter can still train at the wrong intensities, miss recovery, peak two weeks before the goal event, and not understand why. The skill self-coached riders need to develop is not analytical. It's deciding what to ignore, when to do less, and which metrics to act on out of the dozens available. ## Why most self-coached riders plateau in year two The first year of structured riding produces big gains because almost any consistent training works against an untrained baseline. The plateau hits in year two when 'ride more, ride harder' stops paying off. The cause is almost always one of three things: no real periodization, intensity drift to a junk middle, or chronic under-recovery. The fixes are unsexy and they work. The periodization gap. Friel's framework [Friel 2018] organizes the year into Prep, Base, Build, Peak, Race, and Transition phases — each with a different aerobic-versus-specificity emphasis. Self-coached riders often skip the Base phase because it feels slow and unproductive, jumping straight from off-season to high-intensity work in March. The result: a great FTP in May with no aerobic depth to support it through August. Intensity drift. Seiler's foundational research [Seiler 2010] showed that elite endurance athletes spend roughly 80% of training time at low intensity (below the first lactate threshold) and ~20% at high intensity (above the second), with very little time in the moderate tempo zone in between. Descriptive studies of well-trained athletes confirm this distribution [Stöggl & Sperlich 2015]. Amateurs invert it. They live in tempo and threshold because every ride feels productive — and every ride leaves them too fatigued for the next session to be genuinely hard. Under-recovery. Foster's work introduced training monotony — the variance in your daily training load [Foster 1998]. High monotony plus high load is the single strongest behavioral predictor of overtraining syndrome [Meeusen 2013]. The self-coached rider who runs the same Sunday endurance ride and Wednesday threshold session every week, year-round, with no recovery weeks, is climbing this curve whether or not the numbers say so. The fix is unromantic. Add a real Base phase, even if it's only 4–6 weeks. Make most rides actually easy. Schedule a recovery week every 3rd or 4th week. The riders who break through year-two stagnation in year three are usually doing the same individual workouts — just with a calendar around them. ## The 80/20 rule and why most self-coached riders have it backwards Seiler's 80/20 distribution — 80% of training time at low intensity below the first lactate threshold, 20% at high intensity above the second [Seiler 2010] — is the most replicated finding in endurance training. Self-coached cyclists tend to do the opposite: too much in the middle, not enough in either tail. The directional lesson is what matters most. Low intensity in this context is specific: below LT1, roughly Coggan zones 1–2 [Allen et al. 2019], 56–75% of FTP, conversational pace where you can comfortably breathe through your nose. This is the zone Iñigo San Millán has popularized for its mitochondrial adaptations [San Millán & Brooks 2018]. It is also the zone most amateur riders skip because riding 165 watts for two hours feels like nothing is happening. Mitochondria don't care that it feels easy. Why amateurs drift up. A 90-minute ride at zone 2 power feels weak when your training partners are riding 230 normalized power. So you push to 220, ride at upper tempo, and get the worst of both worlds — not enough mitochondrial stimulus to count as base, too much fatigue to support a genuinely hard interval session 48 hours later. A 16-week study of recreational male cyclists [Filipas 2024] showed pyramidal training intensity distribution (heavy on low intensity, moderate on threshold, small slice at VO2max) produced significant improvements in lactate-threshold power and body composition. Less middle, more bottom. How to do it without a lab. Two markers work for self-coached riders: heart rate (below ~75–78% of max for zone 2) and the talk test (full sentences, not paragraphs). On a 90-minute ride, if you can't hold a normal conversation for the entire ride, you're not in zone 2. The flip side is just as important: when you do go hard, go genuinely hard. Seiler's interval research [Seiler 2010] and follow-up systematic reviews [Rosenblat 2024] support 4×8-minute intervals at ~90% HRmax over the more-popular 4×4-minute model for VO2max gains. Most amateurs pick the wrong intensity for both halves of their week. ## What a coach actually does that you have to replace A good cycling coach does five things: manages training load, periodizes the year, enforces recovery, designs the race taper, and provides outside-perspective accountability. The first four are technical and replaceable with the right framework or tool. The fifth — outside perspective — is the hardest to recreate when you are coaching yourself. Training load management. CTL, ATL, and TSB — chronic load, acute load, and balance — are the standard quantitative tools [Allen et al. 2019]. A coach watches these continuously: rising CTL with TSB landing in the right window for race day, ATL spikes that warrant an unscheduled recovery day, monotony scores that flag overtraining risk before symptoms appear [Foster 1998]. Self-coached riders need a tool that surfaces these without manual work. Intervals.icu does it for free. AdaptCycling does it as part of plan adjustment. Spreadsheet warriors can do it themselves. Periodization. The macrocycle structure (Base → Build → Peak → Race → Transition) [Friel 2018] is what keeps year-three athletes improving while year-two athletes stall. The two specific calls a coach makes that self-coached riders most often miss: the recovery week every 3rd or 4th week (volume drops 30–40%, intensity drops further), and the transition phase after a race-block (1–3 weeks of unstructured riding to absorb the work psychologically and physiologically). Recovery enforcement and taper. The Meeusen consensus statement [Meeusen 2013] is explicit: overtraining syndrome can take 6–12 months to recover from. The path in is rarely a single bad week — it's months of high load, high monotony, and life stress. A coach functions as the 'no' in your training week. Replacing that voice when self-coached requires either rules pre-committed in advance (always take a recovery week after three building weeks) or a tool that flags load climbing past safe ramp rate. The taper itself is well-defined: volume drops 40–60% in the final 1–2 weeks, intensity stays high but volume of intensity drops, the last 3 days are very easy [Friel 2018]. Outside perspective. This is the irreplaceable one. A coach sees patterns you can't see in your own training because you're emotionally invested in your own numbers. The closest substitute is rules — pre-committed decisions about when to skip a session, when to bail on a goal, when to add volume — and tools that surface uncomfortable signals you'd otherwise ignore. An adaptive AI coach that says 'your last three weeks are above ramp rate; we're adding a deload' is doing the part of coaching most self-coached riders cannot reliably do for themselves. ## How to structure a year without a coach A defensible self-coached year has four phases: a 4–12 week Base building aerobic capacity, a 6–8 week Build adding race-specific intensity, a 1–2 week Peak/Taper sharpening for the goal event, and a 1–3 week Transition recovering before the next cycle [Friel 2018]. Within each phase, the rules are simple: 80/20 intensity, three weeks on / one week off, retest FTP every 4–6 weeks. Base. Volume up, intensity low. Most rides at zone 2 power, one tempo or sweet-spot session per week as a maximum. The goal is mitochondrial density, capillary growth, and fat oxidation [San Millán & Brooks 2018] — slow adaptations that take weeks of consistent low-intensity work. Length depends on the goal: 12 weeks of Base if you're targeting a June peak; 4–6 weeks if you're using a time-crunched approach for a specific short event [Carmichael & Rutberg 2017]. Skipping Base is the most common self-coached error and the one that costs you in August. Build. Specificity becomes the rule. If your goal is a flat criterium, your hard sessions should look like a flat criterium — short, sharp, repeatable anaerobic efforts. If it's a hilly gran fondo, threshold and sweet spot dominate. 2–3 hard sessions per week, with total weekly intensity time around 20% of total training time. Continue the 3-week-build, 1-week-recovery cadence. The recovery weeks are not optional — they're where supercompensation actually happens. Peak and taper. Volume drops 40–60% over the final 1–2 weeks; intensity stays high but volume of intensity drops sharply. Detraining research [Mujika & Padilla 2000] shows that VO2max only begins meaningful decline after ~10 days of complete cessation, so a 14-day taper does not threaten fitness — but it does require maintaining neural intensity (short, fast efforts in the final week) so you arrive at the start line feeling sharp rather than flat. Transition. After the goal event, take 1–3 weeks of unstructured riding. Ride for fun, ride less, do other sports. This phase is where year-three athletes recover physically and psychologically. Skipping it is the most reliable way to start the next training cycle already 5% behind. Strava's 2024 data noted that 82% of pro cyclists logged something other than a ride at some point during the year [Strava 2024]. Cross-training is not just off-season weight room work — it's a category of recovery. ## When hiring a human coach is still the right call Self-coaching works for the vast majority of amateur cyclists whose goals are personal — local races, gran fondos, getting faster, staying fit through life changes. Hire a human coach when the math actually justifies the price: you're chasing a specific result, you've genuinely plateaued for two years despite trying, you have a complex injury or medical history, or you don't enjoy the analytical part and won't engage with it. When the cost makes sense. A masters athlete chasing a national-level result, a road racer trying to upgrade categories, or a triathlete prepping for a specific Ironman qualifier all benefit from the marginal gains an experienced human coach can extract. The same goes for riders with a complex medical history — recovering from overtraining syndrome, for example, can take 6–12 months under the Meeusen ECSS/ACSM consensus [Meeusen 2013] and is not a problem you want to triage from a forum thread. At $400–$600 per month [CTS 2024], a year of 1-on-1 coaching costs $5,000–$7,200 — if your goal is worth that, it's worth that. Below that level of stakes, the value gets harder to justify against good self-coaching with the right tools. When AI coaching is enough. For most self-coached riders, an adaptive AI coach connected to your real ride data is closer to a personal coach than to a static training app. AdaptCycling, for example, reads your ride history, generates plans that respect periodization invariants — the Base/Build/Peak macro structure Friel codified [Friel 2018] and the CTL ramp-rate ceilings Coggan and Allen formalized [Allen et al. 2019] — and restructures the week when life disrupts it. It costs roughly 1/13th of 1-on-1 coaching and responds in seconds. It will not catch every subtle thing a great coach catches. It will catch the obvious things — the missed recovery weeks and the intensity-drift patterns Foster's monotony work [Foster 1998] identifies as the dominant drivers of year-two plateaus. The honest answer. The right choice is rarely human OR AI OR self-coached. Most cyclists move between modes as life and goals change. The dad who trains 6 hours a week for fitness is happy with self-coached and a power meter. The same rider chasing a state championship two years later might want a coach for one season. The point is to know what you're actually optimizing for, then pick the cheapest tool that gets you there. The worst outcome is paying for a service whose questions you wouldn't have asked anyway. ## Common questions **How many hours per week should a self-coached cyclist train?** Frequency and total volume drive most of the adaptation [Seiler 2010]. Below 6 hours per week, focus on consistency and quality over structure. Between 6–10 hours, basic periodization (base/build/peak) starts paying off. Above 10 hours, intensity distribution and recovery become the limiting factors. The right answer is whatever you can sustain for 12 months without resentment. **Do I need a power meter to coach myself well?** Not strictly, but it makes everything easier. Heart rate works for zone 2 endurance and RPE works for hard intervals, but neither tracks fatigue and progression as cleanly as power. Used power meters are sub-$300 in 2026, and most smart trainers come with one built in. The bigger gap is learning what to do with the numbers, not whether you have them. **How often should I retest my FTP?** Every 4–6 weeks during Base and Build phases is the standard recommendation [Allen et al. 2019]. Less often than that and your zones drift; more often and you're spending recovery capacity on tests instead of training. AdaptCycling estimates FTP from your power curve continuously, which removes the test entirely — but if you're testing manually, a 20-minute test or a ramp test every 5–6 weeks is plenty. **What's the biggest mistake new self-coached cyclists make?** Skipping recovery weeks. The pattern is consistent: someone discovers structure, gets motivated, builds for 8–12 weeks straight, peaks at week 6, then declines through weeks 7–12 wondering why their numbers are worse. Recovery weeks every 3rd or 4th week — volume down 30–40%, intensity dropped — are the single most leveraged change a self-coached rider can make. **How do I know if I'm overtrained versus just tired?** The Meeusen consensus [Meeusen 2013] distinguishes functional overreaching (a few days of feeling flat after a hard block, full recovery in 2–3 weeks) from non-functional overreaching (weeks-to-months of underperformance) and full overtraining syndrome (months-to-a-year of system-wide burnout). The pattern that matters: if 5–7 days of complete rest doesn't restore your usual numbers, you've crossed from tired into something serious. See a sports doctor. **Is the 80/20 rule oversimplified for amateurs?** Slightly. Research on recreational and well-trained athletes [Stöggl & Sperlich 2015, Filipas 2024] suggests pyramidal distributions (a bit more time at threshold/tempo than strict 80/20 prescribes) work nearly as well for sub-elite riders. The directionally-correct lesson is the same: most amateurs are stuck in the middle and would benefit from moving more time to either tail of the intensity curve. **Can I train without a structured plan and still improve?** Yes, for a while. Untrained and lightly-trained riders gain meaningfully from any consistent riding for 12–18 months. After that, the gap between unstructured and structured training widens fast. Most year-two plateaus are caused by the absence of structure, not the absence of effort. ## References 1. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 2. **Stöggl & Sperlich 2015.** [The training intensity distribution among well-trained and elite endurance athletes](https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2015.00295/full). Frontiers in Physiology. 3. **Filipas 2024.** [Effects of a 16-Week Training Program with a Pyramidal Intensity Distribution on Recreational Male Cyclists](https://pubmed.ncbi.nlm.nih.gov/38251291/). Sports (MDPI). 4. **Meeusen 2013.** [Prevention, diagnosis and treatment of the overtraining syndrome: Joint consensus statement of the European College of Sport Science and the American College of Sports Medicine](https://pubmed.ncbi.nlm.nih.gov/23247672/). European Journal of Sport Science. 5. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 6. **Mujika & Padilla 2000.** [Detraining: Loss of Training-Induced Physiological and Performance Adaptations. Part I: Short Term Insufficient Training Stimulus](https://link.springer.com/article/10.2165/00007256-200030020-00002). Sports Medicine. 7. **San Millán & Brooks 2018.** [Assessment of Metabolic Flexibility by Means of Measuring Blood Lactate, Fat, and Carbohydrate Oxidation Responses to Exercise in Professional Endurance Athletes and Less-Fit Individuals](https://link.springer.com/article/10.1007/s40279-017-0751-x). Sports Medicine. 8. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 9. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. 10. **Carmichael & Rutberg 2017.** [The Time-Crunched Cyclist: Race-Winning Fitness in 6 Hours a Week (3rd ed.)](https://www.velopress.com/books/the-time-crunched-cyclist-3rd-ed/). VeloPress. 11. **Rosenblat 2024.** [The Effect of Polarized Training Intensity Distribution on Maximal Oxygen Uptake and Work Economy Among Endurance Athletes: A Systematic Review](https://pmc.ncbi.nlm.nih.gov/articles/PMC11679080/). Sports Medicine. 12. **Strava 2024.** [Strava Year in Sport Trend Report 2024](https://press.strava.com/articles/strava-releases-annual-year-in-sport-trend). Strava. 13. **PeopleForBikes 2024.** [U.S. Bicycling Participation Report 2024](https://www.peopleforbikes.org/news/bicycling-participation-report-2024). PeopleForBikes. 14. **USAC 2024.** [USA Cycling Participation and Membership Statistics](https://usacycling.org/article/usa-cycling-seeing-unprecedented-numbers-in-participation-and-support-from-pro-and-amateur-cyclists-and-partner-brands). USA Cycling. 15. **CTS 2024.** [CTS Coaching Pricing](https://trainright.com/coaching/pricing/). Carmichael Training Systems. --- # Training with Strava: how to actually use the data layer every cyclist already has URL: https://www.adaptcycling.com/guides/training-with-strava Updated: 2026-04-30 Author: Jim Camut Strava has won the data-layer war for amateur cycling: over 135 million athletes, virtually every smart trainer and head unit syncing automatically, the de facto activity log for the self-coached [Strava 2024]. What Strava is not is a coach. Its Fitness/Freshness/Form chart is a downstream surface of Andrew Coggan's TSS and Eric Banister's 1975 fitness-fatigue model [Hellard et al. 2007, Allen et al. 2019]. Knowing that — and knowing what Strava measures honestly versus where it leaves you guessing — is the difference between a year-three breakthrough and another year of segment hunting. ## What Strava actually measures (and what it doesn't) Strava measures three things well — power, time-in-zone, and GPS — and three things poorly or not at all: rate of perceived exertion, sleep, and the context of your day. Knowing which is which is the difference between using Strava as a training tool and using it as a journal. What Strava records is whatever your head unit uploads. Power (if you have a meter), heart rate (if you wear a strap or watch), GPS, elevation, and time. From those it derives every secondary number on the ride — average watts, normalized power, weighted average heart rate, time-in-zone splits, and the segment efforts. None of this requires Strava Premium; the free tier captures and displays all of it. The post-ride basics — what you did, how hard, where, for how long — are honest and well-implemented. What Strava doesn't measure: perceived exertion (it asks for it on uploads, but most riders never fill it in), sleep (no native ride-record integration with WHOOP, Oura, or watch sleep tracking), the difference between a 220-watt ride after a normal night and a 220-watt ride on three hours of sleep with a sick toddler. Carl Foster's session-RPE work [Haddad et al. 2017] is clear that internal load — what the work felt like — explains training response better than external load alone. Strava captures the external; the internal is on you. What Strava measures unevenly: heart rate. HR data shows on a Strava ride and Strava uses it to compute Relative Effort [Meyer 2018], but the sensor itself drifts upward during prolonged work even when power output is constant. The phenomenon is called cardiovascular drift, well-characterized in physiology research [Coyle & González-Alonso 2001]: at moderate intensity in warm conditions, HR rises 5-10% over the first 60-90 minutes without any change in actual work rate. Strava shows the drift but doesn't correct for it. A self-coached rider who reads HR as ground truth will conclude every long ride got harder when in fact the body adapted normally. ## Fitness, Freshness, Form: the chart that came from Banister Strava Premium's Fitness/Freshness/Form chart is the same thing TrainingPeaks calls the Performance Manager Chart and Intervals.icu calls Fitness/Form: a downstream view of Eric Banister's 1975 fitness-fatigue model [Hellard et al. 2007]. Three lines, four decisions you actually make from them. The hard part is not reading the chart; it's resisting the urge to optimize it. The lines map to specific definitions. Fitness is a 42-day exponentially-weighted average of daily training load — Coggan and Allen call it CTL, Chronic Training Load [Allen et al. 2019]. Freshness is a 7-day EWMA — ATL, Acute Training Load. Form is the difference, fitness minus freshness — TSB, Training Stress Balance. The window lengths are not arbitrary; they came from Banister's curve-fitting against athlete performance data in the 1970s and have held up across four decades of validation work [Hellard et al. 2007]. What the lines actually tell you. Rising Fitness over weeks means you're accumulating adaptation faster than you're shedding it — the desired state in Base and Build phases. Rising Freshness means the body is in repair mode — the desired state heading into a goal event. Form crosses zero on the way down when you're loading hard enough to be tired (negative TSB), and crosses back up when you taper (positive TSB). For a goal event, decades of taper research and Friel's framework [Friel 2018] put the optimal range somewhere between +5 and +25 TSB — fresh enough to perform, not so fresh you've lost edge. Where Strava's version differs from TrainingPeaks. Strava uses Relative Effort as its load input when you don't have a power meter, and TSS-equivalent values when you do. TrainingPeaks insists on TSS from power data. For most amateur riders the difference is academic: both produce an honest fitness curve as long as you ride consistently with the same sensor stack. The dangerous mistake is mixing inputs — a month of HR-only outdoor rides followed by a month of indoor power rides will produce a discontinuity in the chart that doesn't reflect your actual fitness change. Four decisions to make from the chart. One: when Fitness has been climbing 4-6 weeks straight, you owe yourself a recovery week. Two: when Freshness spikes for no good reason — you've been in bed, traveling, or sick — drop the next planned hard session and substitute zone 2. Three: target Form between +10 and +20 in the final week before any goal event. Four: ignore the chart on a daily basis. It's a multi-week tool. Reading it daily is how you talk yourself into bad decisions. ## Relative Effort and TSS: when Strava's HR-based load works and when it doesn't Relative Effort is Strava's HR-zone-weighted load score, developed in 2018 with Marco Altini to replace the older Suffer Score [Meyer 2018]. It correlates well with TSS for steady aerobic and sub-threshold work and badly for short, sharp, anaerobic intervals where heart rate lags behind actual stress. The methodology. Strava takes your HR stream, slices time-in-each-HR-zone using your max HR, weights each zone, and sums to a single number. Different sport types are weighted differently so a 60-minute run and a 60-minute ride at equivalent strain return comparable numbers. The math is closer to Banister's TRIMP than to Coggan's TSS [Hellard et al. 2007, Allen et al. 2019], which means it answers a slightly different question: how much aerobic strain did this session produce, scaled to your individual zones? Where Relative Effort is honest. Steady zone 2 endurance rides, sweet-spot intervals, threshold work over 5+ minutes — anywhere HR has time to settle into the zone the body is actually working in. A 90-minute zone 2 ride at 165 watts and one at 180 watts will produce different Relative Efforts because HR responds to the difference. Foster's session-RPE validation work [Haddad et al. 2017] supports HR-zone-based load metrics as a reasonable proxy for internal load when the work is sustained. Where Relative Effort lies. Short anaerobic intervals — 30-second on, 30-second off, repeated for 8 minutes — produce massive metabolic stress and very little time-in-zone-5 because HR can't keep up. Cardiovascular drift [Coyle & González-Alonso 2001] also pushes Relative Effort upward late in long rides without any actual increase in work rate. The fix self-coached riders need: when the workout is anaerobic or above-threshold, trust power-based TSS over Relative Effort. When it's aerobic and sustained, either is fine. What this means in practice. If you train indoors with power and outdoors with only HR, your Strava load will look like two different athletes. AdaptCycling resolves this by reading both data streams from the same rides — through Wahoo, Hammerhead, Polar or intervals.icu, since Strava seats are capped — and computing an internal training-stress estimate that doesn't whipsaw between sensors. Intervals.icu does similar work for free. TrainerRoad sidesteps the problem by being indoors-only with power. The do-it-yourself answer is to commit to one primary metric and hold it across the season. ## Strava segments without poisoning your training Segments are Strava's most original contribution — a leaderboard for every climb, every road, every loop. They are also the single most reliable way for self-coached riders to ruin a periodized plan. Used right, they are an occasional test. Used wrong, they convert every endurance ride into a tempo ride. The pull. Segments turn every ride into a contest, even one you don't intend to enter. A friend takes a KOM you've held for two years and the temptation to redline a planned recovery ride to win it back is enormous. Strava's social design [Strava 2024] uses the same pattern that makes social media compulsive — variable rewards, public ranking, low friction. The behavior is predictable and the cost to training is real. What segment chasing does to your week. Most amateur cyclists already drift toward the dreaded no-man's-land tempo zone — too hard to count as base, too easy to count as quality work [Seiler 2010]. Segment chasing accelerates that drift. A planned zone 2 endurance ride that includes one all-out segment effort is no longer zone 2. The 80% of training time that's supposed to be easy got nibbled away one PR attempt at a time. How to use segments productively. Pick three or four segments that map cleanly to test efforts you'd do anyway: a 5-minute climb for VO2 testing, a 20-minute climb for FTP estimation, a flat 8-minute stretch for sweet-spot benchmarking. Use them every 4-6 weeks as your structured intensity work, scheduled the way a coach would schedule a test [Allen et al. 2019]. The rest of the year, ignore them. The riders who do this get the data benefit segments offer with none of the intensity drift cost. ## The social trap: kudos, PRs, and the year-two plateau The pattern is consistent across self-coached riders: year one shows steady gains, year two stalls, year three comes back if and only if the rider stops optimizing for Strava and starts optimizing for the season's goal. Most year-two plateaus are behavioral, not physiological. The behavior is segment chasing, kudos hunting, and following-feed comparison. The mechanism. Carl Foster's training monotony work [Foster 1998] showed that high-load weeks with low day-to-day variability — every ride looking the same intensity — drive overtraining risk faster than total volume does. Riders who structure their week around the social feed tend to ride at one intensity: hard enough to look like they're working, never easy enough to count as recovery. Monotony scores climb. The threshold of overtraining gets closer. The compounding effect. Compare your ride to the riders you follow. They're on a 90-minute group ride averaging 230 watts. You went on a 90-minute zone 2 ride averaging 165 watts. Both are correct training; only one looks impressive in the feed. The behavior most self-coached riders fall into is matching the feed, which means inverting the 80/20 rule [Seiler 2010, Stöggl & Sperlich 2015] and producing exactly the year-two plateau the research describes. Practical decoupling. The riders who break through year-two stagnation are the ones who treat Strava as a private training log. Disable the feed in the app settings if you have to. Hide your zone 2 rides from public view if seeing the kudos count of a slow ride bothers you. The point of Strava for a self-coached rider is the data, not the social proof. The latter is fun; the former is what makes you faster. ## How AI coaches actually read Strava data (and why most apps don't) Reading Strava data well is a non-trivial engineering problem. Strava's API allows full read access to a connected athlete's activities; webhooks notify connected services in seconds when a new ride uploads [Strava Developers]. Most cycling apps don't actually use this data — they sync activity files for display and ignore the implications. The few that act on it are the ones worth paying for. What 'connected to Strava' usually means. The bare minimum is OAuth at sign-in plus periodic activity polling. The activity files appear on the dashboard. Power, heart rate, and time-in-zone show up. The plan, however, is built from your inputs at sign-up — typed answers about hours per week and goal events — and never updates from the rides you actually did. JOIN, Athletica, and most static training apps work this way. What reading Strava data actually means. Reading means deriving fitness from your last 6-12 months of rides without making you take a test. It means recognizing that Saturday's spirited group ride was zone 4 work and Tuesday's planned VO2 session is now redundant. It means seeing a 7-day gap and rebuilding the next two weeks at 70% of pre-break volume rather than picking up where the script left off. The infrastructure is webhook-driven [Strava Developers]: every new activity is parsed within seconds, summary metrics are computed, and the plan adjusts. Where AdaptCycling sits in this. We connect at sign-in — Wahoo, Hammerhead, Polar or intervals.icu, with Strava seats capped and Strava riders joining a short waitlist — read your ride history, estimate FTP from your power curve, and pick up every new ride on upload. The plan adapts daily to what you actually rode rather than what was scheduled. TrainerRoad's Adaptive Training does similar work for indoor power workouts. Intervals.icu is a free analytics layer that surfaces the data without coaching on top of it. The honest answer is that the gap between 'syncs Strava' and 'reads Strava' is the gap between an app and a coach. ## Common questions **Is Strava Premium worth it for a self-coached cyclist?** Mostly yes for the Fitness/Freshness/Form chart and segment leaderboards; mostly no for the route-planning and training-plan features, which are decent but not differentiated. The chart alone replaces a separate paid TrainingPeaks subscription for many riders. If you train indoors with TrainerRoad or use Intervals.icu, the value of Premium drops considerably. **Should I trust Relative Effort or TSS more?** Both, in different contexts. TSS is more accurate when you have power data and the ride is mostly steady or anaerobic. Relative Effort is reasonable for sustained aerobic work scored against your individual HR zones [Meyer 2018]. The trap is comparing a Relative Effort number from a HR-only outdoor ride to a TSS number from an indoor power ride — they aren't the same units. **How long does it take Strava to update my Fitness number after a ride?** Once an activity uploads to Strava, the Fitness/Freshness/Form chart updates within minutes for Premium subscribers. The 42-day window means a single ride moves your Fitness slowly — usually 1-3 points per high-load ride. Watch the trend across 2-4 weeks; daily readings are noise. **Can I use Strava without ever paying for Premium?** Yes. Free Strava captures all the raw data — power, HR, GPS, segments. What you lose is the Fitness/Freshness/Form chart, segment leaderboard filtering, and route analytics. Intervals.icu reads your Strava history for free and produces an equivalent fitness chart. Most self-coached riders can run on free Strava plus a free Intervals.icu account indefinitely. **Should I make my rides private?** If the social feed is causing you to ride harder than your plan calls for, yes. Privacy controls let you hide individual rides from your followers without losing the data on your own dashboard. The behavioral change typically lasts about two weeks before riders return to public mode with healthier habits. **How does AdaptCycling use my Strava data differently from other apps?** We read your ride history at sign-in to estimate FTP and seed the plan, then pick up every new ride on upload to adjust the rest of the week. We connect through Wahoo, Hammerhead, Polar or intervals.icu; Strava seats are capped, so Strava riders join a short waitlist. Other apps display your activities; we use them as the input the plan adapts to. The technical difference is small; the behavioral difference is the whole product. **What metrics should I actually look at after every ride?** Time-in-zone (did the workout do what it was supposed to), normalized power if you have it, perceived exertion (Foster's session-RPE [Haddad et al. 2017] is the cheapest valid metric in cycling), and the chart trend over the last 4 weeks. Skip the kudos count and the segment leaderboards on a routine ride. ## References 1. **Strava 2024.** [Strava Year in Sport Trend Report 2024](https://press.strava.com/articles/strava-releases-annual-year-in-sport-trend). Strava. 2. **Hellard et al. 2007.** [Assessing the limitations of the Banister model in monitoring training](https://pmc.ncbi.nlm.nih.gov/articles/PMC1974899/). Journal of Sports Sciences. 3. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 4. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. 5. **Meyer 2018.** [Quantifying Effort through Heart Rate Data](https://medium.com/strava-engineering/quantifying-effort-through-heart-rate-data-e6a0e3dd6a52). Strava Engineering. 6. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 7. **Haddad et al. 2017.** [Session-RPE Method for Training Load Monitoring: Validity, Ecological Usefulness, and Influencing Factors](https://pmc.ncbi.nlm.nih.gov/articles/PMC5673663/). Frontiers in Neuroscience. 8. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. 9. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 10. **Stöggl & Sperlich 2015.** [The training intensity distribution among well-trained and elite endurance athletes](https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2015.00295/full). Frontiers in Physiology. 11. **Strava Developers.** [Webhook Events API](https://developers.strava.com/docs/webhooks/). Strava Developers. --- # FTP without a test: how to estimate threshold power from data you already have URL: https://www.adaptcycling.com/guides/ftp-without-a-test Updated: 2026-04-30 Author: Jim Camut I've done dozens of formal threshold tests in labs and on the road across my career. The dirty secret is that almost none of them produced a number that held up across the season — and the 20-minute test most amateurs run on Zwift was never a definition of FTP, only a protocol [Allen et al. 2019]. In 2026, the data your power meter has already collected over your last 90 days of riding is enough to estimate your threshold within a few percent — without a forced test. Here is how the math actually works, which apps do it well, and when you still owe yourself a real effort. ## What FTP actually is (and what it isn't) Functional Threshold Power was defined by Andrew Coggan as the highest power a rider can sustain in a quasi-steady state for approximately one hour [Allen et al. 2019]. The 20-minute test estimating it as 95% of average power is a protocol, not the definition itself. The distinction matters because the protocol fails for a meaningful share of riders. The original concept dates to the early 2000s. Coggan and Hunter Allen formalized FTP in Training and Racing with a Power Meter [Allen et al. 2019] as the upper boundary between the heavy and severe intensity domains — the wattage you can hold for roughly an hour before fatigue forces you off. Physiologically, it sits near maximal lactate steady state (MLSS), the highest power at which blood lactate stabilizes rather than climbing [Borszcz et al. 2019]. It is not the same number as MLSS, but it is in the same neighborhood for most trained cyclists. The 20-minute test came later, as a workaround. A genuine 60-minute time trial is brutal and rare, so the protocol multiplies 20-minute power by 0.95 to estimate it. Group-level studies show this works well on average — Borszcz and colleagues [Borszcz et al. 2019] reported a bias of 1.4% versus MLSS with a near-perfect correlation across trained cyclists. But Borszcz and colleagues [Borszcz et al. 2018] showed that on an individual basis, the limits of agreement between the 20-minute estimate and a real 60-minute effort spanned 40-60 watts. For a 250-watt rider, that is the difference between a productive Tuesday session and a disaster. What FTP isn't: a fixed physiological constant that only one test reveals. It is an estimated parameter of the power-duration curve, and like any estimate it has confidence intervals. Treating it as a single sacred number — the way most apps' onboarding flows do — bakes false precision into every workout that follows. ## Why ramp tests and 20-minute tests are flawed (and when they are useful) Ramp tests measure maximal aerobic power and approximate FTP at 75% of MAP — a relationship that varies between 72% and 77% across riders. The 20-minute test is highly pacing-dependent and warm-up-dependent. Both produce a number; neither produces ground truth. They are useful as periodic reality checks, dangerous as the only input to a training plan. The ramp test's central assumption is that FTP equals 75% of the peak one-minute power achieved at the top of an incremental ramp [Allen et al. 2019]. The 75% figure is an average — published validation work places the actual ratio between 72% and 77% depending on the rider's anaerobic capacity. A pure sprinter with a high MAP and a modest aerobic ceiling gets a flattering ramp number that overstates their true FTP by 8-15 watts. A diesel-engine rider with a low MAP relative to their threshold gets the opposite. TrainerRoad's ramp test is the most-used FTP protocol in indoor cycling and it is honest about being a starting estimate — the criticism is the apps that consume the number as if it were the truth. The 20-minute test is worse than its reputation. Tramontin and colleagues [Tramontin et al. 2022] showed that warm-up structure alone shifts the resulting FTP estimate by clinically meaningful amounts. Pacing matters — a rider who blows up at minute 14 produces a falsely low number, and a rider who paces conservatively the entire 20 minutes produces a falsely low number for the opposite reason. The test demands not just fitness but pacing experience, which most amateurs do not have. When tests are useful: as a periodic reality check on whatever model you are using. A power-curve-derived FTP estimate that disagrees with a clean 20-minute effort by more than 5% means one of the two is wrong, and the field test is usually the tiebreaker. Once a season — typically at the end of base, before the build phase — running an actual 20-minute or 8-minute test [Allen et al. 2019] is good hygiene. Doing it every six weeks because an app demands it is mostly a recovery tax. ## The Critical Power model: where modeled FTP estimates come from The Critical Power model is the mathematical backbone of every modern FTP-without-a-test approach. Hugues Monod and Jacques Scherrer first formalized it in 1965 [Monod & Scherrer 1965]; David Poole, Andrew Jones and colleagues canonized it for endurance science in 2016 [Poole et al. 2016]. Two parameters describe almost all of your sustainable power output: Critical Power and W'. The model is simple in form. For any duration above roughly two minutes, the highest power you can sustain follows a hyperbolic curve with two parameters: Critical Power (CP), the asymptote, and W' (W-prime), a fixed amount of work in kilojoules you can do above CP before you fail [Poole et al. 2016, Burnley & Jones 2018]. Below CP, physiological responses stabilize and you can ride for hours. Above CP, lactate, phosphocreatine and pH spiral until fatigue ends the effort. The threshold between those two domains is what FTP approximates. What makes CP useful for FTP-without-a-test is that the model can be fit from any three or more maximal efforts of different durations — a 5-minute climb, a 12-minute time trial, a 20-minute KOM attempt — without forcing you to do all of them on the same day. Vanhatalo, Doust and Burnley [Vanhatalo et al. 2007] validated a single 3-minute all-out test that estimates CP within 6 watts of conventionally-measured values. Software with months of your ride history can do better than that, because it has dozens of near-maximal efforts of varying duration to fit against rather than relying on one acute test. Once you have a working FTP, a power zones calculator turns that single number into the seven training zones you actually ride to. The relationship between CP and FTP is close but not identical. Critical Power tends to run a few percent higher than 60-minute power for highly-trained athletes, and the offset is athlete-specific [McGrath et al. 2021]. Practical apps account for this by reporting an FTP that is CP scaled down by a small individual factor, calibrated against the rider's own time-trial efforts when available. The core insight is that the data needed to estimate FTP already exists in any rider's power file — it just has to be parsed correctly. ## How AdaptCycling estimates FTP from your ride history We read the rides already in your connected account at sign-in — a Wahoo, a Hammerhead, a Polar or intervals.icu — build your mean-maximal-power curve from every ride, and fit a Critical Power model to it. The output is an FTP estimate accurate to within roughly 3-5% for most riders — usable on day one with no forced test. The estimate updates as new rides come in. The pipeline is straightforward. On connection, we pull every activity with power data from your ride history. From each ride, we extract the mean-maximal power for durations from 5 seconds to 60 minutes — the same MMP curve Allen and Coggan describe [Allen et al. 2019] and that GoldenCheetah, Intervals.icu and TrainingPeaks have all surfaced for years. With dozens to hundreds of rides feeding the curve, the longer-duration points are statistically reliable rather than dependent on whether you happened to feel good last Tuesday. We then fit the curve to a Critical Power model and derive FTP from CP, applying a small individualizing offset based on your longest sustained efforts — the same correction McGrath and colleagues quantified when they showed CP runs systematically above 60-minute power in highly-trained athletes [McGrath et al. 2021]. The result skips the test entirely. Riders who have been training with power for at least three months get a usable FTP within seconds of connecting their ride history. Riders with thinner histories get a more conservative estimate that updates as they accumulate quality efforts, leaning on the broader Critical Power validation literature [Poole et al. 2016] until the personal curve fills in. Two things matter for accuracy. First, you need at least one near-maximal effort in the longer-duration range — a 12-minute climb, a 20-minute time trial, a hard 30-minute solo segment. Without that, the curve is anchored only by short bursts and over-estimates FTP. Second, the data has to be clean — power-meter spikes, indoor sessions with broken calibration, and ERG-mode workouts at suppressed power all distort the curve. We filter the obvious outliers; the rest is on the rider's own data hygiene. ## How Xert, Intervals.icu, and TrainingPeaks each derive FTP without a test Three competitors have credible no-test FTP approaches. Xert builds a three-parameter Fitness Signature from breakthrough efforts. Intervals.icu fits a Critical Power model to your MMP curve and reports an eFTP. TrainingPeaks detects a recent best 20-minute power and suggests 95% of it. Each method has tradeoffs worth understanding before you choose a primary tool. Xert's model is the most ambitious. Their Fitness Signature combines Threshold Power, High Intensity Energy (their W' analogue) and Peak Power into a three-parameter description of your power-duration curve [Xert MPA]. The system updates from breakthrough efforts — rides where you push your limits — and predicts your real-time Maximum Power Available second by second. The strength is the granularity. The weakness is that breakthrough rides are required for the model to update; passive zone-2 weeks leave the signature drifting. Intervals.icu is the free standard. The platform fits a CP model against your MMP curve, reports an eFTP based on the best fit, and lets you set the model parameters manually if you disagree [Intervals.icu]. The eFTP value updates whenever a new ride extends or lifts the curve. For self-coached riders who already use Intervals.icu for analytics, eFTP is essentially free FTP estimation that runs as a background process. TrainingPeaks' threshold suggestion is more conservative. The system watches for any 20-minute effort that exceeds your previous best, computes 95% of it, and proposes that as a new FTP. It will not estimate from short or long efforts, only from a clean 20-minute peak. That makes it the most predictable but also the least proactive — riders who do not produce 20-minute peak efforts in normal training will see the suggestion go stale for months. The honest comparison: Xert is the most sophisticated; Intervals.icu is the best free option; TrainingPeaks' threshold suggestion is the safest if you distrust modeled estimates and just want a small bump when you actually go faster. AdaptCycling's approach is closest to Intervals.icu — CP-derived, continuously updated, no test required — with the difference being we use the FTP estimate to drive an adaptive plan rather than just displaying it on a dashboard. ## When you actually do need a test (and which test to do) You need an actual FTP test in three situations: at the start of a structured block when you have less than 90 days of power data, after a long layoff that invalidates your MMP curve, and as a sanity check before a goal event when the modeled FTP feels off. Outside those three, modeled FTP is sufficient for nearly all amateur self-coached training. The first situation is the most common. A new rider, or a rider who has just bought their first power meter, has nothing for the model to fit against. Two weeks of zone 2 and one ramp test produces a usable starting FTP within a 5% margin of the truth, and the modeled estimates take over once the rider has accumulated 60-90 days of varied riding. Skipping the early test in favor of pure modeling means accepting wider error bars for the first month. The second is the long-layoff case. After 6+ weeks off — illness, injury, life crisis — the MMP curve no longer reflects current fitness. Mujika and colleagues [Mujika & Padilla 2000] showed that VO2max and threshold-related performance markers begin meaningful decline after roughly 10 days of complete cessation, and the rate accelerates after three weeks. Coming back from a long break, a fresh test re-anchors the model and prevents the first six weeks of the new block from being run at the wrong intensity. The third is the pre-goal sanity check. Two weeks before a target event is a good time to do a clean 20-minute or 8-minute effort [Allen et al. 2019] and compare it to the modeled FTP. If the two agree within 5%, you taper with confidence. If they disagree by more, you have time to investigate before the event rather than discovering on race day that the plan was built against a number 15 watts off the truth. The protocol that holds up best in research is the 3-minute all-out test [Vanhatalo et al. 2007] for direct CP estimation, paired with a separate 20-minute effort for FTP cross-check. It is unpleasant but short. For most amateurs the friction-free alternative is the standard 20-minute test — well-warmed-up, well-paced, ideally outdoors on a known climb. Whatever you do, do it twice a year, not every six weeks. ## Common questions **How much ride history does AdaptCycling need before it can estimate FTP?** We need at least 30 days of riding with power data and one near-maximal effort longer than 8 minutes. With 90 days of varied training, the estimate stabilizes to within 3-5% of a clean test result. With less than 30 days of data, we ask for a quick benchmark effort to anchor the curve. **Why does my modeled FTP differ from my Zwift ramp test number?** Ramp tests assume FTP equals 75% of peak one-minute power, but the actual ratio varies between 72% and 77% across riders [Allen et al. 2019]. Anaerobically gifted riders get inflated ramp numbers; aerobic specialists get deflated ones. A modeled FTP from your full power curve usually corrects this individual bias, which is why the two numbers can differ by 5-10 watts and the modeled one is often closer to your actual sustained power. **Can I trust an FTP estimate if I don't have a power meter?** Less so, but it is workable. Heart-rate-based threshold estimation uses your max heart rate and Karvonen-style zones to bracket FTP-equivalent intensities. The result is roughly a 10% margin of error rather than 3-5%. AdaptCycling supports HR-only riders during onboarding; the moment you add a power meter or ride with one borrowed, accuracy improves significantly. **How often should the modeled FTP update?** Continuously, but slowly. A single hard ride should not move FTP more than 1-2%. The 90-day rolling window of MMP data smooths daily noise — a great workout boosts the curve incrementally; a flat day does not drag it down. Coggan and Allen recommend reviewing FTP every 4-6 weeks during structured blocks [Allen et al. 2019]; modeled approaches deliver this as a background process. **Is critical power the same thing as FTP?** Close but not identical. Critical Power is the asymptote of the power-duration curve and tends to run a few watts above 60-minute power in trained athletes [McGrath et al. 2021]. FTP is operationally defined as approximately 60-minute power. Most apps that report FTP from a CP model apply a small downward offset, calibrated per-rider when there is enough data to do so. **Why do most training apps still demand an FTP test on signup?** Engineering simplicity. A single number from a single test is easier to consume than a power-duration model fit to historical data. TrainerRoad, JOIN and Athletica all default to a forced test because their plan generators expect FTP as a fixed input. Tools that read your ride history first — Xert, Intervals.icu, AdaptCycling — make the test optional because they have the data to estimate without it. **Should I bother with FTP at all if I don't have a power meter?** Yes. FTP is useful even as a heart-rate-anchored or RPE-anchored target. The point of any threshold concept is to organize training intensity into zones below, at, and above sustainable effort. Whether the number is in watts or heart-rate beats matters less than having a stable reference for what 'sweet spot' or 'tempo' or 'easy' actually means in your riding. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Borszcz et al. 2018.** [Functional Threshold Power in Cyclists: Validity of the Concept and Physiological Responses](https://pubmed.ncbi.nlm.nih.gov/29801189/). International Journal of Sports Medicine. 3. **Tramontin et al. 2022.** [Functional Threshold Power Estimated from a 20-minute Time-trial Test is Warm-up-dependent](https://pubmed.ncbi.nlm.nih.gov/34749416/). International Journal of Sports Medicine. 4. **Borszcz et al. 2019.** [Is the Functional Threshold Power Interchangeable With the Maximal Lactate Steady State in Trained Cyclists?](https://pubmed.ncbi.nlm.nih.gov/30676826/). International Journal of Sports Physiology and Performance. 5. **Monod & Scherrer 1965.** [The Work Capacity of a Synergic Muscular Group](https://www.tandfonline.com/doi/abs/10.1080/00140136508930810). Ergonomics. 6. **Poole et al. 2016.** [Critical Power: An Important Fatigue Threshold in Exercise Physiology](https://pmc.ncbi.nlm.nih.gov/articles/PMC5070974/). Medicine & Science in Sports & Exercise. 7. **Burnley & Jones 2018.** [Power-duration relationship: Physiology, fatigue, and the limits of human performance](https://pubmed.ncbi.nlm.nih.gov/27806677/). European Journal of Sport Science. 8. **Vanhatalo et al. 2007.** [Determination of Critical Power Using a 3-min All-out Cycling Test](https://pubmed.ncbi.nlm.nih.gov/17473782/). Medicine & Science in Sports & Exercise. 9. **McGrath et al. 2021.** [Do Critical and Functional Threshold Powers Equate in Highly-Trained Athletes?](https://pubmed.ncbi.nlm.nih.gov/34055164/). International Journal of Exercise Science. 10. **Mujika & Padilla 2000.** [Detraining: Loss of Training-Induced Physiological and Performance Adaptations. Part I: Short Term Insufficient Training Stimulus](https://link.springer.com/article/10.2165/00007256-200030020-00002). Sports Medicine. 11. **Xert MPA.** [Maximal Power Available (MPA) — Methodology](https://www.baronbiosys.com/maximal-power-available/). Xert / Baron Biosystems. 12. **Intervals.icu.** [Power and eFTP — Intervals.icu Documentation](https://forum.intervals.icu/t/power-zones-and-eftp/). Intervals.icu. --- # Training with Wahoo: running a real plan on an ELEMNT and KICKR URL: https://www.adaptcycling.com/guides/training-with-wahoo Updated: 2026-08-01 Author: Jim Camut A Wahoo ELEMNT renders a structured workout better than most head units on the market — a workout graph in the page stack, an alert about ten seconds before every interval transition, LEDs showing whether you are above or below target, and ERG control of a KICKR without a laptop in the room. What it will not do is decide what the workout should be. Wahoo's own documentation names the sources: SYSTM, TrainingPeaks, TrainerRoad, or another third-party provider [Wahoo ELEMNT Support]. The device is the delivery layer. Which provider you pick determines how much of your week actually reaches the handlebars. ## What the device records, and what it leaves to something else An ELEMNT records the inputs a training plan is built from — power, heart rate, cadence, speed, position — and hands them to the Wahoo Cloud, which forwards them to whatever apps you have authorized. What no part of that chain does is conclude that this week's plan was wrong. Recording and deciding are different jobs. The recording half is solid and it is yours by default. Completed rides sync to the Wahoo app over Bluetooth, or straight to the cloud over Wi-Fi when you finish in range of a saved network, and from there upload automatically to any third-party app you have authorized [Wahoo ELEMNT Support]. That means the ride file — not a set of dashboard averages, but the recorded session itself — exists in a place an analysis layer can read. For a self-coached rider, that is the whole substrate: everything a coach would want to know about Tuesday is in the file Tuesday produced. The targets the device understands are broader than most riders assume. A planned workout's intervals can carry power, heart rate, rate of perceived exertion, or speed targets [Wahoo ELEMNT Support]. This matters more than it sounds. A rider with no power meter is not locked out of structured training on an ELEMNT — the interval structure, the alerts, and the on-screen graph all work against heart rate. What changes is the precision of the target, not the availability of the workout. What has to happen somewhere else is the decision. Nothing on the head unit models your accumulated fatigue, notices that a four-hour Saturday group ride made Tuesday's VO2 session redundant, or rebuilds the following fortnight after a week off sick. The device faithfully displays a plan that something upstream committed to. When riders say their plan on the ELEMNT stopped making sense, this is nearly always the layer at fault — the delivery worked exactly as designed. ## How planned workouts actually reach the head unit Wahoo names four routes: SYSTM, TrainingPeaks, TrainerRoad, or another third-party provider. They are not equivalent. How many days of your plan appear on the device depends entirely on which one you use, and the spread runs from a single day to a full week [Wahoo ELEMNT Support]. The sync windows are specific and worth memorizing. A free TrainingPeaks account syncs today's workout and nothing else. A TrainerRoad account syncs today plus the next three days. Every other account — SYSTM and paid TrainingPeaks included — syncs today plus the next six [Wahoo ELEMNT Support]. The same table applies to the current generation: ELEMNT ACE, BOLT 3, and ROAM 3 follow identical rules [Wahoo ELEMNT ACE Support]. So the seven-day week on your bars is a property of your provider, not of your computer. The subscription detail behind the folklore. Riders repeat that TrainingPeaks is the only way to get workouts onto an ELEMNT, and the useful version of that claim is narrower: a TrainingPeaks Basic account syncs today's workout, and after the first thirty days of a trial a Premium subscription is required to sync a week of planned workouts [Wahoo ELEMNT Support]. The monopoly was never on structured workouts reaching the device. It was on how many days of them arrived, and what that cost. The transport explains most of the failures. Wi-Fi pulls multiple days directly from the Wahoo Cloud to the computer; Bluetooth through the phone app moves exactly one workout at a time [Wahoo ELEMNT Support]. When a rider reports that only today's session is on the device despite a full week on the calendar, the usual cause is that the computer itself — not just the phone — never joined the Wi-Fi network. The computer shows a solid Wi-Fi symbol in its top bar when it has. One trap worth naming because it looks like a bug. Power targets arrive keyed to the zones set in the provider's app, and those may differ from the zones stored in your Wahoo account [Wahoo ELEMNT Support]. Update your FTP in one place and not the other, and the device will faithfully display targets computed from a threshold you no longer believe in. The fix is to reconcile the two, not to scale the workout every time. ## 4DP, Full Frontal, and what a test actually costs SYSTM's 4DP profile is the best-known route to personalized targets inside the Wahoo ecosystem, and it is gated behind a one-hour maximal test. Wahoo's coaching blog advises against repeating Full Frontal more often than every 12 to 16 weeks, with the shorter Half Monty available every 6 to 8 [Wahoo Sports Science]; its SYSTM support pages ask for it a little sooner. Full Frontal is a real protocol, not a marketing number. It runs an hour: two 5-second sprints, a 5-minute maximal effort, a 20-minute maximal effort, and a closing 1-minute effort, each separated by 5-minute recovery blocks [Wahoo Sports Science]. Out of it come four numbers rather than one — neuromuscular power, maximal aerobic power, functional threshold power, and anaerobic capacity. A four-dimensional profile genuinely tells you more about which sessions will hurt than a lone FTP figure does, and Wahoo deserves credit for building it. The cost is that the number is only as current as the last time you rode the test, and the recommended interval means it is allowed to go stale for three to four months. There is also a deeper problem no testing cadence fixes: the threshold a 20-minute protocol produces is a good group-level estimate carrying real individual spread. Measured against maximal lactate steady state, FTP tracks the average closely — a bias of 1.4% — but its 95% limits of agreement run to roughly 9%, so any single rider can sit well off it [Borszcz et al. 2019]. Critical power and functional threshold power do not equate either, even in highly-trained athletes [McGrath et al. 2021]. The test gives you a clean, comparable number. It does not give you a truth. The alternative is the data you have already produced. The critical-power model derives sustainable power from the hyperbolic relationship between power output and time to exhaustion [Jones et al. 2010], and a rider who does group rides, climbs, and interval sessions is generating near-maximal efforts across exactly the durations that model needs. Estimating threshold from a rolling power curve costs nothing, updates continuously, and never asks you to spend a Saturday on a maximal hour. What it needs is a system that actually reads your rides rather than waiting for you to schedule a test. ## ERG mode, and why the target can feel impossible ERG holds the prescribed watts regardless of what you do with the pedals. That is the feature and the failure mode. If the target was set from a threshold that is even 5% too high, ERG will not let you off, and the session collapses in the last two reps rather than being adjusted in the first. The mechanism is well characterized. Above critical power you are spending a finite work capacity, W-prime [Jones et al. 2010], which only reconstitutes when you drop back below it [Skiba et al. 2012]. Skiba's balance model tracks that reservoir through intermittent work and predicts the point of exhaustion from the pattern of efforts and recoveries [Skiba et al. 2012]. The practical consequence for a KICKR session is specific: a rep scheme pitched slightly too hard does not fail on rep one, when you would still have the composure to change it. It fails on rep five, after you have already paid for the first four. The device gives you a better response than quitting. A planned workout can be scaled from the workout page before or during the session, in 10% or 1% steps [Wahoo ELEMNT Support]. Intervals can be skipped or replayed, and the workout can be paused while the ride keeps recording. A session finished at 95% is training; a session abandoned at rep five is a data point about your FTP setting and nothing else. Riders who learn the scale control get more completed quality work per month than riders who treat the prescribed number as a verdict. Passive mode is the other escape hatch, and it is underused: the ELEMNT keeps reading the KICKR without driving resistance [Wahoo ELEMNT Support]. That matters beyond salvaging bad targets, because perfectly constant power is not what riding a bike outdoors involves — variable and constant power output at matched averages produce measurably different physiological responses [Kolsung et al. 2020]. ERG is a tool for hitting a prescription precisely, not a superior way to ride. ## One training load, two environments The same rider produces different numbers indoors and outdoors. In a controlled comparison of 40-km efforts at matched perceived exertion, both power output and heart rate were higher outdoors, with cooler skin temperature and a wider core-to-skin gradient [Mieras et al. 2014]. If your indoor and outdoor rides live in different apps, your training record has a seam in it. The study is worth knowing in detail because it reframes a common anxiety. Twelve recreationally trained male cyclists rode 40-km trials in a laboratory and outdoors, instructed to give the same perceived effort in both, and averaged 208 W outdoors against 163 W indoors [Mieras et al. 2014]. Same riders, same effort, higher outdoor power. An indoor number that sits below your outdoor one is an environmental effect before it is evidence you have lost fitness since spring. Manage the fan and the room temperature before you revise the training plan. Heart rate carries its own indoor complication. Cardiovascular drift raises heart rate over prolonged work even when power output holds constant [Coyle & González-Alonso 2001]. On a long indoor session with heart-rate targets, that drift means the last intervals ask you to reach a number your body is already producing at lower work — so you push harder to satisfy a target that has quietly moved. Power targets are immune to this; heart-rate targets need the drift read into them. The reason to keep a single ledger is that intensity distribution, not any individual session, is what decides whether a training year works [Seiler 2010]. Split the record — SYSTM indoors, something else for the road, a third thing for the gravel weekends — and no system holds the whole picture, which means nobody is checking whether the easy rides stayed easy. Whatever tooling you choose, the requirement is that indoor and outdoor rides land in the same account and count toward the same load. ## What has to sit on top of the hardware Everything above is a delivery system, and an accurate one. The gap in the Wahoo ecosystem is not rendering — it is adaptation. Nothing in the chain from calendar to head unit notices that Saturday's unplanned four hours changed what Tuesday should be, because that decision was never any of their jobs. The honest read on each provider. TrainingPeaks has the deepest structured-workout builder in the business and is the default for coached athletes, with a week on the device gated behind Premium [Wahoo ELEMNT Support]. SYSTM pairs the strongest indoor production in cycling with the 4DP profile, and is indoor-first by design — your outdoor ride history does not reshape the plan. TrainerRoad does genuine within-plan adaptation for indoor power workouts, and syncs a four-day window to the device [Wahoo ELEMNT Support]. Each is good at the thing it was built for, and none of them is built to rebuild your next fortnight around the week you actually had. That is where AdaptCycling sits. Connect a Wahoo account and we read your ride history, estimate threshold from your power curve rather than a test, build a plan, and send today plus the next six days back to your Wahoo account as structured workouts. When the plan changes — a missed week, an illness, an unplanned four-hour ride — the revised sessions replace the old ones on the device rather than waiting for you to notice. Targets go out as low-high ranges rather than a single midpoint, because a threshold estimate is an interval and a target that pretends otherwise is false precision. This runs on the free tier. Two limits worth stating plainly. The seven-day window is Wahoo's, not ours: nobody can put week nine on your handlebars today, and any tool claiming otherwise is describing the phone app, not the computer. And the ELEMNT RIVAL watch is a different pipeline entirely — it takes planned workouts from TrainingPeaks only, synced through the companion app, for today plus the next five days [Wahoo RIVAL Support]. On the RIVAL, the TrainingPeaks monopoly is real. ## Common questions **Can I get structured workouts on my ELEMNT without TrainingPeaks?** Yes. Wahoo's documentation lists SYSTM, TrainingPeaks, TrainerRoad, or another third-party provider as valid sources of planned workouts [Wahoo ELEMNT Support]. Providers other than TrainingPeaks free and TrainerRoad sync today's workout plus the next six days, which is the full rolling week the device can display. **Do I need a paid TrainingPeaks account to get a week of workouts on the device?** To get a week from TrainingPeaks, yes. A Basic account syncs today's workout only; after the first thirty days of a trial, Premium is required to sync a week of planned workouts [Wahoo ELEMNT Support]. Other providers are not subject to that split — they sync the full seven-day window regardless. **How many days of my plan show up on an ELEMNT?** It depends entirely on the provider. TrainingPeaks free: today only. TrainerRoad: today plus three days. Everything else, including SYSTM and paid TrainingPeaks: today plus six days [Wahoo ELEMNT Support]. The rule is identical on ELEMNT ACE, BOLT 3, and ROAM 3 [Wahoo ELEMNT ACE Support]. **Why don't my power targets match my FTP?** Because workout power targets are keyed to the zones set in the provider's app, which can differ from the zones stored in your Wahoo account [Wahoo ELEMNT Support]. If you updated your threshold in one place only, the device is faithfully rendering targets from the other. Reconcile both rather than scaling every workout. **Do I need a power meter to use planned workouts on a Wahoo?** No. Interval targets can be power, heart rate, rate of perceived exertion, or speed [Wahoo ELEMNT Support]. Heart-rate targets need one adjustment in your head: cardiovascular drift pushes heart rate up over long sessions at constant work [Coyle & González-Alonso 2001], so late intervals will feel easier to satisfy than early ones. **How often should I re-test with Full Frontal?** Its coaching blog advises against repeating Full Frontal more often than every 12 to 16 weeks, with Half Monty available every 6 to 8 [Wahoo Sports Science]; the SYSTM support pages ask for it before each new plan, which is sooner. The alternative is not testing at all: critical-power modelling estimates sustainable power from efforts you already produce in normal riding [Jones et al. 2010], which updates continuously instead of quarterly. **Does the ELEMNT RIVAL watch work the same way as the computers?** No, and this catches people out. RIVAL takes planned workouts from TrainingPeaks only, syncs them through the companion app rather than over Wi-Fi, and holds today plus the next five days [Wahoo RIVAL Support]. Third-party providers that reach an ELEMNT computer do not reach a RIVAL. ## References 1. **Wahoo ELEMNT Support.** [Planned Workouts [ELEMNT]](https://support.wahoofitness.com/hc/en-us/articles/115001223770-Planned-Workouts-ELEMNT). Wahoo Fitness Support. 2. **Wahoo ELEMNT ACE Support.** [Planned workouts with ELEMNT ACE, BOLT 3 and ROAM 3](https://support.wahoofitness.com/hc/en-us/articles/22710662013458-Planned-workouts-with-ELEMNT-ACE-BOLT-3-and-ROAM-3). Wahoo Fitness Support. 3. **Wahoo RIVAL Support.** [Planned Workouts on ELEMNT RIVAL](https://support.wahoofitness.com/hc/en-us/articles/360021974640-Planned-Workouts-on-ELEMNT-RIVAL). Wahoo Fitness Support. 4. **Wahoo Sports Science.** [Power Tests: What, When and Why Do Them?](https://www.wahoofitness.com/blog/power-testing-what-when-and-why-do-them/). Wahoo Fitness. 5. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. 6. **Skiba et al. 2012.** [Modeling the expenditure and reconstitution of work capacity above critical power](https://pubmed.ncbi.nlm.nih.gov/22382171/). Medicine & Science in Sports & Exercise. 7. **Borszcz et al. 2019.** [Is the functional threshold power interchangeable with the maximal lactate steady state in trained cyclists?](https://pubmed.ncbi.nlm.nih.gov/30676826/). International Journal of Sports Physiology and Performance. 8. **McGrath et al. 2021.** [Do Critical and Functional Threshold Powers Equate in Highly-Trained Athletes?](https://pubmed.ncbi.nlm.nih.gov/34055164/). International Journal of Exercise Science. 9. **Mieras et al. 2014.** [Physiological and psychological responses to outdoor vs. laboratory cycling](https://pubmed.ncbi.nlm.nih.gov/24476776/). Journal of Strength and Conditioning Research. 10. **Kolsung et al. 2020.** [Physiological Response to Cycling With Variable Versus Constant Power Output](https://pmc.ncbi.nlm.nih.gov/articles/PMC7481374/). Frontiers in Physiology. 11. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 12. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. --- # Training with a Hammerhead Karoo: getting a real plan onto the bars URL: https://www.adaptcycling.com/guides/training-with-hammerhead Updated: 2026-08-04 Author: Jim Camut The Karoo 3 is one of the better structured-workout displays on a bike: a workout drawer you can customize per ride profile, a zone-by-zone preview of the session before you start it, on-device controls to scale or skip an interval, and a new lap opened automatically at every interval boundary. What it does not do is decide what the workout should be. Hammerhead ships no training platform of its own, which makes the choice of what feeds the device the single decision that determines whether a plan survives contact with your week. ## What the Karoo does once it has a workout It renders the session properly and gives you real control over it. Workouts can be sorted by newest, TSS, or duration, and before starting one you can see which zones it targets and how long you will spend in each [Hammerhead Karoo Workouts]. The gap is upstream of all of this. The workout drawer is customizable per ride profile, the same way the rest of the Karoo's pages are — and it is current-generation hardware only. Hammerhead's own note is that the drawer is available on the Karoo, directing Karoo 2 owners to a separate Workouts page instead [Hammerhead Karoo Workouts]. One limit applies to both generations and is worth knowing before you choose a plan source: Hammerhead does not support RPE-based workouts at all [Hammerhead Karoo Workouts]. You choose which data fields appear while a workout is active, and you can show the target as a visual field, as a numeric one, or turn it off entirely [Hammerhead Karoo Workouts]. Three dedicated control fields can be added to any profile: workout control for skipping and rewinding intervals, workout scale for raising or lowering intensity, and workout pause. The pause is genuinely independent — you can pause the workout without pausing the ride, while pausing the ride pauses both, including under auto-pause. One detail deserves more attention than it gets, because it changes what your ride file is worth afterwards. The Karoo automatically starts a new lap at the beginning of every workout interval, and pressing the lap button yourself does not disturb the workout at all — it simply records another lap [Hammerhead Karoo Workouts]. The consequence is that a Karoo-recorded interval session arrives with its structure already marked. Any analysis reading that file can report what you actually held on rep four rather than inferring rep boundaries from a power trace, which is guesswork on a ride with variable terrain. You can also add a workout to a ride already in progress. It counts as a new workout and has to be started from the workout page, while the Karoo keeps recording the session as one continuous effort [Hammerhead Karoo Workouts]. That is the right behaviour for a rider who decides at the turnaround that the day is going better than expected. None of this touches the question of whether today's session is the correct one. The device faithfully renders a decision that something upstream already made. When a plan on a Karoo stops making sense after a heavy weekend or a week off, the delivery layer worked exactly as designed — the fault is in the layer that never revisited the decision. ## Where workouts come from, and why the popular answer is out of date Search this question and nearly every result says TrainingPeaks. That was true for years and is no longer the whole picture. Hammerhead's Dashboard now links to ten third-party accounts, accepts direct file uploads, and the device pulls new workouts the next time it reaches the internet [Hammerhead Account Linking, Hammerhead Workout Import]. The old constraint is still quoted in Hammerhead's own documentation, which is a large part of why it persists. The smart-trainer article still carries the line: currently, only workouts synced from Training peaks or uploaded to the Dashboard in a .FIT or .ZWO file will work [Hammerhead Trainer Support]. The Workouts article, updated more recently, states it more broadly — workouts can be imported from your connected third-party accounts or through the Dashboard [Hammerhead Karoo Workouts]. Both are Hammerhead's; they disagree; the second is the current one. A rider who lands on the first reasonably concludes the door is still shut. What the Dashboard actually links to today is a longer list: AXS, Intervals.icu, Komoot, Ride with GPS, Sentiero, Strava, Suunto, TrainerRoad, TrainingPeaks, and Xert [Hammerhead Account Linking]. Not all of those send workouts — several are route or activity destinations — but the set is no longer a single name, and Intervals.icu in particular closed a feature request that had been open since 2021. Treat any page that still says TrainingPeaks-only as evidence of when it was written. The manual route is the one most riders overlook, and it has no subscription attached to it at all. The Dashboard imports .ZWO and .FIT workout files directly — add the workout, and the next time your Karoo connects to the internet it syncs across [Hammerhead Workout Import]. If a platform can export a structured workout as a file, it can reach your Karoo, whether or not it appears on any partner list. Worth naming the cost of that route honestly: a file you upload by hand is a snapshot. It does not update when your week changes, and doing it daily is a chore that quietly stops after a fortnight. The manual path proves the device is not locked down. It does not solve the problem of keeping a plan current. ## The import rules nobody writes down Hammerhead publishes exactly what its importer accepts, and the rules are stricter than most workout exporters assume. A file that breaks any of them is rejected outright rather than partially imported [Hammerhead Workout Import]. This is the source of most mystifying import failures. For .ZWO files, every step must be a warmup, cooldown, free ride, repeating interval, or steady state. Maximum-effort and ramp steps are explicitly unsupported [Hammerhead Workout Import]. Every step must also carry at least one target — power, a power range, cadence, or a cadence range — and where both power and cadence are supplied, power wins. Zone-based power targets are not supported at all, so a workout expressed as ride zone 4 for ten minutes rather than as a percentage will not import. The .FIT rules are different in a way that matters more than it first appears. A FIT workout's steps may target speed, heart rate, cadence, or power — heart rate and power as absolute values or percentages, speed and cadence as absolute values only [Hammerhead Workout Import]. Heart rate is on that list. It is not on the ZWO list. The two formats are therefore not interchangeable: the same session exported as ZWO carries power and cadence only, and exported as FIT can carry heart rate. If your targets are heart-rate based, the format you choose is the whole question. Two smaller rules catch people out. Distance-based workouts are not supported and are converted to time-based ones on import [Hammerhead Workout Import]. And a FIT file that is an activity or a route rather than a workout is rejected — an easy mistake, because all three share an extension and a rider exporting from an analysis platform often has all three available. The practical takeaway for choosing a provider is that structured-workout export is not a single capability. A platform that exports ZWO can reach your Karoo with power targets. A platform that exports FIT workouts can reach it with heart-rate targets. A platform that only publishes to its own calendar can reach it only if Hammerhead lists it. ## Structured training without a power meter Harder on a Karoo than on some rivals, and worth understanding before you buy into a plan. Hammerhead states plainly that RPE-based workouts are not supported [Hammerhead Karoo Workouts], and following a workout requires a trainer or sensors that can report your effort against the target. That rules out one common fallback. On some head units a workout can prescribe an interval by perceived exertion, which lets a rider with neither a power meter nor a heart-rate strap follow the structure of a session. The Karoo does not offer that route, and it needs a smart trainer or sensors reporting effort against the selected workout in order to follow one at all [Hammerhead Karoo Workouts]. The route that does work is heart rate, delivered as a FIT workout. FIT steps accept heart-rate targets as absolute values or percentages [Hammerhead Workout Import], so a rider with a strap and no power meter can run genuinely structured intervals. The structure, the interval boundaries, the automatic laps, and the on-device controls all behave the same way. What changes is the precision of the target, not the availability of the session. Heart-rate targets need one physiological adjustment read into them. Cardiovascular drift raises heart rate over prolonged exercise even when work rate is held constant [Coyle & González-Alonso 2001]. On a long session, that means the later intervals ask you to reach a number your body is already producing at a lower power output, so satisfying the target late in a workout is not the same achievement as satisfying it early. Power targets are immune to this; heart-rate targets are not, and a plan built on them should be read with that in mind. ## ERG, ANT+ FE-C, and why a target set too high fails late The Karoo controls a trainer over ANT+ FE-C — and Hammerhead notes that not all smart trainers use it [Hammerhead Trainer Support]. Once paired, ERG holds the prescribed watts no matter what you do with the pedals, which is simultaneously the feature and the failure mode. The control surface is better stocked than most riders realize. While following a workout you can switch between ERG, where targets adjust up or down in 1% increments, and Read-Only, which records power from the trainer without controlling it — so the Karoo becomes a display and recorder while another platform drives the trainer [Hammerhead Trainer Support]. Outside a workout there are four modes: ERG in 5-watt steps, Resistance as a percentage of the trainer's maximum, Grade in 1% increments, and Read-Only. Spindown calibration runs from the Control Center. The reason a slightly-high target fails late rather than early is physiological, not psychological. Work above critical power draws down a finite capacity, W-prime [Jones et al. 2010], which only reconstitutes below that intensity [Skiba et al. 2012]. ERG removes the self-selection that would otherwise have shed a few watts on rep two, so each rep drains slightly more than the recovery restores. Nothing feels wrong until the reservoir is empty, which arrives several reps in rather than at the start — long after the point where adjusting would have saved the session. That is what the scale control is for. A session completed at 95% of target is training; a session abandoned on rep five is a data point about your threshold setting and nothing else. And the threshold itself is a likelier culprit than most riders assume: measured against maximal lactate steady state, FTP tracks the group average closely but its 95% limits of agreement run to roughly 9% [Borszcz et al. 2019], and critical power and functional threshold power do not equate even in highly-trained athletes [McGrath et al. 2021]. A target you cannot hold may be a wrong number rather than a bad day. It is also worth not treating ERG as the superior way to ride. Variable and constant power output at matched averages produce measurably different physiological responses [Kolsung et al. 2020], and Read-Only mode exists precisely so the Karoo can record a session it is not dictating. ## What has to sit on top of the hardware Everything above is delivery, and the Karoo does it well. The gap in the ecosystem is adaptation: nothing between a training calendar and the handlebars notices that Saturday's unplanned four hours changed what Tuesday should be, because Hammerhead never built a platform whose job that was. The honest read on the providers. TrainingPeaks has the deepest structured-workout builder available and is the default for coached athletes. Intervals.icu offers genuinely strong analysis for free and now syncs directly. Xert and TrainerRoad both market adaptation, and both are worth a look if that is what you are after — each adjusts what it prescribes as your training changes, within the scope it was built for. Each is good at what it was built for. What none of them was built to do is rebuild the next fortnight around the week you actually had, outdoors, on the rides you actually did. Keeping one ledger matters more than which tool you pick. The same rider produces different numbers indoors and outdoors — in matched-perceived-exertion 40-km trials, riders averaged 208 W outdoors against 163 W indoors, with higher heart rates outdoors too [Mieras et al. 2014]. Split your record across an indoor app and an outdoor one and no system holds the whole picture, which means nothing is checking the thing that actually decides whether a training year works: whether the easy rides stayed easy and the hard ones stayed hard [Seiler 2010]. That is the layer AdaptCycling is built for. We read your ride history from the source you already use, estimate threshold from your power curve rather than asking you to schedule a test [Jones et al. 2010], build a plan from what you have actually been doing, and revise it when the week goes sideways. Rides recorded on a Karoo arrive with their interval laps intact, so the analysis reports what you held rep by rep instead of estimating it. Two limits stated plainly, including one of ours. A rolling window of upcoming days is what any provider can put on a head unit — nobody can place week nine on your bars today, and a tool claiming otherwise is describing its own calendar rather than the device. And the format constraint above is real: a workout delivered as ZWO carries power and cadence targets only. Our own Karoo push emits ZWO, so until we ship a FIT workout encoder a rider training to heart rate is better served uploading a FIT workout to the Dashboard than waiting on us. ## Common questions **Can I get structured workouts on a Karoo without TrainingPeaks?** Yes. Hammerhead's Workouts documentation states that workouts can be imported from your connected third-party accounts or through the Dashboard [Hammerhead Karoo Workouts], and the Dashboard links to ten services including Intervals.icu, TrainerRoad, and Xert [Hammerhead Account Linking]. You can also upload a .ZWO or .FIT workout file directly, which requires no subscription anywhere [Hammerhead Workout Import]. **Why was my workout file rejected on import?** Almost always a rule violation, and the file is rejected whole rather than partially. For .ZWO, every step must be a warmup, cooldown, free ride, repeating interval, or steady state — maximum-effort and ramp steps are unsupported — and every step must carry a power or cadence target, with zone-based power targets not accepted [Hammerhead Workout Import]. For .FIT, the file must be a workout rather than an activity or route. **Can I follow heart-rate workouts on a Karoo without a power meter?** Yes, but the file format decides it. FIT workout steps accept heart-rate targets as absolute values or percentages; the ZWO rules list only power and cadence [Hammerhead Workout Import]. So a heart-rate session must arrive as a FIT workout. Note that RPE-based workouts are not supported in any format [Hammerhead Karoo Workouts]. **How do I get a workout onto the device once it is in the Dashboard?** It syncs on its own. Once a workout is imported to the Dashboard, the Karoo picks it up the next time it connects to the internet [Hammerhead Workout Import]. A notification appears in the Control Center whenever you have a workout scheduled for today, and it carries a Today tag in the workouts list [Hammerhead Karoo Workouts]. **Will my smart trainer work with the Karoo?** Only over ANT+ FE-C, and Hammerhead notes explicitly that not all smart trainers use it [Hammerhead Trainer Support]. Once paired you get ERG, Resistance, Grade, and Read-Only modes outside a workout, and ERG or Read-Only while following one, with workout targets adjustable in 1% increments. **Why do I fail the last intervals of a workout but not the first?** Because work above critical power spends a finite capacity, W-prime [Jones et al. 2010], which only refills below that intensity [Skiba et al. 2012]. ERG prevents the small self-correction you would make on rep two, so a target pitched a few percent too high drains more each rep than the recovery restores and nothing feels wrong until it is gone. Scale the workout down rather than abandoning it — and consider that the threshold may be wrong, since FTP carries roughly 9% limits of agreement against maximal lactate steady state [Borszcz et al. 2019]. **Does the Karoo mark my intervals in the ride file?** Yes, automatically — a new lap starts at the beginning of every workout interval, and pressing lap yourself only records an extra lap without affecting the workout [Hammerhead Karoo Workouts]. That means an interval session recorded on a Karoo carries its own structure, so analysis can report each rep rather than inferring where the reps were. ## References 1. **Hammerhead Karoo Workouts.** [Karoo OS - Workouts](https://support.hammerhead.io/hc/en-us/articles/18001019006363-Karoo-OS-Workouts). Hammerhead Support. 2. **Hammerhead Workout Import.** [Dashboard - Importing ZWO & FIT based Workouts](https://support.hammerhead.io/hc/en-us/articles/4416694669851-Dashboard-Importing-ZWO-FIT-based-Workouts). Hammerhead Support. 3. **Hammerhead Trainer Support.** [Karoo OS - Smart Trainer Integration](https://support.hammerhead.io/hc/en-us/articles/25559624673179-Karoo-OS-Smart-Trainer-Integration). Hammerhead Support. 4. **Hammerhead Account Linking.** [Dashboard - Linking Third-Party Accounts](https://support.hammerhead.io/hc/en-us/articles/360006755614-Dashboard-Linking-Third-Party-Accounts). Hammerhead Support. 5. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. 6. **Skiba et al. 2012.** [Modeling the expenditure and reconstitution of work capacity above critical power](https://pubmed.ncbi.nlm.nih.gov/22382171/). Medicine & Science in Sports & Exercise. 7. **Borszcz et al. 2019.** [Is the functional threshold power interchangeable with the maximal lactate steady state in trained cyclists?](https://pubmed.ncbi.nlm.nih.gov/30676826/). International Journal of Sports Physiology and Performance. 8. **McGrath et al. 2021.** [Do Critical and Functional Threshold Powers Equate in Highly-Trained Athletes?](https://pubmed.ncbi.nlm.nih.gov/34055164/). International Journal of Exercise Science. 9. **Kolsung et al. 2020.** [Physiological Response to Cycling With Variable Versus Constant Power Output](https://pmc.ncbi.nlm.nih.gov/articles/PMC7481374/). Frontiers in Physiology. 10. **Mieras et al. 2014.** [Physiological and psychological responses to outdoor vs. laboratory cycling](https://pubmed.ncbi.nlm.nih.gov/24476776/). Journal of Strength and Conditioning Research. 11. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 12. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. --- # Training with a Polar: the measurement is world-class, the cycling plan is missing URL: https://www.adaptcycling.com/guides/training-with-polar Updated: 2026-08-27 Author: Jim Camut Polar has spent two decades building the best physiological measurement stack in endurance sport. Training Load Pro scores three separate kinds of strain and grades the result against your own tolerance. Nightly Recharge reads how your nervous system settled overnight. Recovery Pro runs an orthostatic heart-rate-variability test and answers with a training recommendation. What none of it does is write you a cycling plan. Polar has built exactly one adaptive, sport-specific training program, and it is for runners. ## What Polar actually measures, and how good it is Better than almost anyone, and with published thresholds rather than a proprietary score. Training Load Pro splits a session into Cardio Load, Muscle Load and Perceived Load, then grades your accumulated strain against your own tolerance on a ratio with four named bands [Polar Training Load Pro]. Very few platforms publish the cut points. Cardio Load is TRIMP. Polar states the mechanism plainly: it is calculated after every workout from your heart rate and session duration, and it is affected by your physical settings, resting and maximal heart rate, and gender [Polar Training Load Pro]. A typical 60-minute session lands between 70 and 130. Muscle Load is the mechanical side, average power multiplied by session duration, with a typical hour of cycling landing between 360 and 720. Perceived Load is your own 1-to-10 rating. Three numbers, three different questions, and the platform does not pretend one of them answers all three. The status layer is where it gets genuinely useful. Cardio Load Status is a strain-to-tolerance ratio with published boundaries: below 0.8 is Detraining or Recovering, 0.8 to 1.0 is Maintaining, 1.0 to 1.3 is Productive, and above 1.3 is Overreaching [Polar Training Load Pro]. Compare that to the fitness-and-fatigue charts most platforms show, where the equivalent judgement is left to you and a form number with no documented interpretation. Polar drew the lines and published them. Nightly Recharge adds an overnight read. ANS charge combines heart rate, heart-rate variability as RMSSD, and breathing rate, measured optically from the wrist during roughly the first four hours of sleep, and compared against your own 28-day baseline [Polar Nightly Recharge]. Sleep charge measures how you slept against your normal pattern. The 28-day baseline matters more than the sensor: it means the number is relative to you rather than to a population, which is the correct design for a recovery metric. Recovery Pro is the most demanding and the most direct. It needs an H10 or H9 chest sensor and an Orthostatic test on at least three mornings per week, and it returns a short training advice from a fixed set: Train more, Go for it!, Train light, Train light or rest, Rest or train light, Injury and illness risk, and Take it easy! [Polar Recovery Pro]. It runs on the Grit X Pro, Grit X2, Grit X2 Pro, Street X, Vantage M3, Vantage V2 and Vantage V3. That is a daily direction, produced from a real physiological measurement, and it is the closest Polar comes to telling a cyclist what to do. Read that list again, though. Train light is a direction. It is not a session. Nothing in the measurement stack decides whether today's ride is four by eight minutes at threshold or two hours of endurance, or what next week should look like given the last six. Polar tells you how hard you have been going and how recovered you are. The prescription is left to you. ## The split that leaves most cyclists with a half-complete ledger Cardio Load needs heart rate. Muscle Load needs power. Polar's own definitions make both requirements explicit [Polar Training Load Pro], and most cyclists ride with only one of the two sensors, which means half of Training Load Pro is dark on every ride they do. Work through the two common cases. A rider doing an indoor session on a smart trainer with no chest strap produces power for every second and no heart rate at all. Muscle Load computes; Cardio Load, which is calculated from heart rate and duration, does not. Now the same rider goes outside with a strap and no power meter, which is the far more common Polar configuration. Cardio Load computes; Muscle Load, which is average power multiplied by duration, does not. Neither ride is wrongly scored. Each is half-scored, and the halves are different halves. The consequence lands on the status ratio rather than on any single ride. Cardio Load Status compares accumulated strain against tolerance [Polar Training Load Pro], so a training week whose hardest session was a power-only indoor effort contributes almost nothing to the cardio side of that ratio. The chart then reads Maintaining or Detraining after a week that was neither. The metric is behaving exactly as documented. The input was incomplete. The practical fix is unglamorous and worth stating anyway: wear the strap on the trainer even when the trainer measures power. Cardio Load is the one of the three that feeds the status ratio, and it costs a chest strap you already own. If you ride without power outdoors, accept that Muscle Load will be sparse and do not read its gaps as easy weeks. This matters beyond tidy charts, because the thing that actually decides whether a training year works is whether the easy rides stayed easy and the hard ones stayed hard [Seiler 2010]. You cannot audit an intensity distribution from a ledger that scores half your sessions on one axis and half on another. Any system reading your Polar data, ours included, inherits whatever holes the sensor setup left. ## The plan layer Polar never built for cycling Polar publishes a page enumerating every training plan it offers. There are five: Flow manual planning, the Running Program, TrainingPeaks integration, the Fitness Program, and FitSpark [Polar Training Plans]. Not one of them is a cycling-specific training plan, and Polar has said in its own words why. The Running Program is the proof that Polar knows how to build an adaptive plan. It is free, it runs on Flow web, it builds a personalized plan from your race distance and target date, and it schedules two to five sessions a week across nine to twenty weeks [Polar Training Plans]. That is a real training plan by any definition. Asked in its own FAQ why the programs are only for running, Polar answers: we have had the most requests for a running program, so we started with that, and in the future training programs for other sports like cycling are possible [Polar Running Program FAQ]. Polar is not hiding the gap. It is describing a roadmap. One correction to a claim you will see made carelessly, including by people arguing our side of this. Polar does ship an adaptive program that covers cycling: the Fitness Program, a subscription at 9,99 euro a month, running on four-week periods, which suggests a new level after each period if it detects you have progressed enough or if the level was too demanding [Polar Fitness Program]. You can choose any sport except Multisport. So the honest claim is not that Polar has no adaptive program. It is that Polar has no cycling-specific one. The distinction is substantive rather than pedantic. The Fitness Program is based on training in the right heart rate zone [Polar Fitness Program]. There are no power targets, no cycling interval structures, and no periodization toward a cycling goal. It is general-fitness cardio that will accept a bike as the vehicle. A rider training for a gran fondo in fourteen weeks is not served by a rolling four-week cardio ladder, and Polar does not claim otherwise. FitSpark is the fifth option and the one most often mistaken for a plan. It offers two to four ready-made workouts a day chosen against your recovery, fitness level and training history, and its cardio sessions can be completed with any sport profile on your watch [Polar FitSpark]. So it does work on the bike. What it does not do is look forward. A daily suggestion has no memory of a goal date and no obligation to make this week build on the last one, which is the specific job a training plan exists to do. ## Getting a structured session onto the watch Two routes exist and both are real. Build the session yourself as a phased training target in Flow, or let TrainingPeaks push it. Everything else is blocked at the API layer, because Polar's public AccessLink API has no endpoint that writes a workout [Polar AccessLink API]. Phased training targets are the free route and they are better than their reputation. In Flow you add phases, set each one by duration or distance, choose whether phases advance automatically or manually, and select the intensity of each phase based on heart rate, speed or power [Polar Training Targets]. The watch then guides you through the session. This is a genuine structured-workout system and it costs nothing. The catch is power, and it is narrower than most riders expect. Polar states that power-based phased training targets are currently supported only by Grit X Pro, Grit X2 Pro, Street X, Vantage V2 and Vantage V3 [Polar Training Targets]. Five models. Note who is missing: the Vantage M3 and Grit X2 both run Polar's own Cycling Performance Test to establish your FTP [Polar Cycling Performance Test] and then cannot follow a phase expressed in watts. Heart-rate and speed phases work far more widely, so on those two watches an HR-based build of the same session is the working route. The TrainingPeaks door is the other one, and it deserves credit rather than a grudging mention. After you link the accounts, every compatible TrainingPeaks training plan and all target workouts sync to Polar Flow automatically once a day, across a device list that runs from the current Vantage V3 back to the M400 and V800 [Polar TrainingPeaks Sync]. It is genuinely a push of structured workouts onto a Polar watch, and we cannot do that. It is also not paywalled: Polar's own wording is that if you are using TrainingPeaks with a premium or a basic account, you can now sync your training plans made in TrainingPeaks to Polar Flow diary [Polar TrainingPeaks Sync]. Basic is TrainingPeaks' free tier. The sync carries documented constraints worth knowing before you build a season around it. Flow always uses five training zones while TrainingPeaks does not restrict the count, so zones are remapped on the way across. A training target with more than twenty unique phases cannot sync at all. TrainingPeaks does not separate indoor from outdoor cycling the way Flow does. And once a workout is in your Flow diary you cannot edit it in Flow, only in TrainingPeaks [Polar TrainingPeaks Sync]. Everything else is closed, and the reason is architectural rather than commercial. The public AccessLink API exposes read endpoints for exercises, daily activity and physical information, and nothing that creates a training target. That is why no third-party service other than TrainingPeaks pushes workouts to a Polar watch. It is also, read the other way, a guarantee: a service reading your Polar data through the public API cannot alter your Flow account, because the API has no endpoint that could. ## FTP on a Polar, and the wall the heart-rate rider hits Polar's Cycling Performance Test is a proper FTP protocol, offering 20, 30, 40 or 60-minute options and returning FTP in watts along with maximum heart rate and VO2max [Polar Cycling Performance Test]. It also requires a compatible cycling power sensor, which is the wall, because heart rate is this brand's centre of gravity. The test runs on eight watches: Grit X Pro, Pacer Pro, Vantage V2, Vantage V3, Grit X2 Pro, Vantage M3, Grit X2 and Street X [Polar Cycling Performance Test]. If you own one of those and a power meter, take it. Twenty minutes is enough for a usable estimate, the results sync to Flow and update your power zones, and it is a better anchor than a number you guessed. For everyone else the requirement is absolute. No power sensor, no test, and no power zones on the watch either. That is a larger population than it sounds, because Polar is the H10 brand and a great many of its cyclists have trained to heart rate for years by preference rather than by accident. It is also worth knowing that Polar is Bluetooth-only for power: its compatible-sensor list annotates the PowerTap G3 entry with the note that Bluetooth-only and Dual models are supported while the ANT+-only model is not [Polar Power Sensors]. A rider who already owns an ANT+-only power meter cannot reach a Polar watch with it. Threshold from the power curve is the alternative, and it is not a consolation prize. A model fitted to your mean maximal power across durations estimates the same physiological boundary the test is aiming at [Jones et al. 2010], updates itself as you ride, and never asks you to schedule a maximal effort you then have to recover from. The honest limit is that it needs efforts of varying length in your history to anchor against; a rider whose last eight weeks are all steady endurance gives any estimator very little to work with. Keep the number's precision in proportion either way. Measured against maximal lactate steady state, FTP tracks the group mean closely but its 95% limits of agreement run to roughly 9% [Borszcz et al. 2019]. A test result and a modelled estimate that differ by fifteen watts are not in meaningful disagreement, and neither justifies rebuilding a training plan around the delta. ## What has to sit on top of the hardware Everything above is measurement and delivery, and Polar does both unusually well. The vacancy is the layer that decides: nothing in the ecosystem looks at the last six weeks of your cycling, writes the next eight, and rewrites them when the week falls apart. Polar's own plan menu is the evidence [Polar Training Plans]. The honest read on the alternatives. TrainingPeaks has the deepest structured-workout builder in the sport and genuinely pushes sessions to your watch, which is a real advantage over us. Its calendar does not restructure itself when you miss Tuesday and ride four hours on Saturday instead. The Fitness Program adapts, within a general-fitness cardio scope that was never meant to periodize a cycling goal. FitSpark answers today well and has no opinion about the twelve weeks after it. Each is good at what it was built for. That gap is the layer AdaptCycling occupies. We read your rides from Polar, estimate threshold from your power curve rather than asking you to schedule a test [Jones et al. 2010], build a plan from what you have actually been doing, and rebuild it when the week goes sideways. The Polar data is unusually good raw material for this: Cardio Load and Muscle Load are computed against your own settings, and Recovery Pro's advice is grounded in a real orthostatic measurement rather than an activity-ring guess. Three limits stated plainly, all of them ours. First, we cannot send a workout to your watch, because the public API has no endpoint that writes one [Polar AccessLink API], so the plan lives in our app and the phased-target route above is how you put a session on the wrist. Second, connecting Polar gives us no history. Polar's API returns only exercises uploaded to Flow in the last 30 days, and only those uploaded after your account was registered with our client [Polar AccessLink API] — both conditions at once, which at the moment you connect means nothing at all. Polar is a live feed from today forward, not an archive, so a rider with years of history should connect Strava as well and let Polar handle the new rides. Third, the sensor gaps from the section above are inherited rather than solved. If your rides carry heart rate and no power, our plan anchors to heart rate, and the estimates it makes carry the wider error bars that implies. We would rather say that than imply a power-grade plan from data that has no power in it. ## Common questions **Does Polar have a cycling training plan?** Not a cycling-specific one. Polar's own menu lists five options — Flow manual planning, the Running Program, TrainingPeaks integration, the Fitness Program and FitSpark — and none of them is built for cycling [Polar Training Plans]. Polar's Running Program FAQ answers the question directly: in the future training programs for other sports like cycling are possible [Polar Running Program FAQ]. The Fitness Program does adapt and does accept cycling, but it is heart-rate-zone general fitness with no power targets [Polar Fitness Program]. **Can any app send structured workouts to my Polar watch?** TrainingPeaks can, and it is the only one. Once the accounts are linked, every compatible plan and target workout syncs to Flow automatically once a day, and it works on a basic TrainingPeaks account as well as a premium one [Polar TrainingPeaks Sync]. Nothing else can, because Polar's public AccessLink API exposes read endpoints only and has none that creates a training target [Polar AccessLink API]. **Why does my Polar show no Cardio Load for an indoor power session?** Because Cardio Load is TRIMP, calculated from your heart rate and session duration [Polar Training Load Pro]. No heart rate, no Cardio Load. The same session still produces Muscle Load, which is average power multiplied by duration. The reverse happens outdoors with a strap and no power meter: Cardio Load computes and Muscle Load does not. Wearing the chest strap on the trainer closes the more consequential of the two gaps, because Cardio Load is what feeds the status ratio. **Can I do an FTP test on a Polar without a power meter?** No. Polar states that the Cycling Performance Test requires a compatible cycling power sensor [Polar Cycling Performance Test], and Polar is Bluetooth-only for power, so an ANT+-only meter will not pair either [Polar Power Sensors]. The alternative is estimating threshold from your power curve as it accumulates [Jones et al. 2010], which needs no test. Bear in mind that FTP itself carries roughly 9% limits of agreement against maximal lactate steady state [Borszcz et al. 2019], so precision beyond a few watts is not real precision. **Which Polar watches can follow a power-based workout?** Five: Grit X Pro, Grit X2 Pro, Street X, Vantage V2 and Vantage V3 [Polar Training Targets]. The Vantage M3 and Grit X2 are the sharp cases — both run the Cycling Performance Test to establish your FTP [Polar Cycling Performance Test] and then cannot follow a phase expressed in watts. On those, build the same session with heart-rate phases instead, which are supported far more widely. **How much of my ride history does Polar hand over when I connect an app?** None at the moment you connect. Polar's API returns only exercises uploaded to Flow in the last 30 days, and only those uploaded after your account was registered with the connecting client [Polar AccessLink API]. Both conditions apply together, so nothing you rode before connecting satisfies them. The 30 days is a ceiling on how long the live feed keeps a ride, not a backfill. Connect Strava alongside Polar if you want your archive read. **Is Recovery Pro worth the daily test?** If you already own an H10 or H9, probably. It needs the Orthostatic test on at least three mornings a week and returns a specific training recommendation from a fixed set that runs from Train more through Take it easy! [Polar Recovery Pro]. That is a real measurement rather than a wrist estimate. What it gives you is a direction for the day, not a session — the decision about what to actually ride still has to come from somewhere else. ## References 1. **Polar Training Load Pro.** [Training Load Pro](https://support.polar.com/en/training-load-pro). Polar Support. 2. **Polar Nightly Recharge.** [Nightly Recharge recovery measurement](https://support.polar.com/us-en/nightly-recharge-recovery-measurement). Polar Support. 3. **Polar Recovery Pro.** [Recovery Pro](https://support.polar.com/en/recovery-pro). Polar Support. 4. **Polar Training Plans.** [Which Polar Training Plan Is Right for You?](https://www.polar.com/en/guide/polar-training-plans). Polar. 5. **Polar Running Program FAQ.** [Polar Running Program FAQ](https://support.polar.com/en/support/polar_running_program_faqs). Polar Support. 6. **Polar Fitness Program.** [Polar Fitness Program FAQ](https://support.polar.com/en/fitness-program-faqs). Polar Support. 7. **Polar FitSpark.** [FitSpark daily training guide](https://support.polar.com/us-en/fitspark-daily-training-guide). Polar Support. 8. **Polar Cycling Performance Test.** [Cycling Performance Test](https://support.polar.com/en/cycling-performance-test). Polar Support. 9. **Polar Training Targets.** [How do I create a training target in Polar Flow web service?](https://support.polar.com/en/support/how_do_i_create_a_training_target_in_polar_flow). Polar Support. 10. **Polar TrainingPeaks Sync.** [Using Polar Flow with TrainingPeaks](https://support.polar.com/en/support/how_can_i_automatically_sync_my_data_from_polar_flow_to_trainingpeaks). Polar Support. 11. **Polar Power Sensors.** [Which third-party power sensors are compatible with Grit X/Pacer/Street X/Vantage?](https://support.polar.com/en/which-third-party-sensors-are-compatible-with-polar-vantage). Polar Support. 12. **Polar AccessLink API.** [Polar AccessLink API](https://www.polar.com/accesslink-api/). Polar Developers. 13. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. 14. **Borszcz et al. 2019.** [Is the functional threshold power interchangeable with the maximal lactate steady state in trained cyclists?](https://pubmed.ncbi.nlm.nih.gov/30676826/). International Journal of Sports Physiology and Performance. 15. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. --- # Indoor cycling training: how to build an indoor season that adapts when winter life does URL: https://www.adaptcycling.com/guides/indoor-cycling-training Updated: 2026-09-03 Author: Jim Camut Indoor cycling training is not outdoor training with the scenery removed. The trainer takes away airflow, coasting and distraction, and each of those changes the physiology of the session: the same watts cost more heart rate, feel harder, and drift upward over an hour. Run well, an indoor season is the most controllable training a self-coached rider ever gets — every interval hits its number, nothing is lost to junctions. Run badly, it is three sweet-spot sessions a week in a hot room until February. This guide covers what actually changes indoors, how much is enough, the intensity mix the evidence supports, and what to do when the season breaks. ## What changes when the ride moves indoors Three things change on a trainer, and heat is the biggest. Outdoors, forward motion pushes air across the skin and carries heat away; on a trainer that flow is zero, so core temperature and heart rate climb at a wattage that would be comfortable on the road. Riders given the same perceived-effort instruction produced 27% more power outdoors than in the lab [Mieras et al. 2014]. That study deserves its numbers. Twelve recreationally trained cyclists rode 40 km outdoors and 40 km in a laboratory, told to hold the same perceived effort in both. Outdoors they averaged 208 W; indoors, 163 W. Heart rate was higher outdoors (152 vs 143 bpm) because the work was higher; core temperature was the same in both, but skin temperature was 1.6 °C warmer indoors, which shrinks the gradient the body uses to dump heat. The only environmental difference was wind: 2.5 m/s outdoors, zero in the lab [Mieras et al. 2014]. The trainer did not make the riders weaker. It made the same effort produce fewer watts. Temperature sets the ceiling on top of that. In near-still air, cyclists riding at 70% of VO2max lasted 93.5 minutes at 10.5 °C and 51.6 minutes at 30.5 °C — an inverted-U relationship with the optimum in the cold, not at room temperature [Galloway & Maughan 1997]. A spare room at 21 °C with the door shut and no fan is closer to the hot end of that curve than most riders assume, because the air 20 cm in front of a sweating rider warms and saturates within minutes. The third change is perceptual. Cycle commuters rated the same relative intensity as easier in the field than on an ergometer — by 1.4 to 1.6 RPE units for breathing and 2.5 to 2.7 units for the legs [Olsson et al. 2024]. Traffic, terrain and visual flow occupy attention that a blank wall hands straight back to the sensation of effort. Zone 2 in particular, which feels effortless on a road, feels like work on a trainer, and that is a large part of why indoor seasons drift toward intensity. The lower indoor FTP number most riders see is this whole section compressed into one figure; the indoor-vs-outdoor FTP spoke covers whether to keep two numbers, so it is not re-derived here. ## The fan is equipment, not comfort A fan is the single highest-return purchase in indoor training, and it has to be running from the first minute. Airflow of roughly 33 km/h cut heat storage, body temperature and perceived exertion during two hours of cycling compared with still or lightly moving air, and raising it to 50 km/h added nothing [Saunders et al. 2005]. One strong fan, positioned on the torso, is the whole answer. The mechanism behind the fan is cardiovascular drift: after 10 to 20 minutes of steady work, stroke volume falls and heart rate rises to hold cardiac output, and the fall is driven primarily by the rising heart rate itself rather than by blood diverted to the skin [Coyle & González-Alonso 2001]. Every rider has seen it — the same 180 W that sat at 135 bpm in minute five reads 150 bpm in minute fifty. On the road the airflow that comes free with speed limits the drift. On a trainer nothing limits it but the fan, and drift is not cosmetic: as it progresses, the same heart rate corresponds to a higher fraction of a now-reduced VO2max, so a heart-rate-anchored session silently gets harder. Timing matters more than most riders know. Wingo and colleagues let drift develop over 45 minutes of cycling in the heat and then switched on a 4.5 m/s fan; skin temperature dropped about 1 °C and VO2max still fell 15%, statistically the same as the 17% fall with no fan at all [Wingo et al. 2019]. Once drift has happened, cooling does not undo it. The fan goes on before the warm-up, not when the sweat starts. The same logic applies to the room: crack a window or a door before the session so the air the fan moves is not already at skin temperature. The useful signal from all this is decoupling — the gap that opens between heart rate and power across a steady ride. Some drift is physiology and unavoidable; a lot of drift on a trainer is a cooling problem. A few percent over a well-cooled 90-minute endurance ride is ordinary; double digits usually says the fan is too small, the room is too warm, or the bottle is empty. Treat the number as a diagnostic of the setup before reading it as a statement about fitness. We report it on every ride long enough to measure it, for the same reason. ## How much indoor training is enough Less than fear suggests, provided intensity stays. In Hickson's maintenance studies, riders who trained six days a week for ten weeks held their full VO2max gain for a further 15 weeks on two hard sessions a week [Hickson & Rosenkoetter 1981] — but cutting intensity by a third while keeping frequency and duration lost it [Hickson et al. 1985]. Winter hours are negotiable. The hard session is not. Those two studies define the floor and the trap. Two sessions of high-intensity work per week held VO2max at trained levels for 15 weeks, essentially the length of an indoor season. In the intensity study the same 40-minute, six-day schedule at two-thirds of the trained work rate lost VO2max and cut long endurance by 21%, and at one-third of the work rate by 30% [Hickson et al. 1985]. The rider who keeps riding every day at a conversational pace because winter is for base is running the second experiment. The rider who rides twice a week and makes one of them hard is running the first. Translated into a week, three tiers cover most self-coached riders. Four hours: two 60-minute sessions with one carrying the week's intensity, plus one longer endurance ride of 90 to 120 minutes; this holds fitness. Six hours: the same skeleton with the endurance ride at two hours and a second short interval session; this holds and adds a little. Eight hours and up: two interval sessions, two endurance rides, and the long one at two and a half hours or more; this builds. The winter cycling training plan spoke lays out the November-to-February arc in detail; the point here is that every tier keeps the hard session and scales the easy volume, never the reverse. Session length follows from the fact that a trainer hour has no coasting. Sixty minutes on a trainer delivers roughly the pedalling time of 75 to 90 minutes on a road with junctions and descents — a rule of thumb, not a measurement, but a fair one for most terrain. Interval sessions need 60 to 75 minutes including warm-up and cool-down; endurance rides indoors rarely need to exceed two hours, and past 90 minutes the heat problem from the previous section starts to dominate the training stimulus. Forty-five minutes is a real session. Twenty minutes is a warm-up. ## The intensity mix that works on a trainer Mostly easy, some very hard, little in between — indoors as outdoors. In trained cyclists on 6.4 hours a week, six weeks of a polarized distribution (80% low, 20% high, no middle) raised peak power 8% and lactate threshold 9%, against 3% and 2% for a threshold-heavy model on more hours [Neal et al. 2013]. The trainer makes the middle zone convenient. Convenience is not the same as effective. The polarized result is not one study. Across 48 well-trained endurance athletes over nine weeks, the polarized group improved VO2peak 11.7% and time to exhaustion 17.4%, while the threshold group improved VO2peak not at all [Stöggl & Sperlich 2014]. The pattern is consistent enough that the burden of proof sits on the sweet-spot winter, not on the easy one. Sweet spot has an honest place — for a rider with four hours, one 2×20 at 88 to 94% of FTP is a defensible way to get a threshold stimulus and a load stimulus in one 60-minute slot — but three of them a week with no easy riding around them is the threshold model that lost in both trials. For the hard session, the trainer is where the interval literature was written, and it favours accumulated time at high intensity over maximal intensity. In 35 cyclists training about six hours a week, 4×8 minutes at 90% of peak heart rate produced roughly double the gain in VO2peak and power (11.4%) of either 4×4 at 94% or 4×16 at 88% [Seiler et al. 2013]. Short intervals earn their place too: ten weeks of 30-second work with 15-second recoveries, effort-matched against 4×5 minutes, raised VO2max 8.7% against 2.6% [Rønnestad et al. 2015]. Both formats are easy to execute in ERG mode and miserable to execute on an open road, which is the trainer's real advantage. Zone 2 is the part of the distribution the trainer fights. It feels harder than it is [Olsson et al. 2024], it takes the longest, and the temptation to nudge the target up to make the hour pass is constant. Two rules hold it in place. First, the endurance ride is prescribed by power and heart-rate ceiling, and the ceiling wins: if drift pushes heart rate above the zone, the watts come down, not the fan. Second, the easy ride is where the platform earns its subscription — a group ride on Zwift or a route on Rouvy makes 90 minutes at 65% of FTP tolerable in a way a blank wall does not, and that is a legitimate use of the money, not a weakness. ## Structure without a subscription A structured indoor session needs a trainer, a target and a controller, and a modern head unit is a controller. A Wahoo ELEMNT or a Hammerhead Karoo will drive a smart trainer in ERG mode from a planned workout with nothing running on a laptop or a phone. What that buys is training structure on hardware already owned. What it does not buy is a world. The mechanism is documented by both vendors and covered in the head-unit spokes of this site, so it is only sketched here: the plan pushes tomorrow's workout to the device, the device pairs the trainer over ANT+ FE-C, and the intervals run with the trainer setting resistance to each power target. A workout without power targets — a heart-rate or feel session — renders its intervals and leaves the flywheel free, which surprises riders the first time. ERG has its own failure mode when the target is set too high, which the KICKR ERG spoke explains. None of this needs a subscription of any kind. The trade-off is real and should be stated plainly. No world, no group rides, no racing, no gamification — a rider, a set of numbers and the intervals. For a shoulder-season Tuesday when the Zwift subscription is paused, that is fine, and it is the reason we built the push. For January motivation it is probably not enough for most riders, and the evidence on why is in the first section: distraction lowers perceived effort at the same intensity [Olsson et al. 2024], and Zwift is, honestly, the best indoor riding experience available. Most riders will want it back in winter. A paused subscription should not cost anyone their training structure in the meantime; it also should not be mistaken for a training plan. Keep the roles straight and the two coexist. The platform makes the session happen; the plan decides what the session should be. Zwift, TrainerRoad, Wahoo SYSTM and Rouvy all execute workouts well, and TrainerRoad's Adaptive Training is a genuinely good progression engine inside its own structure. What none of them do is read the ride that actually happened across every device and rebuild the rest of the week around it. We also ship a free Zwift workout generator for riders who want our sessions inside Zwift's world, which is the coherent position: the world is worth paying for, the plan should not be locked to it. ## When the indoor season breaks — and what the plan should do Winter concentrates every force that breaks plans: darkness, cold-and-flu season, holidays, year-end work, and the monotony of the same room and the same hour. Foster's 25-athlete study tied illness to individual thresholds of training strain, the product of load and monotony [Foster 1998]. The indoor season fails on monotony more often than on fitness. Monotony is a specific risk indoors because the trainer removes the variance that outdoor riding adds for free. Foster's index is daily load divided by its standard deviation over the week; a week of five identical 60-minute sessions scores badly even at modest load, and strain — load multiplied by monotony — was the index that best predicted when athletes got sick [Foster 1998]. The fix is cheap: the hard days harder, the easy days genuinely easy, one day off, and the long ride long enough to be different. A plan that prescribes the same 2×20 on Tuesday, Thursday and Saturday is not conservative. It is monotonous, and monotony is the thing the evidence flags. The other breakages are life. A normal cold costs 5 to 10 days between November and February. The holidays remove one to two weeks from most Decembers. Work compresses around the year end. A static plan meets all of this with the next page in the schedule, and the rider either compresses the missed block into January — the wrong answer, because it stacks load into the monotony trap above — or concludes the plan is broken and stops. This is where the word adaptive has to earn something. The right response to a lost week is a restructure: the return week rebuilt at reduced load, the next recovery week pulled forward, the spring arc re-anchored to the dates that now exist. We built AdaptCycling to run that restructure without being asked. Connect the rides — Wahoo, Hammerhead, Polar, intervals.icu — and the plan reads the week that actually happened, notices the ride that did not, and reshapes what remains instead of grading the rider against what was scheduled. An indoor season that adapts when winter life does is the whole thesis of this guide, and it is the difference between January being week ten of something and day one of nothing. The workouts were never the hard part of winter. Keeping the plan and the calendar in the same month is. ## Common questions **How long should an indoor trainer ride be?** Sixty to 75 minutes for an interval session including warm-up and cool-down, 90 minutes to two hours for an endurance ride, rarely longer. A trainer hour carries roughly the pedalling of 75 to 90 road minutes because nothing is lost to coasting, and past 90 minutes heat starts to dominate the stimulus unless the fan and room are very good [Galloway & Maughan 1997]. Forty-five minutes is a real session. **Why is my heart rate different on the trainer than outside?** It can go either way, and both are heat. At the same perceived effort riders produce fewer watts indoors and heart rate reads lower because the work is lower [Mieras et al. 2014]. At the same watts, the missing airflow drives cardiovascular drift and heart rate reads higher over the hour [Coyle & González-Alonso 2001]. A strong fan running from minute one closes most of the gap; a fan switched on late does not [Wingo et al. 2019]. **Do I need Zwift to train indoors in winter?** No. A trainer, a plan and one effort signal — power, heart rate or perceived exertion — cover every session in a winter block, and a Wahoo ELEMNT or Hammerhead Karoo will drive a smart trainer in ERG from a planned workout with no subscription. What Zwift adds is the world, and the world lowers the perceived cost of easy volume in a way a blank wall does not. Most riders will want it for January. Nobody needs it for structure. **Should I ride sweet spot or zone 2 on the trainer?** Mostly zone 2, with the week's hard work done properly hard. Polarized distributions beat threshold-heavy ones in trained cyclists on the same or fewer hours [Neal et al. 2013, Stöggl & Sperlich 2014]. One sweet-spot session a week is a defensible compromise for a four-hour rider; three a week with no easy riding is the model that lost. The trainer makes the middle zone convenient, which is exactly why it needs a ceiling. **How many trainer sessions a week does it take to keep my fitness?** Two, if they are hard. Riders who cut from six sessions a week to two held their VO2max gains for 15 weeks as long as intensity stayed [Hickson & Rosenkoetter 1981]; riders who kept the sessions and cut the intensity by a third did not [Hickson et al. 1985]. Add easy volume as time allows — four to six hours a week holds fitness, six to ten builds it — but the hard session is the part that cannot be dropped. **How do I move back outside in spring without losing the winter?** Expect the reverse of the autumn adjustment: the same perceived effort produces more watts outside [Mieras et al. 2014], so the first outdoor rides feel easy and tempt riders into too much too soon. Keep the interval session indoors for the first two or three weeks where its execution is exact, move the endurance ride outside first, and let the plan read the outdoor rides as they arrive rather than replacing the structure with the weather. ## References 1. **Mieras et al. 2014.** [Physiological and psychological responses to outdoor vs. laboratory cycling](https://pubmed.ncbi.nlm.nih.gov/24476776/). Journal of Strength and Conditioning Research. 2. **Galloway & Maughan 1997.** [Effects of ambient temperature on the capacity to perform prolonged cycle exercise in man](https://pubmed.ncbi.nlm.nih.gov/9309637/). Medicine & Science in Sports & Exercise. 3. **Olsson et al. 2024.** [Perceived exertion can be lower when exercising in field versus indoors](https://pmc.ncbi.nlm.nih.gov/articles/PMC11135770/). PLoS One. 4. **Saunders et al. 2005.** [The effects of different air velocities on heat storage and body temperature in humans cycling in a hot, humid environment](https://pubmed.ncbi.nlm.nih.gov/15743384/). Acta Physiologica Scandinavica. 5. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. 6. **Wingo et al. 2019.** [Fan cooling after cardiovascular drift does not reverse decrements in maximal oxygen uptake during heat stress](https://pmc.ncbi.nlm.nih.gov/articles/PMC6773240/). Temperature. 7. **Hickson & Rosenkoetter 1981.** [Reduced training frequencies and maintenance of increased aerobic power](https://pubmed.ncbi.nlm.nih.gov/7219129/). Medicine & Science in Sports & Exercise. 8. **Hickson et al. 1985.** [Reduced training intensities and loss of aerobic power, endurance, and cardiac growth](https://pubmed.ncbi.nlm.nih.gov/3156841/). Journal of Applied Physiology. 9. **Neal et al. 2013.** [Six weeks of a polarized training-intensity distribution leads to greater physiological and performance adaptations than a threshold model in trained cyclists](https://pubmed.ncbi.nlm.nih.gov/23264537/). Journal of Applied Physiology. 10. **Stöggl & Sperlich 2014.** [Polarized training has greater impact on key endurance variables than threshold, high intensity, or high volume training](https://pmc.ncbi.nlm.nih.gov/articles/PMC3912323/). Frontiers in Physiology. 11. **Seiler et al. 2013.** [Adaptations to aerobic interval training: interactive effects of exercise intensity and total work duration](https://pubmed.ncbi.nlm.nih.gov/21812820/). Scandinavian Journal of Medicine & Science in Sports. 12. **Rønnestad et al. 2015.** [Short intervals induce superior training adaptations compared with long intervals in cyclists — an effort-matched approach](https://pubmed.ncbi.nlm.nih.gov/24382021/). Scandinavian Journal of Medicine & Science in Sports. 13. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. --- # What your Strava Fitness score actually means (and whether yours is good) URL: https://www.adaptcycling.com/guides/what-strava-fitness-number-means Updated: 2026-07-30 Author: Jim Camut The integer Strava labels Fitness is CTL — Chronic Training Load, a 42-day exponentially-weighted average of your daily training load [Allen et al. 2019]. It is not a fitness test score, a VO2max, or a leaderboard rank. It is one rider's rolling load trend, on a scale that only means anything against their own history and goal event. There is no universal good number, because riders respond to the same training differently [Mann et al. 2014]. Here is what the number represents, why the good question is personal, and the decisions worth making from it. ## What the Fitness number represents: CTL, a rolling weighted load Strava Fitness is CTL — a 42-day exponentially-weighted average of daily training load, the same metric TrainingPeaks calls CTL and Intervals.icu calls Fitness [Allen et al. 2019]. It is an accumulated-load index, not a measured fitness value. A 60 means your last six weeks averaged 60 load units a day, nothing more. The definition is specific. Each day Strava takes that day's training load, blends in a sliver of it, and lets the prior value decay slightly — a 42-day exponentially-weighted moving average that Coggan and Allen named Chronic Training Load [Allen et al. 2019]. The 42-day window descends from Eric Banister's 1970s fitness-fatigue model, whose decay constants were curve-fit against athlete performance data and have survived four decades of validation [Hellard et al. 2007]. Strava, TrainingPeaks, and Intervals.icu all compute the same underlying number; only the label on screen differs. What feeds the average is daily training load — roughly intensity multiplied by duration. Strava uses TSS-equivalent values when you ride with power and Relative Effort when you don't [Allen et al. 2019]. That input choice changes the number, which is its own topic; our companion piece on whether to trust Relative Effort or TSS for each workout type covers it. For understanding what the integer means, the point is narrower: Fitness is the running average of whatever load Strava scored, not a direct readout of your engine. It rises slowly, because you can only add a little load per day against a six-week average, and it falls faster than it climbed once recent load drops below that baseline — which is why a falling Fitness line can coexist with hard training. ## Why it is a personal trend, not a number to compare Fitness has no universal good value because the same training produces meaningfully different physiological responses across riders [Mann et al. 2014]. The number is scaled to your own load inputs and your own history, so comparing your CTL to another rider's — or to a feed full of them — is comparing two rulers with different units. Two riders can run the identical eight-week block and arrive at different fitness gains. Mann and colleagues reviewed the literature on high and low responders and found wide inter-individual variation in adaptation to standardized endurance training, driven by genetics, training history, recovery, and measurement factors [Mann et al. 2014]. A CTL of 70 that represents peak form for one rider can be an easy maintenance week for another. The integer carries no information about where it sits on anyone else's scale. The number is also only as comparable as its inputs. A rider on power and a rider on heart rate are feeding the average different load metrics, so their CTLs are not on the same axis even before physiology enters. Strava's own data shows the platform is fundamentally social — over 135 million athletes, a feed built on kudos and segments [Strava 2024]. That design quietly invites the wrong comparison: reading your Fitness integer against the riders you follow, when the number was never a cross-rider statistic. This sits inside the broader discipline of using Strava as a training tool rather than a journal. The data layer every cyclist already has only helps if you read your own trend in its own context. Strava syncs and displays the integer; it does not tell you that a 55 is good for your event or a 90 is more than your life can sustain. We built AdaptCycling to read the full history behind the number — your power curve, your consistency, your gaps — and turn it into a plan, rather than leaving you to guess what the integer is worth. ## Is my Fitness number good? The honest answer is event-relative A good Fitness number is the CTL that matches the demands of your goal event, built sustainably and held without breaking you [Friel 2018]. There is no single target. A 50 can be ample for a local crit and thin for a multi-day gravel race, so the useful question is always good for what, not good in the abstract. Friel's periodization framework anchors the right way to read the number: you reverse-engineer the load your goal event demands and build toward it across a season [Friel 2018]. A flat 90-minute crit asks less accumulated load than a 200-kilometer gravel event with 3,000 meters of climbing. The CTL that leaves you race-ready for one is under-built or over-built for the other. The number is good when it lines up with what you are training for — not when it clears some round threshold you saw in a forum. Coaching convention offers loose ballparks, and they are worth stating with the caveat that they are coaching-grade reference points, not validated cutoffs. Many time-crunched amateurs sustain a CTL in the 40 to 60 range; serious enthusiasts training 8 to 12 hours a week often live in the 70 to 90 range; and figures above 100 generally belong to riders with the time and recovery capacity of near-professionals. These are rough orientation only — your sustainable number depends on your inputs, your hours, and how you personally respond [Mann et al. 2014]. The more useful frame than any single target is a readiness band: roughly where your Fitness should sit in the final weeks before your event, given how it has trended. A number climbing steadily toward an event-appropriate level weeks out is good. The same number reached by a panicked late spike, or drifting well below where your event demands, is not — regardless of the digits. Good is a trajectory relative to a goal, which is exactly the judgment the chart cannot make for you. ## The decisions actually worth making from it The Fitness number earns its place by informing four decisions: whether your load is trending toward your event, whether you are building at a sustainable rate, when to back off, and when you are ready to peak [Friel 2018]. Read it as a multi-week slope, never as a daily verdict. First, direction relative to the goal. In base and build phases you want the line climbing, because a rising CTL means you are accumulating load faster than you are shedding it [Allen et al. 2019]. If your event is twelve weeks out and the number is flat, the plan is not progressing. That is a real signal worth acting on — distinct from a single low ride, which moves the 42-day average by only one to three points and is noise. Second, rate. The number going up is good only if it is going up sustainably; a CTL that spikes too fast is a load you have not earned the recovery for, and that is where injury and illness risk climb [Foster 1998]. Chase the integer upward too aggressively and you buy a number you cannot hold. How fast is too fast is its own question, owned by the broader self-coached training literature rather than the Strava data layer. Third and fourth, backing off and peaking. When Fitness has climbed for several weeks, the number itself is your cue that adaptation is banked and a lighter week is due — and when you taper, the integer is supposed to fall while you arrive fresh, which is the plan working, not failing. Both of those readings depend on pairing Fitness with Form, the freshness side of the same chart, which the parent pillar on training with Strava lays out. The single rule that survives all four decisions: the number is a slope you read over weeks, and reading it daily is how riders talk themselves into bad training. ## Common questions **Is the Strava Fitness score the same as a fitness test result?** No. Despite the name, your Strava Fitness score is not a test result, a VO2max, or a ranking — it is CTL, a 42-day exponentially-weighted average of your daily training load [Allen et al. 2019]. "Score" is just the everyday word riders use for the Fitness number Strava plots; it reflects accumulated load over the last six weeks, not how fast you are on any single day. **What does the Fitness number on Strava actually measure?** It measures CTL — Chronic Training Load — a 42-day exponentially-weighted average of your daily training load [Allen et al. 2019]. It is not a fitness test, a VO2max, or a power figure. A Fitness of 60 means your last six weeks averaged roughly 60 load units a day. It is an accumulated-load index, and the same metric appears as CTL in TrainingPeaks and Fitness in Intervals.icu. **Is a Strava Fitness of 50 good?** It depends entirely on your goal event and how you respond to training [Friel 2018, Mann et al. 2014]. A 50 can be ample for a short flat criterium and thin for a long mountainous gravel race. There is no universal good number, because the integer is scaled to your own load inputs and your own physiology — riders respond to identical training differently. Judge it against your event and your own history, never against another rider. **Why is my Fitness number different from my friend's at the same level?** Because the number is not a cross-rider statistic. Your CTL is the average of whatever load Strava scored from your rides, so a power rider and a heart-rate rider are not even on the same scale [Allen et al. 2019]. On top of that, riders show wide individual variation in how they adapt to the same training [Mann et al. 2014]. Comparing two Fitness integers is comparing two rulers with different units. **What is a good Fitness number for an amateur cyclist?** As loose coaching-grade orientation only: many time-crunched amateurs sustain a CTL of 40 to 60, serious enthusiasts on 8 to 12 hours a week often sit at 70 to 90, and figures above 100 usually require near-professional time and recovery. These are reference points, not validated cutoffs. Your sustainable number depends on your training hours, your load inputs, and how you personally respond [Mann et al. 2014]. **Should I try to maximize my Fitness number?** No. The goal is the CTL your event demands, built at a rate you can recover from — not the highest possible integer [Friel 2018]. Chasing the number upward too fast buys load you have not earned the recovery for, and that is where injury and illness risk rise [Foster 1998]. Read Fitness as a slope trending toward your goal over weeks, and pair it with the freshness side of the chart rather than treating the number as a score to push. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Hellard et al. 2007.** [Assessing the limitations of the Banister model in monitoring training](https://pmc.ncbi.nlm.nih.gov/articles/PMC1974899/). Journal of Sports Sciences. 3. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. 4. **Mann et al. 2014.** [High responders and low responders: factors associated with individual variation in response to standardized training](https://pubmed.ncbi.nlm.nih.gov/24807838/). Sports Medicine. 5. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 6. **Strava 2024.** [Strava Year in Sport Trend Report 2024](https://press.strava.com/articles/strava-releases-annual-year-in-sport-trend). Strava. --- # Apps that connect to Strava: how to tell which ones read your data and which just display it URL: https://www.adaptcycling.com/guides/strava-syncs-but-doesnt-read Updated: 2026-06-01 Author: Jim Camut Almost every training app lists a Strava connection, but connecting and reading are different things. The bare minimum is OAuth at sign-up plus activity polling, which puts your rides on a dashboard and stops there. Reading means the app parses each new ride and changes what it tells you to do next. The tell is behavioral: if Saturday's hard group ride never alters Tuesday's planned session, the app is displaying Strava, not reading it. Here are the buyer-side questions to ask and the technical signals — webhooks, same-day plan changes, FTP from history — that separate the two. ## What "connects to Strava" usually means For most apps, "connects to Strava" means OAuth at sign-up and periodic polling that copies your activity files onto their dashboard. Power, heart rate, and time-in-zone appear; the plan was built from a sign-up questionnaire and never moves. Display is the default because it is the cheap, one-time integration [Strava Developers]. The minimum viable Strava integration is a single OAuth handshake and read access to your activity list. After that, the app pulls your rides on a schedule, renders them, and computes the same summary numbers Strava already shows — average and normalized power, weighted-average heart rate, time-in-zone. This is useful as a log, and it is where most of the market lives. What it does not do is feed those numbers back into anything that decides your training. The giveaway is that the plan is static. You answered questions at sign-up — hours per week, goal event, current FTP — and the app generated a block from those answers. The rides you actually do afterward are displayed alongside the plan but do not change it. A six-hour week that turned into a two-hour week leaves next week's prescription untouched. The data syncs; the plan does not read it. This is the broader point the pillar on training with Strava keeps making: Strava is the data layer every cyclist already has, and most apps treat it as a journal rather than an input. None of this is a knock on display-only apps. A clean activity log against a fixed plan is exactly what some riders want, and it is honest about what it is. The problem is only that the Strava-connection badge looks identical whether the app reads your data or merely shows it, so the buyer cannot tell which job the app actually does. ## What "reading Strava" looks like from the outside Reading means the app derives something from your rides that it could not have known at sign-up, then acts on it. Three observable behaviors prove it: FTP estimated from your power history rather than typed in, a plan that changes after an unplanned hard ride, and recognition that a missed week needs a rebuilt next week [Allen et al. 2019, Hellard et al. 2007]. The first behavior is FTP from data. An app that reads your history can estimate threshold power from your last 60-90 days of rides — the shape of your power-duration curve — without asking you to take a test [Allen et al. 2019]. TrainerRoad's AI FTP Detection does this across indoor and outdoor rides with power, including files imported from Strava. If an app only ever uses the FTP you typed at sign-up, it is not reading your power data; it is storing your guess. The second behavior is plan response to unplanned efforts. Saturday's spirited group ride was zone 4 work whether or not it was on the calendar. An app that reads your rides notices the load and can flag that Tuesday's planned VO2 session is now redundant or risky. An app that only displays will happily prescribe the hard interval on top of the accidental hard ride, because the two events never touched. This is the fitness-fatigue accounting the Banister model formalizes [Hellard et al. 2007] — and it only works if every ride actually moves the ledger. The third behavior is gap recovery. Miss a week to a sick kid or a work trip and a reading app rebuilds the next two weeks at a reduced volume rather than resuming the original script as if nothing happened. A display app has no mechanism to know the week was missed, because the missing rides are simply absent from a chart, not a signal that triggers a rebuild. The point here is only that the rebuild happening at all is the proof the app read the gap. ## The technical tell: webhooks and same-day plan changes The cleanest objective signal is timing. Strava's Webhook Events API pushes a notification to a connected app within seconds of a ride uploading, and the app must acknowledge within two seconds [Strava Developers]. An app wired for webhooks can change your plan the same hour you finish a ride. A polling-only app waits for its next sync cycle. Strava offers two ways to get your data: poll the API on a schedule, or subscribe to webhooks that fire on activity create, update, and delete events [Strava Developers]. Polling is simpler to build and fine for a log — the activity shows up whenever the next pull runs, minutes or hours later. Webhooks are the infrastructure of an app that intends to act fast: the moment your ride uploads, the service is notified, parses the file, and adjusts the rest of the week before you have showered. You can observe this without reading code. Finish a ride, let it upload to Strava, and watch the app. If your plan, your fitness estimate, or tomorrow's session updates within minutes, the app is almost certainly webhook-driven and reading each ride as it lands. If the change takes hours, or only appears after you manually trigger a sync, the integration is polling and probably display-first. Webhooks are necessary but not sufficient — an app can receive the event and still do nothing useful with it. So pair the timing test with a content test: did the update actually change a decision, or just refresh a number on a chart? An app that webhooks in a ride and then alters tomorrow's prescription is reading and acting. One that webhooks in a ride and only updates a fitness line is reading and displaying. We connect at sign-in — Wahoo, Hammerhead, Polar or intervals.icu, with Strava seats capped and Strava riders joining a short waitlist — read your ride history to seed the plan, and pick up every new ride on upload so the plan adapts the same day; Intervals.icu reads your history to build a free analytics layer without coaching on top; the difference between those two is the difference between a coach and a dashboard. ## The questions to ask before you trust an app with your training Four buyer-side questions separate readers from displayers, and none requires a trial. Does it estimate FTP from my history or ask me to type it? Does an unplanned hard ride change my next session? Does a missed week rebuild the plan? And does it capture how the ride felt, not just what the sensors recorded [Foster 1998, Haddad et al. 2017]? Start with FTP sourcing, because it is the fastest to check. If onboarding asks for your FTP and never revisits it from your rides, the app is not reading your power curve. If it estimates a number from your history and re-estimates as you train, it is [Allen et al. 2019]. The same logic extends to heart-rate zones: an app that reads your data can flag a max-HR setting that looks wrong from your actual peaks rather than trusting a formula. Then probe adaptation directly. Ask the app — or its docs — what happens when you do a hard ride that was not on the plan, and what happens when you miss three days. A reading app has a specific answer: the next session is rescaled, the week is rebuilt, the 80/20 easy-hard balance the plan is built to protect is restored [Seiler 2010]. A display app has a vaguer answer, usually some version of "your activities will sync" — which describes the log, not the response. Finally, ask whether the app captures context. Power and heart rate are external load; they cannot see the three hours of sleep or the sick toddler behind a flat-looking ride. Carl Foster's session-RPE work is clear that internal load — what the work felt like — explains training response better than external load alone [Foster 1998, Haddad et al. 2017], and the validation literature backs a one-to-ten rating as a metric that stands on its own [Haddad et al. 2017]. An app that asks how the ride felt, and changes its advice based on the answer, is reading a dimension Strava itself does not record. As our companion piece on whether Strava Premium is worth it for a self-coached rider puts it, the gap between syncing your data and acting on it is the gap between a data tool and a coach. ## Common questions **How can I tell if a training app actually reads my Strava data?** Finish a ride, let it upload to Strava, and watch the app. If your plan or fitness estimate changes within minutes, it is likely webhook-driven and reading each ride [Strava Developers]. If nothing changes until the next scheduled sync or a manual refresh, it is polling and probably display-only. Then check whether the change altered a decision or just refreshed a chart number. **Do all apps that connect to Strava use the same data?** They have access to the same data — Strava grants connected apps read access to your activities — but they use it very differently. Most copy the activity files onto a dashboard and stop there. A few parse each ride and feed it into an FTP estimate, a fitness model, or a plan that adapts [Allen et al. 2019]. Same input, very different output. **What is a Strava webhook and why does it matter for training apps?** A webhook is a real-time notification Strava sends a connected app the moment you create, update, or delete an activity, and the app must acknowledge within two seconds [Strava Developers]. It matters because it lets an app react the same hour you ride rather than waiting for a polling cycle. Webhooks are the infrastructure behind same-day plan changes. **Is a display-only Strava app worse than one that reads my data?** Not worse, different. A clean log against a fixed plan is the right tool for a rider who wants structure they manage themselves. A reading app is the right tool for a rider who wants the plan to absorb unplanned hard rides and missed weeks automatically. The mistake is paying for a reading app and getting a display one because the Strava badge looked the same. ## References 1. **Strava Developers.** [Webhook Events API](https://developers.strava.com/docs/webhooks/). Strava Developers. 2. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 3. **Hellard et al. 2007.** [Assessing the limitations of the Banister model in monitoring training](https://pmc.ncbi.nlm.nih.gov/articles/PMC1974899/). Journal of Sports Sciences. 4. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 5. **Haddad et al. 2017.** [Session-RPE Method for Training Load Monitoring: Validity, Ecological Usefulness, and Influencing Factors](https://pmc.ncbi.nlm.nih.gov/articles/PMC5673663/). Frontiers in Neuroscience. 6. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. --- # Is Strava Premium worth it for a self-coached cyclist? URL: https://www.adaptcycling.com/guides/strava-premium-worth-it-self-coached Updated: 2026-06-01 Author: Jim Camut For a training-focused rider, Strava Premium at $11.99 a month or $79.99 a year buys mostly one thing worth paying for: the Fitness and Freshness chart, a downstream view of Eric Banister's fitness-fatigue model [Hellard et al. 2007, Allen et al. 2019]. It does not buy coaching or a plan that adapts to your rides. Intervals.icu rebuilds the same chart from your Strava history for free. Whether Premium is worth it comes down to one question: do you want the chart inside the app you already open, or are you fine running it elsewhere for nothing? ## What you actually pay for Strava's free tier already records every ride and displays power, heart rate, time-in-zone, segments, and Relative Effort. Premium adds the Fitness and Freshness chart, filtered and live segment leaderboards, route planning, and AI workout summaries [Strava 2024]. For a self-coached rider, only the chart changes how you train. Separate the training-relevant features from the rest. Free Strava captures and shows your raw data: watts, weighted-average heart rate, time-in-zone splits, segment times, and the Relative Effort score on every ride. None of that sits behind the paywall. Premium's training-facing additions are the Fitness and Freshness chart, segment leaderboards you can filter by age, weight, or club, Live Segments on your head unit, and route building. At $79.99 a year, you are paying roughly $6.67 a month, and for most training-focused riders the chart is the only line item that touches how the week is structured. The chart is the real product. It plots Fitness as a 42-day average of training load, Freshness as a 7-day average, and Form as the difference between them [Allen et al. 2019] — the same three lines TrainingPeaks renders as its Performance Manager Chart. The window lengths trace back to Banister's 1970s curve-fitting and have held up across four decades of validation work [Hellard et al. 2007]. If you make four or five decisions a season from that curve — when to take a down week, when you are fresh enough to race — Premium is paying for a tool you use. If you glance at it once a month, you are renting a screensaver. The rest of the Premium bundle is real but not differentiated for training. Route planning is good and other tools do it. Segment-leaderboard filtering feeds the competitive pull that derails periodized weeks more than it helps them, a trap the broader pillar on training with Strava covers at length. None of it reads your rides and changes tomorrow's session, which is the line that separates a data tool from a coach. ## What Premium does not give you: coaching or adaptation Strava Premium describes your training; it does not direct it. The Fitness and Freshness chart shows what happened. Athlete Intelligence, Strava's subscriber-only AI, summarizes a finished ride in plain language [Strava 2024]. Neither builds a plan, neither restructures your week when you miss three days, and neither tells you what to do tomorrow. Athlete Intelligence, launched to subscribers in 2025, is the feature most often mistaken for coaching. It reads a completed activity and writes a short summary — your longest ride this month, a heart-rate or power highlight, a segment note. That is descriptive summarization of one ride, not prescription across a block. It does not know your goal event, your available hours, or that Saturday's group ride already covered Tuesday's planned intensity. The technical gap matters here: reading Strava data to coach on it is webhook-driven [Strava Developers], and summarizing a single ride after the fact is a different and far smaller problem than rebuilding a plan from it. The chart has the same ceiling. Form crossing zero tells you that you are tired; it does not tell you whether to push through or back off, because that answer depends on context the chart cannot see — your sleep, your deadline, whether the fatigue is planned overload or the start of illness. Carl Foster's session-RPE work is clear that internal load, what the work felt like, explains training response better than external load alone [Haddad et al. 2017], and Strava captures almost none of the internal side. A negative Form reading and a great night's sleep is a green light; the same reading on three hours of sleep is a stop sign. The chart shows one number for both. This is the honest boundary of Premium for a self-coached rider. It is an analytics layer, and a good one. It answers what happened and roughly where your fitness sits. It does not answer what to do next, and paying $79.99 a year expecting it to is the most common way riders end up disappointed in the subscription. ## The free alternative: what Intervals.icu replicates for nothing Intervals.icu connects to your Strava account, imports your ride history, and builds the same Fitness, Fatigue, and Form chart from CTL, ATL, and TSB — for free, with no credit card [Hellard et al. 2007, Allen et al. 2019]. For a rider who wants the chart but not the subscription, it removes the main reason to pay. Intervals.icu reads your Strava data as a connected source, pulls in your historical activities, and computes the same three lines Premium charges for. It plots Fitness, Fatigue, and Form on the same fitness-fatigue model [Allen et al. 2019], and it lets you configure the CTL and ATL time constants rather than locking you to 42 and 7 days — useful if you want to experiment, irrelevant if you do not. It also surfaces time-in-zone, power-curve, and load distribution views. The cost is free; the developer funds it through optional donations. The honest tradeoffs are workflow, not capability. Intervals.icu is a second tab — you upload or sync to Strava, then it pulls across, so your chart lives outside the app you already open after a ride. Its interface is denser and less polished than Strava's; it rewards a rider who wants to see the math and punishes one who wants a clean glance. And it is a small independent project rather than a funded company, which is a durability consideration over a multi-year horizon, not a function gap today. What it does not lack is the chart itself. For the buy-or-skip decision, this is the pivot. If the Fitness and Freshness chart is the one Premium feature you would use, Intervals.icu gives you an equivalent for free, and many self-coached riders run on free Strava plus a free Intervals.icu account indefinitely. We read both your power and heart-rate streams the same way Intervals.icu does, but the difference is what we do with them: the chart is the input to an adapting plan, not the output you read and interpret yourself. ## When Premium is worth it and when it is not Premium is worth $79.99 a year if you will make real decisions from the chart and want it inside the app you already live in, or if you ride segments and leaderboards socially. Skip it if you only want the chart — Intervals.icu is free — or if you want a plan, which no version of Strava provides. Buy Premium if two things are true: the Fitness and Freshness chart will drive actual decisions in your season, and the friction of a second app is worth avoiding for you. Friel's framework puts the useful taper window around +5 to +25 Form before a goal event [Friel 2018], and if you are timing peaks off that curve, having it one tap away in the app you open after every ride has genuine value. Riders who race segments, want filtered leaderboards, or use Live Segments on a head unit also get real use from the bundle. For them, $6.67 a month on the annual plan is a fair price for a tool they touch weekly. Skip Premium if either of these is true. First, if the chart is the only feature you want, Intervals.icu rebuilds it from your Strava history for free and the only thing you lose is the convenience of one app. Second, and more important, if what you actually want is a plan — a coach that reads your rides and restructures the week when life disrupts it — no tier of Strava provides that, and paying for Premium hoping it will is a category error. Strava syncs and displays your data well; reading it and acting on it is a different product. TrainerRoad and Athletica build adaptive plans around power; Intervals.icu gives you the analytics free; AdaptCycling reads your ride history — through Wahoo, Hammerhead, Polar or intervals.icu, with Strava seats capped — and adapts the plan to the rides you actually did. Decide which of those three jobs you are hiring for before you decide whether $79.99 a year is worth it. ## Common questions **Is Strava Premium worth it just for the Fitness and Freshness chart?** Only if you value having it inside the app you already use. The chart is a real tool — three lines from Banister's fitness-fatigue model [Hellard et al. 2007] that support a handful of season decisions. But Intervals.icu rebuilds the same Fitness, Fatigue, and Form chart from your Strava history for free [Allen et al. 2019]. If the chart is the sole reason you would pay $79.99 a year, the free alternative likely makes more sense. **Does Strava Premium include a training plan or coaching?** No. Premium gives you analytics — the Fitness and Freshness chart and AI workout summaries through Athlete Intelligence [Strava 2024] — but no plan and no adaptation. It describes what you did; it does not prescribe what to do next or restructure your week when you miss sessions. For a plan that reads your rides, you need a coaching product, not a Strava subscription. **What does Strava's Athlete Intelligence AI actually do?** It summarizes a single completed ride in plain language — a power or heart-rate highlight, a milestone, a segment note. It is descriptive, not prescriptive. It does not know your goal event or weekly hours, and it does not build or adjust a plan. Reading Strava data to coach on it is a webhook-driven problem [Strava Developers]; summarizing one finished ride is a much smaller one. **Can I get everything I need for training on free Strava?** For raw data, yes — free Strava records power, heart rate, time-in-zone, segments, and Relative Effort on every ride. What free leaves out is the Fitness and Freshness chart, which Intervals.icu replaces at no cost. What no version of Strava provides, free or paid, is an internal load measure like session-RPE [Haddad et al. 2017] or a plan that adapts to your rides. ## References 1. **Strava 2024.** [Strava Year in Sport Trend Report 2024](https://press.strava.com/articles/strava-releases-annual-year-in-sport-trend). Strava. 2. **Hellard et al. 2007.** [Assessing the limitations of the Banister model in monitoring training](https://pmc.ncbi.nlm.nih.gov/articles/PMC1974899/). Journal of Sports Sciences. 3. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 4. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. 5. **Haddad et al. 2017.** [Session-RPE Method for Training Load Monitoring: Validity, Ecological Usefulness, and Influencing Factors](https://pmc.ncbi.nlm.nih.gov/articles/PMC5673663/). Frontiers in Neuroscience. 6. **Strava Developers.** [Webhook Events API](https://developers.strava.com/docs/webhooks/). Strava Developers. --- # Relative Effort vs TSS: which to trust for each workout type URL: https://www.adaptcycling.com/guides/relative-effort-vs-tss-which-to-trust Updated: 2026-06-01 Author: Jim Camut Relative Effort and TSS are not two readings of the same thing. Relative Effort is a heart-rate-zone score, closer to Eric Banister's TRIMP than to a power meter's output [Meyer 2018, Hellard et al. 2007]. TSS is mechanical work, scaled so one hour at threshold equals 100 [Allen et al. 2019]. They answer different questions on different rulers, which is why no single number is right for every ride. The rule that holds across workout types: trust power-based TSS when heart rate cannot keep up with the effort, and Relative Effort when the work is steady and aerobic enough for heart rate to settle into the zone it belongs in. ## What each metric actually measures Relative Effort sums weighted time-in-heart-rate-zone against your max HR; TSS scales normalized power against your FTP [Meyer 2018, Allen et al. 2019]. One counts heartbeats, the other counts watts. Neither is a corrected version of the other, and there is no conversion factor between them. Relative Effort sums weighted time-in-heart-rate-zone against your configured max HR, then scales it so efforts across sports stay comparable. The math descends from Banister's TRIMP, an exponential weighting of heart-rate reserve [Hellard et al. 2007]. TSS does something structurally different: it takes normalized power, scales it against your FTP, and multiplies by duration, calibrated so 60 minutes at threshold returns exactly 100 [Allen et al. 2019]. One counts heartbeats; the other counts watts. The split matters because heart rate is a lagging, capacity-bounded signal and power is an instantaneous, unbounded one. Heart rate takes 30 to 90 seconds to climb to the level an effort demands, and it cannot exceed your maximum no matter how hard you push past it. Power registers a 700-watt surge the instant your legs produce it. On steady work the lag washes out and the two signals track each other closely. On short, sharp work the lag is the whole story, and a metric built on heart rate systematically under-counts the stress [García-Ramos et al. 2015]. This is the practical core of using Strava as a training tool rather than a journal: the data layer every cyclist already has only helps if you read each ride with the metric that fits it. The parent pillar gives the rule of thumb in one line. Below is the full per-workout-type version, because the right metric flips depending on what the session was built to do. ## A decision table by workout type The right metric flips with the session. Either works for steady endurance; power edges sweet-spot and threshold; power is the only honest read for anaerobic and VO2 work; and long or hot rides need a discount on Relative Effort for cardiovascular drift [García-Ramos et al. 2015, Coyle & González-Alonso 2001]. Steady zone 2 endurance: either metric is honest, and Relative Effort is genuinely useful because heart rate has hours to settle into the aerobic zone the ride lives in. Two 90-minute endurance rides at 165 and 180 watts post different Relative Efforts, and the gap reflects real internal strain [Meyer 2018]. The one caveat is duration: past about 90 minutes, cardiovascular drift lifts heart rate 5 to 10 percent at unchanged power [Coyle & González-Alonso 2001], so a long ride's Relative Effort runs high for a reason that is not extra training stress. Sweet-spot and tempo: both work, with a slight edge to power. These efforts sit just below threshold, sustained for 8 to 20 minutes, long enough that heart rate reaches the right zone, so Relative Effort tracks them sensibly. But sweet-spot is precisely the zone where small power differences matter most to adaptation, and heart rate is too blunt to distinguish 88 percent of FTP from 94 percent. If you have a meter, read the watts [Allen et al. 2019]. Threshold intervals over 5 minutes: power, clearly, but Relative Effort is not useless. A 3-by-12-minute set holds heart rate near its ceiling long enough to capture most of the work. The problem is the recovery valleys: the classic TRIMP method that Relative Effort descends from does not separate work from recovery, and a modified version that does runs about 9 percent higher on interval sessions, with the gap widening as intensity climbs [García-Ramos et al. 2015]. Power-based TSS counts each interval at its true intensity and is not fooled by the rests. Anaerobic and VO2 max work: Relative Effort breaks here, and the only honest answer is power. Thirty-second on, thirty-second off repeated for 8 minutes produces enormous metabolic stress and almost no time in the top heart-rate zone, because heart rate cannot climb and fall fast enough to register the spikes. The TRIMP lineage under-counts interval stress structurally, not occasionally [García-Ramos et al. 2015]. If a session has efforts under two minutes or anything above threshold, trust TSS. Long or hot rides are the mirror case: read Relative Effort with a discount, since drift inflates it on a ride your legs found easy [Coyle & González-Alonso 2001]. ## Why you cannot compare a Relative Effort number to a TSS number Comparing the two numbers directly is meaningless. A Relative Effort of 140 and a TSS of 95 are no more comparable than 140 pounds and 95 kilograms, except there is no conversion at all [Meyer 2018, Allen et al. 2019]. The error compounds when Strava splices both into one Fitness curve. A rider sees a Relative Effort of 140 on Saturday's group ride and a TSS of 95 on Tuesday's trainer session and concludes Saturday was harder. That comparison is meaningless. The numbers are not on the same axis, because the two metrics are scaled to different physiology with no conversion between them [Meyer 2018, Allen et al. 2019]. The error compounds in Strava's Fitness curve. Strava feeds Relative Effort into the chart when a ride has no power and a TSS-equivalent when it does, then averages both into one exponentially-weighted line [Meyer 2018]. A month of HR-only outdoor rides followed by a month of indoor power rides produces a step that no change in fitness created. Each ride is scored honestly on its own ruler; the trend is the lie, because the trend splices two rulers together — the failure mode our companion piece on mixing power and heart rate in the Strava chart walks through in detail. This is also why chasing the bigger number is a trap. Relative Effort rewards long steady volume and under-rewards short intensity [García-Ramos et al. 2015], so a rider optimizing for the higher score drifts toward more tempo and less of the hard, brief work that drives top-end adaptation. The metric quietly nudges your training toward what it happens to measure well, which is the opposite of what a structured plan needs [Seiler 2010]. ## Pick one primary metric and hold it Accuracy on any single ride matters less than consistency of input across the season [Allen et al. 2019]. Pick the metric you can apply to most rides and feed your fitness trend only that one. If heart rate is primary, pair it with session-RPE to cover its anaerobic blind spot [Haddad et al. 2017]. Pick the metric you can apply to most of your rides and feed your fitness trend only that one. If you have power on more than half your rides, make TSS primary and estimate it for the heart-rate-only rides rather than letting Relative Effort into the same curve [Allen et al. 2019]. If you are mostly outdoors on heart rate, make Relative Effort primary, set your max HR from a real observed peak, and accept the known blind spot on anaerobic work [Meyer 2018]. If heart rate is your primary metric, pair it with a one-to-ten effort rating after every ride. Carl Foster's session-RPE method is validated across continuous and interval work alike and costs nothing [Haddad et al. 2017], and it catches exactly the sessions Relative Effort under-counts. The hard 40-minute interval session that posts a modest Relative Effort will post a 9 on the effort scale, and that disagreement is the signal. Two independent reads beat one read you already know is blind in places. ## Common questions **Can I convert Relative Effort to TSS or compare the two numbers?** Not interchangeably, no. TSS is power-based, scaled so an hour at threshold equals 100 [Allen et al. 2019]. Relative Effort is a heart-rate-zone score descended from Banister's TRIMP [Meyer 2018, Hellard et al. 2007]. They are scaled to different physiology with no conversion factor between them. Compare each metric only to its own past values, and never read a Relative Effort number as if it were a TSS number. **Which should I trust for VO2 max and anaerobic intervals?** Power-based TSS, every time. Anaerobic and VO2 efforts spike metabolic stress faster than heart rate can climb, so the heart-rate-zone score that Relative Effort uses misses most of the load. The classic TRIMP method it descends from averages work and recovery together and under-counts interval sessions, running about 9 percent low versus a method that separates them [García-Ramos et al. 2015]. If the session has efforts under two minutes or anything above threshold, read the watts. **Is Relative Effort accurate for easy endurance rides?** For steady aerobic work, yes, with one caveat. On a zone 2 ride heart rate has time to settle into the right zone, so Relative Effort reflects real internal strain and tracks power closely [Meyer 2018]. The caveat is duration: past about 90 minutes, cardiovascular drift lifts heart rate 5 to 10 percent at unchanged power [Coyle & González-Alonso 2001], so a long or hot ride's score runs high for reasons unrelated to training stress. **Why does my Relative Effort look higher than my TSS?** Because they are different units. A Relative Effort of 140 and a TSS of 95 are no more comparable than 140 pounds and 95 kilograms, except there is no conversion between them at all [Meyer 2018, Allen et al. 2019]. One counts weighted heartbeats, the other counts normalized watts against your FTP. Track each against its own history; the only meaningless comparison is one metric's number against the other's. **Which metric should I make primary for tracking fitness over a season?** Whichever you can apply consistently to most of your rides. Consistency of input beats accuracy on any single ride, because your fitness trend is an exponentially-weighted average that assumes every day is on the same scale [Allen et al. 2019]. If most rides have power, make TSS primary; if most are heart-rate-only outdoors, make Relative Effort primary and pair it with session-RPE on your hard days to cover its anaerobic blind spot [Haddad et al. 2017]. ## References 1. **Meyer 2018.** [Quantifying Effort through Heart Rate Data](https://medium.com/strava-engineering/quantifying-effort-through-heart-rate-data-e6a0e3dd6a52). Strava Engineering. 2. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 3. **Hellard et al. 2007.** [Assessing the limitations of the Banister model in monitoring training](https://pmc.ncbi.nlm.nih.gov/articles/PMC1974899/). Journal of Sports Sciences. 4. **García-Ramos et al. 2015.** [Training load quantification in elite swimmers using a modified version of the training impulse method](https://pubmed.ncbi.nlm.nih.gov/24942164/). European Journal of Sport Science. 5. **Haddad et al. 2017.** [Session-RPE Method for Training Load Monitoring: Validity, Ecological Usefulness, and Influencing Factors](https://pmc.ncbi.nlm.nih.gov/articles/PMC5673663/). Frontiers in Neuroscience. 6. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. 7. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. --- # How to set up Strava for training: the one-time configuration that makes every metric honest URL: https://www.adaptcycling.com/guides/set-up-strava-for-training-workflow Updated: 2026-06-01 Author: Jim Camut Setting up Strava for training is a four-step configuration you do once: enter a real FTP and a measured max heart rate instead of formula estimates, anchor your zones to those numbers, commit to one consistent sensor stack so your load data is comparable, and set the feed and privacy so they protect your training instead of steering it. Strava syncs your data faithfully but does not audit whether the inputs are right [Meyer 2018]. The thirty minutes you spend on these settings is what keeps every downstream number — time-in-zone, Relative Effort, the Fitness curve — describing your physiology rather than an arithmetic guess. ## Enter a real FTP and max HR, not a formula estimate The two anchor numbers Strava scales everything else to are your FTP and your max heart rate. Both default to estimates — max HR from the 220-age formula, which carries a standard error near 10 bpm and can miss an individual by 10-20 [Tanaka et al. 2001]. Replace both with observed values before you trust a single downstream metric. Functional Threshold Power is the highest power you can hold in a quasi-steady state for roughly an hour, the boundary Andrew Coggan defined between heavy and severe intensity [Allen et al. 2019]. It is the denominator of intensity factor and the reference your power zones scale to, so a wrong FTP miscalibrates every workout. You do not need a formal test to seed it honestly — your last 90 days of power files already contain a best 20-minute and best 60-minute effort to anchor a defensible estimate. The point here is not which test to run; it is to put a real number in the field rather than leaving a placeholder. Max heart rate is the other master input, and its default is statistically weak. Tanaka and colleagues analyzed 351 studies covering 18,712 subjects and found 220-age systematically underestimates max HR in older adults; even their improved 208 minus 0.7-times-age equation carries a standard error near 10 bpm [Tanaka et al. 2001]. For any one rider, a formula-derived max can sit 10-20 bpm from the true ceiling. Set it instead from the highest heart rate you have actually seen in a hard, warmed-up effort over the last few months — the top of a 5-minute climb or the final seconds of a maximal sprint. Enter both before your first structured week, not after. Every ride Strava processes is scored against whatever anchors are in place at upload time, and correcting them later generally does not rewrite scores already sitting on past activities [Meyer 2018]. A few minutes spent on two fields up front saves you a Fitness curve with a seam in it three months from now. ## Set your zones so every downstream metric is honest Once the anchors are real, decide how your zones derive from them. Strava builds heart-rate zones as percentages of max HR by default; the more individualized option anchors the threshold zones to lactate threshold heart rate, which Joe Friel's framework treats as the better reference because it tracks a metabolic event rather than an age-based ceiling [Friel 2018]. Power zones scale off FTP and heart-rate zones scale off the max or threshold anchor, so the zones are only as honest as the numbers from the previous step. If your max HR is 15 bpm low, an endurance effort that should read mid-zone-2 logs as low zone 3, and your time-in-zone splits misreport where you actually trained. That distortion then propagates: Relative Effort sums weighted time-in-zone with higher zones counting for far more per minute [Meyer 2018], so a shifted boundary mechanically inflates or deflates the score, and that score feeds the Fitness curve when you ride without power. We walk through that full cascade in our companion piece on what happens when your Strava heart-rate zones are misconfigured. For the configuration itself, two defensible paths exist. The simpler one: correct the max HR in settings and let Strava regenerate its percentage-based zones — most of the accuracy gain for a few minutes of work. The more precise one, if you have done a threshold field test, is to enter custom boundaries from your lactate threshold heart rate using Strava's draggable zone editor [Friel 2018]. Either beats the default. Do not blend them across the season — pick one zone model and hold it, so a ride from March and a ride from August are measured against the same ruler. Heart-rate zones are not a substitute for what power measures, either. Heart rate lags intensity and drifts with heat, fatigue, and caffeine; power records mechanical work directly. If you have a meter, treat power zones as your primary reference and heart-rate zones as the backup for outdoor rides — but configure both, because Strava scores whatever stream a ride carries. ## Commit to one sensor stack so your load data is comparable Load data is only comparable if it comes from the same instruments. Two validated power meters can disagree by a couple of percent in steady work and far more in sprints [Nimmerichter et al. 2017], so switching stacks mid-season introduces an offset that can rival a real fitness change. Pick one primary stack and hold it. Nimmerichter and colleagues compared a Garmin Vector against the SRM laboratory standard and found no significant difference in steady power output, with both showing good reliability at a coefficient of variation under 3% — but 1-second peak power differed by limits of agreement near 19 W, and sprint agreement was meaningfully lower [Nimmerichter et al. 2017]. The lesson for a self-coached rider is not that one meter is wrong; it is that even two accurate devices are not perfectly interchangeable, and the discrepancy lands hardest on the short, sharp efforts. The practical failure mode is mixing data types, not just devices. If you train indoors on a smart trainer's power and outdoors on heart rate only, your Strava load looks like two different athletes — the chart shows a discontinuity that reflects the sensor switch, not your body. Decide on one primary load metric: power-based TSS-equivalent if you have a meter on every ride, or Relative Effort scored against correct HR zones if you do not. Hold it across the block so the Fitness curve measures training, not instrumentation. Reading both streams correctly is exactly the gap between syncing data and using it, the broader case our pillar on training with Strava keeps returning to. Strava records faithfully but does not reconcile a power ride against a heart-rate ride. AdaptCycling reads the same two streams from your rides — through Wahoo, Hammerhead, Polar or intervals.icu, since Strava seats are capped — and computes an internal load estimate that does not whipsaw when the sensor stack changes, so one indoor-power block followed by an outdoor-HR block reads as continuous training rather than a fictional fitness jump. ## Set the feed and privacy so they protect your training The last configuration step is social, and it is load-bearing. Strava's feed uses the same variable-reward design that makes social apps compulsive [Strava 2024], which steers self-coached riders toward matching the feed instead of their plan — quietly inverting the roughly 80/20 easy-hard balance endurance training is built on [Seiler 2010]. The mechanism is behavioral, not physiological. A 90-minute zone 2 ride at 165 watts is correct training that earns fewer kudos than a 230-watt group ride, so riders who structure around the feed drift toward one intensity: hard enough to look like work, never easy enough to count as recovery. Stephen Seiler's intensity-distribution research is clear that well-trained endurance athletes spend roughly 80% of training time easy [Seiler 2010]; matching the feed inverts that ratio one ride at a time. Configure against it once. Strava's privacy controls let you set activity visibility — followers-only or no-feed-post — without losing any data on your own dashboard. If seeing the kudos count on a slow ride pushes you to ride harder than the plan calls for, hide those rides from the feed by default. The data stays; only the social proof goes. Some riders disable the activity feed in the app entirely and report the behavioral pull drops within about two weeks. ## Common questions **What should I configure in Strava before I start training with it?** Four things, once: enter a real FTP and a measured max heart rate instead of the defaults, anchor your power and heart-rate zones to those numbers, commit to one sensor stack so your load data stays comparable, and set the feed and privacy so they do not steer your intensity. The 220-age max HR default alone can miss you by 10-20 bpm [Tanaka et al. 2001], and every downstream metric scales off these inputs. **Do I have to take an FTP test to set up Strava properly?** No. You need a real FTP number in the field, but your last 90 days of power files already contain the best efforts to estimate one without a forced test. FTP is the highest power you can hold for roughly an hour [Allen et al. 2019] and it scales your power zones and intensity factor, so the priority is putting a defensible number there rather than leaving a placeholder. **Why does Strava ask me to pick max HR or lactate threshold for zones?** Those are the two ways to anchor heart-rate zones. Max HR percentages are Strava's default; lactate threshold heart rate is more individualized because it tracks a metabolic event rather than an age-based ceiling [Friel 2018]. If you have done a threshold field test, enter custom boundaries from your LTHR; if not, correcting the max HR to a real observed peak captures most of the accuracy gain. **Does it matter if I switch power meters or use HR sometimes and power other times?** Yes. Two validated power meters can differ by a couple of percent in steady work and by limits of agreement near 19 W in sprints [Nimmerichter et al. 2017], and mixing power rides with HR-only rides produces a chart discontinuity that looks like a fitness change but is not. Pick one primary load metric and hold it across the block. **Should I make my training rides private?** If the feed pushes you to ride harder than your plan calls for, yes. Strava's social design uses variable rewards that steer riders toward matching the feed [Strava 2024], which inverts the roughly 80/20 easy-hard balance [Seiler 2010]. Privacy controls hide rides from followers without losing any data on your own dashboard; the behavioral pull typically eases within about two weeks. ## References 1. **Meyer 2018.** [Quantifying Effort through Heart Rate Data](https://medium.com/strava-engineering/quantifying-effort-through-heart-rate-data-e6a0e3dd6a52). Strava Engineering. 2. **Tanaka et al. 2001.** [Age-predicted maximal heart rate revisited](https://pubmed.ncbi.nlm.nih.gov/11153730/). Journal of the American College of Cardiology. 3. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 4. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. 5. **Nimmerichter et al. 2017.** [Validity and Reliability of the Garmin Vector Power Meter in Laboratory and Field Cycling](https://pubmed.ncbi.nlm.nih.gov/28460405/). International Journal of Sports Medicine. 6. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 7. **Strava 2024.** [Strava Year in Sport Trend Report 2024](https://press.strava.com/articles/strava-releases-annual-year-in-sport-trend). Strava. --- # Strava segments as fitness tests: which efforts map to which test, and how to make them comparable URL: https://www.adaptcycling.com/guides/strava-segments-as-fitness-tests Updated: 2026-06-01 Author: Jim Camut A Strava segment is a fixed stretch of road timed every time you ride it, which makes it a usable test surface — if you treat three or four of them as scheduled benchmarks instead of KOM hunts. The trick is matching segment profiles to the test: a 4-to-6-minute climb for VO2-range work, a 12-to-20-minute climb for threshold, a flat 8-minute drag for sweet-spot. Run them every 4-6 weeks, hold the variables that move the result constant, and the segment becomes a repeatable measurement [Allen et al. 2019, Leo et al. 2021]. This is the deeper how-to behind the pillar's rule that segments belong in your plan as tests, not as a daily contest. ## Which segment profile maps to which test Match the segment to the energy system you want to read. A 4-to-6-minute climb tracks VO2-range power, a 12-to-20-minute climb estimates threshold, and a flat 6-to-10-minute drag benchmarks sweet-spot. Each duration samples a different point on the power-duration curve, so one segment can't stand in for another [Leo et al. 2021]. Power profiling rests on the power-duration relationship: the best power you can hold falls predictably as the effort gets longer, and sampling that curve at a few durations characterizes a rider [Leo et al. 2021]. A 4-to-6-minute maximal effort tracks power near VO2 max; a 12-to-20-minute effort sits near the threshold between heavy and severe and is what the 20-minute FTP protocol approximates [Allen et al. 2019]. A rider who is strong over 5 minutes and weak over 20 has a profile you only see by testing both. Pick segments whose terrain enforces the duration rather than inviting you to surge. A steady, uninterrupted climb is the best test surface because gradient holds your power honest — you can't coast a climb the way you can soft-pedal a flat. For the threshold test, use a 12-to-20-minute climb with no descents or junctions mid-segment. For the VO2-range test, a 4-to-6-minute climb steep enough that you can't sprint it away in the first 30 seconds. For sweet-spot, a flat or shallow 6-to-10-minute drag out of the wind. Three or four segments is the right number, and the pillar's rule is to keep that set small and fixed. One short climb, one long climb, and one flat effort cover the curve points an amateur trains against. Adding a fourth — a 60-to-90-second segment for anaerobic capacity — is reasonable if you race criteriums. Beyond that you are collecting leaderboards, not data, and the variable-reward pull the pillar warns about creeps back. ## Scheduling segment tests every 4-6 weeks Run the test set every 4-6 weeks, at the seam between training blocks rather than mid-block, and treat the test day as the hard session it is — not an effort bolted onto an endurance ride. Testing more often than every four weeks mostly measures freshness; less often than six and you miss the adaptation you trained for [Friel 2018]. A 4-to-6-week interval matches the rhythm of a training block. Fitness accrues on a roughly six-week timescale — the 42-day window behind the Fitness curve is built on that physiology — so a test spaced shorter than four weeks captures noise more than signal [Friel 2018]. Schedule it at the end of a block, after a couple of easier days, so the number reflects adaptation rather than fatigue. Most periodized plans put a lighter week roughly every fourth week; the back end of that week is the natural test slot. Treat the test as your quality work for that day, not an addition to it. A common mistake is doing a planned VO2 session and then hitting the test segment on the way home, which means the test runs on pre-fatigued legs and reads low. The test is the session: warm up, hit the segments, then ride home easy. This also keeps the test from poisoning your easy days — the failure mode the pillar describes, where one all-out segment effort turns a zone 2 endurance ride into a tempo ride and inverts the roughly 80/20 easy-hard balance endurance training depends on [Seiler 2010]. Run the segments in a fixed order with real recovery between them. If you test all three in one ride, do the shortest and most demanding first — the VO2-range climb while you are freshest — then the threshold effort, then sweet-spot, with enough easy spinning between that each starts recovered. Or split them across two days in the same test week. What you must not do is compare a fresh-legs result from one cycle to a tired-legs result from the next. ## Controlling the variables so results compare A segment time only means something if the conditions around it are roughly constant. Wind, fatigue, fueling, temperature, and even tire pressure move a result by more than real fitness changes over six weeks. Outdoor power output is noisier than indoor — within-rider variability runs about twice as high outside [Jeffries et al. 2019]. Wind is the biggest uncontrolled variable on any non-climb segment, which is the main argument for testing on climbs. On a flat segment a 15 km/h tailwind can flatter your time by more than a full training block of real improvement. If you must use a flat segment, test only on near-calm days and note the wind. On a steep climb, gravity dominates aerodynamics and a moderate wind barely moves the result — another reason climbs are the better test surface. Standardize what you bring to the test. Outdoor cycling produces far more power-output variability than indoor riding — one study measured within-rider standard deviation of about 69 watts outdoors against 33 watts indoors [Jeffries et al. 2019], so the conditions you can control deserve discipline. Test at a similar time of day, after similar sleep, fueled the same way, with the same warm-up and tire pressure. Heat is a hidden confounder: cardiovascular drift pushes heart rate up 5-10% over the first 60-90 minutes of moderate work in warm conditions [Coyle and Gonzalez-Alonso 2001], so a hot-day threshold test shows a higher heart rate at the same power — read the power, not the heart rate, when comparing. Use power as the comparison metric whenever you have it. A segment's elapsed time bundles fitness, conditions, drafting, and equipment into one number; the power you held isolates the engine. Compare the 5-minute and 20-minute power from each test cycle, not the leaderboard placing. If you only have heart rate, compare average heart rate at a fixed perceived effort, and accept it is coarser. This is the same lesson as our companion piece on reading the metrics that matter after every ride: the leaderboard is entertainment, the power number is the measurement. ## Reading the trend across a season One test is a data point; the trend across three or four tests is the signal. Expect non-linear progress — a jump after a base block, a plateau during a hard build, a bump after a taper. A single low result is usually conditions or fatigue, not lost fitness [Allen et al. 2019]. Read the sequence, not the single number. A 4-to-6-week test cadence gives you eight to ten data points across a season, enough to see a real trajectory through the noise. A 3% rise in 20-minute power across two test cycles is a meaningful gain; a 3% wobble between any two adjacent tests is within the day-to-day variability of an outdoor effort [Jeffries et al. 2019]. The honest read comes from the slope across several tests, the same way the Fitness curve is a multi-week tool, not a daily one. Different curve points move at different rates, which is the diagnostic payoff of testing more than one duration. A base block typically lifts your 20-minute power while leaving 5-minute power flat; a VO2 block does the reverse. If both stall for two cycles in a row while training continues, that is a signal worth acting on — usually accumulated fatigue or a stale plan rather than a true ceiling [Friel 2018]. Watching which segment moves tells you what training is changing. This is the broader point of using Strava as a training tool rather than a journal: the segments you have ridden for years can become a free, repeatable test battery without any extra subscription. Strava stores the efforts but does not read them as a test series. AdaptCycling reads your full segment and power history, estimates your power curve from the efforts you have already done, and folds the trend into the plan — so the test result changes the next block instead of sitting on a leaderboard. ## Common questions **Which Strava segments work best as fitness tests?** Steady, uninterrupted climbs. A 4-to-6-minute climb tests VO2-range power and a 12-to-20-minute climb estimates threshold, because gradient holds your effort honest where a flat lets you coast or get a tailwind [Allen et al. 2019]. Add a flat 6-to-10-minute drag for sweet-spot only if you test it on near-calm days. Pick three or four and keep the set fixed across the season. **How often should I test fitness on a segment?** Every 4-6 weeks, at the end of a training block rather than mid-block. That spacing matches the roughly six-week timescale on which fitness adapts [Friel 2018], so the result reflects real change instead of noise. Testing weekly mostly measures how fresh your legs are that day, and the constant near-maximal efforts start poisoning your easy training [Seiler 2010]. **Why does my segment time vary so much when my fitness hasn't changed?** Conditions. Outdoor power output is intrinsically noisy — within-rider variability runs about twice as high outdoors as indoors [Jeffries et al. 2019] — and wind, heat, fatigue, and fueling each move a result by more than six weeks of real improvement. Compare power rather than elapsed time when you can, test in similar conditions, and read the trend across several tests, not any single one. **Should I compare segment time or power across tests?** Power, whenever you have a meter. Elapsed time bundles fitness, wind, drafting, and equipment into one figure, while the power you held over the segment isolates the engine [Allen et al. 2019]. Compare the 5-minute and 20-minute power from each cycle. Heart rate is a weaker proxy because it drifts with heat and fatigue [Coyle and Gonzalez-Alonso 2001]. **Do I need Strava Premium to use segments as tests?** No. Free Strava records and times every segment effort and stores your history, which is all the test battery needs. Premium adds leaderboard filtering and the Fitness chart, but the test itself is your own power or time on a fixed stretch of road compared cycle to cycle — that runs on the free tier plus a meter or a heart-rate strap. ## References 1. **Leo et al. 2021.** [Power profiling and the power-duration relationship in cycling: a narrative review](https://pmc.ncbi.nlm.nih.gov/articles/PMC8783871/). European Journal of Applied Physiology. 2. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 3. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 4. **Jeffries et al. 2019.** [An Analysis of Variability in Power Output During Indoor and Outdoor Cycling Time Trials](https://pubmed.ncbi.nlm.nih.gov/30859858/). International Journal of Sports Physiology and Performance. 5. **Coyle and Gonzalez-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. 6. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. --- # What to look at on Strava after a ride: the four metrics that matter and the ones to ignore URL: https://www.adaptcycling.com/guides/what-metrics-to-check-after-every-ride Updated: 2026-06-01 Author: Jim Camut After most rides you need about thirty seconds and four numbers: time-in-zone (did the ride do its job), normalized power or weighted average heart rate (how hard the work actually was), a one-to-ten perceived-effort rating (the cheapest valid load metric there is [Haddad et al. 2017]), and your Fitness trend over the last four weeks. That is the whole post-ride routine. Everything else on the activity page — kudos, segment placings, the single-ride Relative Effort score — is either entertainment or a number you are about to misread. This is the hub of using Strava as a training tool, not a journal. ## The four metrics that matter after every ride Check four things and stop: time-in-zone against what the ride was supposed to be, normalized power or intensity factor for how hard it was, a one-to-ten perceived-effort note, and the Fitness trend over four weeks. The first three take seconds on the activity page; the fourth is a weekly glance, not a daily one. Time-in-zone answers the only question a workout really poses: did it do what it was for? A planned zone 2 endurance ride should show most of its minutes in zone 2; if half the ride landed in zone 3, it was a tempo ride wearing an endurance label. Open the activity, look at the power or heart-rate distribution, and confirm the shape matches the intent. One mis-zoned long ride a week is enough to invert the roughly 80/20 easy-hard balance that endurance training is built around [Seiler 2010]. Normalized power and intensity factor tell you how hard the work was, not just how long. Andrew Coggan defined normalized power to weight the surges a flat average hides, and intensity factor — normalized power divided by FTP — puts the ride on a single 0-to-1 scale you can compare across days [Allen et al. 2019]. An IF near 0.65 is a recovery or endurance ride; 0.85 and up is real quality. If you ride on heart rate only, weighted average HR plays the same role at lower fidelity. One number, glanced at, tells you whether today was easy or hard. Perceived effort is the third number and the one almost nobody records. Rate the ride one to ten before you close the app — Carl Foster's session-RPE multiplies that rating by duration to estimate internal load, and it validates against heart-rate-based TRIMP at correlations from roughly 0.49 to 0.97 across sports [Haddad et al. 2017]. It costs nothing, needs no sensor, and captures the day Strava can't see: the 220-watt ride on three hours of sleep that felt like 260. Jot it in the activity description if nothing else. The fourth metric is not a per-ride number at all. It is the Fitness trend over the last four weeks, the 42-day CTL curve [Allen et al. 2019]. A single ride moves it 1-3 points, so reading it daily is reading noise; read the slope every week or two. Whether your Fitness is climbing, flat, or sliding is a decision input. Whether today's single ride nudged it up half a point is not. ## What to ignore after a routine ride Skip the kudos count, the segment leaderboard, and the single-ride Relative Effort or Suffer score on any ordinary training day. None of the three tells you whether the ride did its job, and all three pull you toward riding for the feed instead of the plan [Strava 2024]. Kudos and the social feed are entertainment, not data. Strava's design uses the same variable-reward loop that makes social apps compulsive [Strava 2024], and the predictable cost is riding to look like you are working rather than to train correctly. A 90-minute zone 2 ride at 165 watts is correct training that earns fewer kudos than a 230-watt group ride; matching the feed is how riders quietly invert their intensity distribution [Seiler 2010]. Close the feed after a routine ride. Segment placings are the same trap with a leaderboard attached. Chasing a KOM on a planned recovery spin converts an easy day into a hard one, and a planned endurance ride with one all-out segment effort is no longer an endurance ride. The productive use is the opposite of after-every-ride checking: pick three or four segments and treat them as scheduled tests every four to six weeks [Allen et al. 2019], then ignore the rest. On a normal Tuesday, the leaderboard is not a post-ride metric. The single-ride Relative Effort or Suffer score is the subtlest distraction because it looks like load data. It is a heart-rate-zone score, honest for steady aerobic work and unreliable when your zones are mis-set or heart rate drifts late in a long ride [Meyer 2018]. Our companion piece on why Relative Effort runs high on easy rides walks through how one low max-HR setting inflates the number on genuinely easy days. Glancing at a single ride's score and concluding you overdid it is exactly the misread to avoid; trust the four metrics above instead. ## Why perceived effort is the cheapest valid metric A one-to-ten effort rating is the highest-value number in the post-ride routine because it captures internal load — what the work cost you — which no sensor on your bike records. Foster's session-RPE method is simple, free, and validates against objective load measures well enough to stand on its own [Haddad et al. 2017]. Power and heart rate measure external work and the body's cardiac response to it. Neither sees context: sleep, stress, heat, illness, the cumulative fatigue of the last ten days. Two rides at identical 200-watt averages can sit worlds apart in cost, and only the rider knows which. Foster developed session-RPE precisely to capture that internal load with one rating and a stopwatch [Foster 1998], and the validation literature backs it — correlations with heart-rate-based TRIMP run from about 0.49 to 0.97 depending on the sport [Haddad et al. 2017]. The practical payoff is a cross-check that does not depend on your sensors being configured correctly. When time-in-zone, intensity factor, and your RPE note all agree, the ride was what it looked like. When they disagree — an easy-feeling ride that scored hard, or a brutal session the numbers undersell — the disagreement itself is the signal, usually pointing at a mis-set zone or a sensor gap rather than a fitness change. RPE is the leg of that triangle that costs nothing and lies least. Recording it is the hard part, not the rating. Strava asks for perceived exertion on upload, but most riders never fill it in, so the cheapest valid metric in cycling sits unused on most accounts. Type a single digit into the activity description before you close the app. Thirty rides in, you have a log of internal load that the power file alone could never reconstruct. ## Building a 30-second post-ride habit The routine works because it is short and fixed: open the activity, confirm time-in-zone matched the plan, note intensity factor and a one-to-ten effort, and glance at the four-week Fitness trend only on the weekend. Reading the data this way — not logging it for the feed — is the difference between Strava as a training tool and Strava as a journal. Keep it to four checks in a fixed order so it becomes automatic. Did the zones match the intent. How hard was it (IF, or weighted HR). What did it feel like (one to ten). And, once a week, which way is Fitness trending. The first three are per-ride and take seconds; the fourth is a multi-week tool that punishes daily reading [Allen et al. 2019]. Anything not on this list — kudos, segments, single-ride scores — is optional entertainment, not part of the routine. This is the sub-question the broader case for training with Strava keeps raising: the data layer every cyclist already has only helps if you read it on purpose. Strava syncs your data; it does not read it. Intervals.icu is a free analytics layer that surfaces these same numbers cleanly, and TrainingPeaks does it for power-based athletes who pay for it. The habit, not the tool, is what converts a sync into a signal. AdaptCycling reads the full stream — time-in-zone, power, heart rate, and the context you log — and adapts the next sessions to what you actually rode, so the thirty-second check informs the plan instead of just decorating a feed. ## Common questions **What should I look at on Strava after a ride?** Four things: time-in-zone (did the ride match its intent), normalized power or intensity factor (how hard it was), a one-to-ten perceived-effort rating (the cheapest valid load metric [Haddad et al. 2017]), and your Fitness trend over the last four weeks. The first three take seconds; the fourth is a weekly glance. Skip kudos, segments, and the single-ride score on a routine ride. **Should I check my Fitness number after every ride?** No. Fitness is a 42-day average, so one ride moves it only 1-3 points [Allen et al. 2019] — reading it daily is reading noise. Check the slope over the last two to four weeks, not the daily value. Whether the line is climbing, flat, or falling is a useful decision input; whether today's ride nudged it half a point is not. **Is the single-ride Relative Effort score worth checking?** Rarely. Relative Effort is a heart-rate-zone score that is honest for steady aerobic work but inflates when your max HR is set too low or heart rate drifts on a long ride [Meyer 2018]. A single ride's score will talk you into misreads. Trust time-in-zone, intensity factor, and your effort rating instead, and use the Fitness trend for the multi-week picture. **Why bother rating perceived effort if I already have power data?** Because power measures external work and misses context — sleep, stress, heat, accumulated fatigue. Session-RPE captures that internal load with one rating and validates against objective measures at correlations from about 0.49 to 0.97 [Haddad et al. 2017]. It is free, needs no sensor, and gives you a cross-check that does not depend on your zones being configured correctly [Foster 1998]. **Do I need Strava Premium to run this post-ride routine?** Mostly no. Free Strava shows time-in-zone, normalized power, and the activity stream; you can record perceived effort in the description for nothing. The one Premium-only piece is the Fitness/Freshness chart, and a free Intervals.icu account reads your Strava history to produce an equivalent trend line. The four-metric habit runs on free tools plus a number you type yourself. ## References 1. **Haddad et al. 2017.** [Session-RPE Method for Training Load Monitoring: Validity, Ecological Usefulness, and Influencing Factors](https://pmc.ncbi.nlm.nih.gov/articles/PMC5673663/). Frontiers in Neuroscience. 2. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 3. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 4. **Meyer 2018.** [Quantifying Effort through Heart Rate Data](https://medium.com/strava-engineering/quantifying-effort-through-heart-rate-data-e6a0e3dd6a52). Strava Engineering. 5. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 6. **Strava 2024.** [Strava Year in Sport Trend Report 2024](https://press.strava.com/articles/strava-releases-annual-year-in-sport-trend). Strava. --- # Strava Fitness going down while you're still training hard: the 42-day decay math URL: https://www.adaptcycling.com/guides/strava-fitness-dropping-still-training Updated: 2026-06-01 Author: Jim Camut If you're riding consistently and your Strava Fitness line is sliding down, the chart usually isn't broken — it's doing exactly what a 42-day exponentially-weighted average does. Fitness is CTL, a decaying average of daily training load [Allen et al. 2019]. It falls whenever your recent load drops below your 42-day baseline: lower volume even at higher intensity, a couple of missed rides, a sensor or input gap that under-scores your work, or a deliberate taper. Sometimes the drop is wrong. More often it's correct, and a few cases mean you're peaking. Here's how to tell which. ## How Strava Fitness decays: the 42-day EWMA math Fitness is a 42-day exponentially-weighted moving average of daily training load — CTL in Coggan and Allen's terms [Allen et al. 2019]. Each day it blends in today's load and lets the rest decay by a fixed fraction. The headline number that explains a falling line: with zero training, you lose about 15% of your Fitness in a week and roughly half in 29 days. The mechanism is one line of arithmetic. A 42-day EWMA carries forward about 41/42 of yesterday's value and adds 1/42 of today's load [Allen et al. 2019]. The window length traces to Eric Banister's 1970s fitness-fatigue model, whose decay constants were curve-fit against athlete performance and have held up across decades of validation [Hellard et al. 2007]. Because the average is exponential, recent rides count more than old ones, and the line is always drifting toward whatever your last few weeks of load actually were. Run the decay forward and the paradox dissolves. Ride nothing and Fitness multiplies by (41/42) each day: about 0.85 after a week, so roughly 15% gone in seven days, and about 0.36 after 42 days — you keep barely a third. The half-life is near 29 days. The point isn't that you should panic; it's that the line falls fast relative to how it climbed. CTL rises slowly because you can only add a little load per day, but it sheds quickly the moment your input drops below the running average. That asymmetry is the whole story. To hold a Fitness number steady, today's load has to roughly match your 42-day average load — not your peak week, the average. A rider who built CTL on 10 hours a week and drops to 6 will watch Fitness fall even if every one of those 6 hours is harder than before, because total daily load fell below the baseline the average is still anchored to. The chart is reporting load, not effort, and the two are not the same. ## The real reasons your Fitness is dropping Four causes account for almost every falling Fitness line in a rider who feels like they're working: volume dropped even as intensity rose, a few rides went missing, a sensor or input gap under-scored real work, or you're tapering on purpose. The first three are fixable misreads; the fourth is the chart working correctly. Cause one, and the most common: volume fell while intensity climbed. Training load is roughly intensity multiplied by duration, so a 45-minute threshold session can score less total load than a 2-hour endurance ride even though it hurt far more [Allen et al. 2019]. Trade three long base rides for three short hard ones and your weekly load drops, dragging the 42-day average down with it. The legs feel trashed; the Fitness line feels insulting. Both are telling the truth about different things. Cause two: missed or shortened rides you've half-forgotten. A skipped weekend long ride is the single biggest one-day hit to CTL most riders have, because it was the largest load input in the week. Two or three of those across a month — travel, weather, a sick kid — quietly reset the baseline the average chases. Cause three is a data problem, not a fitness one: if you train indoors with power and outdoors on heart rate only, or your max HR is mis-set, the load Strava records can badly under-count what you actually did. Our companion piece on mixing power and heart rate in the Strava chart walks through how that sensor switch puts a discontinuity in the curve that looks like lost fitness but isn't. Cause four is deliberate: a taper. Cut volume before a goal event and Fitness will fall by design while freshness rises — that's the trade you want. Bosquet and colleagues' meta-analysis found that reducing training volume by roughly 41 to 60% over about two weeks maximizes performance gains [Bosquet et al. 2007], and a drop that size visibly lowers the Fitness line. Friel's framework treats the same dip as the cost of arriving fresh [Friel 2018]. If your Fitness is falling in the last two weeks before an event, that is the plan working, not failing. ## When a falling Fitness line is exactly right A dropping Fitness line is correct, not alarming, in three situations: a planned taper into a goal event, a recovery week after a hard block, and the first days of any genuine rest you actually needed. In all three, the freshness you gain is the reason the fitness number gives a little back. Form is fitness minus freshness, and you cannot raise one without spending some of the other. During a taper, cutting volume drops Fitness by a few points but lets fatigue clear so Form climbs into the positive range research and Friel associate with peak performance [Friel 2018]. Losing 3 to 5 CTL points over a two-week taper while Form swings from negative to +10 or +20 is not detraining — it's the mechanism by which you show up rested. Reading the Fitness drop as a problem is how riders sabotage their own peak. A recovery week works the same way at smaller scale. Pull volume back for seven days after a long build and Fitness slips a few points; the adaptation you banked in the preceding weeks is still there. Short-term reductions in training don't erase fitness so much as pause its accumulation — Mujika and Padilla's detraining review defines short-term as under four weeks and notes the early losses are dominated by blood-volume and cardiovascular factors that rebound quickly once you train again [Mujika & Padilla 2000]. A one-week dip in the line is noise against that backdrop. The trap is conflating a correct dip with a failure and adding load to fight it. If the chart falls during a planned recovery week and you respond by riding hard to push it back up, you've converted a recovery week into a monotony problem and undone its purpose. The Fitness line going down is not a command to train more. In a taper or a recovery week, it's confirmation the plan is doing what it's supposed to. ## Trend versus noise: how long a drop has to last to mean something A single ride moves Fitness by 1 to 3 points; daily wiggles are noise. A genuine downward trend needs roughly two to four weeks of load consistently below your 42-day baseline before it means your fitness is actually receding rather than just settling. Read the slope over weeks, never the daily number. Because Fitness is a 42-day average, one big ride barely moves it and one missed ride barely dents it — typically 1 to 3 points either way [Allen et al. 2019]. That smoothing is the feature: it filters the day-to-day chaos of real life so you can see the multi-week signal. The cost is lag. The line keeps falling for days after you've resumed normal training, because the average is still digesting the low-load stretch behind it. Judging the chart day by day is reading the lag as if it were news. The honest threshold is two to four weeks. A drop that persists that long, with your weekly load genuinely below the baseline the whole time, is a real downtrend worth acting on. A drop that reverses inside a week or two was a blip — a missed weekend, a sensor gap, a deliberate easy week. This sits inside the broader discipline of using Strava as a training tool rather than a journal: the data layer every cyclist already has only helps if you read its trend, not its daily mood. Strava syncs the number every few minutes; whether the trend means anything is a judgment the chart can't make for you. ## Common questions **Why is my Strava Fitness going down even though I'm training hard?** Fitness is a 42-day average of training load, which is intensity times duration — not intensity alone [Allen et al. 2019]. Hard short rides can score less total load than long easy ones, so trading volume for intensity drops the average even though the sessions feel brutal. The chart is reporting load, not how much a ride hurt. **How fast does Strava Fitness drop if I stop riding?** Roughly 15% in the first week and about half in 29 days, because the 42-day average multiplies by about 41/42 each day you add no load [Allen et al. 2019]. Fitness falls noticeably faster than it climbs. The early loss is dominated by blood-volume and cardiovascular factors that rebound quickly once you resume training [Mujika & Padilla 2000]. **Is it normal for Fitness to drop during a taper?** Yes — that's the taper working. Cutting volume by roughly 41 to 60% over about two weeks maximizes performance while lowering the Fitness line by a few points [Bosquet et al. 2007, Friel 2018]. You trade a small Fitness dip for the freshness that lets you perform. Don't add load to fight it. **How long does a Fitness drop have to last before I should worry?** About two to four weeks of load consistently below your 42-day baseline. A single ride moves Fitness only 1 to 3 points [Allen et al. 2019], so daily dips are noise. If the line keeps falling for weeks while your weekly load is genuinely low, that's a real downtrend; a drop that reverses inside a week or two was a blip. **Could my Fitness be dropping because of a data problem, not lost fitness?** It can. If you switch from outdoor heart rate to indoor power, or your max HR is mis-set, Strava can under-score real work and the Fitness line falls without any actual loss [Allen et al. 2019]. Check whether a sensor or zone change lines up with the drop before concluding your fitness receded. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Hellard et al. 2007.** [Assessing the limitations of the Banister model in monitoring training](https://pmc.ncbi.nlm.nih.gov/articles/PMC1974899/). Journal of Sports Sciences. 3. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. 4. **Mujika & Padilla 2000.** [Detraining: loss of training-induced physiological and performance adaptations. Part I: short term insufficient training stimulus](https://pubmed.ncbi.nlm.nih.gov/10966148/). Sports Medicine. 5. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. --- # Strava indoor power, outdoor heart rate: why your Fitness chart jumps when you mix load inputs URL: https://www.adaptcycling.com/guides/mixing-power-and-hr-strava-chart Updated: 2026-06-01 Author: Jim Camut If you ride indoors with a power meter and outdoors on heart rate only, your Strava Fitness/Freshness/Form chart will show a step that no physiology produced. The cause is not a bug — it is a units problem. Power-based TSS and heart-rate-based Relative Effort answer different questions and are scaled to different rulers [Allen et al. 2019, Meyer 2018], so feeding both into one exponentially-weighted curve splices two incompatible signals. The fix is not to abandon either sensor. It is to commit to one primary load metric and hold it across the season. ## Power-based load and HR-based load are not the same units TSS and Relative Effort measure different things on different scales. TSS is mechanical work normalized so one hour at FTP equals 100 [Allen et al. 2019]. Relative Effort is weighted time-in-heart-rate-zone, closer to Banister's TRIMP than to a power score [Meyer 2018, Hellard et al. 2007]. Putting both on one axis compares watts to heartbeats. TSS is built from external load. It takes your normalized power, scales it against your FTP, and multiplies by duration, with the calibration point that 60 minutes at threshold returns exactly 100 [Allen et al. 2019]. The number is anchored to mechanical output — what your legs did to the pedals — regardless of heart rate. Relative Effort is built from internal load. Strava slices your heart-rate stream into time-in-zone against your configured max HR, weights higher zones more heavily, and sums [Meyer 2018]. The math descends from Eric Banister's TRIMP, which applies a sex-specific exponential weighting to heart-rate reserve — roughly e^1.92 for men and e^1.67 for women [Hellard et al. 2007]. There is no point on that scale that corresponds to one hour at FTP. The two metrics were never calibrated to each other. This is why the same physiological effort can post different numbers depending on which sensor recorded it. A steady 90-minute aerobic ride might log a TSS in the 70s indoors with power and a Relative Effort in the 50s or 60s outdoors on heart rate, or the reverse if your max HR is set low. Neither is wrong. They are different rulers, and the gap between them is not a fitness signal — it is a measurement artifact from switching sensors. ## The discontinuity it creates in your Fitness/Freshness/Form chart Both Fitness and Freshness are exponentially-weighted moving averages of daily load — 42 days for Fitness, 7 for Freshness [Allen et al. 2019, Hellard et al. 2007]. When the daily load you feed them jumps by 20-30 units overnight because you switched from heart rate to power, the curve reads that step as a real change in training. It is not. The chart's mechanics make this unavoidable. Fitness updates each day as roughly 97.6% of yesterday's value plus 2.35% of today's load, the math of a 42-day exponential average; Freshness uses a 7-day constant, so today's load counts for about 13.3% [Allen et al. 2019]. The model assumes every daily input is on the same scale. It has no way to know that Monday's number came from a heart-rate ride and Tuesday's came from a power meter. Work a concrete case. Suppose your outdoor HR rides have been logging around 60 units of daily load, and then you move indoors for winter where the same effort logs around 85 TSS. That is a 25-unit step with no change in your actual training. Freshness, weighting each day at 13.3%, climbs fast — within a week it has absorbed most of the jump and your Form (Fitness minus Freshness) dips into territory that looks like accumulated fatigue. Fitness, at 2.35% per day, drifts upward more slowly but just as falsely. The chart now tells a story about a fatigue-then-fitness swing that your legs never lived. The reverse hurts more. A rider who trains hard indoors on power all winter and then moves outdoors to heart-rate-only rides in spring can watch Fitness sag for weeks, because Relative Effort on easy outdoor base rides logs lower daily load than the structured power sessions it replaced. The rider reads a falling Fitness line and panics, adds intensity, and inverts the 80/20 distribution that base season is supposed to protect [Seiler 2010]. The chart drove a bad training decision off a units change. This is the sub-question the broader case for treating Strava as a training tool, not a journal, keeps surfacing: the data layer every cyclist already has only helps if the inputs are commensurable. A Fitness curve spliced from two metrics is not lying about any single ride — each TSS and each Relative Effort is honest on its own scale. It is lying about the trend, which is the one thing the chart exists to show. ## How to commit to one primary load metric across the season Pick one load metric and feed the chart only that. If you have power on most rides, make power-based TSS primary and estimate it for the HR-only rides. If you are mostly heart rate, make Relative Effort primary and accept it under-reports anaerobic work. Consistency of input beats accuracy of any single ride [Allen et al. 2019]. Power-primary is the stronger default if you have a meter on more than half your rides. Power-based TSS is immune to the heart-rate problems that distort the other metric — it does not depend on a max HR that the 220-age formula can miss by 10-20 bpm [Tanaka et al. 2001]. For the HR-only rides, estimate a TSS from duration and perceived intensity rather than letting Relative Effort feed the same curve. An honest estimate on the same ruler beats an exact number on a different one. Heart-rate-primary is fine if power is the exception, not the rule. Many self-coached riders ride outdoors most of the year and only see power on a borrowed trainer. If that is you, make Relative Effort your season-long metric, set your max HR from a real observed peak rather than a formula, and accept the known blind spot: Relative Effort under-reports short anaerobic intervals because heart rate lags the metabolic stress [Meyer 2018]. Foster's session-RPE is a cheap, validated companion for exactly those sessions [Haddad et al. 2017]. What you must not do is let Strava average the two automatically. Strava uses Relative Effort as its load input when there is no power and a TSS-equivalent when there is, and it will happily build one Fitness curve across both [Meyer 2018]. That is the splice. The discipline is boring but it works: one metric, one ruler, all season, so the trend reflects your training rather than your sensor stack. ## How tools that read both streams resolve it Some platforms read both the power and heart-rate streams off every activity and compute one internal load estimate that does not whipsaw when the sensor changes. That is the difference between an app that displays your rides and one that reads them [Strava Developers]. Several tools do a version of this; the honest gap is between syncing and reading. Intervals.icu reads your full Strava history for free and lets you choose how load is computed, which lets a disciplined rider hold one metric without paying. TrainerRoad sidesteps the problem by being indoors-only with power, so there is never a second ruler. TrainingPeaks insists on power-derived TSS and treats HR-only rides as estimates. Each is a defensible answer to the same problem: a fitness curve is only meaningful when its inputs share a scale. Reading both streams means deriving a single training-stress estimate from whichever data each ride carried, rather than passing two different scores into one average. The infrastructure is webhook-driven — every new activity is parsed within seconds of upload [Strava Developers] — so the estimate comes from the raw power and heart-rate data, not Strava's display-layer numbers. The point is one consistent internal load, whichever sensor was on the bike that day. This is where AdaptCycling sits. We read both your power and heart-rate streams from the same rides — through Wahoo, Hammerhead, Polar or intervals.icu, since Strava seats are capped — and compute an internal training-stress estimate that stays on one scale whether you rode indoors on power or outdoors on heart rate, so a winter spent on the trainer does not splice a false step into your fitness trend. It is the same principle behind why a single mis-set max HR should not be allowed to rewrite a week of history — covered in our companion piece on a Relative Effort that reads high on easy rides. The data is the input the plan adapts to; keeping it commensurable is what makes the trend trustworthy. ## Common questions **Why does my Strava Fitness chart jump when I switch from outdoor HR to indoor power?** Because the two rides are scored on different scales. Outdoor rides without power log Relative Effort, a heart-rate-zone score [Meyer 2018]; indoor power rides log TSS, where one hour at FTP equals 100 [Allen et al. 2019]. Feeding both into one exponentially-weighted Fitness curve splices incompatible units, so the chart shows a step that your training never produced. **Should I trust TSS or Relative Effort for tracking fitness over a season?** Whichever one you can apply consistently to most of your rides. If you have power on more than half your rides, make TSS primary and estimate it for HR-only rides [Allen et al. 2019]. If you are mostly heart rate, make Relative Effort primary and set your max HR from a real observed peak [Tanaka et al. 2001]. Consistency of input matters more than the accuracy of any single ride. **How much does one mismatched ride actually move the chart?** More than you would think on Freshness, less on Fitness. Freshness is a 7-day exponential average, so a single day's load counts for about 13.3%; Fitness is a 42-day average, so it counts for about 2.35% [Allen et al. 2019]. A 25-unit units mismatch repeated over a week of switched sensors can swing your Form into apparent fatigue that is purely an artifact. **Can I make Strava use the same load metric for every ride?** Not fully — Strava uses Relative Effort when there is no power and a TSS-equivalent when there is, and builds one curve across both [Meyer 2018]. The workaround is to standardize the input yourself: estimate a TSS for HR-only rides if you are power-primary, or rely on Relative Effort plus session-RPE if you are heart-rate-primary [Haddad et al. 2017]. **Does estimating FTP differently fix the discontinuity?** No. The discontinuity is a units mismatch between power-based and heart-rate-based load, not an FTP error. A more accurate FTP makes your TSS values truer but does nothing to reconcile them with Relative Effort, because the two metrics are scaled to different physiology in the first place [Allen et al. 2019, Meyer 2018]. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Meyer 2018.** [Quantifying Effort through Heart Rate Data](https://medium.com/strava-engineering/quantifying-effort-through-heart-rate-data-e6a0e3dd6a52). Strava Engineering. 3. **Hellard et al. 2007.** [Assessing the limitations of the Banister model in monitoring training](https://pmc.ncbi.nlm.nih.gov/articles/PMC1974899/). Journal of Sports Sciences. 4. **Tanaka et al. 2001.** [Age-predicted maximal heart rate revisited](https://pubmed.ncbi.nlm.nih.gov/11153730/). Journal of the American College of Cardiology. 5. **Haddad et al. 2017.** [Session-RPE Method for Training Load Monitoring: Validity, Ecological Usefulness, and Influencing Factors](https://pmc.ncbi.nlm.nih.gov/articles/PMC5673663/). Frontiers in Neuroscience. 6. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 7. **Strava Developers.** [Webhook Events API](https://developers.strava.com/docs/webhooks/). Strava Developers. --- # Heart rate climbing while power stays flat on a long ride: cardiovascular drift, decoupling, and what to do URL: https://www.adaptcycling.com/guides/strava-cardiovascular-drift-long-rides Updated: 2026-06-01 Author: Jim Camut Three hours into a long ride your heart rate is 12 beats higher than it was at the start, but your power has not moved. The ride did not get harder and you did not lose fitness. This is cardiovascular drift, a well-characterized response to prolonged work: stroke volume falls after 10-20 minutes and heart rate rises to compensate [Coyle & González-Alonso 2001]. The useful part is that the size of the gap between power and heart rate — aerobic decoupling — is one of the better durability signals Strava's data can give you, if you know how to read it. ## What cardiovascular drift actually is Cardiovascular drift is the steady upward creep of heart rate during prolonged steady work, even when power output never changes. The mechanism is mechanical, not a loss of fitness: stroke volume — the blood pumped per beat — declines after the first 10-20 minutes, and heart rate rises to hold cardiac output constant [Coyle & González-Alonso 2001]. The classic picture blamed rising skin blood flow as the body shed heat, but Coyle and González-Alonso reframed it: the progressive decline in stroke volume after 10-20 minutes of exercise is driven primarily by the rising heart rate itself, not by blood pooling in the skin as core temperature climbs [Coyle & González-Alonso 2001]. The practical upshot is the same either way. To keep delivering the same oxygen at the same 200 watts, a heart pumping less per beat has to beat more often. Over a two-to-three-hour ride that compensation routinely shows up as a 5-10% rise in heart rate at unchanged power. Heat and dehydration amplify it sharply. Montain and Coyle had trained cyclists ride two hours at a fixed sub-threshold intensity in a 33C room while replacing different fractions of their sweat losses; the heart-rate rise and the stroke-volume decline both scaled almost perfectly with how dehydrated each rider got, with correlations of r=0.99 [Montain & Coyle 1992]. A rider who finishes a long summer ride 3% down on body weight will see far more drift than the same rider on a cool day with a full bottle every 45 minutes. The drift is real physiology responding to a real thermal and fluid load — it is not your power meter or your fitness lying to you. ## How to see drift in a Strava ride (heart rate up, power flat) On a Strava activity, open the analysis view and overlay the heart-rate and power streams. Drift looks like two lines diverging: power holds a flat band while heart rate ramps slowly upward across the back half. The cleaner the power line, the more obvious the divergence — and the more confidently you can call it drift rather than effort. Strava records both streams honestly and plots them on the same time axis, but it does not interpret the gap between them — which is the recurring theme of the broader case for using Strava as a training tool rather than a journal: it syncs your data faithfully but does not read it. To see drift, scrub to a stretch where the terrain and your effort were steady, then compare the first hour to the last. If power is sitting at a flat 190 watts in both windows while average heart rate has climbed from 138 to 150, that 12-beat gap at identical power is drift, not a harder ride. The trap is heart-rate-only riders reading the climbing line as effort. Strava's Relative Effort is built from weighted time-in-heart-rate-zone [Meyer 2018], so drift mechanically pushes late-ride minutes into higher zones and inflates the score on a ride that felt easy throughout — the exact pattern our companion piece on why Relative Effort runs high on easy rides walks through. If you only have heart rate, lean on perceived exertion as the tiebreaker; Foster's session-RPE is the cheapest valid internal-load check in cycling [Haddad et al. 2017], and a ride that drifted but felt like a 4 out of 10 was a 4. If you have a power meter, the read is unambiguous. Normalized power flat across the ride plus heart rate trending up is the textbook signature [Allen et al. 2019]. The presence of power data is what lets you separate drift from a genuine intensity increase, because power answers the question heart rate cannot: did the actual work rate change, or only the cost of producing it? ## Aerobic decoupling as a durability test, not a warning Quantify the gap and it becomes a fitness signal. Aerobic decoupling — Pw:HR — compares the power-to-heart-rate ratio in the first half of a steady ride to the second half. A long ride that stays coupled (typically under about 5% drift) indicates good aerobic durability at that duration; a ride that decouples badly flags a ceiling you have not yet trained past. The logic runs straight out of the drift mechanism. A well-trained aerobic engine maintains stroke volume longer, so heart rate rises less for the same watts and the power-to-heart-rate ratio stays stable across the ride. A 2025 analysis using machine learning to track training response in cyclists found exactly this — that attenuated cardiovascular drift, a smaller heart-rate rise for a given power output, is interpreted as a marker of robust aerobic endurance, and that drift shrinking over a block corresponds to genuine cardiovascular adaptation [Barsumyan et al. 2025]. Decoupling is drift, measured and turned into a number you can track. Used as a test, it answers a question FTP cannot: not how much power you can hold for an hour, but how long you can hold a given aerobic power before the wheels start coming off. A rider whose two-hour endurance rides decouple at 3% has durability a rider who decouples at 12% at the same power does not, even if their FTPs match. For anyone training toward a long event — a gran fondo, a four-hour road race, a gravel day — that durability is the thing the event actually tests. TrainingPeaks surfaces Pw:HR and an Efficiency Factor for this directly; Intervals.icu computes decoupling on rides for free. The honest caveat: a single decoupled ride is weak evidence on its own, because heat, a missed bottle, a poor night's sleep, or simply going out too hard all inflate it. Read decoupling the way you read any drift number — as a trend across several comparable rides in similar conditions, not a one-ride verdict. Drift trending down across a base block is the signal worth chasing; one ugly number on a hot, under-fueled afternoon is mostly noise. ## What to do about it (usually nothing) On a normal long ride, the correct response to drift is nothing. Do not log the ride as harder than it was, do not skip tomorrow's quality session, and do not add recovery you do not need. The two things drift should change: how you fuel and hydrate long rides, and whether you start treating decoupling as a durability metric to track. The training decision is the whole point. A drifted heart rate is not evidence of fatigue or lost fitness, so it should not reclassify a zone 2 endurance ride as a hard day or talk you out of a planned interval session. Trust the inputs that do not depend on heart rate — power-based load if you have it, perceived exertion if you do not [Allen et al. 2019, Haddad et al. 2017]. A ride long enough to drift is doing the aerobic work a base block is built around, and the only mistake is reading the climbing heart-rate line as a problem to solve. Where drift should change behavior is fuelling and heat strategy. Because dehydration scales the drift almost linearly [Montain & Coyle 1992], a rider who consistently sees large drift on long rides has a concrete, actionable fix: drink more and earlier, manage heat, and start long rides topped up. That is the rare case where a high drift number points at something you can act on — not your training plan, but your bottles. ## Common questions **Why does my heart rate climb on a long ride when my power stays the same?** Cardiovascular drift. Over a prolonged ride your stroke volume — blood pumped per beat — declines after the first 10-20 minutes, so heart rate rises to hold the same output [Coyle & González-Alonso 2001]. At constant power, expect a 5-10% heart-rate rise across a couple of hours, more in heat or when dehydrated. It means the cost of producing the watts went up, not that the ride got harder or you lost fitness. **Is cardiovascular drift a sign I'm losing fitness or overtraining?** No. Drift is normal physiology on any long or warm steady ride, and the work rate that drives adaptation never changed. A ride long enough to drift is doing exactly the aerobic work a base block wants. Do not add recovery or skip your next session based on a drifted heart rate; judge fatigue from power-based load, perceived effort, and the multi-week trend instead. **What is aerobic decoupling and what is a good Pw:HR number?** Decoupling, or Pw:HR, compares the power-to-heart-rate ratio in the first half of a steady ride to the second half — it is drift turned into a number. Under roughly 5% over a long aerobic ride indicates good durability at that duration; well-trained riders show attenuated drift for a given power [Barsumyan et al. 2025]. Read it as a trend across similar rides, not a one-ride verdict, since heat and poor fueling inflate it. **How do I see cardiovascular drift in a Strava ride?** Open the activity analysis and overlay the heart-rate and power streams on the same time axis. Find a stretch of steady terrain and effort, then compare the first hour to the last: flat power with heart rate trending upward is drift. With power, normalized power flat plus heart rate rising is the textbook signature [Allen et al. 2019]. Strava plots both lines but does not interpret the gap. **Does cardiovascular drift inflate my Relative Effort or Fitness on long rides?** Yes. Relative Effort sums weighted time-in-heart-rate-zone [Meyer 2018], so drift pushes late-ride minutes into higher zones and lifts the score on a ride that felt easy throughout. The fix is to trust power-based load and perceived exertion over a heart-rate score on long rides [Haddad et al. 2017], and to expect drift rather than read it as a harder session. ## References 1. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. 2. **Montain & Coyle 1992.** [Influence of graded dehydration on hyperthermia and cardiovascular drift during exercise](https://pubmed.ncbi.nlm.nih.gov/1447078/). Journal of Applied Physiology. 3. **Barsumyan et al. 2025.** [Quantifying training response in cycling based on cardiovascular drift using machine learning](https://pmc.ncbi.nlm.nih.gov/articles/PMC12271085/). Frontiers in Artificial Intelligence. 4. **Meyer 2018.** [Quantifying Effort through Heart Rate Data](https://medium.com/strava-engineering/quantifying-effort-through-heart-rate-data-e6a0e3dd6a52). Strava Engineering. 5. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 6. **Haddad et al. 2017.** [Session-RPE Method for Training Load Monitoring: Validity, Ecological Usefulness, and Influencing Factors](https://pmc.ncbi.nlm.nih.gov/articles/PMC5673663/). Frontiers in Neuroscience. --- # Strava heart rate zones wrong: how one bad max HR cascades through your whole dashboard URL: https://www.adaptcycling.com/guides/misconfigured-strava-hr-zones Updated: 2026-06-01 Author: Jim Camut If your Strava heart-rate zones are wrong, the damage doesn't stop at the zone chart. Strava derives your zones from a single max heart rate, defaulting to the 220-age formula that carries a standard error near 10 bpm [Tanaka et al. 2001]. Get that one number wrong and the error flows downstream into your time-in-zone splits, your Relative Effort score [Meyer 2018], and the Fitness curve Strava builds on top of them. This is the broader version of a config problem self-coached riders hit constantly: the dashboard looks authoritative while every number on it is scaled to a ruler that doesn't fit you. ## How Strava sets your heart-rate zones Strava derives all five heart-rate zones from one anchor: your maximum heart rate. By default it estimates that max as 220 minus your age, or 190 bpm if no age is on file. Every zone boundary is then a fixed percentage of that single number, so the anchor decides where every zone line falls. There are two common ways to build heart-rate zones, and Strava's default is the weaker one. The first method anchors zones to maximum heart rate and sets each boundary as a percentage of it. The second anchors to lactate threshold heart rate (LTHR) — the heart rate at your sustainable threshold — which Joe Friel's framework treats as the more individualized reference because it tracks a metabolic event rather than an age-based ceiling [Friel 2018]. Strava ships with the max-HR method on by default and offers a custom-zone editor with draggable boundaries for mapping in LTHR-based zones. Out of the box, though, the numbers are percentages of an estimated max. That makes the max-HR value the master input. If Strava has your max at 185, every zone scales to 185; if your true max is 198, every line should sit roughly 13 bpm higher than where Strava drew it. The 220-age estimate that seeds this is statistically convenient and individually unreliable — Tanaka and colleagues analyzed 351 studies covering 18,712 subjects and found it systematically underestimates max HR in older adults, with a standard error near 10 bpm even in their improved 208 minus 0.7-times-age equation [Tanaka et al. 2001]. A 10-to-20 bpm miss is ordinary, not a worst case. Heart-rate zones are also not a substitute for what power measures. Heart rate is an input-lagging signal: intensity drives it, it doesn't define intensity, and it drifts with heat, fatigue, caffeine, and dehydration. Strava records the stream honestly; it just scales it against whatever max you gave it. The configuration is the difference between zones that describe your physiology and zones that describe an arithmetic guess about your age. ## The cascade: wrong max HR becomes wrong zones, wrong Relative Effort, wrong Fitness A wrong max HR doesn't stay contained in the zone chart. It rescales your time-in-zone splits, which inflates or deflates Relative Effort [Meyer 2018], which is the load input Strava feeds its Fitness curve when you ride without power. One bad number propagates through three layers of the dashboard. Start at the source. A max HR set 15 bpm too low pushes every zone boundary down, so an honest endurance effort that should read mid-zone-2 gets logged as low zone 3. Your time-in-zone splits now misreport where you actually trained. The watts and the perceived effort didn't change; the ruler did. Relative Effort is the next domino. It sums your weighted time-in-each-zone, with higher zones counting for far more per minute [Meyer 2018], so shifting minutes from zone 2 up into zone 3 mechanically lifts the score. This is the focused case our companion piece on why Relative Effort runs high on easy rides walks through in detail — a low max HR inflating the number on genuinely easy days. The broader point is that the same misconfiguration can also deflate the score if the max is set too high, reading hard work as easy. The direction depends on which way the max is wrong; the unreliability is the constant. The last domino is the Fitness curve. When you ride without a power meter, Strava uses Relative Effort as the load input to the 42-day Fitness average and 7-day Freshness average that underlie its Banister-derived chart [Hellard et al. 2007, Allen et al. 2019]. If Relative Effort is systematically inflated by a low max HR, your Fitness line climbs faster than your actual training justifies; if it's deflated, your hard blocks barely register. The chart that's supposed to tell you whether you're building or burying yourself is only as honest as the max HR three layers upstream. ## How to set your zones from real data, not a formula Replace the formula-derived max with the highest heart rate you've actually hit in a hard, warmed-up effort over the last few months. Then either let Strava rebuild zones from that max or enter custom boundaries from a threshold test. The goal is zones anchored to observed physiology, not to 220 minus your age [Tanaka et al. 2001]. Find your real max first. Scroll your last few months of hard rides and look for the peak heart rate at the end of a maximal effort — the top of a 5-minute climb, the final 30 seconds of an all-out sprint, the last rep of a VO2 set. That observed peak beats any formula because the formula's standard error is roughly 10 bpm and its individual miss can be double that [Tanaka et al. 2001]. If your hardest efforts keep topping out at 192 while Strava has you at 178, your max is wrong by 14 bpm and so is every zone built on it. Then decide how to anchor the zones. The simpler path is to correct the max HR in Strava's settings and let it regenerate the percentage-based zones. The more individualized path, if you've done a threshold field test, is to set custom zone boundaries from your LTHR — Friel's framework anchors the threshold zones to that value rather than to max [Friel 2018], and Strava's custom-zone editor lets you drag each boundary to match. Either is a defensible upgrade over the default; the threshold-anchored version is more precise if you have the test data, but a corrected max is most of the win for most riders. This is the data-layer discipline the broader case for training with Strava keeps returning to: the platform syncs your numbers faithfully but doesn't audit whether their inputs are right, and using Strava as a training tool rather than a journal means owning the configuration it leaves on autopilot. Setting your zones from a real observed peak takes a few minutes and quietly corrects every future ride at the source. And if you train with a power meter, you can sidestep the max-HR problem entirely — a power zones calculator maps your FTP straight to watt targets, which don't drift with heat or fatigue the way heart rate does. ## What to re-check after you fix the zones Correcting your max HR fixes future rides, but it generally won't rewrite the scores already sitting on past activities. Re-check your time-in-zone on the next few rides, watch whether Relative Effort now matches your perceived effort, and treat the Fitness curve's pre-fix history as scored against the old ruler. Past scores mostly stay as they were. Strava computes Relative Effort when an activity is processed, so changing your max HR generally does not retroactively recompute the numbers on rides you've already uploaded. That means your Fitness curve carries a seam: the weeks before the fix were scored against the wrong max, the weeks after against the right one. Don't read a discontinuity at the changeover as a real fitness change — it's the ruler switching, not your body. On the next handful of rides, run three independent checks that don't depend on the zones being perfect. Does time-in-zone now match what the ride actually felt like? Does Relative Effort track your perceived effort, where Foster's session-RPE is the cheapest valid internal-load metric in cycling [Haddad et al. 2017]? If you have a power meter, does normalized power agree with the heart-rate story [Allen et al. 2019]? When all three line up, your zones are anchored correctly. One caveat the new zones won't fix: cardiovascular drift. Even with a perfect max, heart rate climbs 5-10% over the first 60-90 minutes of a long or warm ride at constant power [Coyle & González-Alonso 2001], so a long endurance ride will still post a higher Relative Effort than its steady effort suggests. That's expected physiology, not a configuration error. Correcting your zones fixes the systematic bias; it doesn't make a heart-rate score immune to the things heart rate does on long days. ## Common questions **How do I know if my Strava heart rate zones are wrong?** Check whether your max HR was set from a formula or a real effort. Strava defaults to 220 minus your age, which carries a standard error near 10 bpm and can miss an individual by 10-20 [Tanaka et al. 2001]. If your hardest rides peak well above Strava's max, or if easy rides keep scoring as zone 3 when your power and perceived effort say easy, the zones are anchored to the wrong number. **Does a wrong max HR really affect my Fitness score, not just my zones?** Yes. When you ride without power, Strava uses Relative Effort as the load input to its Fitness and Freshness curves [Meyer 2018, Hellard et al. 2007]. Relative Effort is built from weighted time-in-zone, so a wrong max HR rescales the zones, distorts the score, and then flows into the Fitness line. One bad configuration value propagates through all three layers. **Should I anchor my Strava zones to max HR or lactate threshold?** Threshold (LTHR) is more individualized because it tracks a metabolic event rather than an age-based ceiling [Friel 2018], but it requires a field test. Strava's default uses max HR percentages. If you've tested your LTHR, enter custom zone boundaries from it; if not, correcting the max HR to a real observed peak captures most of the accuracy gain for far less effort. **Will fixing my max HR correct my past Strava rides?** Generally no. Strava computes Relative Effort when an activity is processed, so changing your max HR usually won't retroactively recompute scores on rides you've already uploaded. The fix is forward-looking — every future ride is scored against zones that match you. Expect a seam in your Fitness curve at the changeover and don't read it as a real fitness shift. **After I fix my zones, why is Relative Effort still high on long rides?** Cardiovascular drift. Heart rate climbs 5-10% over the first 60-90 minutes of a long or warm ride at constant power [Coyle & González-Alonso 2001], pushing weighted time-in-zone upward even with perfect zones. That's normal physiology, not a config error. Correcting your max HR removes the systematic bias; it doesn't stop heart rate from drifting on long days. ## References 1. **Tanaka et al. 2001.** [Age-predicted maximal heart rate revisited](https://pubmed.ncbi.nlm.nih.gov/11153730/). Journal of the American College of Cardiology. 2. **Meyer 2018.** [Quantifying Effort through Heart Rate Data](https://medium.com/strava-engineering/quantifying-effort-through-heart-rate-data-e6a0e3dd6a52). Strava Engineering. 3. **Hellard et al. 2007.** [Assessing the limitations of the Banister model in monitoring training](https://pmc.ncbi.nlm.nih.gov/articles/PMC1974899/). Journal of Sports Sciences. 4. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 5. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. 6. **Haddad et al. 2017.** [Session-RPE Method for Training Load Monitoring: Validity, Ecological Usefulness, and Influencing Factors](https://pmc.ncbi.nlm.nih.gov/articles/PMC5673663/). Frontiers in Neuroscience. 7. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. --- # Why your Strava Relative Effort is high on easy rides URL: https://www.adaptcycling.com/guides/relative-effort-high-on-easy-rides Updated: 2026-06-01 Author: Jim Camut You finish a genuinely easy zone 2 ride, legs fresh the whole way, and Strava hands you a Relative Effort higher than last week's interval session. The number feels wrong because it usually is — not as a bug, but as a misread input. Relative Effort is a heart-rate-zone score [Meyer 2018], and the most common reason it inflates on easy rides is a max heart rate that's set too low, which shoves a normal aerobic heart rate up into a higher-weighted zone. The second cause is cardiovascular drift on long or warm rides. Neither means you lost fitness, and neither should change tomorrow's plan. ## Why Relative Effort climbs on an easy ride Relative Effort scores how much time you spent in each heart-rate zone, weighted so higher zones count for more [Meyer 2018]. A high number means your heart rate sat in higher zones, not that the ride was hard. The two come apart whenever your zones are wrong or your heart rate ran high for a reason unrelated to effort. The mechanism is purely heart-rate-based. Strava takes your heart-rate stream, slices it into time-in-each-zone using your configured max HR, weights each zone — zone 4 counts for far more per minute than zone 2 — and sums the result [Meyer 2018]. The math is closer to Banister's TRIMP than to a power-based load score [Hellard et al. 2007], which means it measures internal cardiac strain, not mechanical work. A ride where your heart rate averages 145 bpm produces a higher score than one at 130 bpm even if both held the same 165 watts. That zone-weighting is exactly why an easy ride can outscore a hard one. A 90-minute endurance ride spent entirely in what Strava thinks is zone 3 will out-score a 45-minute interval session that spent 35 of its minutes recovering in zone 1 between efforts. The total time in elevated zones is what drives the number, and a long, steady ride accumulates more of that time than a short, spiky one. The score is doing precisely what it was designed to do — the problem is upstream, in what counts as zone 3 for you. This is the inverse of the more famous complaint, that Relative Effort under-reports short anaerobic intervals because heart rate lags behind metabolic stress. Here the score over-reports because heart rate, scored against the wrong zones, reads higher than the actual physiological demand. Same root cause, opposite direction: heart-rate-zone load is only as honest as the zone boundaries you feed it. ## The max-HR setting that inflates every number The single most common cause of an inflated Relative Effort is a max heart rate set too low — usually from the 220-age formula, which has a standard error of roughly 10 bpm and can miss an individual by 10-20 [Tanaka et al. 2001]. A max that's 15 bpm low makes your zone 2 read as zone 3, multiplying the weighted score. The 220-age formula is convenient and statistically weak. Tanaka and colleagues analyzed 351 studies covering 18,712 subjects and found 220-age systematically underestimates max HR in older adults, proposing 208 minus 0.7 times age as a better fit — and even that improved equation carries a standard error of estimate near 10 bpm [Tanaka et al. 2001]. For any one rider, a formula-derived max can sit 10-20 bpm away from their true ceiling. Strava seeds your zones from this kind of estimate unless you've entered a measured max. Watch what a 15-bpm error does to the zones. Suppose your true max is 190 but Strava has it at 175. Every zone boundary scales down with the max, so a steady aerobic effort that should land mid-zone-2 now reads as low zone 3. Strava weights zone 3 minutes more heavily than zone 2 minutes, so an honest two-hour endurance ride gets scored as if it carried meaningfully more strain than it did. The watts didn't change; the ruler did. The tell is a pattern, not a single ride. If easy rides keep scoring high while your power numbers and perceived effort say easy, your max HR is the first thing to check and the cheapest thing to fix. Set it from the highest heart rate you've actually seen in a hard, well-warmed-up effort over the last few months — the peak of a 5-minute climb or the last 30 seconds of a maximal sprint — not from a formula. Correcting the max re-scores every future ride and quietly fixes the inflation at its source. ## Cardiovascular drift: when a high reading is real but harmless Even with zones set perfectly, a long or warm ride can post a high Relative Effort because heart rate drifts upward over time at constant power. Coyle and González-Alonso documented this drift at 5-10% over the first 60-90 minutes of moderate work in the heat [Coyle & González-Alonso 2001]. The score is correct; it just isn't telling you the ride got harder. Cardiovascular drift is well-characterized physiology, not sensor error. As a steady ride goes on — especially in warm conditions — plasma volume drops, core temperature rises, and the heart compensates with a higher rate to maintain the same output, climbing 5-10% over the first hour to ninety minutes without any increase in work rate [Coyle & González-Alonso 2001]. Because Relative Effort sums time-in-zone, that late-ride drift pushes minutes from zone 2 up into zone 3, lifting the final score on a ride your legs experienced as entirely easy. This is why a three-hour endurance ride on a hot afternoon can post a Relative Effort that rivals a structured interval session. The first hour scores honestly; the back half inflates as drift carries your heart rate upward at unchanged effort. If you have power, you can see this plainly — normalized power flat, heart rate climbing. The body adapted normally; the heart-rate-based score simply can't distinguish drift from a genuine intensity increase, and it isn't designed to. The right reading is to treat drift-driven Relative Effort as expected on long warm rides and otherwise ignore it. It does not mean you went too hard, lost fitness, or need extra recovery. If anything, a ride long enough to drift is doing exactly the aerobic work a base block is built around. The number is real; the conclusion most riders draw from it is not. ## What to do when Relative Effort disagrees with your legs When an easy ride scores high, do not log it as hard and do not skip tomorrow's quality work off one inflated number. Trust power-based load, perceived exertion, and time-in-zone over a heart-rate score whose inputs you may have wrong. Then fix the upstream cause — almost always the max-HR setting — and commit to one primary load metric for the season. The training decision is the whole point, and it's the sub-question the broader case for using Strava as a training tool, not a journal, keeps raising: the data layer every cyclist already has only helps if you read it correctly. The pillar's rule for anaerobic intervals — trust power-based TSS when heart rate can't keep up — has a mirror image here. On easy aerobic rides, when heart rate runs high against suspect zones, trust the things that don't depend on those zones: normalized power if you have a meter, Foster's session-RPE if you don't [Haddad et al. 2017], and whether the ride actually felt easy. Concretely, an inflated Relative Effort on an easy ride should change nothing downstream. Don't reclassify a recovery spin as a hard day in your own log, don't let it talk you out of Tuesday's VO2 session, and don't add a rest day you don't need. One number scored against a possibly-wrong max HR is not evidence of fatigue. Power-based TSS [Allen et al. 2019], a glance at time-in-zone, and a one-to-ten effort rating are three independent checks that will agree with your legs when the heart-rate score doesn't. Then fix the cause once. Set your max HR from a real observed peak, confirm your zones, and commit to a single load metric across the season rather than comparing a heart-rate-only outdoor ride to a power-based indoor one — those aren't the same units [Allen et al. 2019]. This is where reading the data beats logging it: AdaptCycling reads both your power and heart-rate streams and computes an internal load estimate that doesn't whipsaw when one sensor's configuration is off, so a single mis-set max HR can't quietly rewrite a week of training history. ## Common questions **Why is my Relative Effort higher on an easy ride than a hard one?** Relative Effort sums weighted time-in-heart-rate-zone [Meyer 2018], so a long steady ride that sits in elevated zones the whole time out-scores a short interval session that spends much of its time recovering in zone 1. Most often the easy ride is being scored against a max HR set too low, which reads your aerobic effort as a higher zone than it really is. **How do I know if my max heart rate is set wrong in Strava?** If easy rides repeatedly score high while your power and perceived effort say easy, suspect the max. Formula estimates like 220-age carry a standard error near 10 bpm and can miss you by 10-20 [Tanaka et al. 2001]. Compare Strava's max to the highest heart rate you've actually hit in a hard, warmed-up effort over the last few months. If the gap is more than a few beats, reset it to the observed peak. **Should I trust Relative Effort or my power meter on an easy ride?** Power on an easy ride. Power-based load doesn't depend on your heart-rate zones being correct, so it isn't fooled by a mis-set max HR or by cardiovascular drift [Allen et al. 2019]. Use Relative Effort as a rough internal-strain proxy for sustained aerobic work, but let normalized power or TSS settle any disagreement. **Is a high Relative Effort on a long ride a sign I overdid it?** Usually not. On long or warm rides, heart rate drifts up 5-10% over the first 60-90 minutes at constant power [Coyle & González-Alonso 2001], pushing the score higher without any real increase in effort. A ride long enough to drift is doing exactly the aerobic work a base block wants. Don't add recovery or skip your next session based on it. **Does fixing my max HR change my past Relative Effort scores?** Mostly no. Strava computes Relative Effort when an activity is processed, so correcting your max HR generally won't retroactively rewrite the scores already sitting on past rides. The win is forward-looking: every ride after the change is scored against zones that match your physiology, so easy rides finally read as easy. ## References 1. **Meyer 2018.** [Quantifying Effort through Heart Rate Data](https://medium.com/strava-engineering/quantifying-effort-through-heart-rate-data-e6a0e3dd6a52). Strava Engineering. 2. **Tanaka et al. 2001.** [Age-predicted maximal heart rate revisited](https://pubmed.ncbi.nlm.nih.gov/11153730/). Journal of the American College of Cardiology. 3. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. 4. **Haddad et al. 2017.** [Session-RPE Method for Training Load Monitoring: Validity, Ecological Usefulness, and Influencing Factors](https://pmc.ncbi.nlm.nih.gov/articles/PMC5673663/). Frontiers in Neuroscience. 5. **Hellard et al. 2007.** [Assessing the limitations of the Banister model in monitoring training](https://pmc.ncbi.nlm.nih.gov/articles/PMC1974899/). Journal of Sports Sciences. 6. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. --- # How to restart cycling training after two weeks off — the periodization math URL: https://www.adaptcycling.com/guides/missed-two-weeks-cycling-restart Updated: 2026-05-04 Author: Jim Camut Two weeks off the bike — illness, travel, work, a sick kid — is the single most common disruption a self-coached cyclist will face in a season. The honest answer: you have lost less than you fear and more than zero. Resume at last week's intensity and you risk illness, injury, or the year-three plateau. Spend six weeks back in base and you forfeit the season. This is the actual math, the safe ramp rate, and what to do with the rest of your plan. ## What you actually lost in two weeks After 14 days off, VO2max has dropped roughly 3-7%, blood volume is down ~5-12%, and stroke volume has fallen measurably — but capillary density and most enzymatic adaptations are still essentially intact. The honest framing is partial cardiovascular detraining on top of a still-high baseline, not a fitness wipe. The Mujika & Padilla detraining series remains the cleanest synthesis of the literature [Mujika & Padilla 2000a]. Short-term cessation (≤4 weeks) in trained athletes produces a rapid decline in VO2max driven primarily by a fall in blood volume — total blood volume drops within days, stroke volume follows, and maximal cardiac output declines. The Coyle group's foundational study quantified the early curve: in endurance-trained subjects, VO2max fell 7% within the first 21 days of complete inactivity and stabilized at ~16% below baseline only after 56 days [Coyle et al. 1984]. The good news is what does not move in two weeks. Skeletal muscle capillarization remained ~50% above sedentary controls even after 84 days of cessation in the Coyle cohort [Coyle et al. 1984]. Mitochondrial enzyme activity falls faster than VO2max but stays well above untrained baselines at two weeks. Muscle glycogen replenishes within 24-48 hours of resumed feeding. The structural fitness — capillary bed, mitochondrial density, fiber type — is the slow-detraining substrate that two weeks off does not erase. Translated into power-meter terms: most riders see FTP estimates drop in the 3-8% range after a clean two-week break, mostly from cardiovascular drift rather than peripheral loss. A 280W FTP becomes a working 260-270W FTP. That is a setback, not a season. Pretending otherwise — either by ignoring it and slamming back into threshold work, or by treating the rider as untrained — is where most restart protocols go wrong. ## Why most self-coached riders restart wrong Two failure modes dominate. The first is the panic-resume: drop straight back into the planned week-3 intensity, pile a load spike on top of a decompressed system, and earn an illness or an overuse injury inside ten days. The second is the over-cautious six-week base reset that costs the season for fitness already mostly intact. Both are avoidable with a defensible ramp. The panic-resume failure is well-documented in the training-load literature. Foster's monotony work showed that illness onset in trained athletes clusters around weeks where load and monotony spike together [Foster 1998]. Returning from two weeks off with a 100% volume week at planned intensity is, mathematically, a workload spike on a chronic load that has decayed. The acute:chronic workload ratio framework places elevated injury and illness risk above ratios of ~1.5; a typical post-break full-resume week pushes that ratio to 2.0 or higher [Hulin et al. 2016]. This is the TrainerRoad-plan-on-autopilot failure mode: the calendar has not paused, and the athlete pays the bill. The over-cautious failure is less obvious and arguably more expensive. Riders who treat two weeks off as 'starting from scratch' and grind out 4-6 weeks of zone 2 base before reintroducing intensity are working against the detraining curve. Mujika & Padilla's long-term review is explicit that with even moderately reduced training, intensity preserved and volume cut, most cardiovascular adaptations can be largely maintained over weeks [Mujika & Padilla 2000b]. The corollary on resumption: you do not need a multi-week aerobic-only block to safely reintroduce intensity that the body still remembers. You need a graded ramp, not a re-build. This is one of the hardest moments for a self-coached cyclist because the right answer is uncomfortably middle-of-the-road — and the broader playbook of training without a coach (covered in our pillar on the self-coached cyclist) is what most riders fall back on. A coach functions as the calibrated 'no' here. Replacing that voice requires either a pre-committed rule or a tool that flags load climbing too fast. ## The 60/80/100 restart protocol Week 1: 60% of pre-break weekly volume, zone 2 only, no structured intervals. Week 2: 80% of pre-break volume, reintroduce one tempo or sweet-spot session at 85-92% of pre-break FTP. Week 3: ~100% of pre-break volume with normal structure restored. This keeps the acute:chronic ratio inside the safe band while respecting that two weeks off is partial, not full, detraining. The arithmetic is anchored in Gabbett's training-injury prevention paradox: high chronic workloads built progressively are protective, but acute spikes on a decayed chronic base are where injuries cluster. Keeping the weekly acute load between 0.8 and 1.3 times the trailing 28-day chronic average is the operational sweet spot [Hulin et al. 2016]. After two weeks at zero, that chronic average has dropped — so a 60% volume Week 1 lands inside the safe band rather than spiking out of it. By Week 3, the chronic average has caught up enough that a full-volume week with intensity is no longer a spike. Week 1 is exclusively zone 2 (roughly 56-75% FTP, conversational) for two reasons: it rebuilds plasma volume quickly without neuromuscular load, and it reads how recovered the system actually is. If zone 2 rides feel disproportionately hard at the same wattage, extend Week 1 by 3-4 days before progressing. Heart rate at fixed power is the cleanest single marker; expect it 5-10 bpm elevated initially and trending back to baseline by week's end. Week 2 reintroduces exactly one structured session — tempo or sweet-spot, 2x15 to 3x12 minutes at 85-92% of pre-break FTP. Not threshold, not VO2max. The point is to reawaken the muscular and neural pathways without stacking glycolytic stress on a still-recovering aerobic system. Week 3 returns to the prescribed plan, with one caveat: if the next-up block is a peak VO2max or anaerobic block, slide an additional aerobic week in before it. Hard-intensity blocks layered on a recently restarted base are a near-guarantee of the panic-resume failure. ## What to do with the rest of your plan Three honest options: shift the goal event back two weeks (the cleanest call when the date is movable), keep the date and accept a smaller peak (acceptable for B and C events), or pivot to a later target and use the original date as a tune-up. The wrong move is to compress the plan — running the same blocks faster to 'catch up' is the highest-injury-risk path on the menu. If the goal event is movable — most self-coached riders' A events are gran fondos, local stage races, or self-defined fitness peaks — pushing the date two weeks is mathematically equivalent to inserting a recovery week and resuming. The structural fitness substrate is intact, the cardiovascular re-build is fast, and the original peak is reachable. This is the default recommendation when the calendar permits. If the date is fixed and inside 4-6 weeks, accept the smaller peak. Bosquet's taper meta-analysis is unambiguous: intensity preserved and volume cut 41-60% over the final two weeks captures most available performance [Bosquet et al. 2007]. A rider arriving at 95% of the planned peak through a clean restart and taper outperforms one who crammed the missed work and arrived overtrained. Meeusen's consensus warns that overload beyond recovery capacity slides functional overreaching into the non-functional kind, where recovery takes weeks to months [Meeusen et al. 2013]. If the date is fixed and inside 2-3 weeks, treat the original event as a tune-up and re-pick the A event. This is the call most self-coached riders resist and most coaches make immediately. A 280W rider crammed back to 255W has a worse race and a worse next block than the same rider arriving at 270W having ridden the original target as a B event. Compressing the plan reliably produces the worst outcome of the three. ## Common questions **Is two weeks off enough to lose my FTP?** Partially. Expect a 3-8% drop in functional threshold power after a clean two-week break, driven mostly by reduced blood volume and stroke volume rather than muscular loss [Mujika & Padilla 2000a, Coyle et al. 1984]. Most of that comes back inside 2-3 weeks of structured riding. Skip a full FTP test on the restart — re-test no earlier than the end of Week 3, and use perceived exertion plus heart-rate-at-power drift to set targets in the meantime. **Should I do a fitness test before restarting structured intervals?** No. A maximal effort on a decompressed system is the most reliable way to flag a respiratory infection or trigger an injury. Run the 60/80/100 protocol first; if you need a power anchor for Week 2's tempo work, take 90% of your pre-break FTP and adjust by feel. Re-test at the end of Week 3 once the cardiovascular re-build has stabilized. **What if the two weeks off was illness, not just travel or work?** Use the above-the-neck rule and the 24-hour test. Head-and-sinus symptoms with no fever generally permit a graded return; chest, fever, or systemic symptoms require full resolution plus a buffer. Day 1 back: 30 minutes zone 2. If the next 24 hours show no symptom rebound, progress to a full Week 1 of the 60/80/100. Any rebound — sore throat, fatigue spike, lingering cough — adds 3-5 days. Persistent post-illness underperformance beyond two weeks of normal training is the threshold for seeing a sports doctor [Meeusen et al. 2013]. **Could two weeks off actually help me?** Sometimes, yes. If the break was preceded by 6-10 weeks of climbing load, high training monotony, and accumulating fatigue, the forced rest may have unloaded a system that was heading toward non-functional overreaching [Foster 1998, Meeusen et al. 2013]. Riders in that pattern often post personal bests in Weeks 3-5 after the restart. The marker: if your pre-break weeks felt grindy at normal power and your zone 2 heart rate was creeping up, the rest was probably overdue. **How does this fit into a self-coached training year?** Two-week disruptions are the rule for any rider with a job and a family — plan the year assuming 2-3 of them. The broader framework for training as a self-coached cyclist treats adaptive structure as the default; rigid 16-week plans break on contact with real life. Pre-commit to the 60/80/100 protocol and the goal-event decision tree before you need them, and the disruption costs a fortnight, not a season. ## References 1. **Mujika & Padilla 2000a.** [Detraining: loss of training-induced physiological and performance adaptations. Part I: short term insufficient training stimulus](https://pubmed.ncbi.nlm.nih.gov/10966148/). Sports Medicine. 2. **Mujika & Padilla 2000b.** [Detraining: loss of training-induced physiological and performance adaptations. Part II: long term insufficient training stimulus](https://pubmed.ncbi.nlm.nih.gov/10999420/). Sports Medicine. 3. **Coyle et al. 1984.** [Time course of loss of adaptations after stopping prolonged intense endurance training](https://pubmed.ncbi.nlm.nih.gov/6511559/). Journal of Applied Physiology. 4. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 5. **Meeusen et al. 2013.** [Prevention, diagnosis, and treatment of the overtraining syndrome: joint consensus statement of the European College of Sport Science and the American College of Sports Medicine](https://pubmed.ncbi.nlm.nih.gov/23247672/). Medicine & Science in Sports & Exercise. 6. **Hulin et al. 2016.** [The acute:chronic workload ratio predicts injury: high chronic workload may decrease injury risk in elite rugby league players](https://pubmed.ncbi.nlm.nih.gov/26511006/). British Journal of Sports Medicine. 7. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. --- # How often a self-coached cyclist should take a recovery week URL: https://www.adaptcycling.com/guides/recovery-week-cadence-self-coached Updated: 2026-05-04 Author: Jim Camut The default answer is every fourth week — three weeks of progressive load, one week of cut volume — and for most amateur cyclists that cadence is right. But the 3-on/1-off schedule is a starting position, not a rule. Masters athletes, novices, and high-stress weeks deserve different cadences, and the recovery week itself is the part most self-coached riders get wrong. This is where the cadence comes from, what the down-week actually looks like, and the four signals that say take one sooner. ## Where the 3-on/1-off cadence actually came from The three-week build, one-week recovery cadence is not a tradition — it is the operational expression of three converging research lines: supercompensation timing, Foster's training-monotony curve, and the chronic-workload ramp rate. Every framework that produced it lands in the same place because the underlying physiology does. Three weeks is roughly how long an aerobic system can absorb progressively increasing load before the recovery cost outruns the adaptive return. Foster's 1998 monotony work [Foster 1998] formalized what coaches had been working from intuitively: across 25 trained athletes, the strongest behavioral predictor of illness, injury, and overreaching was not absolute load but the product of weekly load and daily-load variance — what Foster termed training strain. High load with high monotony — week after week of similar TSS, no break — is the curve that bends toward non-functional overreaching. A scheduled recovery week breaks the monotony numerator directly. Skipping it for an extra build week is mathematically the highest-risk decision a self-coached rider can make. The supercompensation literature says the same thing from the other direction. Issurin's block-periodization review [Issurin 2010] synthesized the training science behind concentrated workloads in 2-4 week blocks followed by reduced-load recovery, which exploit the lag between fatigue dissipation and fitness retention. Aubry and colleagues' 2014 trial in endurance athletes [Aubry et al. 2014] showed the practical end of this curve: eleven of 23 athletes who completed a deliberate overload block became functionally overreached, and the taper-week supercompensation those athletes saw was real but smaller than the acutely-fatigued group's. Push too far past the three-week mark without unloading and the supercompensation window collapses into non-functional overreaching, where recovery takes weeks to months [Meeusen et al. 2013]. The chronic-workload framework operationalizes the same cadence. Hulin and colleagues' rugby-league data [Hulin et al. 2016] placed elevated injury risk above acute:chronic workload ratios of roughly 1.5; the 28-day chronic load behaves like a four-week rolling average, which means a build that increases weekly load 5-10% for three weeks and then drops 30-40% in week four lands inside the safe band on every week. Skip the recovery week and the ratio climbs out of bounds in week four or five. The 3-on/1-off cadence is, in a sense, the schedule the math forces — and it is part of the broader playbook of training without a coach (covered in our pillar on the self-coached cyclist), where periodization rules that a coach used to enforce now have to be enforced by the rider. ## What a recovery week actually looks like Volume drops 30-40%. Intensity does not disappear — it gets sharper and shorter. No structured intervals over threshold, no long endurance rides at the upper end of zone 2, no group rides where the pace is not yours to set. The recovery week is not a rest week and it is not zero. It is the dose that lets supercompensation actually happen. The volume number is non-negotiable: weekly TSS should fall to roughly 60-70% of the preceding build-week peak. A rider averaging 600 TSS across the build should land near 360-420 TSS in the recovery week. Cutting less than that is the most common amateur error — the rider drops one ride, calls it a recovery week, and arrives in the next build still fatigued. Halson's review of training-load monitoring [Halson 2014] is explicit that load monitoring exists primarily to determine whether an athlete is adapting to the program, and an under-cut recovery week is the most reliable way to fail that test. Intensity is the more interesting variable. The instinct is to cut intensity entirely and ride only zone 2 — but there is good evidence that maintaining short, sharp efforts during the down-week preserves neuromuscular and high-end fitness while the cardiovascular system unloads. The standard prescription is one short opener mid-week — 4-6 x 30-second efforts at well over threshold with full recoveries, or a few 2-minute efforts at sweet spot — and otherwise keep rides conversational. Issurin's block-periodization data [Issurin 2010] showed that maintenance doses of high intensity prevent detraining of glycolytic and neuromuscular qualities while overall load drops. Cutting all intensity for seven days flattens the rider; cutting volume less than 30% leaves the rider undertrained going into the next block. Structurally, the week reads: one or two off days, two or three short zone 2 rides (45-75 minutes), one of those with a small set of sharp openers, and a moderate weekend ride at the lower end of zone 2 — no long sweet-spot, no group hammer-fest. Sleep, calorie intake, and protein should not drop with the volume; the recovery week is when most of the adaptive work happens, and an undernourished recovery week is a wasted one. If the rider arrives at the end of week four with heart rate at power back to baseline, mood normalizing, and an itch to train hard, the dose was right. If they feel sluggish or stale, the cut was probably not deep enough. ## The four signals that say take a recovery week now The 3-on/1-off cadence is the default, not a contract. Four signals override the calendar and say take the recovery week early: heart-rate-at-power drift, RPE creep at fixed wattage, sleep and mood changes, and an acute:chronic workload ratio climbing past 1.5. Two or more concurrent signals is the threshold to act on, regardless of where you are in the build. Heart-rate-at-power drift is the cleanest single marker. On a repeated zone 2 ride at the same wattage, in similar conditions, heart rate that has crept 5-10 bpm above its baseline over the preceding seven days is meaningful sympathetic-system fatigue. Halson's monitoring review [Halson 2014] places submaximal heart-rate response among the most practical and well-validated daily fatigue indicators — non-invasive, repeatable, and sensitive to accumulating load. A rider seeing this pattern on Tuesday and Thursday of build week three should not do Saturday's threshold session; they should pull the recovery week forward. RPE creep is the second signal and the one self-coached riders are most likely to dismiss. The same prescribed wattage feeling subjectively harder week-over-week, even when heart rate has not yet drifted, is an early central-fatigue marker [Foster 1998]. The third signal is a pair: sleep quality dropping, and resting mood flattening. The Meeusen consensus [Meeusen et al. 2013] is explicit that mood disturbance and sleep degradation are among the earliest reliable indicators of non-functional overreaching, often preceding any performance drop. The fourth signal is quantitative: the acute:chronic workload ratio crossing roughly 1.5 [Hulin et al. 2016]. Most modern training platforms — TrainingPeaks, Intervals.icu, AdaptCycling — surface this number directly. A self-coached rider can sanity-check by asking whether the current week's TSS is more than 1.5x the trailing four-week rolling average. If it is, the build is steepening faster than the chronic load can absorb, and a recovery week is the cheapest way to reset the ratio. The trap is treating a single signal as noise; two signals concurrent — say, RPE creep and sleep degradation — is the threshold where the calendar loses and the signals win. A separate scenario sits next to this one: when illness or travel pulls 10-14 days of training off the calendar entirely, the resume-into-next-build approach does not apply — the protocol there is a graded ramp covered in our spoke on restarting after two weeks off. ## Common questions **Should I take a recovery week every 3 weeks or every 4 weeks?** Every fourth week is the right default for most amateur cyclists with 1-3 years of structured training. Move to a 3-on/1-off cadence (a recovery week every third week) if you are over 45, returning from an extended layoff, in a deliberate overload block, or already showing two of the four warning signals. Younger riders with low total weekly load (under ~6 hours) can sometimes carry 4-on/1-off, but the cost of taking an extra recovery week is small and the cost of skipping one is large. **Is a recovery week the same as a rest week?** No, and the distinction matters. A rest week is zero or near-zero load — appropriate at the end of a season, after a goal event, or during forced disruption. A recovery week is roughly 60-70% of preceding build-week volume with intensity preserved as short maintenance efforts. Taking a true rest week mid-block detrains the high-end qualities a 30-40% volume cut would have preserved [Issurin 2010]. Save full rest for the transition phase between mesocycles or after a peak event. **How do I know my recovery week worked?** Three markers. First, heart rate at a fixed zone 2 wattage drops 3-8 bpm by the end of the week. Second, perceived exertion at the same wattage drops noticeably. Third, you finish the week wanting to train hard again, not dreading it. If any of those three are missing on day seven, extend by 2-3 days before starting the next build — riding into a build still fatigued is the fastest way to convert functional overreaching into the non-functional kind [Aubry et al. 2014, Meeusen et al. 2013]. **What if my schedule does not allow a recovery week every fourth week?** Most self-coached riders will have weeks where life delivers an unplanned recovery week — a work crunch, a sick kid, a travel block. Those count. The cadence is a default, not a contract; the goal is roughly one cut week per four weeks averaged across the season, not perfect calendar adherence. Tracking whether your trailing 28-day TSS includes at least one week 30-40% below the others is a more useful check than counting calendar weeks. ## References 1. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 2. **Meeusen et al. 2013.** [Prevention, diagnosis, and treatment of the overtraining syndrome: joint consensus statement of the European College of Sport Science and the American College of Sports Medicine](https://pubmed.ncbi.nlm.nih.gov/23247672/). Medicine & Science in Sports & Exercise. 3. **Halson 2014.** [Monitoring training load to understand fatigue in athletes](https://pubmed.ncbi.nlm.nih.gov/25200666/). Sports Medicine. 4. **Hulin et al. 2016.** [The acute:chronic workload ratio predicts injury: high chronic workload may decrease injury risk in elite rugby league players](https://pubmed.ncbi.nlm.nih.gov/26511006/). British Journal of Sports Medicine. 5. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization: block periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. 6. **Aubry et al. 2014.** [Functional overreaching: the key to peak performance during the taper?](https://pubmed.ncbi.nlm.nih.gov/25134000/). Medicine & Science in Sports & Exercise. --- # Am I overtrained or just tired? How to tell the difference as a self-coached cyclist URL: https://www.adaptcycling.com/guides/am-i-overtrained-or-just-tired Updated: 2026-09-03 Author: Jim Camut Most riders who type this question into a search bar are not overtrained. They are under-recovered, under-fueled, or stacked under life stress that shows up on the bike before it shows up anywhere else. Real overtraining syndrome is rare and serious, and the canonical sports-medicine framework distinguishes three states with very different timelines and very different responses [Meeusen et al. 2013]. This is how to tell which one you are in, what to self-monitor, and the threshold where self-coaching ends and a sports doctor starts. ## The three categories: functional overreaching, non-functional overreaching, and OTS The European College of Sport Science and ACSM joint consensus separates training fatigue into three states by recovery timeline. Functional overreaching resolves in days to about two weeks. Non-functional overreaching takes weeks to months. Overtraining syndrome takes months to a year, sometimes longer. Recovery time, not symptom severity at any single moment, is the variable that distinguishes them [Meeusen et al. 2013]. Functional overreaching (FOR) is the deliberate end of a hard training block. Performance drops, perceived exertion rises, and a brief unloading period (typically 5-14 days) restores baseline and often produces supercompensation above it. Aubry and colleagues' overload-and-taper trial documented this directly: 11 of 23 endurance athletes who completed a deliberate three-week overload block became functionally overreached, and after a two-week taper most rebounded to or above pre-overload performance [Aubry et al. 2014]. This is the state most periodized plans are designed to produce on purpose at the end of a build block. It is uncomfortable. It is not pathological. Non-functional overreaching (NFOR) is functional overreaching that did not resolve. The Meeusen consensus describes NFOR as a state where the athlete continues to underperform, fatigue persists, and the supercompensation never arrives — recovery now takes weeks to several months [Meeusen et al. 2013]. The clinical picture and the hormonal picture begin to overlap with OTS: mood disturbance, sleep degradation, elevated infection susceptibility. Bellinger's 2020 review notes that the NFOR state is associated with measurable cardiovascular, hormonal, and metabolic consequences and that the line between FOR and NFOR is the line at which intentional overload becomes a problem [Bellinger 2020]. Overtraining syndrome (OTS) is rare. The consensus describes it as prolonged maladaptation across multiple biological systems — neuroendocrine, autonomic, immune — with severe and persistent performance decrement and recovery times measured in months to a year [Meeusen et al. 2013]. The diagnosis is one of exclusion. Before OTS is named, the consensus requires ruling out organic disease, infection, low energy availability, iron deficiency, magnesium deficiency, allergies, and depression. For practical purposes, an amateur cyclist is almost always on the FOR-to-early-NFOR spectrum, not in OTS — but the difference matters because the response to each is different. ## What ordinary fatigue actually looks like — and why most riders are here Ordinary training fatigue is the dominant explanation for feeling flat. Three or four hard weeks without a real recovery week, a 50-hour work sprint, two nights of bad sleep, and a lean week of eating will all flatten power output and elevate perceived exertion without any of it qualifying as overreaching. The first move is to rule out the boring explanations before reaching for a diagnosis. Carl Foster's foundational monitoring work showed that illness and overreaching cluster around weeks where load and monotony spike together — same hard sessions repeated, no variance, no recovery week [Foster 1998]. A self-coached rider posting 600 TSS weeks for four weeks straight on a TrainerRoad plan, without a deliberate cut week, is producing exactly the pattern Foster's data flagged. The fix is a 30-40% volume cut for seven days, not a six-month diagnostic workup. Most riders who think they are overtrained are actually overdue for the recovery week they skipped. Energy availability is the other underappreciated explanation. Stellingwerff and colleagues' synthesis of the OTS and Relative Energy Deficiency in Sport (RED-S) literature is striking: 18 of 21 studies of training overload showed concurrent reductions in energy or carbohydrate availability, and the symptom profiles of OTS and RED-S overlap heavily because both originate at the same hypothalamic-pituitary axis [Stellingwerff et al. 2021]. Many self-diagnosed overtraining cases in endurance athletes are a fueling problem, not a training-load problem. The rider who cut calories to lose winter weight while continuing the same training schedule is the canonical pattern. Practically, the test is whether a real recovery week — 30-40% volume cut, intensity preserved as short maintenance efforts, full eating, full sleep — restores baseline power and mood inside seven days. If it does, the rider was tired. If it does not, the next category is in play. A separate but neighboring case is the rider whose recovery-week signals fired early but the cadence held to the calendar — the four signals that say take the recovery week now are covered in our spoke on recovery-week cadence. ## The five self-monitoring signals — and which combinations actually mean something The validated submaximal markers of functional overreaching, in roughly descending reliability, are heart-rate-at-power drift, rating of perceived exertion creep, performance drop on a standardized submaximal test, sleep degradation, and mood flatness. Roete and colleagues' systematic review found consensus across studies for these markers; HRV and VO2max changes were not consistently validated [Roete et al. 2021]. Two or more concurrent signals over 7-10 days is the threshold that means something. Heart rate at fixed submaximal power is the single cleanest marker, and its direction is counter-intuitive. In acute fatigue, heart rate at a given wattage is elevated. In established functional overreaching, the autonomic nervous system shifts and heart rate at a given submaximal wattage often drops, while maximal heart rate falls and heart rate recovery accelerates [Roete et al. 2021]. Garmin's daily resting HR, a Whoop strap, or a simple repeated 20-minute zone 2 ride at the same wattage will all surface the trend. A 5-10 bpm sustained departure from baseline at the same power, in either direction, over 7-10 days is the signal. RPE creep is the second-line signal and the one self-coached riders most often dismiss. The same prescribed wattage feeling subjectively harder week-over-week is an early central-fatigue marker that often precedes power drop [Foster 1998, Roete et al. 2021]. Sleep and mood are the third pair. The Meeusen consensus is explicit that sleep disturbance and depressed mood are among the earliest reliable indicators of NFOR — frequently appearing before any objective performance change [Meeusen et al. 2013]. Heart-rate variability scores from Whoop, Oura, and Garmin track autonomic state but did not consistently differentiate functionally overreached athletes from controls in the systematic-review evidence [Plews & Laursen 2013, Roete et al. 2021]. Use HRV as one input — not a verdict. ## When to stop self-coaching and see a sports doctor The threshold is straightforward: if 14 days of complete or near-complete rest with full eating and sleep does not restore baseline power, mood, and resting heart rate, it is no longer a training-load problem to self-resolve. At that point the differential includes thyroid dysfunction, iron deficiency, RED-S, persistent infection, mononucleosis, and clinical depression. A self-coached rider should hand the problem to a sports physician [Meeusen et al. 2013]. The Meeusen consensus is specific about the workup the athlete cannot do alone. OTS is a diagnosis of exclusion, which means the work is to rule out organic causes first: thyroid panel, ferritin and full iron studies, vitamin D, testosterone in male athletes, menstrual function and energy availability assessment in female athletes, EBV and CMV serology, and a screen for depression [Meeusen et al. 2013]. A standard primary-care physical will miss most of these unless the athlete arrives with the right vocabulary; a sports-medicine physician familiar with endurance athletes is the right specialist. This is the moment where the broader playbook of training without a coach (covered in our pillar on the self-coached cyclist) has reached its limit — self-coaching is about training-load decisions, not differential diagnosis. The single most under-recognized scenario in this population is RED-S. Stellingwerff's review found that the majority of suspected overtraining cases show concurrent low energy availability and the symptom profile is nearly identical to NFOR [Stellingwerff et al. 2021]. RED-S is treated by eating more, not training less, and getting that diagnosis wrong wastes months. The rider's job at this stage is to escalate cleanly: sit out structured training for at least two weeks, eat at maintenance, sleep without an alarm, document the symptom timeline and the preceding 12 weeks of training load, and bring it to a sports physician. ## Common questions **How long should it take to feel normal again after a hard training block?** If the block was a deliberate overload, 5-14 days of reduced volume with intensity preserved should restore baseline and often push performance above it [Aubry et al. 2014]. Past two weeks without recovery, you have crossed from functional into non-functional overreaching, where recovery measures in weeks to months [Meeusen et al. 2013]. The single cleanest test is whether a true recovery week (30-40% volume cut, full eating, full sleep) restores power-at-heart-rate by day seven. **Is a low resting heart rate a sign of overtraining?** Possibly, but not on its own. In established non-functional overreaching, autonomic shifts can produce a paradoxically lower submaximal heart rate alongside a faster HR recovery and a lower maximum heart rate [Roete et al. 2021]. The pattern that matters is a 5-10 bpm sustained departure from your personal baseline over 7-10 days combined with at least one other signal — RPE creep, performance drop, mood flatness, or sleep degradation. A single morning reading of an unusually low resting HR after a hard week is not diagnostic. **Should I trust my Whoop or Oura recovery score?** As one input. HRV-based recovery scores from Whoop, Oura, Garmin, and similar wearables track autonomic state, which is one component of the picture — but the systematic-review evidence is that HRV alone does not reliably differentiate functional overreaching from acute fatigue or healthy adaptation [Plews & Laursen 2013, Roete et al. 2021]. Use a sustained downward drift (more than ~1 SD below your baseline for 5-7 days) as a flag to check the other signals, not as a verdict on its own. Heart-rate-at-power on a repeated submaximal ride is more informative for cyclists than any wearable score. **Could what feels like overtraining actually be under-fueling?** Frequently, yes. The OTS and Relative Energy Deficiency in Sport (RED-S) literature shows the symptom profiles overlap heavily and 86% of training-overload studies showed concurrent low energy availability [Stellingwerff et al. 2021]. Before assuming an overtraining diagnosis, audit the last 4-6 weeks honestly: weight trend, hunger cues, recovery between rides, libido and morning energy in male athletes, menstrual regularity in female athletes. RED-S is treated by eating more, not by resting more, and getting the diagnosis wrong wastes months. ## References 1. **Meeusen et al. 2013.** [Prevention, diagnosis, and treatment of the overtraining syndrome: joint consensus statement of the European College of Sport Science and the American College of Sports Medicine](https://pubmed.ncbi.nlm.nih.gov/23247672/). Medicine & Science in Sports & Exercise. 2. **Aubry et al. 2014.** [Functional overreaching: the key to peak performance during the taper?](https://pubmed.ncbi.nlm.nih.gov/25134000/). Medicine & Science in Sports & Exercise. 3. **Bellinger 2020.** [Functional overreaching in endurance athletes: a necessity or cause for concern?](https://pubmed.ncbi.nlm.nih.gov/32064575/). Sports Medicine. 4. **Roete et al. 2021.** [A systematic review on markers of functional overreaching in endurance athletes](https://pubmed.ncbi.nlm.nih.gov/34108275/). International Journal of Sports Physiology and Performance. 5. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 6. **Plews & Laursen 2013.** [Training adaptation and heart rate variability in elite endurance athletes: opening the door to effective monitoring](https://pubmed.ncbi.nlm.nih.gov/23852425/). Sports Medicine. 7. **Stellingwerff et al. 2021.** [Overtraining syndrome (OTS) and relative energy deficiency in sport (RED-S): shared pathways, symptoms and complexities](https://pubmed.ncbi.nlm.nih.gov/34181189/). Sports Medicine. --- # Junk intensity in cycling: what it is and how to fix it as a self-coached rider URL: https://www.adaptcycling.com/guides/junk-intensity-fix-self-coached Updated: 2026-05-05 Author: Jim Camut Most amateur weeks drift the same way. The Tuesday hammer lands around 82 percent of FTP — too hard to be aerobic base, too soft to drive threshold. The Saturday group ride averages 220 normalized. The Wednesday solo loop creeps up because riding 165 watts for ninety minutes feels like nothing. Three rides, three days, all parked in the same wattage band — and that band is exactly where adaptation goes to die. This is what junk intensity actually is, why amateurs default into it, and the structural levers that fix it without willpower. ## What junk intensity actually is — and the wattage band that defines it Junk intensity is the wattage band sitting between the top of zone 2 and the bottom of sweet spot — operationally about 76 to 87 percent of FTP, roughly Coggan zone 3 tempo without the deliberate prescription [Allen et al. 2019]. It is below LT2, above LT1, and produces neither tail's adaptation. Sweet spot at 88 to 94 percent is intentional; junk intensity is drift into the same neighborhood without the structure that makes the dose useful. The boundary is anchored by the two lactate thresholds. LT1 is roughly 75 percent of FTP for most trained cyclists — the ceiling of conversational zone 2, where blood lactate first departs from baseline. LT2 sits near 95 to 100 percent of FTP, the threshold sweet-spot work targets from below. The band between them is metabolically real estate where the rider is already producing lactate faster than at LT1 but not enough to drive the lactate-clearance and threshold adaptations sweet spot delivers. It is the gap between two prescribed doses, not a third dose. Self-coached riders confuse this band with sweet spot constantly. The wattage gap between a 76-percent ride and an 88-percent ride is small — 30 watts apart for a 250-watt FTP — but the physiology gap is large. Sweet spot is prescribed in 10 to 20 minute blocks with full warm-up and recovery on either side. Junk intensity is what happens when an unstructured ride averages the same effective intensity factor without the structure. The number on the head unit looks similar; the training stimulus is not. ## Why the middle is mechanically the worst place to be The middle band fails on both ends of the adaptation spectrum simultaneously. It is too hard for the mitochondrial and fat-oxidation gains zone 2 produces, and too easy for the lactate-clearance and VO2max gains threshold and above produce. The fatigue cost is real; the adaptive return is the smallest of any wattage band a cyclist can choose. Zone 2 produces specific adaptations: mitochondrial density, capillary growth, and fat oxidation [San Millán & Brooks 2018]. The required dose is volume — two to four hours at 56 to 75 percent of FTP, type I fibers carrying the work. Once intensity rises into the 76 to 87 percent band, type IIa recruitment increases, lactate production rises above clearance, and the metabolic substrate shifts toward carbohydrate. The mitochondrial signal that two hours at 165 watts would have produced does not happen at 200 watts — different fiber pool, different signaling pathway. The high-end adaptations require a dose junk intensity cannot deliver. Neal and colleagues ran a 6-week crossover in 12 trained cyclists: a polarized block (80 percent low / 20 percent high) produced 8 percent peak-power gains versus 3 percent on threshold, and 85 versus 37 percent improvements in high-intensity exercise capacity [Neal et al. 2013, Seiler 2010]. Muñoz et al. found the same pattern in 30 recreational runners over 10 weeks — between-thresholds distribution improved 10K times less than polarized, with effect sizes widening when adherence held [Muñoz et al. 2014]. What junk intensity does deliver is fatigue. Time-in-the-middle is the wattage band that elevates training monotony — every ride costs the rider in recovery; none drives the adaptation either tail produces. Riders stuck here need recovery weeks more often than the calendar suggests (covered in our spoke on recovery-week cadence) and disproportionately end up asking whether they are overtrained (covered in our spoke on telling overtraining from ordinary fatigue). ## Why amateur weeks drift into the middle (and willpower does not fix it) Amateur weeks drift into the middle for structural reasons, not character ones. The Tuesday hammer is set by the group, not the prescription. The head unit reinforces wattage drift. Strava segments reward a hard 12 minutes anywhere on the ride. Riders who would never program a 90-minute tempo session end up doing one most weeks — by accident, in pieces, against their stated plan. Group-ride gravity is the largest cause. A two-hour Saturday scheduled at 65 percent of FTP routinely turns into a 220-NP effort because the group sets the pace and dropping off feels worse than the fatigue cost. The ride is logged as endurance and treated as a hard day by the body — the worst-of-both-worlds Filipas et al. flagged in their 16-week recreational-cyclist trial, where pyramidal distribution outperformed the middle-heavy default [Filipas 2024]. Head-unit anchoring is second: live wattage drifts up whenever the rider feels good, and segment hunting embeds threshold efforts inside zone 2 rides. The third cause is productivity bias. Self-coached riders feel guilty about easy rides — riding 165 watts for two hours feels indistinguishable from not training, so they push to 200 watts, log a productive ride in the app, and arrive at Saturday's threshold session too cooked to do it well. The willpower fix — "just ride easier" — fails because the structure of the week pulls the rider into the middle every time. The fix has to be structural, which is the broader question of training without a coach we cover in our pillar on the self-coached cyclist. ## Three structural fixes that do not require willpower Three structural fixes do the work willpower cannot: a power gate that caps zone 2 rides at LT1, hard-day/easy-day separation by 48 hours, and an anchored interval session that owns the week's intensity dose. Each removes a cause rather than fighting a symptom. Fix one — power gates on easy rides. Set a wattage cap at LT1 (around 75 percent of FTP) on every prescribed zone 2 ride and treat going over as the actual error. Most head units support an upper-bound power alert. The gate works because it converts a feel decision into a beep, which is easier to obey. Riders who try to "just ride easier" without a gate drift up predictably; riders with the alert set at 188 watts on a 250-watt FTP stay within 5 watts of that ceiling. Fix two — separate hard days. The polarized-versus-threshold studies that produced the cleanest gains all separated intensity blocks by at least 48 hours of low-intensity riding [Neal et al. 2013, Stöggl & Sperlich 2014]. The amateur pattern of stacking a Tuesday group hammer next to a Wednesday "easy" ride that drifts into tempo collapses both sessions: Tuesday is incomplete because the next day was not recovery, and Wednesday is junk because the body was not fresh enough to ride it actually easy. Fix the calendar before fixing the ride. Fix three — anchor intensity to prescribed intervals, not group rides. If the week needs 60 minutes above 90 percent of FTP, schedule it as a structured session (4×8 at 92 percent, or 5×4 at VO2max) and complete it before joining any group ride. Stöggl and Sperlich's 9-week study across 48 well-trained athletes showed polarized training produced the largest gains in VO2peak and time-to-exhaustion versus threshold, high-intensity, or high-volume [Stöggl & Sperlich 2014]. The high-end dose has to be deliberate; a group ride that happens to land at threshold does not count. ## Common questions **Is junk intensity the same as the tempo zone?** The Coggan tempo zone runs 76-90 percent of FTP; sweet spot is the deliberate 88-94 percent slice within it [Allen et al. 2019]. Junk intensity is the same wattage band — roughly 76-87 percent — produced by drift rather than prescription. Sweet spot in 10-20 minute structured blocks is a real dose. The same wattage smeared across an unstructured 90-minute ride is not. The number on the head unit looks identical; the stimulus is not. **How do I audit my last month for junk intensity?** Pull time-in-zone for the last 4 weeks from Strava, Intervals.icu, or your training app. A polarized target is roughly 80 percent in zones 1-2, under 10 percent in zones 3-4, and 10-20 percent in zones 5+. Most amateurs land near 50/35/15 — far too much in the middle. The diagnostic question is what percentage of the last month sat between 76 and 87 percent of FTP. Anything over 15 percent is drift. **Does this matter as much for time-crunched riders?** Yes, more so. Time-crunched riders have less margin for any wattage that does not produce adaptation. A 6-hour week with 90 minutes parked in the middle has lost 25 percent of its training time to the lowest-yield zone. Carmichael's time-crunched approach shifts the distribution toward harder intervals because volume is constrained, but the rule still applies — the middle is the band time-poor riders can least afford to spend in. **What is the difference between sweet spot and junk intensity?** Sweet spot is 88-94 percent of FTP, prescribed in 10-30 minute blocks with structured warm-up and recovery, with a deliberate weekly dose. Junk intensity is 76-87 percent of FTP, produced by drift across an otherwise unstructured ride, with no recovery scaffolding around it. One is a tool; the other is a side effect. Sweet-spot blocks have a place in build phases. Junk intensity does not have a place anywhere. ## References 1. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 2. **Neal et al. 2013.** [Six weeks of a polarized training-intensity distribution leads to greater physiological and performance adaptations than a threshold model in trained cyclists](https://pubmed.ncbi.nlm.nih.gov/23264537/). Journal of Applied Physiology. 3. **Muñoz et al. 2014.** [Does polarized training improve performance in recreational runners?](https://pubmed.ncbi.nlm.nih.gov/23752040/). International Journal of Sports Physiology and Performance. 4. **Stöggl & Sperlich 2014.** [Polarized training has greater impact on key endurance variables than threshold, high intensity, or high volume training](https://pubmed.ncbi.nlm.nih.gov/24550842/). Frontiers in Physiology. 5. **Filipas 2024.** [Effects of a 16-Week Training Program with a Pyramidal Intensity Distribution on Recreational Male Cyclists](https://pubmed.ncbi.nlm.nih.gov/38251291/). Sports (MDPI). 6. **San Millán & Brooks 2018.** [Assessment of Metabolic Flexibility by Means of Measuring Blood Lactate, Fat, and Carbohydrate Oxidation Responses to Exercise in Professional Endurance Athletes and Less-Fit Individuals](https://link.springer.com/article/10.1007/s40279-017-0751-x). Sports Medicine. 7. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. --- # Why am I not getting faster in my second year of cycling? It's almost always junk intensity. URL: https://www.adaptcycling.com/guides/year-two-plateau-junk-intensity Updated: 2026-05-05 Author: Jim Camut Year one of structured riding produced an FTP jump, a leaner body, and the feeling that the training was working. Year two — same hours, same effort, often more — produced nothing. The plateau is rarely a periodization gap or under-recovery in isolation. The signature pattern, the one Strava data shows over and over in self-coached riders, is junk intensity: an inverted training distribution where almost every ride lives in the medium-hard tempo zone. This is the diagnostic piece. The fix is its own toolkit. ## Why year one was easy and year two stops working An untrained physiology adapts to almost any consistent stimulus. Year one's gains came from a low baseline, not from getting the training right. Year two demands distribution: the body is no longer a generalist that improves with any dose, it is a specialist that requires the right dose at the right end of the intensity spectrum. The first 6-12 months of structured riding produce big numbers because the rider is starting from an untrained baseline. Mitochondrial density, capillary network, stroke volume, and lactate-clearance machinery are all far below their genetic ceiling. Almost any consistent training — three rides a week of mixed intensity, a Saturday group ride, a Tuesday-night chain-gang — drives improvement because the body absorbs whatever stimulus it gets. FTP gains of 30-50 watts in a first structured year are common and almost entirely explained by this single fact. Year two is a different problem. The aerobic base now exists; what determines further improvement is whether the rider is overloading the right physiological system at the right time. Seiler's review of elite training [Seiler 2010] frames this as the difference between dose-tolerance and dose-specificity: at low fitness, total dose is what matters; at higher fitness, where the dose lands matters more than how big it is. The same 8-hour week that produced a 35-watt gain in year one will produce nothing in year two if 7 of those 8 hours are spent in the same intensity band. The Strava signature of a year-two plateau is recognizable. Weekly intensity factor (NP/FTP) clusters between 0.78 and 0.86, almost no rides land under 0.65, and almost no time accumulates above 0.95. The histogram should look polarized — a tall low-intensity column and a small but real high-intensity column, with very little in between. In year-two plateau riders it looks like a single bulge on top of the tempo zone. ## Junk intensity, defined Junk intensity is the medium-hard zone — roughly 76-90% of FTP, intensity factor 0.78-0.88, the upper end of zone 2 through low sweet spot. It accumulates fatigue without producing the mitochondrial adaptation of true low intensity or the neuromuscular and VO2max adaptation of true high intensity. The label is unkind but the physiology earns it. The mechanism is straightforward in domain terms. Burnley and Jones' work on the power-duration curve [Burnley & Jones 2018] places the boundary between moderate and heavy intensity domains at roughly the first lactate threshold, and the boundary between heavy and severe at critical power. Junk intensity sits in the middle of the heavy domain — above LT1, below CP. It is hard enough that lactate, heart rate, and ventilation rise above steady-state baseline; it is not hard enough to push VO2 to its slow-component ceiling or to drive the intramuscular signaling associated with severe-domain work. The adaptation problem is symmetric. San Millán and Brooks' work on mitochondrial function [San Millán & Brooks 2018] showed that the lower half of the heavy domain — true zone 2, below LT1 — is where mitochondrial biogenesis, fat oxidation, and lactate-clearance capacity improve fastest. Push above LT1 and the dominant signal becomes glycogen utilization rather than mitochondrial loading; the ride feels productive but the specific stimulus that builds aerobic ceiling is gone. At the other end, severe-domain intervals (above critical power, roughly 95-110% of FTP) are what drive VO2max and maximal lactate steady-state forward. Tempo riding is too easy to be one and too hard to be the other. The fatigue cost is the part most amateurs miss. A 90-minute ride at 0.82 IF feels good in the moment but writes 60-70 TSS into the system and leaves the legs measurably less capable of high-quality intervals 24-48 hours later. Stack three or four such rides into a week and the rider has produced enough fatigue to need a recovery day, with no single session that actually moved the ceiling. ## The 80/20 inversion that produces the year-two signature Elite endurance athletes spend roughly 80% of training time below LT1 and 15-20% above LT2, with very little in between [Seiler 2010, Stöggl & Sperlich 2014]. Self-coached amateurs invert it. They live in the middle and underdose both ends. The inversion is the single most common, most replicable cause of the year-two stall. The descriptive evidence is unusually consistent. Seiler's review of well-trained and elite athletes [Seiler 2010] and Stöggl and Sperlich's later synthesis [Stöggl & Sperlich 2015] both place the dominant pattern at 75-90% of training time at low intensity. The prospective trial that put numbers on the directional difference [Stöggl & Sperlich 2014] compared four distributions — high-volume low-intensity, threshold, HIT, and polarized — across 48 well-trained endurance athletes. The polarized group (68% low / 6% threshold / 26% high) produced the largest gains in VO2peak, time-to-exhaustion, and peak velocity. The threshold group, training closest to where most amateurs spend their week, gained the least. The recreational-cyclist data points the same direction with one nuance. A 16-week pyramidal-distribution trial in middle-aged recreational male cyclists [Filipas 2024] held riders to roughly 60% Z1, 30% Z2, 10% Z3 and produced a 38-watt gain at the 4 mMol lactate marker, with zone 2 time correlating most strongly with the improvement. The signal across both elite-polarized and recreational-pyramidal evidence is the same: the bottom of the distribution does the heavy lifting. Riders who skip it cap their own ceiling. This is the deeper-dive question one section of our pillar on the self-coached cyclist raises but cannot fully answer. Among the three usual year-two plateau causes — periodization gap, intensity drift, under-recovery — intensity drift is the most common and the most invisible. A self-coached rider can have a passable Base-Build-Peak structure, take recovery weeks, sleep well, and still stall, because every individual ride within the structure lands in the same medium-hard band. The structure was right. The intensity inside it was not. ## Why year two specifically: three converging traps The drift to junk intensity is not random. Three forces compound in year two that did not exist in year one: group rides with stronger riders, segment chasing now that the rider has fitness to hunt, and FTP-anchored zones drifting upward as fitness rises but lactate clearance does not. Each pulls the same direction. Group rides are the largest single contributor for most riders. A weekly chain-gang sets the pace at the front rider's threshold, not yours; sitting in still produces normalized power in the 0.85-0.95 IF range for two hours, which is a session, not a ride. One per week is fine and arguably useful. Two or three per week, with no genuine zone 2 between them, is a textbook junk-intensity profile. The rides feel hard and Strava's relative-effort number rewards the behavior — none of which is the same as adaptation. Segment chasing and FTP creep close the loop. A rider whose FTP rose from 220 to 260 watts in year one now has the fitness to push KOMs and the data to track them; every endurance ride becomes punctuated. Meanwhile the FTP itself, often estimated from a 20-minute test run hot, drifts upward faster than steady-state lactate clearance does, which means the prescribed zone 2 ceiling (roughly 75% of FTP) creeps into what is physiologically zone 3. The rider thinks they are riding zone 2 at 195 watts; their actual LT1 is closer to 175. Foster's monotony work [Foster 1998] catches the downstream effect — high training strain from week-after-week sameness — but the upstream cause is the intensity calibration drifting silently. The structural fixes are covered in our spoke on spotting and fixing junk intensity. ## Common questions **How do I know if my plateau is junk intensity and not under-recovery?** Look at the histogram of weekly time-in-zone for the last 8 weeks. Junk intensity shows a tall column in tempo (76-90% of FTP), a thin column under 65%, and very little above 95%. Under-recovery shows up differently: heart rate at fixed power has drifted up 5-10 bpm, mood and sleep degrade, and time-in-zone may even look reasonable. The two can coexist, but the distribution check is the first cut. **Is sweet spot training junk intensity?** Sweet spot (88-94% of FTP) is genuinely productive as a structured interval block one or two days a week with the rest of the week skewed low. It becomes junk when it is the default intensity of every ride. The label is about the distribution, not the zone — sweet spot inside an 80/20 frame is fine; sweet spot as the entire frame is the problem. **Can I still get faster in year two without going polarized or pyramidal?** Possibly, if total volume rises substantially. The threshold-heavy distribution underperforms polarized and pyramidal models in head-to-head trials [Stöggl & Sperlich 2014], but it is not zero. The realistic path for time-limited self-coached riders is not adding hours; it is reallocating the existing hours toward the bottom and top of the distribution and away from the middle. ## References 1. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 2. **Stöggl & Sperlich 2014.** [Polarized training has greater impact on key endurance variables than threshold, high intensity, or high volume training](https://pmc.ncbi.nlm.nih.gov/articles/PMC3912323/). Frontiers in Physiology. 3. **Stöggl & Sperlich 2015.** [The training intensity distribution among well-trained and elite endurance athletes](https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2015.00295/full). Frontiers in Physiology. 4. **Filipas 2024.** [Effects of a 16-Week Training Program with a Pyramidal Intensity Distribution on Recreational Male Cyclists](https://pubmed.ncbi.nlm.nih.gov/38251291/). Sports (MDPI). 5. **San Millán & Brooks 2018.** [Assessment of Metabolic Flexibility by Means of Measuring Blood Lactate, Fat, and Carbohydrate Oxidation Responses to Exercise in Professional Endurance Athletes and Less-Fit Individuals](https://link.springer.com/article/10.1007/s40279-017-0751-x). Sports Medicine. 6. **Burnley & Jones 2018.** [Power-duration relationship: Physiology, fatigue, and the limits of human performance](https://pubmed.ncbi.nlm.nih.gov/27806677/). European Journal of Sport Science. 7. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. --- # How to plan cycling training around shift work, a baby, or weekly travel URL: https://www.adaptcycling.com/guides/plan-around-shift-work-baby-travel Updated: 2026-05-05 Author: Jim Camut I went from racing professionally in Europe to training around two daughters and a job that owns my calendar. The standard 7-day cycling plan — Tuesday hard, Thursday VO2, Saturday long — assumes a normal week. If your week rotates 12-hour shifts, includes 3am feeds, or puts you on a Tuesday red-eye, that plan is broken before Monday. This is the planning approach that works when Tuesday is sometimes Saturday, sleep is the actual limiter, and the season still has to happen. ## Why the standard 7-day plan assumes a life you do not have Standard plans are built on a 7-day microcycle with fixed hard and easy days, which assumes a stable sleep-wake schedule and predictable Saturday availability. Riders on rotating shifts, with infants, or on weekly travel violate every one of those assumptions. The plan does not break because you are weak; it breaks because its load and recovery model is calibrated for someone you are not. Every off-the-shelf plan — TrainerRoad's plan-builder, TrainingPeaks templates, JOIN's calendar, Wahoo SYSTM blocks — schedules intensity on fixed weekdays. The architecture assumes Tuesday and Thursday have similar sleep and similar fatigue, Saturday is reliably long-ride day, and recovery between sessions is roughly constant. For a rotating-shift nurse, an infant's parent, or a Monday-Friday business traveler, none of those constants hold. Saturday with a teething 7-month-old is not Saturday at home alone with a coffee. The plan-time-irregularity problem also has a sleep cost the plan never sees. Roberts and colleagues' 2019 cycling-specific trial put endurance athletes through self-paced time trials with and without a night of total sleep deprivation; the post-deprivation trial was meaningfully slower than the rested baseline [Roberts et al. 2019]. Van Dongen's two-week dose-response study found that chronic restriction to 6 hours per night produced cognitive deficits equivalent to two nights of total sleep deprivation, and the subjects rated themselves as essentially unimpaired throughout [Van Dongen et al. 2003]. The ratings matter because most self-coached riders use perceived effort to gauge whether to do the planned session. If the perception is calibrated wrong, the decision is wrong. This is the part of self-coaching that gets harder when life is irregular: the broader playbook of the self-coached cyclist relies on the rider being a good judge of readiness, and irregular sleep specifically degrades that judgment. The plan that works for a regular week becomes actively misleading when the week is not. ## Sleep is the limiter — what shift work, an infant, and red-eyes actually cost Shift work, parental sleep disruption, and transmeridian travel are not three flavors of busy. They are three flavors of sleep and circadian disruption, and the performance cost is well-quantified. Treat them as physiology problems first and scheduling problems second; the planning approach falls out of the physiology. Parental sleep loss persists longer than most cyclists assume. Richter and colleagues tracked thousands of parents in the German Socio-Economic Panel across pregnancy and the postpartum period; sleep duration in the first three months after a first birth dropped roughly an hour per night for mothers and around 13 minutes for fathers, and sleep satisfaction did not return to pre-pregnancy levels for up to six years [Richter et al. 2019]. The functional implication for a self-coached new parent: not just less time to train, but a different physiological substrate to train on — submaximal endurance, glucose handling, immune function, and cognition all move in the wrong direction when sleep is chronically short [Roberts et al. 2019, Van Dongen et al. 2003]. Shift work compounds the disruption with circadian misalignment. Rotating between days and nights forces the body to keep restarting its circadian clock, and the clock realigns roughly one time-zone-equivalent per day. A nurse coming off four nights trying to do Saturday's tempo session is doing that ride with circadian timing several hours out of alignment. Forbes-Robertson and colleagues' 2012 review documents the same effect in transmeridian travel — peak power, reaction time, and submaximal heart rate all show measurable circadian disruption that does not normalize for several days after large schedule shifts [Forbes-Robertson et al. 2012]. The travel case is the cleanest version of the problem. A Monday red-eye eastbound to a client meeting plus a Thursday return is, physiologically, two circadian shifts plus two short nights inside one week. Stacking a Saturday VO2 session on top of that produces lower power output, higher RPE at the same wattage, and elevated injury risk — the same fatigue spike on a decompressed chronic base that drives the acute:chronic workload pattern Hulin and colleagues documented in elite rugby league [Hulin et al. 2016]. The plan does not know any of this happened. The rider has to. ## From the 7-day week to the 14-day rolling block The fix is structural: stop treating the week as the planning unit and start treating a rolling 10-14 day block as the unit. Pre-commit to weekly volume and intensity targets — not which day they land on. Then place the work into the block as the schedule reveals itself, prioritizing frequency of low-intensity rides and protecting one or two harder sessions wherever the gap is. The principle is grounded in the training-intensity-distribution literature. Seiler's 2010 review of how successful endurance athletes actually train converged on roughly 80% low-intensity, 20% high-intensity time across descriptive studies of athletes training 10-13 times per week — the ratio matters more than which calendar day each session falls on [Seiler 2010]. A 10-day block targeting 8 hours of zone 2 and 90 minutes of high-intensity work can be assembled in many valid orders: Tuesday 4x8 plus Thursday VO2, Saturday 4x8 plus Tuesday VO2, or three short zone 2 rides plus a weekend hard session plus a mid-block VO2 set on whichever morning had real sleep behind it. Inside the rolling block, three rules keep the structure honest. First, frequency over duration when minutes are unpredictable — six 45-minute zone 2 rides in 10 days produces more aerobic adaptation than two 2-hour rides plus four skipped sessions, and the smaller bites survive contact with a baby's nap schedule. Second, place hard sessions where sleep was actually adequate — the operational rule is two consecutive nights of 7+ hours before any threshold or VO2 work, no exceptions for shift workers and new parents. The Roberts and Van Dongen data make this an injury-and-illness call, not a softness call [Roberts et al. 2019, Van Dongen et al. 2003]. Third, the weekly acute:chronic workload ratio still applies — keep the trailing 7-day TSS between 0.8 and 1.3 times the trailing 28-day average [Hulin et al. 2016]. What gets cut first when minutes are short is not the hard work — it is the middle. Tempo and upper-zone-2 endurance rides drop first, because they cost recovery without delivering the high-intensity stimulus. The 80/20 distribution preserves itself naturally when minutes shrink: short easy rides plus one or two short hard sessions. Mujika and Padilla's detraining review documents that maintaining intensity while cutting volume preserves most cardiovascular adaptation across weeks of reduced training [Mujika & Padilla 2000a]. The rolling block is the adult-life version of that finding. ## When AI adaptive coaching beats a static plan for irregular athletes Static plans require the rider to do the rolling-block math themselves every week — calculating remaining work, deciding which session to drop, watching the workload ratio. That is exactly the cognitive overhead a sleep-deprived parent or shift worker has the least of. Adaptive software earns its keep here, not on the regular weeks. TrainerRoad's Adaptive Training adjusts workout difficulty inside a fixed plan structure, but the structure itself stays calendar-anchored. Xert's continuous-fitness-signature model is closer to the rolling-block ideal but tilts toward recommendation rather than coordinated planning. JOIN restructures more aggressively when life intrudes. None are built specifically for the irregular-schedule case, and the gap shows on weeks with three sleep-disrupted nights and a midweek schedule flip. The right product reads what was actually slept and ridden, holds a 10-14 day forward view, and recomputes the placement of remaining hard work every time the schedule changes. AdaptCycling is built for this case. A Wahoo, Hammerhead, Polar or intervals.icu account connects on sign-in. The plan re-shapes when you tap not today, log a poor night, or take an unplanned group ride. The hard session moves to where the sleep was. The acute:chronic ratio is computed live as a guardrail. The rider does not have to do the math; the rider has to ride the bike and tell the coach what happened. For the sleep-short, time-poor, schedule-irregular athlete, that is the difference between the plan surviving the year and getting abandoned in March. ## Common questions **Can I run a real training plan on 5 hours of sleep?** Not the plan you would run on 7-8. The cycling-specific data shows meaningful performance loss on time trials after total sleep deprivation, and Van Dongen's chronic-restriction study showed cognitive deficits build silently across two weeks of restriction without the rider noticing [Roberts et al. 2019, Van Dongen et al. 2003]. The honest move is to keep frequency, drop intensity, and treat any week with two or more 5-hour nights as a recovery week regardless of the calendar. **What is the minimum I need to do to maintain fitness through a hard 6-week stretch?** Three to four rides per week, mostly zone 2, with one short hard session if sleep allows it. Mujika and Padilla's detraining review documents that intensity preserved and volume cut maintains most cardiovascular adaptation across weeks of reduced training; the loss only accelerates when intensity disappears entirely or training stops [Mujika & Padilla 2000a]. Six weeks of three-rides-per-week with one short VO2 session is not a great training block — but it is not a detraining block either. **How do I plan around a rotating shift schedule?** Plan in 14-day blocks aligned to the shift rotation, not 7-day calendar weeks. Place hard work on the second day after a stretch of days off or on the back end of a day-shift run — never on a swing day or the first day after nights. Two consecutive nights of 7+ hours before any threshold or VO2 work is the operational rule. The acute:chronic workload ratio still applies; compute it across the rolling 28 days rather than fixed week boundaries [Hulin et al. 2016]. **Should I move my long ride if I am traveling Saturday?** Yes — and the long ride is the cheapest session to relocate inside a rolling block. The aerobic adaptation comes from accumulated zone 2 minutes across the block, not from a single weekend ride. Forbes-Robertson's review documents that a circadian shift of 3+ time zones disrupts power output for several days after arrival [Forbes-Robertson et al. 2012], so a Tuesday-arrival traveler's first useful long ride is often Friday or after the return. ## References 1. **Roberts et al. 2019.** [Effects of total sleep deprivation on endurance cycling performance and heart rate indices used for monitoring athlete readiness](https://pubmed.ncbi.nlm.nih.gov/31526108/). Journal of Sports Sciences. 2. **Van Dongen et al. 2003.** [The cumulative cost of additional wakefulness: dose-response effects on neurobehavioral functions and sleep physiology from chronic sleep restriction and total sleep deprivation](https://pubmed.ncbi.nlm.nih.gov/12683469/). Sleep. 3. **Richter et al. 2019.** [Long-term effects of pregnancy and childbirth on sleep satisfaction and duration of first-time and experienced mothers and fathers](https://pubmed.ncbi.nlm.nih.gov/30649536/). Sleep. 4. **Forbes-Robertson et al. 2012.** [Circadian disruption and remedial interventions: effects and interventions for jet lag for athletic peak performance](https://pubmed.ncbi.nlm.nih.gov/22299812/). Sports Medicine. 5. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 6. **Hulin et al. 2016.** [The acute:chronic workload ratio predicts injury: high chronic workload may decrease injury risk in elite rugby league players](https://pubmed.ncbi.nlm.nih.gov/26511006/). British Journal of Sports Medicine. 7. **Mujika & Padilla 2000a.** [Detraining: loss of training-induced physiological and performance adaptations. Part I: short term insufficient training stimulus](https://pubmed.ncbi.nlm.nih.gov/10966148/). Sports Medicine. --- # Weekly TSS target for an amateur cyclist: the number to aim for and the number to ignore URL: https://www.adaptcycling.com/guides/weekly-tss-target-self-coached Updated: 2026-06-16 Author: Jim Camut The honest answer to 'what weekly TSS should I target' is that the question is usually backwards. TSS is a derived metric — duration multiplied by intensity-squared, scaled to FTP — not a primary target. Chase a TSS number directly and the cheapest way to hit it is to ride harder, which is the failure mode that defines an amateur plateau. The right targets are weekly hours and an intensity distribution that respects the 80/20 rule. The TSS range falls out of those two. ## What 'weekly TSS target' actually answers — and what it doesn't TSS is a load proxy, not a load goal. Coggan and Allen built it to score what an athlete already did, so historical loads could be compared and chronic-load curves modeled [Allen et al. 2019]. The published category benchmarks — 300-450 TSS/week for a Cat 4, 500-700 for a Cat 2, 700-1000+ for an elite — are descriptive averages of what those riders happened to produce, not prescriptions of what an amateur should target on Monday. The category-by-TSS charts circulating on TrainingPeaks and YouTube descend from observational ranges Coggan published alongside the original Performance Manager Chart [Allen et al. 2019]. They are useful as sanity checks — a Cat 3 averaging 200 TSS/week is undertrained for the category; one averaging 900 is probably riding too hard — but they describe the steady-state output of riders training appropriately, not the lever an athlete should pull to get there. Mujika's 2017 review of training-load quantification methods [Mujika 2017] is explicit that external load metrics like TSS are designed to estimate the biological stress an athlete experienced, not to prescribe the stress they should impose. Inverting the relationship is a category error. The practical consequence is that targeting a TSS number directly invites the wrong fix. Below your target on Sunday? The fastest way to close the gap is to ride harder rather than longer — adding 30 minutes of zone 2 produces about 30 TSS, while replacing a 60-minute zone 2 ride with 60 minutes at sweet spot produces about 80. The TSS scoreboard rewards the second behavior even though the second behavior is precisely the intensity drift Seiler's polarized-distribution research [Seiler 2010] and Stöggl and Sperlich's descriptive analysis of well-trained athletes [Stöggl & Sperlich 2015] identify as the amateur's most common training error. This sits inside the broader self-coached cyclist playbook: when there is no coach holding the periodization line, the rider has to know which metrics to act on and which to read as feedback. ## Why hours-per-week is the better primary target for amateurs Weekly hours is the input variable an amateur actually controls. It is bounded by life, scales linearly with low-intensity volume, and is the single strongest correlate of long-term endurance progression in the descriptive literature [Seiler 2010, Stöggl & Sperlich 2015]. A defensible weekly hours target with a fixed 80/20 intensity ratio produces a TSS range automatically — and it produces the right one. Run the math forward instead of backward. Six hours per week at 80% low intensity and 20% high intensity produces about 310 TSS — the bottom of the Cat 4 range. Eight hours produces about 415. Ten hours produces about 520, the bottom of the Cat 3/2 range. The TSS number is the output of the time budget and the distribution, not the input. A rider who fixes hours and distribution lands in a sustainable TSS band by construction; a rider who fixes TSS and lets distribution float hits the number by drifting upward in intensity. Holding that distribution is easiest when the hard days ride to explicit watt targets rather than feel — a power zones calculator gives you the seven bands for your FTP so "20% high intensity" means a specific number, not a vibe. Hours-as-target also handles the realities the TSS framing hides. A 90-minute group ride that averaged a normalized 0.85 IF scores roughly 110 TSS — the same as a structured 75-minute threshold session, even though the cardiovascular and recovery costs differ meaningfully. Counting the hour against a weekly-hours budget treats both honestly; counting the TSS against a weekly-TSS target lets the rider double-book the intensity bucket and call it a productive week. Mujika's quantification review [Mujika 2017] is direct on this point: external load proxies should be paired with internal-load monitoring (heart rate at power, RPE) for exactly this reason. AdaptCycling builds plans against a weekly hours ceiling rather than a TSS target, and re-derives expected weekly TSS from the prescribed sessions. TrainingPeaks lets a coach do the same manually; Intervals.icu surfaces both but enforces neither; TrainerRoad's plan builder is closer to a session-count target. None of these tools are wrong — they optimize different variables. The self-coached rider has to pick which variable to be honest about, and weekly hours is the one life actually constrains. ## How to set a defensible weekly TSS range from your own history Forget the category charts. Pull your last 12 weeks from Strava, Intervals.icu, or TrainingPeaks, drop the highest and lowest weeks, and take the median of the remaining ten. That number — call it your sustainable baseline — is the floor of your weekly range. The ceiling is roughly 1.3-1.4x that baseline, capped by an acute:chronic workload ratio under ~1.5 [Hulin et al. 2016]. The reason to use your own data rather than a Cat 4 chart is that the chart describes a rider whose life, sleep, work stress, and training age are not yours. A 38-year-old parent with a full-time job who has been training consistently for three years has a different sustainable load than a 24-year-old with eight hours of free evenings, even if their FTPs are within 10 watts of each other. Foster's monotony framework [Foster 1998] is built on this directly — illness and overreaching across his 25-athlete sample correlated with the product of weekly load and daily-load variance, not with any absolute TSS number. The relevant baseline is yours, computed from weeks where your training actually went well. Apply the floor and ceiling as guardrails, not targets. A week that climbs above the 1.5x acute:chronic ratio — the threshold for elevated injury risk in Hulin's rugby-league data and one Halson's monitoring review highlights as actionable [Halson 2014] — is the cue to hold the next week flat or pull a recovery week forward. The manual check is whether this week's TSS is more than 1.5x the trailing four-week average. The recovery-week mechanics that follow are covered in our spoke on recovery-week cadence for self-coached riders. What this baseline does not do is grow on its own. Coggan and Allen's framework operationalizes 5-7 TSS/day per week as the upper-safe progression for a build phase, with a deload before resuming the ramp [Allen et al. 2019]. Stack three weeks of progression against a recovery week and the new sustainable baseline is roughly 5-15% above the old one. That is the actual rate at which an amateur's TSS ceiling moves. Anything faster is borrowed. ## When the TSS number is lying to you TSS is a power-meter metric scored against your FTP. Four common scenarios degrade its accuracy enough to mislead a self-coached rider: HR-only rides without a power meter, mountain-bike and gravel rides with prolonged coasting and surges, group rides with high variability index, and a stale FTP that has drifted from reality. In each case the number on the scoreboard is not the load on the body. On HR-only rides, platforms estimate TSS from heart-rate-time-in-zone. The estimate is reasonable for steady-state aerobic work but systematically under-counts very short, very high efforts and over-counts extended hot or dehydrated rides. Halson's monitoring review [Halson 2014] is direct that internal-load methods like HR-derived load are best treated as complements to power-derived load, not equivalents. Treating an hrTSS of 90 and a power-TSS of 90 as the same training stimulus is a small lie that compounds over a season. Off-road riding produces the opposite distortion. A mountain-bike ride with frequent coasting and hard surges produces a normalized power that under-represents the neuromuscular cost — short supra-threshold spikes are flattened by the 30-second rolling average, and TSS often runs 20-30% lower than the rider's perceived load. Group rides have a related problem: high variability index inflates normalized power, which inflates IF, which squares into TSS. The number is technically correct and practically misleading. The fourth distortion is the one self-coached riders miss most often. TSS scales as IF squared, so a stale FTP that is 8% low inflates every TSS score by about 17%. A rider who has not retested in three months and has gained fitness will see weekly TSS climb on identical rides. Allen and Coggan recommend updating FTP whenever a 95%-of-20-minute or 60-minute power best is set [Allen et al. 2019]; AdaptCycling, Intervals.icu, and Xert auto-update FTP from the power curve to defend against this drift. If the TSS line is climbing and perceived effort is not, suspect the denominator before celebrating the numerator. ## Common questions **What weekly TSS should a Cat 4 amateur cyclist target?** The descriptive range is roughly 300-450 TSS/week, but treat that as a sanity check, not a target. Set weekly hours you can sustain (5-8 hours is typical for a working amateur), hold an 80/20 intensity distribution, and let TSS land where it lands. A rider hitting 450 TSS/week from six hours is producing the number through intensity drift — the most common amateur plateau mechanism [Seiler 2010, Stöggl & Sperlich 2015]. **How much can I increase my weekly TSS safely?** Coggan and Allen's chronic-load framework keeps weekly CTL increases inside roughly 5-7 TSS/day during a build phase, with a recovery week every third or fourth week cutting volume 30-40% before the ramp resumes [Allen et al. 2019]. In weekly-TSS terms, that is roughly 35-50 TSS/week of growth across a build before unloading. Crossing an acute:chronic workload ratio of about 1.5 — current week vs trailing four-week average — is the cleanest practical signal that the ramp is too steep [Hulin et al. 2016]. **Should weekly TSS be different in Base vs Build vs Peak?** Yes, but the distribution shifts more than the total. Base weeks are higher hours at lower IF — same TSS, more time, easier intensity. Build weeks compress toward event-specific intensity and TSS climbs as IF rises. Peak weeks cut volume 40-60% during a two-week taper while preserving intensity, so peak-week TSS is meaningfully lower by design. Riders chasing a fixed weekly TSS year-round flatten the periodization curve their fitness depends on. **Why does my Strava TSS look different from my TrainingPeaks TSS?** Three reasons usually. First, the platforms can read slightly different FTP values, and TSS scales with FTP-squared. Second, Strava's relative-effort and TrainingPeaks' TSS use different smoothing windows for normalized power. Third, hrTSS estimates differ because the heart-rate-to-load model varies. Trust the platform with your most current FTP and most reliable power data, and read the TSS column for trends rather than absolute precision. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 3. **Stöggl & Sperlich 2015.** [The training intensity distribution among well-trained and elite endurance athletes](https://pmc.ncbi.nlm.nih.gov/articles/PMC4621419/). Frontiers in Physiology. 4. **Mujika 2017.** [Quantification of training and competition loads in endurance sports: methods and applications](https://pubmed.ncbi.nlm.nih.gov/27918666/). International Journal of Sports Physiology and Performance. 5. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 6. **Hulin et al. 2016.** [The acute:chronic workload ratio predicts injury: high chronic workload may decrease injury risk in elite rugby league players](https://pubmed.ncbi.nlm.nih.gov/26511006/). British Journal of Sports Medicine. 7. **Halson 2014.** [Monitoring training load to understand fatigue in athletes](https://pubmed.ncbi.nlm.nih.gov/25200666/). Sports Medicine. --- # Safe CTL ramp rate for the self-coached amateur cyclist URL: https://www.adaptcycling.com/guides/ctl-ramp-rate-amateur-safe Updated: 2026-05-05 Author: Jim Camut The directional answer is 3-7 CTL points per week, with working-age amateurs landing near the bottom of that band and pros near the top. The number is not a law — it is the operational midpoint of three converging research lines (Banister/Coggan modeling, Foster's monotony curve, and acute:chronic workload data) that all break down past roughly 8 points per week. Where the number comes from, why amateurs need a tighter envelope, what the ramp looks like over twelve weeks, and the signals that say slow down regardless of the calendar. ## Where the 5-7 CTL points per week guidance actually comes from Three independent frameworks land in the same band. The Banister-derived impulse-response math gives a sustainable rate-of-change ceiling, Coggan's Performance Manager Chart operationalizes it as CTL with a 42-day time constant, and the acute:chronic workload literature flags the upper bound. The convergence is around 5-7 CTL points per week as a sustainable amateur ramp — directional, not gospel. CTL is the exponentially-weighted 42-day average of daily TSS, originating in the Performance Manager Chart documented in the canonical power-meter text [Allen et al. 2019]. The 42-day time constant traces back to Banister's original fitness-fatigue impulse-response model [Hellard et al. 2006], where the slow-decay fitness component is modeled with a time constant in the 30-50 day range across endurance studies. The practical implication: a rider's CTL is a roughly six-week rolling window of training stress, and meaningful change in it requires weeks of sustained input. The widely-cited 5-8 CTL/week guidance from coaching literature lands in this range: weekly load increases that drive CTL up faster than 8 points per week generally cannot be sustained without breaking the rider, and rates much under 3 are too slow to drive adaptation in a finite training season. More conservative practitioner guidance for working-age amateurs pegs the band at 3-5 CTL/week, citing life stress, sleep variance, and recovery capacity as the binding constraint rather than power-to-weight. Both numbers describe the same underlying physiology; they differ on where to draw the conservative line for a given athlete population. The third line of evidence comes from acute:chronic workload research. Hulin and colleagues' rugby-league data placed elevated injury risk above an acute:chronic workload ratio of roughly 1.5, with the 1.0-1.3 band sitting in the safe zone [Hulin et al. 2016]. A 5-7 CTL/week ramp on a base of 50 CTL implies an acute load roughly 1.1-1.2x the chronic load — squarely inside that safe band. Push the ramp to 10 CTL/week and the ratio drifts toward 1.4-1.5; push past that and the rider is in territory where every additional point of weekly load is bought at increasing biological cost. The guidance is directional, not the second decimal — TrainingPeaks, Intervals.icu, AdaptCycling, and WKO5 will all show slightly different CTL numbers for the same rider depending on time-constant defaults and load inputs. ## Why amateur ramp rate is different from pro ramp rate Pros at a CTL of 100+ can ramp 7-10 points per week through a build because they sleep nine hours, eat to fuel, train without job stress, and have a decade of cumulative training-age in the bank. Amateurs at a CTL of 40-60 with a job, a family, and 7 hours of sleep cannot. The recovery infrastructure is the binding constraint, not the wattage. The math is symmetric for pros and amateurs. The biology is not. Aubry and colleagues' overload trial in trained triathletes showed that 11 of 23 athletes who completed a deliberate three-week overload block became functionally overreached [Aubry et al. 2014]. Those triathletes were trained and well-rested; the amateur running the same load on six hours of variable sleep with a job and two kids crosses from functional to non-functional overreaching at a much lower stress than the trial population did. Sleep is the largest single moderator. Published ramp-rate guidance assumes 8-9 hours of consistent sleep — a research-condition assumption few self-coached amateurs meet. Life stress is the second moderator and is invisible to CTL entirely. The Foster monotony framework [Foster 1998] flags load-and-monotony product as the strongest behavioral predictor of breakdown, but no smartwatch knows the rider had a brutal work week, a sick kid, and a long flight stacked on the training. Real-life stress compresses the safe ramp envelope without showing up in any standard load metric. Training age is the third compressor. A first-year structured rider on a 35-CTL base does not have the years of cumulative aerobic adaptation that lets an experienced rider absorb 7 CTL/week. The practical translation: novices and returning-after-layoff riders should target the bottom of the 3-5 CTL/week band; consistent multi-year trained amateurs at moderate load can safely sit at 5-7; pros and full-time athletes are the only group for whom 8-10 makes sense, and even then only inside a deliberate overload block followed by a real taper. This is part of the broader skill set of training without a coach (covered in our pillar on the self-coached cyclist) — calibrating the math to your actual life rather than to a textbook research subject. ## What the ramp actually looks like: a 12-week worked example A 12-week build from CTL 45 to CTL 65 illustrates the math. Three weeks of progressive load, one recovery week, repeated. Weekly TSS climbs from ~340 to ~530 across the build weeks; recovery weeks drop to ~60-70% of preceding peak. Average CTL gain is ~1.7 points/week across all twelve weeks, but ramps to 5-6 points/week inside the build weeks themselves. A representative schedule for a rider starting at 45 CTL and 8 hours/week: build weeks 1-3 at 350/400/440 TSS, recovery week 4 at 280 TSS. Build weeks 5-7 at 460/500/530 TSS, recovery week 8 at 350 TSS. Build weeks 9-11 at 510/540/560 TSS, recovery week 12 at 380 TSS. The build-week ramp is roughly 5-6 CTL points/week of growth; the recovery weeks pull the average back to the sustainable ~1.7 points/week mean. This is the block-periodization pattern in practice — concentrated load followed by a deliberate unload that exploits the lag between fatigue dissipation and fitness retention. The recovery week is the part that makes the ramp sustainable, not optional. The reason the schedule reads this way and not as a smooth weekly increase: the acute:chronic workload ratio stays inside the safe band of roughly 1.0-1.3 [Hulin et al. 2016] on every single week of the plan. A continuous 5-CTL/week ramp without recovery weeks pushes the ratio above 1.4 by week four and approaches 1.5 by week six. The 3-on/1-off cadence keeps the ratio in-band by design — the recovery week is the pressure-relief valve for ramp rate, not just a physiological one. Skipping it is the most common amateur error and the one that converts a planned build into non-functional overreaching [Meeusen et al. 2013]. ## The override signals: when to slow the ramp regardless of what the calendar says Four signals override the planned ramp. RPE creep at fixed wattage. Heart-rate-at-power drift over 7-10 days. Foster monotony score above 2.0. Acute:chronic workload ratio crossing 1.5. Two or more concurrent signals is the threshold to flatten the next build week or pull the recovery week forward — regardless of what the calendar says. The cleanest single marker is heart-rate-at-power drift on a repeated zone 2 ride. A 5-10 bpm sustained departure from baseline at the same wattage, in similar conditions, over 7-10 days is meaningful sympathetic-system fatigue and the cheapest early-warning signal a self-coached rider has [Foster 1998]. RPE creep — the same prescribed wattage feeling subjectively harder week-over-week — is the second signal and the one self-coached riders most often dismiss. Both often appear before any measurable performance drop. Foster's monotony score (mean weekly TSS / standard deviation of daily TSS) crossing 2.0 is the quantitative early-warning [Foster 1998]. A rider on the same Tuesday-Thursday-Saturday template with no variance produces the high-load high-monotony pattern Foster's data flagged. The acute:chronic workload ratio crossing 1.5 is the fourth signal [Hulin et al. 2016] — but treat it cautiously. Impellizzeri and Tenan's 2020 critique argues that ACWR has serious statistical problems and should not be treated as a clinical predictor [Impellizzeri & Tenan 2020]. Use it as one input among several, not as a verdict. Two concurrent signals over a 7-10 day window is the action threshold. One signal might be noise. Two is a pattern. The response is not heroic: cap the next build week at the prior week's TSS, or pull the scheduled recovery week forward. Skipped recovery weeks are the dominant path from intentional overload into non-functional overreaching [Meeusen et al. 2013] — covered in our spoke on telling overtraining from ordinary fatigue, and the recovery-week cadence itself is detailed in our spoke on how often a self-coached cyclist should take a recovery week. ## Common questions **Is 5 CTL points per week safe if I am brand new to structured training?** Probably too aggressive for the first 8-12 weeks. Novices building from a CTL under 30 should target 2-4 CTL/week and prioritize ride consistency over absolute load. The aerobic, musculoskeletal, and recovery-system adaptations that let a trained rider absorb 5-7 CTL/week take roughly two seasons to build. Ramping faster than your tissues can adapt is the fastest way to turn a promising first season into the injury that ends it. **What if my CTL drops because I missed a week — can I ramp faster to catch up?** No, and this is the trap most self-coached riders fall into. After a 7-14 day disruption, the safe ramp rate is actually lower than baseline because chronic load has dropped — the ratio climbs faster on the way back up. Target 3-4 CTL/week for the first three weeks back, then resume the normal 5-7 cadence. What matters is the ratio of acute to chronic load, not the raw weekly TSS. **Why do TrainingPeaks, Intervals.icu, and AdaptCycling show different CTL numbers for the same week?** Different time-constants, different handling of non-power rides (TSS vs. hrTSS vs. relative effort), and different rules for filling missed days. The numbers usually fall within ~5% of each other on a rider with consistent power data. Ramp rate within a single platform is more meaningful than the absolute number — pick one tool, stay with it for a season, and judge against its own history. **Should I trust ACWR as the primary signal that my ramp is too fast?** Use it as one input, not the verdict. Hulin's 2016 work established the 1.0-1.3 safe band as a useful directional guide, but Impellizzeri and Tenan's 2020 critique laid out serious statistical problems with ACWR as a clinical predictor [Impellizzeri & Tenan 2020]. An ACWR climbing past 1.5 is a flag worth checking against other signals — not a diagnosis. ACWR drift plus RPE creep plus heart-rate drift is far more meaningful than any one alone. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.penguinrandomhouse.com/books/611232/training-and-racing-with-a-power-meter-by-hunter-allen-and-andrew-coggan-and-stephen-s-mcgregor/). VeloPress. 2. **Hellard et al. 2006.** [Assessing the limitations of the Banister model in monitoring training](https://pubmed.ncbi.nlm.nih.gov/16608765/). Journal of Sports Sciences. 3. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 4. **Hulin et al. 2016.** [The acute:chronic workload ratio predicts injury: high chronic workload may decrease injury risk in elite rugby league players](https://pubmed.ncbi.nlm.nih.gov/26511006/). British Journal of Sports Medicine. 5. **Impellizzeri & Tenan 2020.** [Acute:Chronic Workload Ratio: Conceptual Issues and Fundamental Pitfalls](https://pubmed.ncbi.nlm.nih.gov/32502973/). International Journal of Sports Physiology and Performance. 6. **Aubry et al. 2014.** [Functional overreaching: the key to peak performance during the taper?](https://pubmed.ncbi.nlm.nih.gov/25134000/). Medicine & Science in Sports & Exercise. 7. **Meeusen et al. 2013.** [Prevention, diagnosis, and treatment of the overtraining syndrome: joint consensus statement of the European College of Sport Science and the American College of Sports Medicine](https://pubmed.ncbi.nlm.nih.gov/23247672/). Medicine & Science in Sports & Exercise. --- # The cheapest way to self-coach cycling in 2026 URL: https://www.adaptcycling.com/guides/cheapest-way-to-self-coach-2026 Updated: 2026-05-05 Author: Jim Camut The cheapest credible self-coaching stack in 2026 is mostly free. Strava reads your rides at no cost, Intervals.icu gives you the same CTL/ATL/TSB charts coaches paid four figures for in 2010, GoldenCheetah is open-source, and a used power meter is sub-$300. The honest question is not "how do I get this for free." It's "which one piece is worth paying $10-15 a month for, and where do the false economies hide?" This is the answer. ## What's actually free in 2026 — and good enough The free stack covers most of self-coaching. Strava free reads every ride and computes summary metrics. Intervals.icu does CTL, ATL, TSB, fitness charts, and power-curve analysis at no cost. GoldenCheetah is open-source desktop analytics with critical-power modeling. A used single-sided power meter is roughly $200-$300. None of those line items costs anything monthly. Strava free is the data layer. Every ride uploads, every segment compares, every device syncs. The paid Strava tier (around $12/month or $80/year) adds segment leaderboards and route planning, but for training purposes the free tier is sufficient — the metrics that matter for self-coaching live downstream. Intervals.icu is the analytics layer. Free, no trial limit, no credit card. It pulls from Strava and computes the standard load-and-fitness model [Allen et al. 2019] — CTL, ATL, TSB, weekly intensity distribution, power curve, eFTP estimates from the rides you already did. Optional supporter tier is around $4/month if you want to fund the project. As an analytics surface it is competitive with TrainingPeaks Premium at roughly $20/month. Hardware. Used Stages, 4iiii, and Favero single-sided power meters trade in the $200-$300 range; new entry-level units like the 4iiii Precision sit roughly $300-$500. Most modern smart trainers come with built-in power. A heart-rate strap is $40-$70. None of this needs to be premium for the data to be actionable. Books. The canon is cheap and one-time. Allen and Coggan's Training and Racing with a Power Meter [Allen et al. 2019], Friel's Cyclist's Training Bible [Friel 2018], and Carmichael and Rutberg's Time-Crunched Cyclist [Carmichael & Rutberg 2017] cover almost every periodization decision a self-coached rider has to make. Together they cost less than two months of TrainerRoad. ## Where the false economies hide Free isn't free if you misuse it. Three patterns catch most self-coached riders: treating Zwift or TrainerRoad as a coach, treating Garmin's Daily Suggested Workouts as a periodized plan, and asking ChatGPT for a 12-week plan it cannot adapt. Each looks cheap up front and costs you a season. Zwift and TrainerRoad are workout libraries, not coaches. Both sit in the roughly $20/month range and both ship excellent structured workouts and indoor experiences. TrainerRoad's Adaptive Training adjusts difficulty inside its plans. Neither restructures your week when a sinus infection wipes Wednesday. Treating either as your full coach means you get great individual sessions inside a structure that does not flex around real life. Garmin Daily Suggested Workouts are free with the watch and reasonable as a daily nudge. They are not a periodized plan with a goal event, a base phase, or a taper. Following DSW for a year is closer to unstructured riding with intensity guidance than to coaching yourself. For a self-coached rider chasing an August event, that gap matters — the periodization invariants the parent pillar describes [Friel 2018, Foster 1998] cannot be enforced one day at a time. ChatGPT writing your training plan is the cheapest-feeling option and the most expensive in misallocated effort. A generic LLM with no Strava connection, no memory across weeks, and no awareness of what you actually rode produces a plan that reads great on day one and decays the moment your first session goes off-script. Foster's monotony research [Foster 1998] is exactly the kind of constraint these prose-only plans violate by accident — they will happily stack three threshold sessions back-to-back if you ask nicely. The pattern. False economy is paying nothing for a tool that does not do the job, then paying for it later in stalled fitness, missed peaks, or injury. A free plan you ignore is more expensive than a $12/month plan you actually follow. ## The one paid tool worth its monthly cost If you spend a single dollar on self-coaching in 2026, spend it on the layer the free stack does not cover: adaptive plan generation. Free analytics tell you what happened. Free workout libraries tell you what to do today. Neither answers the question that actually limits self-coached riders — given my goal, my history, and the week I actually had, what's the right plan for the next 8 weeks? What the free stack misses. Intervals.icu shows your CTL ramp climbing past safe range; it does not redesign your week to absorb the disruption. Strava notes the unplanned 90-minute group ride; it does not recompute Sunday's session as recovery. GoldenCheetah models critical power; it does not write a Base-Build-Peak macrocycle. The gap between analytics and coaching is the gap most self-coached riders fall into. Where adaptive planning sits in the cost stack. The major paid platforms cluster between $10 and $30 a month: Xert at roughly $15/month, TrainingPeaks Premium at roughly $20/month, Wahoo SYSTM at roughly $18/month, TrainerRoad at roughly $22/month. Each makes a different bet. Xert bets on its Training Load model. TrainingPeaks bets on charts and structured workouts a coach hands you. SYSTM and TrainerRoad bet on indoor workout libraries. We at AdaptCycling sit in this same range — for the price of a Strava subscription — and bet on the part the others largely don't do: reading your ride history and rebuilding the plan when your week falls apart. The pillar's framing applies here directly: the most leveraged thing self-coaching adds is structure and recovery enforcement [Meeusen 2013], not more workout variety. One paid tool, chosen well, replaces the part of a coach the free stack cannot. The right test for any paid tool. Does it read what you actually rode, respect periodization invariants you cannot violate without consequence [Friel 2018, Allen et al. 2019], and restructure when life disrupts the week? If yes, the $10-15 it costs is the cheapest piece of equipment in your training. If no, you are paying a monthly fee for prettier charts. ## A defensible cheapest stack — three configurations Three honest configurations cover almost every self-coached rider. The zero-budget stack is Strava free + Intervals.icu free + a used power meter + the books. The one-paid-tool stack adds one $10-15/month subscription for adaptive planning. The two-tool ceiling is $20-30/month total. Above that you are paying for things you don't need. Configuration one — zero subscription. Strava free, Intervals.icu free, GoldenCheetah on the desktop, a used power meter, a heart-rate strap, and the canon books on your shelf. You read your own data, you write your own plan from Friel's framework [Friel 2018], you enforce your own recovery weeks. This works — but only for the rider with the discipline and analytical interest to do the coaching themselves. The honest failure mode is a year-two plateau because nothing external surfaces uncomfortable signals. Configuration two — one paid tool, around $10-15/month. Free stack plus one adaptive coaching layer that does what the free stack does not: reads your rides, generates the plan, and rebuilds when life intrudes. This is the configuration the parent pillar argues most self-coached amateurs actually want. You pay roughly 1/13th of 1-on-1 coaching [CTS 2026] for the four technical things a coach does — load management, periodization, recovery enforcement, taper design — and keep the fifth (outside-perspective accountability) as a deliberate self-coaching skill. Configuration three — two tools, capped at $20-30/month. Free stack plus an adaptive coach plus one indoor workout platform if you ride indoors a lot (Zwift for the social pull, TrainerRoad or SYSTM for the workout library). The two tools serve different jobs: one writes the plan, the other delivers the indoor sessions. Avoid the trap of buying both an adaptive coach and an analytics-heavy paid tool — the overlap is large and the monthly bill doubles for marginal gain. What never makes sense at this scale. Paying $200-$700 a month for 1-on-1 coaching [CTS 2026] when your goal is local races or personal fitness. The math only works above a clear performance threshold the parent pillar lays out. For most self-coached cyclists, the cheapest credible stack tops out around $15/month — and the thing it buys is not data, it's the coaching layer the free tools cannot do. ## Common questions **Can I really self-coach with $0/month?** Yes, and a meaningful number of self-coached cyclists do. Strava free, Intervals.icu free, GoldenCheetah, and a used power meter cover the data and analytics layers. The catch is that you have to do the coaching yourself — periodization, recovery enforcement, taper design — using the canon books [Friel 2018, Allen et al. 2019]. The free stack works for the rider who genuinely enjoys the analytical side. For everyone else, one $10-15/month tool is usually a better return on time. **Is Intervals.icu really as good as TrainingPeaks Premium?** For solo self-coached riders, close to it. Intervals.icu's free tier covers CTL/ATL/TSB charts, power curve, intensity distribution, and the standard load-and-fitness frameworks [Allen et al. 2019]. TrainingPeaks Premium adds peer-share with a coach, structured workout calendars, and deeper PMC features. If you do not have a coach and you are not coaching others, Intervals.icu is the better-value option in 2026. **Can ChatGPT write me a training plan for free?** It can write something that looks like a plan. It cannot read your rides, remember your goal across weeks, enforce ramp-rate limits [Foster 1998], or rebuild when you miss a session. The output is a static document the moment it leaves the chat window. For a one-off macrocycle sketch it can be useful as a starting point; as a sustained coaching layer it is the most expensive free option once you account for the seasons it costs you. **What's the single highest-leverage purchase a self-coached cyclist can make?** If you don't already have one, a power meter — a used single-sided unit in the $200-$300 range is sufficient. After that, the highest-leverage monthly spend is whichever tool actually closes the gap between your free analytics and a real plan. For most riders that is an adaptive coaching layer at $10-15/month, not a second analytics platform. ## References 1. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 2. **Meeusen 2013.** [Prevention diagnosis and treatment of the overtraining syndrome: Joint consensus statement of the European College of Sport Science and the American College of Sports Medicine](https://pubmed.ncbi.nlm.nih.gov/23247672/). European Journal of Sport Science. 3. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 4. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. 5. **Carmichael & Rutberg 2017.** [The Time-Crunched Cyclist: Race-Winning Fitness in 6 Hours a Week (3rd ed.)](https://www.velopress.com/books/the-time-crunched-cyclist-3rd-ed/). VeloPress. 6. **CTS 2026.** [CTS Coaching Pricing](https://trainright.com/coaching/pricing/). Carmichael Training Systems. --- # When self-coached stops working: how to know it's time to hire a cycling coach URL: https://www.adaptcycling.com/guides/when-self-coached-stops-working Updated: 2026-06-16 Author: Jim Camut Most riders who Google this question do not need to hire a $400-a-month coach. They need a real recovery week, an honest hours-per-week number, or to fix the junk middle of their intensity distribution. But there is a real category of self-coached cyclist who has hit the actual ceiling of DIY — and the diagnostic for which one you are is concrete, not vibes. This spoke gives you the four signals, the cheaper interventions to try first, and an honest accounting of what a human coach buys you that an AI coach or a power meter doesn't. ## The four signals you've actually outgrown self-coaching Four honest signals: 18-24 months of consistent training without measurable progress, a goal whose stakes justify $5,000-$7,000 a year, a medical or injury history that complicates load management, or the analytical work itself making you hate cycling. Anything short of one is usually a fixable self-coaching problem, not a coaching-tier problem. The first signal is a true plateau, not a perceived one. A genuine plateau means 18-24 months of consistent structured training (50-plus weeks per year, deliberate periodization, recovery weeks honored) with no movement in FTP, durability, or repeatable race results. Most self-described plateaus are 6-9 months long and sit on top of an inconsistent year — illness, travel, two skipped recovery weeks, a winter of unstructured Zwift racing. Stoggl and Sperlich's RCT showed that even well-trained endurance athletes gained 11.7% in VO2peak in nine weeks when they switched from threshold-heavy training to a polarized intensity distribution [Stoggl & Sperlich 2014]. If you have not deliberately tested a different training model, you have not plateaued — you have settled. The second signal is goal stakes that pencil out. Carmichael Training Systems' published 2025 pricing puts a senior-level coach at $350 per month and a pro-level coach at $629 per month [CTS Pricing 2025] — $4,200 to $7,500 per year, before camps and travel. That number makes obvious sense for a masters racer chasing nationals, a road racer trying to upgrade categories, or someone preparing for a specific Ironman qualifier. It does not pencil out for the gran fondo finisher whose goal is being faster on Saturday — the marginal gains a great coach can extract above what an adaptive training tool delivers are real but small, and they are priced for stakes you may not have. The third and fourth signals are softer: medical complexity that genuinely changes the load-management problem — a recurring injury, a cardiac history requiring HR ceilings, a chronic condition where ramp rate matters in ways a generic tool can't model — or the analytical loop itself becoming the friction. The wider playbook for self-coaching well is in our pillar on the self-coached cyclist; this spoke is specifically about the line where that playbook runs out. ## Cheaper interventions to try first — most plateaus are fixable for free Before paying $4,200-$7,500 a year for a coach, run four cheaper interventions in order: take a real recovery week, audit your intensity distribution against the 80/20 rule, capture honest hours-per-week, and try an adaptive tool that catches the obvious patterns. A coach is what you hire after these have failed. First: take a real recovery week. Foster's foundational monitoring work showed that illness and overreaching cluster around weeks where load and monotony spike together — same hard sessions, no variance, no cut week [Foster 1998]. A self-coached rider stacking 600 TSS weeks for four weeks straight on a TrainerRoad plan, without a deliberate recovery week, is producing exactly the pattern the data flagged. A 30-40% volume cut for seven days, intensity preserved as short maintenance efforts, often produces in two weeks what the rider was about to spend $4,000 on a coach to fix. Second: audit your intensity distribution. Seiler's descriptive work across rowers, runners, skiers, and cyclists found elite endurance athletes converge on roughly 80% of sessions at low intensity (below 2 mM lactate) and 20% at high intensity, with very little in the middle [Seiler 2010]. The most common self-coaching error is a distribution that is 50% threshold-and-tempo: too hard for recovery, too easy for adaptation. Stoggl and Sperlich's RCT showed polarized training produced 11.7% VO2peak gains versus essentially zero for threshold-only training in nine weeks [Stoggl & Sperlich 2014]. Intervals.icu computes time-in-zone for free; TrainerRoad and TrainingPeaks both make the audit trivial. Third: capture hours-per-week honestly — average across the last 12 weeks, not what you intended. Most self-coaches anchor to peak weeks and end up training above what their actual schedule supports, the structural cause of the same illness-monotony pattern Foster described. Fourth: try an adaptive tool before a human. AdaptCycling reads your ride history, generates plans that respect periodization invariants, and restructures the week when life disrupts it — catching ramp-rate flags, junk middle, missed recovery weeks, and TSS overshoots, which is most of what we see in year-two plateau cases. ## What a $400-600/month human coach actually buys you (and what they don't) Honest accounting: a $350-$629/month coach gives you 1-2 monthly calls, written plan adjustments, race-day messages, and an outside voice. They do not buy you exemption from the laws of training adaptation. Marginal gains are real at the masters-podium and category-upgrade level, where a small taper improvement can decide results [Bosquet et al. 2007]. What you actually get at $350/month (CTS Premium tier in 2025): a senior-level coach, twice-weekly workout analysis in TrainingPeaks, up to 2 hours of monthly calls, unlimited text and email with 24-hour response, and a personalized strength plan [CTS Pricing 2025]. At $629/month (CTS Ultimate), daily workout analysis, unlimited communication, and up to 3 private camp days. What you do not get at any tier is exemption from training principles: the same Seiler intensity distribution, the same Bosquet taper math (volume reduced ~40-60% over two weeks with intensity preserved produces a small-to-moderate performance lift) [Bosquet et al. 2007], the same Foster monotony rules. The coach is paid to apply these correctly. Where a human coach earns the price difference is the outside-perspective work — the call where they tell you your reported sleep is bad, your nutrition is under-fueled, the plan you keep asking for is wrong, or your goal is unrealistic. Stellingwerff and colleagues' synthesis found that 18 of 21 studies of training overload showed concurrent low energy availability, meaning many self-diagnosed overtraining cases are actually under-fueling masquerading as a training-load problem [Stellingwerff et al. 2021]. A good coach is more likely to catch that on a call than a self-coach is from a power file. An AI coach reading Strava data won't see your face. Where AdaptCycling has limits: we flag obvious patterns from ride data and chat — ramp-rate spikes, missed recovery weeks, monotony scores, fueling inconsistency surfaced from message context. We can't see you on a Zoom call, can't read the family-stress signal in your voice, and don't sit in the team car at the start line. For most self-coached riders that gap doesn't move the needle. For the masters athlete chasing a podium-tier result, sometimes it does — and a one-season trial of human coaching, layered on top of the data tooling, is the cleanest way to find out which side of that line you are on. ## A decision tree for the hire-or-not call Three inputs decide it: goal stakes, weekly hours, and analytical engagement. High stakes plus 10+ hours plus engagement: AI coaching plus a one-season human trial. High stakes plus low hours: a coach earns the most. Moderate stakes at any hours: AI or self-coached. Quadrant one — high stakes (cat-3-and-up racer, masters national-level goal, qualifying event), 10+ hours per week, engaged with data. The math justifies a human coach at $350-$629/month [CTS Pricing 2025], but only after you have run a structured polarized block and a properly-timed taper. The Bosquet meta-analysis found taper alone produces a small-to-moderate performance gain at the elite level [Bosquet et al. 2007] — at podium-tier stakes, that can decide the result, and a coach who has periodized hundreds of athletes is worth the price. AdaptCycling layered underneath as the daily-decisions tool is the cleanest combination. Quadrant two — high stakes, low hours (under 8 per week). Counter-intuitively, this is where a coach earns the most: every session has to count, recovery margins are thin, the wrong call costs a goal. An adaptive AI coach that restructures the week when life disrupts it covers most of the ground for roughly 1/13th of 1-on-1 coaching. A human coach is the upgrade only if the goal is genuinely time-sensitive. Quadrant three — moderate stakes (gran fondo finisher, local Cat-4-5, fitness goals), any hours. Self-coached or AI-coached. The marginal gains a $400/month coach extracts above an adaptive tool are real but small at this level, and they are priced for stakes you don't have. ## Common questions **How long should I be self-coached before deciding it's not working?** Eighteen to twenty-four months of consistent, structured training — meaning 45-plus weeks per year on a real plan, recovery weeks honored, deliberate periodization. Most self-described plateaus are 6-9 months long sitting on top of an inconsistent year (illness, travel, missed recovery weeks). The Stoggl and Sperlich RCT showed even well-trained athletes can produce 11.7% VO2peak gains in nine weeks by switching intensity distribution [Stoggl & Sperlich 2014]. If you have not deliberately tested polarized vs. threshold and run a clean season, you have not plateaued — you have settled. **Is AdaptCycling enough for a Cat 3 or masters racer with a real goal?** For most of them, yes — paired with honest self-monitoring. AdaptCycling reads your ride history, restructures the week, and applies the same periodization invariants (Seiler intensity distribution, Bosquet taper math, Foster monotony bounds) that a coach applies [Seiler 2010, Bosquet et al. 2007, Foster 1998]. Where it falls short is the outside-perspective layer — the Zoom call where a coach catches the under-fueling pattern or the life-stress signal you are minimizing. For a podium-tier specific result, a one-season human coaching trial layered on top of the data tooling is the cleanest way to test whether that gap moves your needle. **I'm tired all the time — is that a sign I need a coach?** Usually no — it's a sign you need a recovery week, more food, or both. The Stellingwerff review found 18 of 21 training-overload studies showed concurrent low energy availability, meaning most fatigue patterns in endurance athletes are fueling problems, not coaching problems [Stellingwerff et al. 2021]. The Meeusen consensus distinguishes functional overreaching (resolves in 5-14 days of reduced volume) from non-functional overreaching (weeks to months) [Meeusen et al. 2013]. Try a real recovery week and a 4-week honest fueling audit before you sign a coaching contract — if those resolve it, the answer was never coaching. ## References 1. **Stoggl & Sperlich 2014.** [Polarized training has greater impact on key endurance variables than threshold, high intensity, or high volume training](https://pmc.ncbi.nlm.nih.gov/articles/PMC3912323/). Frontiers in Physiology. 2. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 3. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. 4. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 5. **Stellingwerff et al. 2021.** [Overtraining syndrome (OTS) and relative energy deficiency in sport (RED-S): shared pathways, symptoms and complexities](https://pubmed.ncbi.nlm.nih.gov/34181189/). Sports Medicine. 6. **Meeusen et al. 2013.** [Prevention, diagnosis, and treatment of the overtraining syndrome: joint consensus statement of the European College of Sport Science and the American College of Sports Medicine](https://pubmed.ncbi.nlm.nih.gov/23247672/). Medicine & Science in Sports & Exercise. 7. **CTS Pricing 2025.** [Coaching pricing tiers and inclusions](https://trainright.com/coaching/pricing/). Carmichael Training Systems. --- # FTP test vs power curve: when a self-coached cyclist actually needs to test URL: https://www.adaptcycling.com/guides/ftp-test-vs-power-curve-self-coached Updated: 2026-05-05 Author: Jim Camut I have done dozens of FTP tests across my racing career, and the dirty truth is most of them never changed a training decision. The self-coached rider sitting at their desk on Tuesday with a modeled FTP on screen is making a decision, not a measurement: trust the curve, or burn Saturday on a 20-minute effort? The math behind tests versus power-curve modeling is covered in our ftp-without-a-test pillar. This spoke is the decision framework — when the model is enough, when a test is non-negotiable, and what testing actually costs. ## When the modeled FTP is enough on its own If you have 90+ days of varied power data, at least one near-maximal effort longer than eight minutes, and the modeled estimate has not jumped more than three percent in the last month, the curve is doing its job. What a fresh 20-minute test would add at that point sits inside the test's own noise band. The honest reason most self-coached riders test is not that the plan needs a better number — it is that Zwift, TrainerRoad, or Wahoo SYSTM scheduled a test and the calendar nudged them. A modeled FTP from a fitted Critical Power curve sits within roughly three to five percent of a clean field test once you have 90 days of mixed riding to fit against [Allen et al. 2019, McGrath et al. 2021]. The 20-minute protocol you would replace it with carries its own measurement noise: Tramontin and colleagues showed warm-up structure alone shifts the resulting FTP estimate by a clinically meaningful amount on the same rider on the same day [Tramontin et al. 2022]. Trading three percent of model uncertainty for five percent of pacing and warm-up uncertainty is not progress. The other under-appreciated point is that the model updates continuously. Intervals.icu re-fits eFTP every time you upload. AdaptCycling re-fits on every new ride you upload. Xert updates on breakthrough efforts. A field test gives you one data point on one Saturday in one set of conditions; the model gives you a rolling estimate that already incorporates dozens of efforts at varying duration. Inside the broader self-coached cyclist playbook — where the rider is making the periodization, intensity, and recovery decisions a coach used to make — chasing the test is one of the cheapest places to stop spending decision energy. ## The three situations where you actually owe yourself a test Three cases break the model and require a real effort: a brand-new power meter with under 30 days of data, a return from 4+ weeks fully off the bike, and the two weeks before a goal event where being 15 watts wrong costs the season. Outside those, the model is doing its job. The first case is the cleanest. A rider who just bought their first power meter has nothing for the curve to fit against. Two weeks of zone 2 followed by a ramp test or a 20-minute effort gives the algorithm an anchor while the power-duration curve fills in. A new rider who skips this accepts a 5-to-10 percent margin of error for the first month of structured training — the difference between sweet-spot intervals that feel hard and sweet-spot intervals that drift into threshold. TrainerRoad, Zwift, and Wahoo SYSTM all default to a forced test on signup for exactly this reason. The second is the long-layoff case. Mujika and Padilla showed VO2max and threshold-related markers begin meaningful decline after roughly 10 days of complete rest, with the rate accelerating past three weeks [Mujika & Padilla 2000]. After 4-to-6 weeks fully off, the power-duration curve is a record of a cyclist who no longer exists. Coming back, the model needs a fresh anchor before the first build week or the rider trains the first month at intensities calibrated to pre-layoff fitness. For shorter layoffs the graded ramp covered in our spoke on restarting after two weeks off applies; past four weeks, retest first. The third is the pre-goal sanity check, and self-coached riders skip it most often. Two weeks before a target event, run a clean 20-minute effort against the modeled FTP. Agreement within 5% means taper with confidence. Disagreement of 10% or more means investigate before race day rather than discovering at the start line that the plan was built against a number 15 watts off. The asymmetry matters: one tired Saturday is the cost of the test; the cost of being wrong about FTP at a goal event is the goal event. ## What testing actually costs you A formal FTP test is not a free measurement. It costs a hard ride that needs 48-to-72 hours of recovery, displaces a quality session in the build, introduces false precision around a single number, and primes the rider to chase the result rather than train against it. Six tests a year is a real chunk of your annual TSS budget burned on a number you mostly already had. The training-load cost is easy to underestimate. A 20-minute test ridden honestly is roughly equivalent to a hard threshold session — 80 to 100 TSS once you include warm-up and the protocol's anaerobic finish. A ramp test is shorter but maximal, and the recovery cost is similar. Testing every six weeks the way Zwift Academy or some TrainerRoad plans suggest is six to eight quality sessions per year displaced by a measurement workout. Halson's training-load review is explicit that load monitoring exists to tell you whether you are adapting [Halson 2014]; trading adaptation work for measurement work to feed that monitoring is the wrong direction of the trade. The second cost is false precision. A single 20-minute test produces a number that looks authoritative but sits inside a wide individual confidence interval. Borszcz and colleagues found the limits of agreement between a 20-minute estimate and a true 60-minute effort spanned 40-to-60 watts on individuals, even when the group-level bias was small [Borszcz et al. 2018]. The rider who walks away with FTP at 268 watts to a stated precision is fooling themselves. A modeled estimate drawn from dozens of efforts has narrower individual error bars than any single test does. The third cost is psychological, and self-coached riders are uniquely vulnerable. A test result becomes a target to defend rather than a parameter to train against. Riders who test every six weeks tend to optimize their training for the test rather than the goal event. The quiet virtue of a continuously-modeled FTP is the rider never has a single sacred number — the curve drifts upward as fitness improves and there is no test day to peak for or to fail. ## How to read disagreement between the model and a recent test When you do test and the result disagrees with the modeled FTP, the size of the gap tells you what to do. Within 3%, treat them as identical. 3-to-7%, accept the lower number for a week of intervals and retest one workout. Past 7%, something is wrong with one of the inputs and the answer is investigation, not averaging. A 3% gap is below the test's own measurement noise [Tramontin et al. 2022] — average them, or stay with the modeled number, and move on. A 3-to-7% gap is the most common case and the one that matters most. Take the lower of the two for the next block of work, run a sweet-spot interval session at the new number on day three or four, and ask whether the prescribed wattage felt right. If it did, the lower number was correct; if it felt easy, the higher one was. One session, dispute resolved. Past 7% is a flag, not a result. Common explanations: a power meter calibration shift, a test ridden fatigued or tapered, a power curve dominated by stale efforts that no longer reflect current fitness, or a model missing a long-duration anchor. Before changing FTP by 15 watts on one test, check the power meter zero-offset, audit the last two weeks for outlier fatigue, and look at the MMP curve to see whether the long-duration points are populated by recent rides. Big gaps almost always have an explainable cause, and the explanation usually invalidates one of the inputs rather than splitting the difference. ## Common questions **Do I need to do an FTP test if I'm using AdaptCycling, Intervals.icu, or Xert?** Not on a regular cadence. All three derive FTP from your power-duration curve, which gives a usable estimate within 3-5% of a clean test once you have 90 days of varied riding [McGrath et al. 2021]. The exceptions: less than 30 days of power data, a return from 4+ weeks off, or the two weeks before a goal event. Outside those, a fresh test is mostly a recovery tax that does not change a training decision. **How often should I test if I am going to test?** Twice a year is plenty for most amateur self-coached riders — once at the end of base before the build phase, once two-to-three weeks before a goal event. The every-6-week cadence baked into TrainerRoad and Zwift Academy plans was built when modeled FTP did not exist; with a continuously-fit curve, six tests a year is six quality sessions displaced by measurement workouts [Halson 2014]. **My modeled FTP jumped 8 watts last week without a test — should I trust it?** Yes, if the jump is anchored by a real effort. Look at the MMP curve: a single hard 12-to-30 minute ride that exceeded your previous best is usually the cause and the model is reading new information correctly. If the jump came from short anaerobic efforts with no new long-duration data point, the algorithm is over-extrapolating and the next long ride will likely pull the estimate back. Track it for a week before changing zones. **Ramp test or 20-minute test for cross-checking modeled FTP?** A 20-minute test, ideally outdoors on a known climb. The ramp test estimates FTP as 75% of peak one-minute power, and that ratio varies between 72% and 77% across riders [Allen et al. 2019] — which makes the ramp unreliable for the exact use case (a sanity check) you are trying to perform. A 20-minute effort is direct, comparable to your power curve in the same duration band, and its limits of agreement are at least known [Borszcz et al. 2018]. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Borszcz et al. 2018.** [Functional Threshold Power in Cyclists: Validity of the Concept and Physiological Responses](https://pubmed.ncbi.nlm.nih.gov/29801189/). International Journal of Sports Medicine. 3. **Tramontin et al. 2022.** [Functional Threshold Power Estimated from a 20-minute Time-trial Test is Warm-up-dependent](https://pubmed.ncbi.nlm.nih.gov/34749416/). International Journal of Sports Medicine. 4. **McGrath et al. 2021.** [Do Critical and Functional Threshold Powers Equate in Highly-Trained Athletes?](https://pubmed.ncbi.nlm.nih.gov/34055164/). International Journal of Exercise Science. 5. **Mujika & Padilla 2000.** [Detraining: Loss of Training-Induced Physiological and Performance Adaptations. Part I: Short Term Insufficient Training Stimulus](https://link.springer.com/article/10.2165/00007256-200030020-00002). Sports Medicine. 6. **Halson 2014.** [Monitoring training load to understand fatigue in athletes](https://pubmed.ncbi.nlm.nih.gov/25200666/). Sports Medicine. --- # How a self-coached cyclist builds a 12-week training plan for a goal event URL: https://www.adaptcycling.com/guides/12-week-goal-event-self-coached Updated: 2026-05-05 Author: Jim Camut The difference between hitting and missing a 12-week peak is almost always the last two weeks. Most self-coached cyclists nail the base, get the build roughly right, and then dismantle eight weeks of work in a panicked taper. This is the operational plan: how to anchor twelve weeks to a single date, what each phase has to deliver, and the taper that holds the peak. ## The 12-week macrocycle anchored to event date Twelve weeks is the shortest block long enough to express a full base-build-peak-taper macrocycle without compromising any phase. Working backward from the event date: weeks 1-6 are base, weeks 7-10 are build, weeks 11-12 are taper. Phase boundaries are non-negotiable; the chronic-load ramp inside each phase is where individual variation lives. Twelve weeks is the smallest block where each phase still delivers what the physiology requires. The classical periodization frame allocates 8-16 weeks of base, 6-10 weeks of build, and 2-3 weeks of peak/taper; compressed into 12, that becomes 6 weeks base, 4 weeks build, 2 weeks taper. The structure works because the time constants line up — aerobic adaptation in weeks, lactate-threshold adaptation in 4-6 weeks, taper supercompensation in 14 days. Run it in 8 weeks and base collapses; run it in 16 and you invite the year-two drift the broader self-coached cyclist playbook covers. Anchor everything to the event date and count backward. If the goal is Saturday in week 12, the taper begins on the Sunday 13 days out and the last hard week ends 14 days before the start. That hard week (week 10) is the highest-load week of the macrocycle — the one whose chronic load you spend the taper banking. Inside base (weeks 1-6), structure as two 3-week mesocycles, each progressing weekly TSS by 4-7 TSS/day per week — a sustainable chronic-load ramp. Week 4 is a recovery week with volume cut 30-40% and intensity preserved. Week 7 starts build at the load you ended base on, not lower. The intensity ratio across all 12 weeks should sit close to 80/20 — roughly 80% of riding time below the first lactate threshold, 20% above the second. Seiler's polarized-training research [Seiler 2010] across two decades of elite descriptive data converges on this distribution as the most reliable predictor of VO2max and threshold gains in trained endurance athletes. Self-coached riders systematically violate it during build, drifting into a tempo middle that produces fatigue without adaptation. ## Picking the right build-phase workouts for the event Base looks identical for every event; build is where the specificity lives. A criterium is a VO2max and anaerobic-repeatability problem, a gran fondo or century is a sub-threshold durability problem, a hill-climb is a 20-60 minute threshold problem, and a stage race is a back-to-back load-tolerance problem. Allocate build's intense work against the event's actual demand. Race-pace and supra-race-pace work is what bridges base fitness to event performance. Mujika's intense-training synthesis [Mujika 2010] documents that elite athletes shift training toward race-specific intensity zones during the competitive build phase while preserving low-intensity volume — exactly the pattern that translates to cycling. For a criterium peaking in week 12, build weeks 7-10 should weight 30-second to 2-minute VO2max repeats and 10-30 second neuromuscular surges; aerobic capacity matters but repeatability above threshold is what wins or loses. For a 100-mile gran fondo, the same four weeks should be 60-90 minute sweet-spot blocks at 88-94% FTP and 4-5 hour endurance rides with the last hour at upper Z2. Hill-climb and time-trial events sit between those poles: 20-40 minute threshold intervals and 2x20 at 95-100% FTP are the highest-yield workouts because the event itself is largely a sustained near-FTP effort. Stage races are the multi-day load-tolerance problem — what matters is back-to-back hard days, so build weeks should include a deliberate hard-Saturday-into-hard-Sunday block at least twice. Mujika's same paper [Mujika 2010] is unambiguous that high-intensity work in the weeks before the taper drives the largest physiological adaptations, and that reducing intensity to make room for volume is the wrong trade. ## The taper — the most-screwed-up two weeks of the plan Bosquet's 27-study taper meta-analysis [Bosquet et al. 2007] is the most-cited number in tapering: a 2-week taper with volume reduced 41-60%, intensity preserved, and frequency held within 20% of pre-taper produces the largest competition performance gain. Most amateur tapers fail by cutting intensity, cutting frequency, or starting too early. The numbers in Bosquet et al.'s 2007 meta-analysis [Bosquet et al. 2007] across 249 swimmers, 80 road cyclists, and 110 runners are tight: a two-week duration, a 41-60% exponential volume reduction (achieved by cutting session duration, not session count), and intensity held at pre-taper values produced the largest aggregated effect on competition performance. Mujika and Padilla's earlier scientific-bases review [Mujika & Padilla 2003] reports the same convergence with a typical 0.5-6.0% performance window and a 3% mean. The taper banks the supercompensation latent in the build's accumulated fatigue while preserving the neuromuscular and metabolic stimulus high-intensity work provides. The amateur failure modes are predictable. Mistake one — cutting intensity along with volume, because cutting volume alone feels like not tapering. The result is detraining: VO2max begins to drop within 10-14 days of insufficient stimulus. Keep the same workouts on the calendar in week 11 and week 12, but make each interval set roughly half the work — three threshold intervals instead of six, two VO2 repeats instead of five — at the same target wattage. Mistake two — cutting frequency hard, riding three days instead of five. Mujika's 2010 paper [Mujika 2010] is explicit that frequency reductions beyond 20% start trading away adaptation. Mistake three — starting the taper four weeks out because race nerves arrive early. Aubry et al.'s 2014 trial [Aubry et al. 2014] settles this directly: across 33 trained triathletes, the acutely-fatigued group (heavy load right up to the taper) produced significantly larger 4-week-taper supercompensation than the control group that had backed off. Heavy training immediately before the taper is what makes the taper work. Mistake four — riding the day before the event easy and short leaves most riders flat. A 45-90 minute ride with two or three 30-second openers near event pace consistently produces good legs on race morning. ## When the plan breaks — adapting a 12-week block when life happens Twelve uninterrupted weeks is fiction for any cyclist with a job and family. The realistic question is what to do when week 5 disappears to a sinus infection or week 8 collapses under a work crunch. Triage rule: protect the taper at all costs, protect the build's specificity second, absorb the disruption inside base if at all possible. A lost week in base (weeks 1-6) is the cheapest to absorb — extend base by a week and compress one mesocycle's recovery week into a shorter unload, or accept a slightly lower starting CTL going into build. Foster's monotony research [Foster 1998] across 25 trained athletes showed that the strongest behavioral predictor of illness and overreaching was high load combined with low daily-load variance — exactly what you produce by trying to compress the missed week into the next available slot. The temptation to make up missed work is statistically the highest-risk move available. Skip it, or stretch the recovery instead. Build-phase disruption is harder. A lost week in weeks 7-10 costs specificity directly. Triage: if one week is lost, drop the planned recovery week and treat the lost week as the recovery, then re-enter build at the load you would have been at. If two weeks are lost, the build is functionally cut from four to two; pull the weakest workout type and preserve only the highest-yield specificity work. Rosenberg et al.'s 2023 systematic review [Rosenberg et al. 2023] confirms that taper performance is more sensitive to the load right before it than to the cumulative load earlier in the cycle. Taper-week disruption — sickness in week 11 or 12 — is the one case where the rule is to stop trying to rescue the cycle and accept the result. Push through a sick taper and the immune-suppression and fatigue effects compound; the literature on functional overreaching crossing into non-functional overreaching is unambiguous that recovery then takes weeks, not days. Ride easy until 48 hours before the event, do a short opener the day before, and race what you have. This is the operational version of the year-structure question covered in the broader self-coached cyclist playbook: a static plan tells you to make up the missed week; a defensible self-coached approach absorbs the cost into the phase that can afford it. ## Common questions **Should I do an FTP test in week 1 of the 12-week plan?** Yes, or a 20-minute power test that gives you the same number. Every workout target is a percentage of FTP, so anchoring it inside week 1 is worth the one hard day. Retest at the end of week 6 and use the new number for build. Skip retesting in build — an extra hard test in weeks 7-10 costs more than it returns. **What weekly TSS should I be hitting in the peak week of build?** Week 10 should sit roughly 15-25% above your average over weeks 7-9 — but the absolute number depends on starting CTL. A rider entering at 60 CTL might peak week 10 at 700-800 TSS; a rider at 90 CTL at 950-1100. The number that matters is the ramp rate inside build, which should stay within the 4-7 TSS/day per week chronic-load progression. **Can I race other events during the 12-week block?** Small B and C events in weeks 6-10 work as build-phase intensity and can replace a planned interval session. Avoid racing in weeks 1-5 (base needs aerobic continuity) and weeks 11-12 (taper). A race in week 7 or 8 is often the best workout in the block. The trade is that the following 2-3 days have to be pure recovery. **How should the week before the event actually look?** Week 12 has roughly half the volume of weeks 7-10, with two short interval sessions Monday/Tuesday at full target wattage but cut to one-third the work, an easy Wednesday, a short opener Thursday or Friday with 2-3 fast 30-second efforts, and full rest the day before if the event is Saturday — or a 45-60 minute ride with openers if the event is Sunday. **What if my goal event is 8 weeks out, not 12?** Compress base, not build or taper. Run a 2-week base, 4-week build, 2-week taper. The build and taper convert fitness into result; cutting either costs more than starting build with less aerobic depth. Below 8 weeks, you are not building a peak — you are sharpening whatever fitness you already have. ## References 1. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. 2. **Mujika & Padilla 2003.** [Scientific bases for precompetition tapering strategies](https://pubmed.ncbi.nlm.nih.gov/12840640/). Medicine & Science in Sports & Exercise. 3. **Mujika 2010.** [Intense training: the key to optimal performance before and during the taper](https://pubmed.ncbi.nlm.nih.gov/20840559/). Scandinavian Journal of Medicine & Science in Sports. 4. **Aubry et al. 2014.** [Functional overreaching: the key to peak performance during the taper?](https://pubmed.ncbi.nlm.nih.gov/25134000/). Medicine & Science in Sports & Exercise. 5. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 6. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 7. **Rosenberg et al. 2023.** [Effects of tapering on performance in endurance athletes: a systematic review and meta-analysis](https://pmc.ncbi.nlm.nih.gov/articles/PMC10171681/). PLOS One. --- # Is my training plan actually adapting? The diagnostic test for self-coached cyclists URL: https://www.adaptcycling.com/guides/signs-training-plan-not-adapting Updated: 2026-05-09 Author: Jim Camut Every training app in 2026 calls itself adaptive. Most are not. The honest test is not the marketing page — it is what the plan does in the week after you go off-script. Three diagnostic signs separate plans that genuinely adapt from plans that just generate. If your plan exhibits any of them, it is a static schedule with a dynamic UI. This is how to tell, why each sign happens, and a 14-day probe that settles the question. ## Sign one: it gives the same plan to a 6h/wk rider and a 12h/wk rider If two riders with similar FTPs but different weekly time budgets receive structurally identical plans — same workout types, same progression, same intensity distribution — the plan is not adapting to either of them. It is fitting both to a population template. Individualization at the level of weekly load is the lowest bar an adaptive plan must clear, and a surprising number of products fail it. Kiely's 2018 review of periodization theory [Kiely 2018] is one of the most direct critiques of how training plans are actually built. Prescriptions are typically derived from group means rather than individual response curves, and the inter-individual variation in adaptive response to identical training is large enough that a plan calibrated to the population mean can be net-negative for a meaningful share of athletes. A plan that does not take weekly hours and recovery capacity as primary inputs is, by construction, optimizing for a population the rider may not belong to. The mechanism this hides is Foster's training-monotony curve [Foster 1998]. The 6h/wk rider on a plan calibrated for 10h/wk lands in a high-monotony zone — too much of the time gets swallowed by intensity because the plan assumes endurance volume is being added on top. The 12h/wk rider on the same plan is under-stimulated and adds unstructured volume to compensate, which then collides with the next prescribed hard day. Both failures look like the rider's problem and are actually the plan's. A genuinely adaptive plan reshapes the entire weekly intensity distribution around hours-per-week before a single workout is selected. ## Sign two: a 90-minute hard group ride does not shift the rest of the week An unscheduled hard ride banks intensity stimulus the plan was going to prescribe later. If the plan reads the ride and changes nothing — Tuesday's threshold session still on the calendar, weekly TSS target unchanged — it is not adapting. The week is a budget, not a sequence, and a plan that cannot rebalance after an unplanned effort is treating it as a sequence. The mechanism here is the weekly intensity distribution, not the daily TSS total. Stoggl & Sperlich's 2014 controlled trial [Stoggl & Sperlich 2014] randomized 48 well-trained endurance athletes across four distribution models for nine weeks; the polarized group — roughly 80% low intensity, 20% high intensity, very little tempo — produced the largest VO2peak gain (+11.7%) and the largest time-to-exhaustion gain. The follow-up across well-trained and elite athletes [Stoggl & Sperlich 2015] confirmed the distribution is what matters, not any single workout. A plan that lets a hard unplanned ride land on top of a normal week of intensity has pushed the ratio outside the band where adaptation happens. Foster's monotony work [Foster 1998] amplifies the same point from the illness side. Across his 25-athlete cohort, illness clustered in weeks where load and daily-load variance were both high — exactly the pattern produced when an unplanned hard ride lands on an unchanged plan. The right reaction is to redistribute. Tomorrow's threshold work becomes recovery; the weekend's prescribed hard ride drops one tier or pushes 24 hours; the weekly TSS target may not change at all. A plan that does this is reading the ride. A plan that reprints the same Tuesday workout is reading the calendar. TrainerRoad's Adaptive Training adjusts the difficulty of the next indoor workout, but the structure of the week stays fixed by the Plan Builder block. Xert's Training Load model continuously updates against every effort, including outdoor rides, and is genuinely strong here. JOIN handles a moderate version of this case well; many TrainingPeaks plans and most generic 12-week PDFs fail it outright. The probe is simple: do an unscheduled 90-minute hard ride on Sunday, then check Monday whether anything in the plan moved. ## Sign three: a sick week makes the plan shrink, not restructure When illness or travel removes a week, a static plan compresses what got missed into the following weeks at lower volume — the plan shrinks. An adaptive plan restructures: skip the missed work, reset the chronic-load ramp, and re-anchor the macro arc to the new dates. Compression is the most common amateur-coaching error and one of the strongest signals that the plan is not adapting at all. Mujika & Padilla's detraining series [Mujika & Padilla 2000] documented that even short interruptions of 7-14 days produce measurable cardiovascular drift — VO2max drops driven mostly by reduced blood volume and stroke volume, with peripheral adaptations largely intact. Resuming the original plan at the original wattages puts the rider back at full chronic load on a decompressed acute base. The acute:chronic workload framework [Hulin et al. 2016] places elevated injury and illness risk at ratios above roughly 1.5, and a normal full-volume week after seven days off lands well outside that band. The diagnostic version of this sign is simpler. Look at what the plan does on the day you log a sick week. If the next week starts at 100% of pre-break target and the prescribed intervals look identical, the plan has not adapted — it has just slid forward by seven days. If the plan compresses the missed peak block into fewer weeks at the same intensity, it has done the worst thing on the menu: increased acute load on a decayed chronic load while shortening the recovery between hard days. A plan that adapts will skip the missed work, set the resumption week at 60-70% of pre-break volume, and either move the goal event back or downgrade it from A to B priority. ## How to probe your plan in 14 days If reading the signs is ambiguous, run the probe. Two weeks, three deliberate perturbations, three things to check. The probe is designed so a genuinely adaptive plan will visibly respond and a static plan will visibly not. It costs nothing in fitness — every perturbation is a normal cycling event that real life produces anyway. Day 1: skip a prescribed key workout — the hardest one of the week. Mark it as missed in whatever app you are using. Check the next prescribed week. A genuinely adaptive plan will have shifted the intensity distribution, dropped the next hard day one tier, or rebuilt the week to absorb the missed stimulus across two or three sessions. A static plan will look identical, with the missed workout either deleted or copy-pasted into the next available slot. Halson's training-load monitoring review [Halson 2014] is direct: load monitoring exists to confirm whether an athlete is adapting to the program, and a plan that does not respond to a missed key session is not running that loop. Day 5: do an unscheduled 90-minute hard group ride at zone 3 to threshold. Check the next 48 hours. Did Tuesday's threshold work shift to recovery? Did the weekly TSS target adjust to keep the intensity ratio inside the polarized or pyramidal band [Stoggl & Sperlich 2015]? Day 12: log a 7-day forced break — most apps support this through a sickness flag or travel block. Check the resumption week. If it starts at 100% of pre-break volume, the plan is not handling the detraining curve [Mujika & Padilla 2000]. If it starts at 60-70% with intensity backed off, the plan is doing the right thing. Two or more failures and the plan is operating as a static schedule with adaptive marketing — at which point the choice is to keep the plan and become your own restructure engine, or to move to a tool that does it natively. ## Common questions **Is TrainerRoad's Adaptive Training actually adaptive?** Partially. TrainerRoad's Progression Levels system genuinely adapts the difficulty of the next workout based on completed indoor sessions, and within that scope the engineering is excellent. What it does not do well is restructure the week when an outdoor ride, a missed session, or a forced break disrupts the Plan Builder block. Mostly indoor with a stable schedule, the limitation rarely bites. Outside or unpredictable weeks hit the seam quickly. **How long should I run the probe before deciding the plan is not adapting?** Two weeks and three perturbations is enough. One missed key workout, one unscheduled hard ride, and one logged 7-day break — spread across 14 days — generates the diagnostic signal. If the plan visibly responds to two of three, it is adapting. If it responds to one or zero, it is not. Running longer rarely changes the verdict and costs real training time. **What if my plan handles missed workouts but not unplanned hard rides?** That is the most common partial-pass. Plans built around indoor power data tend to handle compliance well — they read what you completed and adjust difficulty. Plans connected to Strava and reading every ride tend to also handle the unplanned hard ride correctly. If your plan only reads structured indoor sessions, the workaround is to manually tag the unplanned ride as a key workout, which forces the load accounting to update. **Could a plan be adapting in ways I cannot see?** Possible, but unlikely to matter. The whole point of an adaptive plan is that the rider can see the response and trust it. A plan adapting silently in a recommendation engine you never check has the same operational behavior as a static plan. Halson's monitoring framework [Halson 2014] is explicit that the value of load tracking is in feedback loops the athlete can act on. If you cannot tell the plan adapted, treat it as if it did not. **Where does this fit into the broader question of adaptive training plans?** These three signs — population-template prescription, ignored unplanned efforts, and shrink-not-restructure on missed weeks — are the diagnostic surface of the deeper question of what adaptive cycling training plans actually need to do. The pillar on that topic covers the architecture; this spoke is the field test. If your plan fails the probe, the pillar is the next read. ## References 1. **Kiely 2018.** [Periodization Theory: Confronting an Inconvenient Truth](https://pubmed.ncbi.nlm.nih.gov/29189930/). Sports Medicine. 2. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 3. **Stoggl & Sperlich 2014.** [Polarized training has greater impact on key endurance variables than threshold high intensity or high volume training](https://pubmed.ncbi.nlm.nih.gov/24550842/). Frontiers in Physiology. 4. **Stoggl & Sperlich 2015.** [The training intensity distribution among well-trained and elite endurance athletes](https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2015.00295/full). Frontiers in Physiology. 5. **Mujika & Padilla 2000.** [Detraining: loss of training-induced physiological and performance adaptations. Part I: short term insufficient training stimulus](https://pubmed.ncbi.nlm.nih.gov/10966148/). Sports Medicine. 6. **Hulin et al. 2016.** [The acute:chronic workload ratio predicts injury: high chronic workload may decrease injury risk in elite rugby league players](https://pubmed.ncbi.nlm.nih.gov/26511006/). British Journal of Sports Medicine. 7. **Halson 2014.** [Monitoring training load to understand fatigue in athletes](https://pubmed.ncbi.nlm.nih.gov/25200666/). Sports Medicine. --- # Missed a key workout vs missed a recovery ride: why it matters which one URL: https://www.adaptcycling.com/guides/missed-key-workout-vs-missed-recovery Updated: 2026-05-10 Author: Jim Camut A missed workout is not one event. A missed key session — threshold, VO2max, or sweet spot — and a missed recovery ride are different signals that should produce different plan responses, and most apps treat them identically. The asymmetry is physiological, not philosophical. This is what each loss actually costs, why static plans get both wrong in opposite directions, and what an adaptive plan should do in each case. ## The asymmetry: why these are not the same signal A key session and a recovery ride deliver different stimuli to different systems. The hard session targets VO2max, lactate threshold, or sweet-spot adaptations on a discrete dose-response curve. The recovery ride accumulates low-intensity volume that drives mitochondrial biogenesis and fat oxidation. Skipping one is not interchangeable with skipping the other. Seiler's polarized model [Seiler 2010] formalized the distinction. Roughly 80 percent of training time at low intensity below the first lactate threshold drives a different set of adaptations than the 20 percent above the second threshold. The low band, which San Millán and Brooks characterized through lactate and substrate-oxidation testing in elite cyclists [San Millán & Brooks 2018], is where mitochondrial density and fatty-acid oxidation capacity are built — measured in weeks of cumulative volume, not single sessions. The high band drives stroke-volume and oxygen-utilization gains on a sharper dose-response curve, where one or two well-executed sessions per week is the operative unit. Same week, two clocks running at different speeds. That is why missed-workout accounting cannot be one number. If the plan only tracks weekly TSS, a 90-minute zone 2 ride and a 60-minute threshold session can be roughly equivalent in load — and missing either looks identical in the deficit column. They are not equivalent in adaptation. Filipas et al.'s 16-week trial in recreational male cyclists [Filipas 2024] showed that the pyramidal distribution — heavy on low intensity, moderate on threshold, small slice at VO2max — produced significant lactate-threshold and body-composition gains. The intensity ratio is what produced the result. A plan that smooths over which side of the ratio went missing is, by construction, not adapting to either. ## What you actually lose with each Skip a key workout and you lose a discrete adaptation event the rest of the week was scheduled around. Skip a recovery ride and you lose chronic-load volume that does not show up in next week's FTP test but does show up six weeks from now. Different mechanisms, different timelines. The missed key session. VO2max-targeted intervals — typical prescriptions cluster around 3 to 5 minutes at 106 to 120 percent of FTP, accumulating 15 to 20 minutes of total work — operate on a frequency model where one to two sessions per week is the working dose and three is roughly the upper bound before recovery breaks down [Mølmen et al. 2019]. Miss the only VO2 session in the week and the targeted block loses its central stimulus; miss it twice in three weeks and the block has effectively become a sweet-spot block, regardless of what the calendar says. The cost is not the missed TSS. It is the missed dose-response. The missed recovery ride. Zone 2 looks deceptively cheap to skip because it produces no acute fatigue, no lab-visible threshold movement, no sense of having earned anything. The mechanism it drives — mitochondrial biogenesis, capillary density, fat-oxidation capacity — is built through cumulative low-intensity volume across weeks, and missed volume cannot be back-filled with a harder ride [San Millán & Brooks 2018]. A rider who skips two zone 2 rides per week for six weeks has lost a meaningful share of the chronic aerobic stimulus the plan assumed they were getting. The FTP test in eight weeks looks fine; the three-hour-event ceiling at 12 weeks is where the deficit shows up. The comparative cost depends on context. Inside a six-week build block aimed at a hilly road race, a missed VO2 session is the more expensive miss because the block has a sharp specificity arc and few opportunities to redeliver the stimulus. Inside a 14-week base phase aimed at a long gravel event, a recurring missed zone 2 ride is the more expensive miss because the entire phase is a volume accumulation. The same physical absence — one workout — costs different fitness depending on what the surrounding weeks were trying to build. A plan that does not know which arc it is on cannot price the loss correctly. ## What an adaptive plan should do in each case The right plan response is asymmetric. A missed key workout should trigger redistribution — protect the next prescribed key session, consider reseeding the lost stimulus across two lighter days, and avoid compressing it into the next available slot. A missed recovery ride should trigger ramp-rate accounting — the chronic-load delta needs to be absorbed, not ignored. For a missed key session, the wrong move is the most common one: copy-paste the missed workout into tomorrow. Foster's monotony research [Foster 1998] showed across 25 athletes that the strongest behavioral predictor of illness was high load combined with low daily-load variance — exactly what stacking a missed VO2 session next to the originally prescribed one produces. The acute:chronic workload framework [Hulin et al. 2016] places elevated risk above ratios near 1.5, which a stacked-intensity rebound week routinely crosses. The right response is to defend the next key session at full quality, redistribute some of the missed stimulus across two reduced sessions later in the block, and accept the lost dose where the timing does not allow recovery. For a missed recovery ride, the failure mode is opposite. Most apps simply delete it. The next week resumes as if the chronic-load curve were unbroken, which means the rebuilt week is calibrated against an aerobic base it does not actually have. After 10 to 14 days of substantially reduced volume, Mujika and Padilla's detraining work [Mujika & Padilla 2000] documents measurable cardiovascular drift. Long before that point, the right behavior is to re-anchor the weekly volume target downward by the missed hours, hold the intensity ratio inside the polarized or pyramidal band [Seiler 2010], and resist the temptation to make up missed volume by adding tempo — which moves the rider exactly the wrong direction. This is the practical surface of the broader question of what adaptive cycling training plans actually have to do. TrainerRoad's Adaptive Training adjusts the next workout's difficulty but does not reshape the week around which kind of session went missing. JOIN handles redistribution moderately well for missed key workouts but tends to under-react to missed endurance volume. Xert's Training Load model treats both as load-curve perturbations, which catches the chronic-load effect but does not differentiate the dose-response cost of a missed VO2 session. The sibling spoke on signs your plan is not adapting covers how to probe these behaviors directly; the solver should carry the asymmetry into every reshuffle. ## Common questions **Is one missed VO2 session a bigger deal than one missed zone 2 ride?** Inside a build block, usually yes — VO2 prescriptions operate on a frequency model where one or two sessions per week is the working dose [Mølmen et al. 2019], and missing the only one makes the week sweet spot in disguise. Inside a base phase, the math reverses: cumulative low-intensity volume is the entire point, and a single missed long ride is a measurable hit to chronic load. **Can I just push the missed workout to tomorrow?** For a key session, almost never — stacking intensity collides with the next prescribed hard day and pushes acute:chronic workload outside the safe band [Foster 1998, Hulin et al. 2016]. For a missed recovery ride you can sometimes shift it 24 hours, but only if it does not displace tomorrow's hard session into a fatigued state. The honest answer is usually that the missed work is gone. **Does it matter when in the week I missed the workout?** Yes. A Tuesday miss with four days of plan left is more recoverable than a Saturday miss with one day. For key sessions specifically, missing late in the week with a Sunday hard ride still scheduled tends to produce the worst outcome — riders compress and arrive at Sunday already cooked. The redistribution math is what an adaptive plan is supposed to handle. **How do I tell if my app is handling this asymmetry?** Skip a prescribed VO2 session and a prescribed zone 2 ride in the same week. Check what changes. If the next week looks identical in both cases, the app is treating them as fungible TSS — which they are not. The pillar on adaptive cycling training plans covers the architecture; the spoke on signs your plan is not adapting covers the field test. ## References 1. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 2. **San Millán & Brooks 2018.** [Assessment of Metabolic Flexibility by Means of Measuring Blood Lactate, Fat, and Carbohydrate Oxidation Responses to Exercise in Professional Endurance Athletes and Less-Fit Individuals](https://pubmed.ncbi.nlm.nih.gov/28623613/). Sports Medicine. 3. **Filipas 2024.** [Effects of a 16-Week Training Program with a Pyramidal Intensity Distribution on Recreational Male Cyclists](https://pubmed.ncbi.nlm.nih.gov/38251291/). Sports (MDPI). 4. **Mølmen et al. 2019.** [Block periodization of endurance training — a systematic review and meta-analysis](https://pubmed.ncbi.nlm.nih.gov/31802956/). Open Access Journal of Sports Medicine. 5. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 6. **Hulin et al. 2016.** [The acute:chronic workload ratio predicts injury: high chronic workload may decrease injury risk in elite rugby league players](https://pubmed.ncbi.nlm.nih.gov/26511006/). British Journal of Sports Medicine. 7. **Mujika & Padilla 2000.** [Detraining: loss of training-induced physiological and performance adaptations. Part I: short term insufficient training stimulus](https://pubmed.ncbi.nlm.nih.gov/10966148/). Sports Medicine. --- # Can an adaptive cycling plan work without a goal race? URL: https://www.adaptcycling.com/guides/adaptive-plan-no-goal-race Updated: 2026-05-10 Author: Jim Camut Most adaptive cycling plans are sold as preparation for a specific event — a gran fondo, a goal race, an A-priority date eight to sixteen weeks out. If you do not have one, the obvious question is whether an adaptive plan is the right tool at all. The honest answer is yes, but the structure has to change. Without an event, the plan stops marching toward a peak and starts cycling through one — and what the rider anchors on shifts from race readiness to capacity and consistency. ## What changes structurally when there is no event date An adaptive plan without a goal event runs as a rolling improvement cycle — base, build, peak, short reload, repeat — instead of a linear march toward a single date. The macro arc still exists; it is just shaped like a loop rather than an arrow. The mesocycles get shorter, the peaks get smaller, and the plan re-anchors continuously rather than counting down. Friel's classical macrocycle [Friel 2018] organizes a competitive year into Prep, Base, Build, Peak, Race, and Transition phases pointing at a date. Strip the Race phase and the structure does not collapse — it shortens. Issurin's block-periodization model [Issurin 2010] is the cleaner fit here, because block periodization is already designed around concentrated training emphases rotating through 2-4 week mesocycles. A rider without a date runs the same blocks in sequence — aerobic base, threshold build, a short VO2max emphasis, a recovery block — and the engine schedules the next one when the current one closes. The plan is always inside a cycle, never inside a 12-week countdown that has nowhere to land. Practically the cycle length compresses from 16-24 weeks to roughly 8-12. A typical pattern is a 4-week base block, a 3-week build block, a 1-2 week mini-peak with a deliberate test or fitness-peak window, then a recovery week, then back to base. Each loop is short enough that a missed week costs less than it would inside a 16-week countdown — there is no taper to push, no event to slide. This is also where adaptive cycling training plans behave most differently from event-anchored ones: a plan that survives real life when the goal is rolling capacity rebuilds gently after a missed week because the next peak is six weeks away anyway, not nine days. The aggressive compression that fixed-date plans sometimes attempt is structurally unnecessary. The peak is also smaller. Bosquet and colleagues' 27-study taper meta-analysis [Bosquet et al. 2007] converged on roughly a 41-60% volume reduction over two weeks while preserving intensity for genuine race tapers — a tool that produces measurable competition gains, but also a tool that costs aerobic base. A no-event cycle replaces it with a 4-7 day mini-reload that drops volume modestly, lets freshness surface, and closes the block with a benchmark effort. The bigger taper is held in reserve for when an event eventually shows up. ## What does not change: the periodization invariants are date-agnostic The rules that govern an event-anchored plan apply identically to a rolling one. Seiler's polarized distribution, Foster's monotony rule, the chronic-load ramp, mandatory recovery weeks, and Mujika's detraining curve are all date-agnostic — they describe how an aerobic engine responds to load. Removing the event removes the date, not the physiology. Seiler's polarized model [Seiler 2010] — roughly 80% of training time at low intensity below the first lactate threshold, 20% at high intensity above the second, very little tempo in between — is the most replicated finding in endurance training. It applies to a self-coached rider in a permanent base block exactly as it applies to one tapering for a national championship. Filipas and colleagues' 16-week trial of recreational male cyclists [Filipas 2024] showed a pyramidal distribution produced significant lactate-threshold and body-composition gains over four months without any race-specific peaking — the distribution itself does the work. A rolling plan that drifts the easy days into tempo because there is no race to save up for loses the same fitness an event-anchored plan would. Foster's training-monotony research [Foster 1998] is similarly indifferent to whether a date exists. Across his 25-athlete cohort, the strongest behavioral predictor of illness and overtraining was high load combined with low daily-load variance — an athlete who rides similarly every day climbs this curve regardless of whether the season has a target. Recovery weeks every third or fourth week remain mandatory, and the chronic-load ramp from the Coggan framework still caps weekly increases inside roughly 4-7 TSS/day during a build. The temptation when the goal is just to get faster is to remove the deload because it feels unproductive. The athletes who follow that temptation plateau at month three. Detraining math also stays in force. Mujika and Padilla's foundational work [Mujika & Padilla 2000] documents that VO2max and maximal aerobic power begin meaningful decline within 10-14 days of insufficient training stimulus, with metabolic and neuromuscular changes inside two weeks. A no-event plan still has to handle missed weeks correctly — resuming at roughly 60-70% of pre-break load and ramping back gradually rather than slotting in at the original numbers. The absence of a goal date does not cushion the rider against deconditioning. It just means the plan has more flexibility about when the next peak lands. ## What to anchor on instead of an event Without an event, the plan needs a substitute target. The right anchor is a 12-16 week proxy — an FTP number, a sustained chronic-load band, a specific long hilly ride, or a self-defined fitness peak window. The proxy gives the cycle a closing condition that is not arbitrary, and it gives the plan something to re-anchor on when life disrupts. A useful default is an FTP target. The rider declares a 12-16 week goal — lifting FTP from 245 to 265 watts, or holding 3.8 W/kg for a 20-minute hill — and the plan structures its blocks around delivering it. This works because the FTP delta is a real periodization target with the same downstream consequences as an event: the build block has a measurable success criterion, the mini-peak has a benchmark, and a missed two-week stretch genuinely shifts the date the proxy is hit. Filipas et al.'s 16-week pyramidal trial in recreational male cyclists [Filipas 2024] is itself an example of this structure — recreational athletes with no race date achieving significant FTP and threshold gains across a defined window, anchored on the physiology rather than the calendar. A second option is a fitness band — a target chronic-training-load range the rider commits to building and then defending. A cyclist running CTL around 65 might commit to lifting it to 80 across three blocks and holding it there. The blocks build the number; the mini-peaks defend it. The same chart that monitors fatigue against an event works as a year-round capacity gauge. A third option is a route — a specific long hilly ride the athlete wants to complete cleanly, treated as a self-imposed event. The proxy does not need to be public or competitive; it just needs to be specific enough that the plan can reverse-engineer the build. The frame matters because no proxy at all collapses the cycle. Without a target, the rolling plan tends to flatten into perpetual tempo — every ride becomes generic, every block looks identical, and the rider drifts into the same plateau that hits self-coached cyclists with no periodization. A 12-16 week proxy reintroduces structure. The diagnostic spoke on whether a plan is actually adapting makes a related point from a different angle: a plan with no target to re-anchor on cannot meaningfully restructure either, because there is nothing for the restructure to point at. ## Common questions **Should I just pick a fake event date if I don't have a real one?** It is one valid option, but not the only one. A fake date works because it gives the plan a closing condition; the same purpose is served by an FTP target, a CTL band, or a route. The choice depends on what the rider actually finds motivating to chase. A fake date that feels arbitrary tends to lose pull around week six; a measurable physiological target with a real success criterion holds longer because the rider can see the progress against it. **How short can the cycles get before they stop working?** Below roughly 8 weeks the structure starts losing meaningful adaptation windows. Block periodization research [Issurin 2010] generally treats 2-4 week mesocycles as the working unit, and a full base-build-peak-reload loop wants at least three of them. Cycles shorter than that compress the build phase to the point where the mini-peak arrives before the work has accumulated. The practical floor is around 8 weeks; the comfortable range is 10-14. **Does the polarized 80/20 distribution still apply if I have no race?** Yes, and arguably more strongly. Seiler's distribution research [Seiler 2010] describes how an aerobic engine responds to load, not how an athlete prepares for a date. The 16-week pyramidal trial in recreational male cyclists [Filipas 2024] specifically tested a no-event population and reproduced the same pattern of gains. The more time the rider spends in the rolling structure, the more compounding the cost of drifting the easy days into tempo. **How do I know when to actually hit the bigger taper?** When an event eventually shows up. The rolling structure is built so that mini-reloads keep the engine fresh between blocks; a full Bosquet-style two-week taper [Bosquet et al. 2007] is held in reserve for when there is something to peak for. If a goal event materializes mid-cycle, the plan can pivot — the proxy becomes the real event, the next block becomes the build, and the closing reload becomes a real taper. ## References 1. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. 2. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. 3. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. 4. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 5. **Filipas 2024.** [Effects of a 16-Week Training Program with a Pyramidal Intensity Distribution on Recreational Male Cyclists](https://pubmed.ncbi.nlm.nih.gov/38251291/). Sports (MDPI). 6. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 7. **Mujika & Padilla 2000.** [Detraining: Loss of Training-Induced Physiological and Performance Adaptations. Part I: Short Term Insufficient Training Stimulus](https://pubmed.ncbi.nlm.nih.gov/10966148/). Sports Medicine. --- # Why your adaptive cycling plan keeps prescribing the same workouts URL: https://www.adaptcycling.com/guides/plan-prescribes-same-workouts Updated: 2026-05-10 Author: Jim Camut Self-coached riders ask this every few months: my plan keeps giving me the same threshold session on Tuesday and the same VO2max session on Thursday — is that normal? The honest answer is that it depends. Block periodization repeats workouts on purpose for a defined window, then transitions. A plan that repeats the same wattages and intervals indefinitely without progression is broken — and there are three specific reasons why. ## When repetition is the point: block periodization Issurin's block-periodization framework concentrates one workout type for two to four weeks, on purpose, because adaptation requires a coherent stimulus. Inside a threshold block you should expect 4-6 threshold sessions of similar structure. The repetition is the mechanism. What separates a healthy block from a stuck plan is what happens at week three and what happens at week five. Issurin's 2010 review of block periodization [Issurin 2010] is the cleanest articulation of why a good plan repeats workouts. The argument is that classical concurrent training spreads stimuli too thin to produce a measurable adaptation in any single capacity. Block periodization concentrates one or two compatible targets — threshold and tempo, or VO2max and high-cadence neuromuscular work — for a 2-4 week window so the residual fatigue and the residual training effect both build inside the same physiological pathway. The repetition is the load. A threshold block that runs only one threshold session per week is not a threshold block; it is a maintenance week with a label. Inside that block, the workouts can and should look similar. Allen and Coggan's Progression Levels concept [Allen et al. 2019] formalizes this: the same workout type repeats while the difficulty climbs in small increments — interval count, time-at-target, percentage of FTP, or rest-interval ratio. A 4x8 minute sweet-spot session at 88% of FTP in week one becomes 4x10 at 90% in week two becomes 5x10 at 92% in week three. The structure looks identical because it is the same workout family. The numbers underneath are not. This is the legitimate case the rest of this spoke is contrasted against. The block ends. That is the load-bearing word. Bosquet et al.'s 27-study taper meta-analysis [Bosquet et al. 2007] frames the end of a build block as a transition rather than a continuation — a deload week, then a new block targeting a different physiological system. Friel's Base/Build/Peak/Race/Transition macrocycle [Friel 2018] makes the same point at a longer timescale. The hallmark of healthy repetition is that the block transitions: into a deload, into a different block targeting a different system, or into a goal-event taper. If your plan never transitions, the repetition is not block periodization. It is something else. ## When repetition is broken: three failure modes A plan that repeats the same workouts indefinitely without progression is suffering from one of three causes: a thin workout catalog with nowhere to go, no progression model so the wattages never climb, or no goal-anchored periodization so the block never transitions. Each has a distinct fingerprint. All three are common in generic plans, and they often show up together. Failure mode one: thin workout catalog. Most generic 12-week PDFs and many template-driven apps draw from a library of roughly 20-30 workouts; TrainerRoad's library famously runs into the thousands. Catalog size matters because the rider who sees the same six workouts month after month is correctly perceiving that the variety is not there. The plan is not adapting to the rider; it is rotating through a small fixed set. Filipas et al.'s 16-week pyramidal-distribution trial in recreational male cyclists [Filipas 2024] structured the protocol around shifting workout types across the 16 weeks — base aerobic, then sweet spot, then VO2max, then race-specific — precisely because the stimulus needed to evolve. A plan that cannot evolve the stimulus does not produce the trial's results. Failure mode two: no progression model. The block runs Tuesday threshold and Thursday VO2max for four weeks, but week one and week four prescribe the same wattage, the same interval count, and the same rest. There is repetition without progression. The fingerprint is the wattage column: in a healthy build, the target percentage of FTP, the time-at-target, or the interval count climbs week-over-week inside the block. If those numbers are flat across four weeks, the plan has no Progression Levels equivalent [Allen et al. 2019] running underneath the structure. It is not building toward anything; it is holding station. Failure mode three: no goal-anchored periodization. The block runs and runs and never transitions. There is no deload week, no shift from threshold to VO2max, no taper pointing at a date. Friel's framework [Friel 2018] is structured around a goal event because the goal is what determines when the build ends and the peak begins. A plan with no goal — or a plan whose goal does not actually drive the macro arc — has no reason to ever transition out of the current block. It just keeps prescribing what it prescribed last week. This is where Foster's training-monotony work [Foster 1998] becomes operational: across his 25-athlete cohort, illness clustered in weeks combining high load with high daily-load variance, and indefinite repetition of the same hard structure is a textbook high-monotony pattern. ## How to tell which case you're in The diagnostic is a five-minute audit. Look at four weeks of your plan and check three things: is the wattage progressing inside the block, does the block transition at week four or five, and how many distinct workout types appear across an eight-week window. If the answers are no, no, and fewer than five, the repetition is broken — and which of the three failure modes applies usually becomes obvious in the same audit. Pull up the next four prescribed weeks and write down the Tuesday workout details for each: target percentage of FTP, interval count, interval duration, rest. If those numbers are flat — same wattage, same intervals, same rest week after week — the plan has no progression model. If they climb in small steps, the block is doing its job. Do the same for the Thursday workout. Seiler's intensity-distribution work [Seiler 2010] frames the principle here: well-trained athletes do not just repeat sessions; the distribution and the structure of those sessions evolve across a season. A plan that holds the same numbers indefinitely is not running that loop. Now look at week five. Does the structure change — a deload week, a transition to a different block, a different intensity emphasis? If week five looks identical to weeks one through four, the plan has no goal-anchored periodization. There is nothing telling it to ever transition. This is the failure mode that is hardest to spot inside a single week and easiest to spot across two months. Bosquet et al.'s taper data [Bosquet et al. 2007] only matters if the plan ever decides to taper, and a plan with no transition signal will never get there. Finally, count distinct workout types across eight weeks. A reasonable build cycle should touch 6-10 distinct workout structures: at least two threshold variants, at least two VO2max variants, an endurance ride, sweet spot, an over-under or microburst, and a recovery template. If your eight-week window contains four or five distinct workouts on rotation, the catalog is the problem. This is the broader question the pillar on adaptive cycling training plans returns to: a plan that survives real life has to read your data, restructure when the week breaks, and progress the stimulus across blocks. A plan that fails the four-week audit fails at the third leg before the first two are even tested. The sibling spoke on signs your training plan is not actually adapting covers the unplanned-ride and missed-week probes; this audit covers the inside-the-block question those probes do not reach. ## Common questions **Is repeating the same workout twice a week always a bad sign?** No. Inside a 2-4 week block, a Tuesday threshold session and a Thursday VO2max session can absolutely repeat — that is what Issurin's block periodization [Issurin 2010] prescribes. The question is whether the wattage, interval count, or time-at-target is progressing inside the block, and whether the block transitions to something different at week four or five. Repetition with progression and transition is healthy. Repetition without either is broken. **How long should I let a repeating block run before deciding the plan is stuck?** Six weeks is the right window. A standard build block is 3-4 weeks plus a deload, and Friel's macrocycle [Friel 2018] has the next block looking different from the last. If the same workouts at the same wattages keep appearing past week six with no deload and no transition, the plan is not running a periodization loop — it is rotating a fixed template. At that point, override or switch tools. **What if the workouts repeat but the wattages are climbing?** Then the plan is doing its job. Allen and Coggan's Progression Levels concept [Allen et al. 2019] is built on exactly this pattern: same workout family, climbing difficulty. A 4x8 sweet spot at 88% becoming 5x10 at 92% across three weeks is not a stuck plan; it is a working one. The visual sameness is misleading. The numbers underneath are what matter. **Could a thin catalog still work if the progression model is good?** Partially. A plan with 20 workouts and a clean Progression Levels system can produce real fitness gains for a season — the wattages climb even if the structures repeat. The ceiling shows up at year two and beyond, where Filipas et al.'s 16-week protocol [Filipas 2024] and Seiler's distribution work [Seiler 2010] both depend on shifting workout types across blocks. A thin catalog cannot evolve the stimulus across a full periodized year, so the year-over-year gains slow. **How does this relate to whether my plan is adapting at all?** Workout repetition is the inside-the-block question. The sibling spoke on signs your training plan is not actually adapting covers the cross-block questions: does it respond to a missed key workout, does it rebalance after an unplanned hard ride, does it restructure after a sick week. Both checks together cover the full surface area. A plan can pass one and fail the other, and the fixes are different. ## References 1. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. 2. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 3. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. 4. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 5. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. 6. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 7. **Filipas 2024.** [Effects of a 16-Week Training Program with a Pyramidal Intensity Distribution on Recreational Male Cyclists](https://pubmed.ncbi.nlm.nih.gov/38251291/). Sports (MDPI). --- # Adaptive vs static training plans: how to spot a rebrand URL: https://www.adaptcycling.com/guides/adaptive-plan-vs-rebranded-static Updated: 2026-05-10 Author: Jim Camut Every cycling app in 2026 calls itself adaptive. Most are static plans wearing a dynamic UI. The difference is structural, not aesthetic — a real adaptive plan reads what you rode, restructures when your week breaks, and explains why each session exists. A rebranded static plan ships a fixed schedule, swaps a workout when you fail one, and calls that adaptation. This is the framework for telling them apart before you subscribe. ## The three tiers of adaptation: intensity, load, life Adaptation is not one thing. It is three operational tiers — intensity-adaptive, load-adaptive, and life-adaptive — and almost every plan that markets itself as adaptive lives in tier one. Tier three is what most amateurs actually need, and it is the rarest engineering in the category. Tier one is intensity-adaptive. Inside a fixed weekly skeleton, the next workout gets harder or easier based on how the last one went. TrainerRoad's Adaptive Training is the cleanest implementation — Progression Levels move workout difficulty inside Plan Builder structure, and at $21.99/month it does that job well. Wahoo SYSTM and Garmin's Daily Suggested Workout sit in roughly the same tier. The plan shape is fixed; the dial inside each session moves. Tier two is load-adaptive. The training-load target itself shifts based on current fitness and recent rides. Xert's proprietary Training Load model continuously fits curves to every effort and recommends the next session against a fitness signature; Intervals.icu computes CTL, ATL, and TSB in the Coggan framework [Allen et al. 2019] and lets coaches plan against the chart. Both move the load target, not just the workout difficulty. Tier three is life-adaptive. The plan itself restructures when life intrudes — a sick week, a missed Wednesday VO2 session, a 90-minute group ride that wasn't on the schedule. This requires four systems working together: a memory of your goal and its constraints, a live connection to what you actually rode, a generator that respects periodization invariants the model cannot violate, and a solver that can reshape weeks without abandoning the macro arc. Most apps stop at tier one. We cover the wider framework in our pillar on adaptive cycling training plans, including which tier matches which kind of athlete. ## Five structural tells of a rebranded static plan Marketing pages are useless for distinguishing adaptive from static. The product behavior is not. Five structural signals separate genuinely adaptive systems from static plans wearing dynamic UI: ship date without rider data, weekly templates that ignore Strava, fixed compliance scoring, no transition logic across blocks, and intensity-only autopilot. Tell one — the plan ships before the app reads your training. If a free trial generates a 12-week plan after three onboarding questions and zero historical activities, the plan is a template. A real adaptive system needs a baseline — at minimum the last 6 weeks of CTL/ATL [Allen et al. 2019], ideally 90 days of power data to anchor the FTP estimate. Without it, peak-week volume is a guess against a population mean. Tell two — weekly templates that survive a wildly off-script week. Did a 90-minute group ride at 230 normalized watts on Tuesday when Tuesday was supposed to be recovery? In a real adaptive plan, Wednesday's workout is rebuilt; in a rebranded one, the template still says intervals at 92% of FTP. The intensity stimulus is double-booked, the athlete absorbs it, and Foster's training-monotony work [Foster 1998] across 25 athletes documents that high load combined with low daily-load variance is the single most reliable behavioral predictor of illness — exactly what compounding double-stress weeks produce. Tell three — fixed compliance scoring. If the dashboard shows a green checkmark for completing the workout-as-prescribed and a red X for deviating, the plan rewards adherence to a schedule rather than progress toward a goal. A binary compliance score throws every signal away — the rider who deviated because life intruded looks identical to the rider who deviated because the prescribed work was wrong. Plans that adapt do not punish deviation — they absorb it as input. Tell four — no transition logic across blocks. Block periodization [Issurin 2010] only works if the next block adjusts to what the previous block actually delivered, not what it was supposed to. If a base block ends with the athlete 30% under prescribed volume because of travel and the build block starts on the original schedule anyway, no transition logic exists. The plan moved forward in time; it did not adapt. Tell five — intensity-only autopilot. The most common rebrand is an intensity dial that moves while the weekly structure stays frozen. This is tier one dressed as tier three. The give-away is asking the support team a single question: what does the plan do if I miss an entire week? If the answer is "we drop the next workout one Progression Level" or "pick up where you left off," the plan does not restructure — and Mujika and Padilla's detraining work [Mujika & Padilla 2000] makes clear that picking up where you left off after a 7-14 day break is the wrong dose, since VO2max begins meaningful decline inside that window. ## A 10-minute audit of any plan's free trial You don't need to subscribe to test for adaptation. Three checks inside a free trial reveal whether the plan reads your data, restructures your week, and respects your goal — and they take ten minutes total. Run them before you pay, not after. Check one — connect your ride history, then look at week one. If the prescribed weekly hours match what you actually rode the previous month within ±20%, the plan read your data. If it prescribes 10 hours a week to a rider whose 12-week rolling average is 5, the onboarding questionnaire is the only input — your ride history is decoration. The fix isn't a stretch goal; the math should anchor to your CTL [Allen et al. 2019]. Check two — manually log a fake 90-minute hard ride yesterday. A genuinely adaptive plan rebuilds today and tomorrow; a static one keeps the original sequence. The four-week tapering meta-analysis from Bosquet [Bosquet et al. 2007] across 27 studies and 439 athletes establishes that volume must drop 41-60% over two weeks while intensity is preserved — the principle that hard days require easy days extends to any week, not just race week. A plan that can't adjust 24 hours after an unplanned hard effort cannot adjust a taper either. Check three — read the goal field. Genuinely adaptive systems require a date, an event type, and an hours-per-week ceiling because peak-week volume cannot be inferred from history alone — historical means are poisoned by illness, travel, and taper weeks. If the plan accepts "get faster" as a goal and generates a 16-week structure anyway, the periodization is a template — Kiely's critique [Kiely 2018] on periodization theory is precisely about systems that paper over individual variability with rule-based templates. The diagnostic version of this audit, for plans you've already been training off for weeks, lives in our sibling guide on signs your training plan is not adapting. ## When intensity-adaptive is enough — and when it isn't Tier one is not bad engineering. For some riders it is the right product. The fit depends on schedule variance: how far this week is from a normal week. Below 15% week-over-week variance, intensity-adaptive works; above that, the plan needs to restructure or it falls apart. If your week is a metronome — same five training slots, same hours, same recovery patterns — TrainerRoad-tier intensity adaptation is sufficient and probably better than the alternatives because the engineering is mature and the workout library is deep. The polarized 80/20 distribution Seiler documented [Seiler 2010] runs cleanly inside a fixed weekly skeleton when the skeleton itself is real. The plan does not need to restructure because the week does not change. Above roughly 15-20% schedule variance week-over-week — kids, travel, shift work, an unpredictable job — intensity-adaptive plans accumulate failed sessions. Each one drops a Progression Level, the plan compensates by softening the next workout, and within four weeks the prescribed load is well below what the athlete can absorb. The fix is not training harder; it is a system that reshapes the week. This is where life-adaptive plans like AdaptCycling, JOIN (€16.99/month), and a coach with a deep TrainingPeaks setup pull ahead — they treat the missed Wednesday as input, not failure. ## Common questions **Is TrainerRoad's Adaptive Training actually adaptive?** Yes, in tier one. Adaptive Training adjusts workout difficulty inside a fixed Plan Builder structure based on Progression Levels, and the engineering is solid. It does not restructure the week when life disrupts the schedule — that is a different tier. If your training week is consistent, it is the right tool; if it varies week-to-week, you will outgrow what intensity-only adaptation can do. **Can a coach with TrainingPeaks deliver a life-adaptive plan?** Yes, and historically this was the only way. A coach who reviews your TrainingPeaks calendar daily, restructures your week after a missed session, and understands the Coggan framework delivers tier-three adaptation. The constraint is cost (several hundred dollars per month) and response latency — most amateur coaching contracts review the week once, not daily. Software-based life-adaptive plans aim to deliver the same restructuring at $20-30/month with minute-level latency. **Why don't more apps offer life-adaptive plans?** It is hard engineering. A life-adaptive plan requires four integrated systems — data ingestion from your ride history, persistent memory of goals and constraints, an LLM-based generator constrained by deterministic periodization solvers, and a re-planning engine that runs nightly. Most apps shipped tier-one intensity adaptation in 2020-2022 because the workout-difficulty problem is bounded; tier-three restructuring requires the plan generator and the solver to talk every night, which is a much larger system. **Is JOIN a true adaptive plan?** JOIN sits between tiers two and three. It rebuilds the week when you flag a schedule change and adjusts intensity based on completed workouts. Where it tends to struggle is high week-over-week variance — when the schedule shifts substantially three weeks in a row. For moderate disruption it works well at €16.99/month; for chaotic schedules the restructure logic gets tested. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 3. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. 4. **Mujika & Padilla 2000.** [Detraining: loss of training-induced physiological and performance adaptations. Part I: short term insufficient training stimulus](https://pubmed.ncbi.nlm.nih.gov/10966148/). Sports Medicine. 5. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. 6. **Kiely 2018.** [Periodization Theory: Confronting an Inconvenient Truth](https://pubmed.ncbi.nlm.nih.gov/29189930/). Sports Medicine. 7. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. --- # Unplanned group ride training plan adaptation: what should change in the next 24 hours URL: https://www.adaptcycling.com/guides/unplanned-group-ride-training-plan Updated: 2026-05-10 Author: Jim Camut You went out for a 60-minute zone 2 spin and came home with 90 minutes at 230 normalized watts because someone attacked the city-limit sign. Your plan still says 4×10 at threshold tomorrow. The single most-important question for any adaptive cycling training plan is what it does in the next 24 to 48 hours when life adds a hard ride to the week — and the wrong answer is the most common one. This is what should change, why, and how the mechanism works. ## Why an unplanned hard ride is information, not a problem An unplanned group ride that lands harder than scripted is not a plan failure. It is exactly the kind of disruption a life-adaptive plan exists to absorb. The job is not to pretend the ride did not happen — it is to read what happened and reshape the next 48 hours so the week's intensity budget still lands inside the band the periodization was built around. Strava-connected plans see the ride within seconds of the upload. Normalized power, intensity factor, time-in-zone, and TSS are computed from the standard framework Coggan and Allen formalized for cycling [Allen et al. 2019] — a 90-minute ride at NP 230 on a 250-watt FTP returns IF 0.92 and roughly 127 TSS, comfortably above the threshold-day load that was actually scheduled. From the plan's standpoint, an intensity stimulus has already been delivered for the week. The question is no longer whether to do tomorrow's threshold session as written. It is what tomorrow should be given that the work is already booked. Halson's training-load review [Halson 2014] frames the principle: monitoring exists to convert what actually happened into what the next session should be. An unplanned ride the plan never reads is unmonitored by definition. The ride's physiological cost does not vanish because it was off-script — whatever happens tomorrow has to account for it. ## What the plan should do in the next 24 to 48 hours The right answer depends on what the unplanned ride displaced and what is on the calendar next. Three patterns cover most cases: a recovery day turned into a hard group ride should swap tomorrow's quality session for an easier slot. An easy day pushed into tempo should compress, not delete, tomorrow's quality. A scripted hard day blown into an all-out smashfest should shift the week, not just the day. Pattern one — recovery rewritten as a hard ride. The week budgeted one quality session and you delivered two days early. The right move is to swap tomorrow's threshold prescription for zone 2 endurance and push the threshold work to the next available recovery-buffered slot — typically 48 to 72 hours out. The acute:chronic workload framework [Hulin et al. 2016] places elevated soft-tissue and illness risk above ratios near 1.5, which back-to-back hard days from a recovered baseline routinely cross. Foster's monotony research [Foster 1998] is the same lesson from the other side: 25-athlete data showed high load combined with low daily-load variance was the strongest behavioral predictor of overtraining. Stacking yesterday's NP 230 onto tomorrow's prescribed 4×10 produces both signals at once. Pattern two — easy day pushed into tempo. You planned 90 minutes of zone 2 and ended up holding 78 percent of FTP for an hour because the group surged. The intensity is not far enough into the high band to count as the week's quality session, but the ride is no longer a recovery contributor either. Keep tomorrow's quality session in place and trim 20 to 30 percent off its duration — the aerobic system is fine, but the freshness budget is partially spent. The dose-response of the threshold session does not collapse if you shorten the work without changing the target. Pattern three — scripted hard ride turned all-out. The plan called for 4×8 at sweet spot and the group ride became a 2.5-hour breakaway with three full-gas climbs and a sprint. Acute load can land 60 to 80 percent above the prescribed session. This is a week-shape problem, not a one-day adjustment. The right response is to convert the next two days to easy or off, replace the week's second quality session with endurance, and re-anchor the rest of the week against an acute load that already exceeds what the build was scheduled to absorb. Coggan and Allen's CTL ramp guidance [Allen et al. 2019] keeps weekly chronic-load increases inside roughly 4 to 7 TSS-per-day per week during a build phase. A ride that pushes a week 25 TSS over plan needs absorption time, not bonus work. ## Why the weekly intensity ratio matters more than the daily TSS total An adaptive plan that reshuffles around an unplanned ride is solving for the wrong variable if it only tracks weekly TSS. The variable that drives the adaptation outcome is the time-in-zone ratio across the week. A week that hits its TSS number with the wrong distribution is not a week that hit its target. Seiler's polarized model [Seiler 2010] is the most replicated finding in endurance training: roughly 80 percent of training time below the first lactate threshold, 20 percent above the second, very little tempo in between. Stöggl and Sperlich's descriptive analysis of well-trained athletes [Stöggl & Sperlich 2015] confirmed elites cluster around polarized or pyramidal patterns. Filipas et al.'s 16-week trial in recreational male cyclists [Filipas 2024] found that pyramidal distribution — heavy low intensity, moderate threshold, small slice at VO2max — produced significant lactate-threshold and body-composition gains. Three independent lines of evidence pointing at the same conclusion: the ratio is what produces the result, not the gross load. The implication for an unplanned-ride day is direct. A 90-minute group ride at NP 230 on a 250-watt FTP probably booked 35 to 45 minutes above the second threshold — most of the week's high-band budget in one ride. If tomorrow's 4×10 at 95 percent of FTP runs as written, the week ends with 70 to 90 minutes above LT2 and a substantially undersized low band, which is not a polarized or pyramidal week. The plan's job is to defend the ratio, which usually means making tomorrow easier than the calendar said and adding zone 2 volume on the back half of the week to rebalance. This is also why an adaptive plan cannot just look at the day. A reshuffle that fixes tomorrow but lets the rest of the week drift into too much tempo replaces one ratio failure with another. The reverse error is just as real — too much zone 2 padding to compensate for an accidental hard ride pushes the rider toward the high-monotony pattern Foster's work [Foster 1998] flagged as the strongest predictor of overtraining. The week is the unit of accounting; the day is just where the adjustment shows up. ## What a static plan does wrong here Static and intensity-only-adaptive plans handle this case in three predictable ways: they ignore the ride entirely, they adjust tomorrow's workout difficulty without changing its kind, or they layer tomorrow's threshold work on top of yesterday's unplanned effort. All three produce the same downstream outcome — a week with the wrong intensity distribution and an elevated overtraining signature. Pure static plans — the calendar-PDF model still common in coaching templates — do nothing. The rider rode 127 TSS yesterday and the plan still prints 95 TSS at threshold today, which means a week budgeted for a single quality stimulus delivers two and lands 25 to 40 percent over its acute-load target. Foster's monotony framework [Foster 1998] and the acute:chronic workload work [Hulin et al. 2016] both flag this as the textbook overtraining trigger. The rider does not feel it on day three. They feel it on day 18. Intensity-adaptive plans like TrainerRoad's Adaptive Training are better at the workout-difficulty layer — they correctly sense the rider arrives at tomorrow's session with reduced freshness and dial the prescription down a notch. What they do not do is change which session is on tomorrow. The kind of stimulus is unchanged; only the magnitude moves. Inside a week where the high-band budget is already spent, dialing the threshold session from 95 to 92 percent of FTP is the wrong axis of adjustment. The right axis is swapping the session, not softening it. Plans that track TSS without distribution — the trap most chronic-load-only systems fall into — handle gross load but get the ratio wrong in the opposite direction, padding the week with tempo to top up TSS while quietly inverting the polarized distribution Seiler's, Stöggl and Sperlich's, and Filipas's research [Seiler 2010, Stöggl & Sperlich 2015, Filipas 2024] all converge on. The sibling spoke on signs your plan is not adapting covers the diagnostic probe directly. ## Common questions **Should I just skip tomorrow's planned workout entirely after a hard group ride?** Usually not — the right move is a swap, not a delete. If the unplanned ride covered the week's intensity budget, replace tomorrow's quality session with zone 2 endurance and push the prescribed work to a recovery-buffered slot 48 to 72 hours later [Foster 1998, Hulin et al. 2016]. Deleting altogether under-loads the week; doing both stacks the wrong intensity signature. **How hard does the group ride have to be before the plan should change?** A useful threshold is intensity factor above 0.85 for 30+ minutes, or any ride that books meaningful time above LT2. Below that, the ride is closer to a long endurance day than a quality session and tomorrow can usually run as written. Above it, the high-band budget for the week is partially or fully spent and the plan should reshape [Allen et al. 2019]. **Does this also apply to unplanned long rides that were not hard?** Sometimes. A surprise four-hour endurance ride at zone 2 does not blow the week's intensity ratio, but it does push acute load and freshness. The right adaptation is usually to keep tomorrow's quality session and trim its duration 20 to 30 percent — the aerobic system is unchanged, the freshness is partially spent. This is the same broader question of how much an adaptive cycling training plan should restructure when life adds volume the schedule did not budget. **What if I do unplanned hard rides every weekend?** Then the plan needs to know that and rebuild around it, not treat each one as a one-off disruption. A recurring Saturday group ride is a feature of your training, not noise. The plan should anchor weekend intensity around it and program the rest of the week to support and recover from it — which is what a life-adaptive engine should do automatically once it sees the pattern in the data. ## References 1. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 2. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 3. **Stöggl & Sperlich 2015.** [The training intensity distribution among well-trained and elite endurance athletes](https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2015.00295/full). Frontiers in Physiology. 4. **Filipas 2024.** [Effects of a 16-Week Training Program with a Pyramidal Intensity Distribution on Recreational Male Cyclists](https://pubmed.ncbi.nlm.nih.gov/38251291/). Sports (MDPI). 5. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 6. **Halson 2014.** [Monitoring training load to understand fatigue in athletes](https://pubmed.ncbi.nlm.nih.gov/25200666/). Sports Medicine. 7. **Hulin et al. 2016.** [The acute:chronic workload ratio predicts injury: high chronic workload may decrease injury risk in elite rugby league players](https://pubmed.ncbi.nlm.nih.gov/26511006/). British Journal of Sports Medicine. --- # Training plan after a sick week: what your plan should actually do to itself URL: https://www.adaptcycling.com/guides/plan-restructure-after-sick-week Updated: 2026-05-10 Author: Jim Camut This is not a return-to-riding piece. It is about what the plan does to itself when you have lost a week to a normal illness — flu, sinus infection, COVID, a GI bug. The 5 to 10 day disruption is one of the most common reasons amateur plans break, and the right response — skip, compress, restart, push the goal, downgrade — depends on which week you were in and how far the goal sits. If you have been off the bike longer than 10 days, the sibling guide on returning after time off covers the deeper-hole case. ## What detrains in a week and what does not Inside 5 to 10 days off, the rider has lost cardiovascular adaptations and almost nothing else. Coyle and colleagues' classic detraining time-course study [Coyle et al. 1984] documented a 7% drop in VO2max across the first 21 days of inactivity in trained subjects, stabilizing around 16% below baseline by day 56. The peripheral adaptations — mitochondrial density, capillary density, fat oxidation capacity — barely move in a week. The mechanism is almost entirely cardiovascular. Mujika and Padilla's foundational short-term detraining review [Mujika & Padilla 2000] documents that VO2max decline inside the 4-week window is driven primarily by a rapid reduction in blood volume and stroke volume, with maximal cardiac output dropping accordingly. Recently acquired gains drift fastest — a rider three weeks into a sweet-spot block has more to lose proportionally than a rider in week ten of a long base. The same cardiovascular adaptations rebuild quickly when load resumes, often within 2 to 3 weeks of structured riding. The plan does not need to mourn the missed block; it needs to hand the rider back a week that respects the temporarily lower ceiling. The corollary is that fitness did not vanish. The chronic training load number — CTL in the Coggan and Allen framework [Allen et al. 2019] — drifts down by roughly 5 to 8 points across a 7-day off-week depending on starting CTL, but the underlying aerobic engine is largely intact. This is what the plan has to operate on: a real but modest decay, concentrated in the cardiovascular layer, with a 2 to 3 week rebuild horizon. ## Skip versus compress: the most common amateur-coaching error A static plan, faced with a missed week, compresses. The next week starts at 100% of the original target and the missed peak block is squeezed into fewer remaining weeks at the same or higher weekly intensity. This is the wrong answer in almost every case. The right move is to skip the missed work, accept the phase has shortened, and rebuild forward — not backward. The compression failure is operationally a chronic-load ramp violation. Hulin and colleagues' acute:chronic workload framework [Hulin et al. 2016] places elevated injury risk above ratios near 1.5, with high-chronic-load athletes plus a two-week ratio of 1.54 reaching 28.6% injury risk in the cohort. A normal-volume week dropped onto a 7-day-decayed chronic load lands in that danger band by construction; a compressed week with extra intensity to make up missed work lands well outside it. Foster's training-monotony work [Foster 1998] reinforces the point: across his 25-athlete cohort, the strongest behavioral predictor of next-cycle illness was the combination of high load and low daily-load variance, exactly what stacking made-up sessions produces. This is the implementation-level detail behind sign three from the sibling spoke on whether your plan is actually adapting: a sick week should make the plan restructure, not shrink. The ECSS and ACSM joint consensus on overtraining [Meeusen 2013] is explicit that resuming training without a graded reintroduction is a leading cause of non-functional overreaching — the failure mode amateurs most often produce by compressing out of guilt. The right behavior is asymmetric: skip the missed key session entirely, downgrade and shift if it was endurance volume, and treat the next week as the restart, not the catch-up. ## What an adaptive plan should change: the four moves A genuinely adaptive plan makes four specific moves when a 5 to 10 day illness gap closes. The chronic-load ramp resets to the new lower CTL. The first week back caps intensity time at 60 to 70% of pre-break load. The next deload moves up by one cycle. And the macro arc gets re-anchored to the new dates — not by compressing what got missed, but by accepting the phase is one week shorter. The first two moves are the week-of mechanics. Allen and Coggan's CTL framework [Allen et al. 2019] provides the math: keep weekly TSS climbing inside the safe ramp band of roughly 4 to 7 TSS/day per week starting from the post-illness CTL number, not the pre-break one. The first week back is endurance-heavy by design — most of the missed cardiovascular adaptation rebuilds with low-intensity volume, not stacked threshold work. The graded-reintroduction principle the Meeusen consensus [Meeusen 2013] is built on is explicit on this point: resuming at pre-break load is the leading mechanism behind non-functional overreaching. Foster's monotony rule [Foster 1998] should govern the daily-load variance: the rebuilt week needs at least one genuine recovery day, ideally two, even if the original plan only had one. Move three is structural. If the next deload was three weeks out, it now lands two weeks out — the post-illness week is in load terms already half a deload, and adding another high-load block before recovery violates the periodization invariant the build cycle was structured around. The build-block rhythm [Allen et al. 2019] is 2 to 4 weeks of progressive load, one week of recovery, repeat. A missed mid-block week shifts the rhythm one cycle, not zero. Move four is the macro-arc decision, which is the next section. ## When the goal event has to move — and when it does not The honest rule is mathematical: if more than six weeks separate the rider's first ride back from the event date, the goal almost certainly does not move. If fewer than three weeks separate them and the event is A-priority, the goal usually does. The middle band of three to six weeks is the judgment zone, and the answer depends on what phase the missed week was in. Six-plus weeks out, a 5 to 10 day gap is absorbable. The post-illness CTL drop of 5 to 8 points is roughly one normal build block of recovery; the rider can rebuild, hit the planned peak, and taper into the date with the standard Bosquet protocol — 41 to 60% volume reduction over two weeks while preserving intensity [Bosquet et al. 2007]. The macro arc compresses by one week, the build runs one cycle shorter, the taper math holds. Inside three weeks, the math collapses. The taper window is the rider's cardiovascular-rebuild window, and the two cannot share the same days. An A-priority event inside that window either gets pushed back, downgraded to B priority with reduced expectations, or attended for participation rather than performance. The middle band is where the missed phase matters. If the gap fell during a base block, the phase can usually still finish on time — the missed adaptations rebuild quickly. If the gap fell during a peak or specificity block, where the targeted stimulus is dose-response sharp, the plan should either extend the phase by a week or accept a smaller peak. AdaptCycling treats the goal-event date as a movable parameter inside the restructure flow — when the rider marks an illness, the engine reads the remaining weeks, the affected phase, and the priority level, then either holds the date or proposes a move with the math attached. This is what the parent pillar on adaptive cycling training plans means by life-adaptive: the plan absorbs the gap and reshapes the arc instead of asking the rider to be their own coach in the worst possible week. The single-day missed-workout case follows a different load-decay model — closer to a redistribution problem than a rebuild — and the sibling spoke on missed-key-workout-vs-missed-recovery covers it. ## Common questions **Should I do a fitness test in the first week back?** No. The cardiovascular layer is still rebuilding and blood volume has not recovered; a fresh FTP test in the first 7 days back returns a depressed number the rest of the plan then anchors on. Coyle and colleagues [Coyle et al. 1984] documented stroke-volume and VO2max decline that is reversible within 2 to 3 weeks of structured training. Wait until at least week three back — anything earlier locks in the worst day. **How much should I cut volume in the first week back?** Roughly 40% off pre-break weekly TSS, with intensity time capped at 60 to 70% of normal. The Meeusen consensus [Meeusen 2013] frames graded reintroduction as the prevention mechanism for non-functional overreaching, and the acute:chronic workload framework [Hulin et al. 2016] keeps the rebuilt-week ratio inside the protective 0.85 to 1.35 band rather than the high-risk zone above 1.5. Most of the rebuild week is zone 2 endurance with one short tempo or threshold opener at most. **What if I feel fine on day three of the rebuild week and want to add intensity?** The temptation is the trap. Symptom resolution and cardiovascular readiness are not the same thing — blood volume can lag perceived recovery by a week. Mujika and Padilla [Mujika & Padilla 2000] document that the cardiovascular layer is what moved, not the muscular layer, so the rider feels strong on the bike before the engine has fully rebuilt. Hold the rebuild week as designed and add intensity in week two. **How is this different from the diagnostic question of whether my plan is adapting at all?** The diagnostic spoke on signs your training plan is not actually adapting covers how to detect whether your plan is restructuring or compressing. This guide covers what the restructure should specifically do — the four moves and the goal-event decision tree. The pillar on adaptive cycling training plans frames the broader architecture; the diagnostic spoke is the field test; this spoke is the implementation. ## References 1. **Coyle et al. 1984.** [Time course of loss of adaptations after stopping prolonged intense endurance training](https://pubmed.ncbi.nlm.nih.gov/6511559/). Journal of Applied Physiology. 2. **Mujika & Padilla 2000.** [Detraining: Loss of Training-Induced Physiological and Performance Adaptations. Part I: Short Term Insufficient Training Stimulus](https://pubmed.ncbi.nlm.nih.gov/10966148/). Sports Medicine. 3. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 4. **Hulin et al. 2016.** [The acute:chronic workload ratio predicts injury: high chronic workload may decrease injury risk in elite rugby league players](https://pubmed.ncbi.nlm.nih.gov/26511006/). British Journal of Sports Medicine. 5. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 6. **Meeusen 2013.** [Prevention, diagnosis and treatment of the overtraining syndrome: Joint consensus statement of the European College of Sport Science and the American College of Sports Medicine](https://pubmed.ncbi.nlm.nih.gov/23247672/). European Journal of Sport Science. 7. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. --- # Goal race rescheduled mid-block: what an adaptive training plan should actually do URL: https://www.adaptcycling.com/guides/goal-race-moves-mid-block Updated: 2026-05-10 Author: Jim Camut The gran fondo got pushed three weeks for weather. The USAC calendar shifted your A-race forward by ten days. The gravel event you registered for added an extra week. Your plan still counts down to the original date. Most plans, faced with a moved goal, just slide the calendar — same blocks, same workouts, new labels. That is the wrong answer in both directions. Moving a date earlier collapses the taper window. Moving it later turns an unbuilt rest week into accidental load. The right response is structural, and it depends on which direction the date moved. ## Why a date change is a structural problem, not a calendar problem An adaptive plan is built around a specific arc — base, build, peak, taper, race — with each phase sized to the date. Moving the date does not just shift labels. It changes which phase the rider is in, what residual adaptations are still in play, and whether the taper math can still execute. A static plan that slides the calendar without reshaping the arc breaks every assumption the original build was sized against. Issurin's block-periodization review [Issurin 2010] frames the constraint directly: each training block produces residual adaptations that decay on different timescales, and the sequence of blocks is sized to deliver the targeted abilities at the date. Aerobic endurance residuals persist for roughly weeks; high-end anaerobic glycolytic residuals decay inside days. A peak block placed three weeks before a date is not interchangeable with one placed six weeks before — the residual window for the high-end adaptations has either elapsed or not yet been built. Sliding the calendar without re-sizing the block sequence breaks the alignment between residual decay and event day. The taper itself is the second hard constraint. Bosquet and colleagues' 27-study taper meta-analysis [Bosquet et al. 2007] converged on a two-week taper with a 41-60% exponential volume reduction while preserving training intensity and frequency. The 2023 Wang and colleagues systematic review and meta-analysis [Wang et al. 2023] converged on a similar window — a taper of 21 days or fewer with 41-60% volume reduction is the evidence-anchored recommendation. The taper is a fixed-cost period at the end of the arc; it does not compress, and it does not expand productively. A goal date that moves changes how many weeks remain for the taper to land cleanly, and that arithmetic — not the workout calendar — is what the plan has to re-solve. The pillar on adaptive cycling training plans frames this as one of the four invariants any plan must respect; this is what re-anchoring looks like in practice. ## When the date moves earlier: the taper is the constraint When the goal date moves closer, the taper becomes the binding constraint. The taper window is fixed by physiology — roughly two weeks of reduced volume with preserved intensity — and it cannot share days with the build. The plan has to decide whether the build phase shortens, the peak gets skipped, or the race priority drops. The decision turns on how many weeks remain. If the new date sits more than four weeks out, the build phase can usually compress by one cycle without breaking the arc — the rider loses one mesocycle of progression, the taper math holds, and the residual windows from the block-periodization framework [Issurin 2010] still land inside the event. If the new date sits between two and four weeks out, the build is over by definition. The remaining time is taper plus opener, and any high-load work in the front half of that window is borrowed against the freshness the Bosquet protocol [Bosquet et al. 2007] is supposed to surface. The plan should stop building, drop volume by 30-40% in week one, and run the back week as the standard 41-60% reduction. Inside two weeks the math collapses. The taper is the only thing the remaining days can produce, and a rider trying to squeeze one more sweet-spot block in week one is loading fatigue the taper cannot shed. Mujika's intense-training-during-taper review [Mujika 2010] is unambiguous on the trade: high-intensity work during the taper preserves and sharpens adaptations, but only when total volume is reduced — adding load to a compressed taper window erases the freshness gain. The honest move when the date pulls inside two weeks is to drop expectations to a B-priority effort, run the taper as designed, and ride the residual fitness already in the legs. ## When the date moves later: the extra weeks are not free A later date looks like a gift — more time to prepare. It is not. Inserted weeks become accidental load unless the plan actively reshapes them. Two added weeks at full build intensity violates the chronic-load ramp; two added weeks of unstructured riding decays the peak residuals. The plan has to decide what the new weeks are for. The first decision is whether the inserted weeks land before or inside what was the taper. If the original taper has not started, the cleanest move is to extend the build by one cycle and re-anchor the taper to the new date — but only at the ramp rate the rider was already on. The Coggan and Allen CTL framework [Allen et al. 2019] caps safe weekly chronic-load increases inside roughly 4-7 TSS/day per week during a build phase; adding two weeks of higher load on top of an already-peaked CTL routinely breaches that ceiling. Foster's training-monotony research [Foster 1998] is the same lesson from the variance angle — sustained high load with low daily variance is the strongest behavioral predictor of illness and non-functional overreaching in his 25-athlete cohort. The added weeks need a deload built in, not a third build block bolted on. If the original taper had already started, the added weeks are harder. A rider who has spent four days at 50% volume and now needs to delay the peak two weeks cannot just resume the build — the freshness gain has already started surfacing, and rebuilding load through it is a partial detraining-and-rebuild cycle with the wrong timing. The right structure is a short re-load block of 7-10 days at 70-80% of pre-taper volume, then a second clean taper to the new date. This is also where adaptive cycling training plans diverge from static ones most visibly: a plan that just slides the calendar lets the original taper run, then drops the athlete cold into an unstructured two weeks before the race. That is detraining by accident. ## When to downgrade priority instead of restructuring Sometimes the right answer is not to restructure the plan around the new date but to demote the event. An A race that moves inside the taper window or shifts onto a date the build cannot reach should usually become a B race. The plan keeps marching toward the original peak, the rescheduled date gets ridden on residual fitness, and the next A race holds the arc. The priority downgrade is a tool, not a failure. Friel's cyclist-coaching framework [Friel 2018] sizes the taper and the build around A-priority dates specifically because a true peak is metabolically expensive and the macro arc can only support two or three per year. A rescheduled event that no longer fits the arc does not need to consume that budget. Downgrading it to B priority means a 5-7 day mini-taper instead of a full two-week one, a participation-or-tactics goal instead of a result goal, and the build continues marching toward the next anchor date. The rider rides well, the arc stays intact, and the next A race gets the full Bosquet protocol [Bosquet et al. 2007] it was sized around. Static plans cannot do this because they have no concept of priority beyond the single date they were sold around. TrainingPeaks-style structured plans and most calendar-PDF products slide every workout when the date moves and call it a re-plan; intensity-adaptive systems like TrainerRoad's Adaptive Training adjust workout difficulty inside the new structure but still treat the new date as the single anchor. A genuinely life-adaptive plan reads the date change, the remaining weeks, the priority level, and the rider's current CTL, then proposes whether to restructure or downgrade — with the taper math and the residual-window math attached so the rider can see why. That is the operational difference between an adaptive plan and a calendar that knows how to redraw itself. ## Common questions **My A race got moved two weeks earlier. Should I cut my taper in half?** No. The taper window is fixed by physiology, not by calendar convenience. Bosquet and colleagues' meta-analysis [Bosquet et al. 2007] found the two-week, 41-60% volume reduction protocol is what produces measurable performance gains; a one-week compressed taper does not deliver the same freshness response. The right move is to end the build immediately, run the standard taper into the new date, and accept that one build cycle was lost. **My event got pushed back three weeks. Can I just keep training like nothing changed?** Not at full load. Three extra weeks at the chronic-load ramp you were on usually breaches the safe 4-7 TSS/day weekly increase the Coggan and Allen framework [Allen et al. 2019] flags as the upper bound, and Foster's monotony research [Foster 1998] shows sustained high load with low variance is a leading overtraining trigger. Build in one deload week, hold the second at maintenance, then run a clean taper into the new date. **What if the rescheduled date now overlaps with another event I had planned?** One of them has to become a B race. A true two-week taper costs aerobic base, and the macro arc can only support two or three A-priority peaks per year. The closer event almost always wins as the A race because the taper window is already partially in play; the farther one downgrades to a B-priority effort with a 5-7 day mini-taper. This is the structural decision the broader question of adaptive plans that survive real life is built to handle. **Does this also apply to gran fondos and unsanctioned events, or just structured races?** The physiology is the same. A gran fondo or a gravel event that moves three weeks for weather puts exactly the same constraint on the taper and the build that a USAC date change does. The only thing that changes is the priority calculus — recreational events are more likely to demote cleanly to B priority because the rider has more freedom to choose participation over performance. ## References 1. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. 2. **Wang et al. 2023.** [Effects of tapering on performance in endurance athletes: A systematic review and meta-analysis](https://pubmed.ncbi.nlm.nih.gov/37163550/). PLOS One. 3. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization: block periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. 4. **Mujika 2010.** [Intense training: the key to optimal performance before and during the taper](https://pubmed.ncbi.nlm.nih.gov/20840559/). Scandinavian Journal of Medicine & Science in Sports. 5. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 6. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 7. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. --- # Adding a gravel event mid-block: what an adaptive plan should change in the next 4 weeks URL: https://www.adaptcycling.com/guides/gravel-event-mid-block-adaptation Updated: 2026-05-10 Author: Jim Camut You signed up Tuesday for a gravel event 8 weeks out — BWR, SBT GRVL, Unbound, Mid South, doesn't matter which. Your plan was running a clean sweet-spot build pointed at lifting FTP. Now the target is six hours in the saddle on mixed surfaces, and the question is what the adaptive plan should actually change between now and the start line — and just as importantly, what it should leave alone. ## Why a gravel event is a durability target, not an FTP target Gravel events are durability problems first, intensity problems second. BWR runs roughly 6 hours, SBT GRVL 100 about 5 to 7, Unbound 200 around 10 to 13. The gating physiology is fat-oxidation capacity, fuel tolerance, and time-on-bike — not the 20-minute power number that drives a road build. The acute load math makes the point. A 6-hour mixed-terrain event at moderate average intensity books 250 to 400 TSS in a single ride. For a 10-hour-per-week amateur whose normal week runs 550 to 650 TSS total, one event day is 40 to 70 percent of a normal training week. The day is not the problem the build needs to solve. The problem is the rider arriving at hour four still able to produce useful power, still digesting carbohydrate, still holding position on rough surface — all of which sit downstream of the aerobic engine and the gut, not the threshold ceiling. The physiology is well-mapped. San-Millán and Brooks' metabolic-flexibility work [San Millán & Brooks 2018] documented that elite endurance athletes oxidize fat at roughly double the rate of less-fit subjects at the same relative intensity, with peak fat-oxidation rates landing around 0.6 to 0.9 grams per minute in trained cyclists. Costa and Hoffman's ultra-endurance nutrition review [Costa & Hoffman 2019] is explicit that fueling tolerance — the ability to process 60 to 110 grams of carbohydrate per hour without GI failure — is itself a trained adaptation, not a given. A rider whose normal long ride is 2 hours has never asked the gut to process 600 grams of carbohydrate in one session. The event is the first time the system has been stressed at that load. ## What changes in the next 4 weeks: long-ride absorption and fuel rehearsal The weekend long ride is the lever. In a sweet-spot build it would have been a 2-hour endurance plus tempo session; for a gravel target it absorbs 30 to 50 percent of event duration with the rider fueling at race intake rates. Weekday quality compresses to one session; the second high-intensity slot converts to mid-length endurance. The mechanism is specificity dosing. Issurin's block-periodization framework [Issurin 2010] separates accumulation work — voluminous low-intensity volume targeting mitochondrial biogenesis — from transmutation work targeting event-specific demands. A mid-block gravel target shifts the accumulation-to-transmutation balance toward accumulation, because the missing adaptation is time-on-bike capacity, not threshold power. Practically, the long ride scales to roughly 50, 60, 70, and 75 percent of event duration across the four-week run-in, with the last build long ride landing 10 to 14 days out so it remains absorbable into the taper. Two of those rides should include the surface and equipment combination the event will use — not for fitness, but to surface tire, hydration, and contact-point problems while there is still time to fix them. Fueling is the other adaptation. Jeukendrup's personalized carbohydrate-intake review [Jeukendrup 2014] establishes the dose-response: roughly 60 grams per hour with single-source glucose, climbing to 90 grams per hour with multi-transportable glucose-fructose blends as event duration crosses 2.5 hours. Costa and Hoffman [Costa & Hoffman 2019] document that intakes above 60 grams per hour reliably produce GI symptoms in athletes who have not gut-trained the rate. The four-week run-in is where that training happens. The last two long rides should run at the planned race intake rate from the first hour, not just the last. A rider who has only ever taken 40 grams per hour and shows up at the start line planning to take 90 is rehearsing a bonk or a porta-john stop, not a race. This is exactly what the broader pillar on adaptive cycling training plans means by life-adaptive. The plan absorbs the new target by reshaping the weekly shape — long ride up, second quality day down, weekly TSS within the original ramp band — instead of bolting extra rides on top of the existing schedule. A static plan handed the same event date typically does the latter and stacks the high-intensity work onto an already-loaded build, which is the textbook overtraining setup [Foster 1998]. ## What does not change: the polarized intensity distribution The weekly intensity ratio holds. Roughly 80 percent of training time at low intensity, 20 percent at high intensity, very little tempo in between — the polarized pattern Seiler formalized [Seiler 2010] is not a road-versus-gravel question. Replacing the second weekly quality session with sweet-spot or tempo to feel more 'race-like' is the most common mistake. The reasoning is direct. Seiler's polarized-training synthesis [Seiler 2010] is the most replicated finding in endurance research, and the principle is event-format-agnostic — the ratio of low to high intensity across the week is what produces the adaptation, not the specific surface. Substituting tempo for endurance to chase a 'gravel feel' produces a more monotonous week, a depressed low-intensity volume, and an aerobic engine that arrives at the event with less fat-oxidation capacity than it had before the substitution. Foster's monotony research [Foster 1998] flags exactly this drift as a leading behavioral predictor of illness and overtraining. The single quality session that stays in the week should be the one that supports the event's demand. For most gravel events that means VO2max work — 4 to 6 minute intervals at 105 to 115 percent of FTP — because the surges that decide the race tend to land in that band, not at threshold. Issurin's block-periodization framework [Issurin 2010] supports keeping one concentrated high-intensity stimulus per microcycle rather than splitting it into two diluted sessions. The discipline is to resist the urge to make every weekday ride feel like preparation; most weekday rides during the run-in should be unambiguously easy. ## The event itself is part of the plan: acute load and recovery A 250 to 500 TSS event day is well outside any prescribed week's ramp band. The plan should treat the event as a planned overload and budget recovery accordingly — typically 7 to 14 days at substantially reduced load, with no quality work until the freshness signal recovers. Stacking a normal training week on top of the event is where most amateur post-event detraining starts. Coggan and Allen's chronic-load framework [Allen et al. 2019] keeps weekly TSS climbing inside a roughly 4 to 7 TSS-per-day per-week safe ramp during build phases. A single 350 TSS event day, dropped into a week that already contained a 90-minute opener and a 60-minute easy spin, lands the week 40 to 60 percent over its acute-load target — well above the acute:chronic ratios associated with elevated illness and soft-tissue risk in the broader monitoring literature [Foster 1998]. Treating the event as 'just one ride' and resuming normal load Monday is the operational error. The week containing the event should be planned as the deload, not the week after it. Practically, the post-event seven days run easy and short — zone 2 only, 50 to 60 percent of normal weekly TSS, no intervals. Returning to load while still recovering from a 6 to 12 hour event with high carbohydrate intake, sleep disruption, and contact-point trauma is the textbook non-functional overreaching trigger — the same high-load high-monotony pattern Foster's data [Foster 1998] identified in his 25-athlete cohort. The macro arc the plan was running before the event resumes in week two post-event, anchored to the new chronic-load number, not the pre-event one. ## Common questions **How many weeks out should I tell the plan about the event?** Eight to twelve is ideal; four is the minimum useful window. Below four weeks the long-ride progression and the gut-training rehearsal cannot both fit, and one or both gets compromised. Above twelve weeks the build can usually absorb the target without any meaningful structural change to the next month — the specificity work shifts into the final block. **Should I add a second long ride during the week?** Usually no. The total weekly TSS still has to sit inside the ramp band Coggan and Allen's framework [Allen et al. 2019] established. Adding a midweek 3-hour ride while keeping the existing structure pushes the week 20 to 30 percent over its acute-load target and produces the high-load high-monotony pattern Foster's work [Foster 1998] flagged as the strongest behavioral predictor of overtraining. Length comes from the weekend, not the weekday. **Do I need to train on gravel to race gravel?** Not for fitness — the aerobic engine does not know what surface produced the watts. You do need surface time for tire choice, hydration position, drivetrain noise, and rough-surface contact-point tolerance. Two of the four long rides should match the event's surface and equipment combination. The rest can stay on whatever surface is convenient. **What if the event is my A goal and the build was pointed at a different one?** Re-anchor the macro arc. The previous A goal becomes a B priority or moves; the gravel event becomes the taper target with a standard two-week taper — roughly 40 to 60 percent volume reduction while preserving intensity. Trying to peak twice in eight weeks is the failure mode. Pick the event the calendar actually points at. ## References 1. **San Millán & Brooks 2018.** [Assessment of Metabolic Flexibility by Means of Measuring Blood Lactate Fat and Carbohydrate Oxidation Responses to Exercise in Professional Endurance Athletes and Less-Fit Individuals](https://pubmed.ncbi.nlm.nih.gov/28623613/). Sports Medicine. 2. **Jeukendrup 2014.** [A Step Towards Personalized Sports Nutrition: Carbohydrate Intake During Exercise](https://pmc.ncbi.nlm.nih.gov/articles/PMC4008807/). Sports Medicine. 3. **Costa & Hoffman 2019.** [Considerations for ultra-endurance activities: part 1 — nutrition](https://pubmed.ncbi.nlm.nih.gov/30056753/). Research in Sports Medicine. 4. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. 5. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 6. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 7. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. --- # What 'restructuring the week' actually does inside an adaptive training plan URL: https://www.adaptcycling.com/guides/what-week-restructure-actually-does Updated: 2026-05-10 Author: Jim Camut Every adaptive-plan vendor uses the word restructure. Almost none define it. Operationally, restructuring a week is not a swap and not a shift — it is a constrained solver running on the days the athlete has left, treating the week as a budget rather than a sequence. This is what the engine reads, what it decides, what invariants it honors, and the specific things no restructure can do. The mechanism under the marketing claim. ## What 'restructure' means operationally: swap, shift, restructure are three different operations A swap is one-for-one substitution: Tuesday's threshold becomes Tuesday's endurance. A shift slides days: Tuesday's threshold becomes Wednesday's threshold. A restructure rebuilds the entire remaining week against a new constraint set — load, intensity ratio, recovery placement, and goal arc all recomputed together. Most apps that claim to restructure are actually doing one of the first two. The distinction matters because the three operations solve for different things. A swap defends the day; a shift defends the calendar; only a restructure defends the week. When an athlete logs a missed VO2 on Monday with five days left and eight weeks to a goal event, swap-only logic asks what Monday should have been; shift-only logic asks where the session goes next. A restructure asks what the remaining five days should look like given the missed stimulus, current chronic load, the polarized-distribution target [Seiler 2010], and goal proximity — then rebuilds Tuesday through Sunday as a single decision. Concretely: weekly target was 600 TSS with one VO2 session, one threshold session, and a long endurance ride. The athlete misses the VO2. A restructure skips it rather than rescheduling, demotes Tuesday's threshold to tempo to keep the high-band budget defensible, holds Friday's recovery, and defends Saturday's long ride because endurance volume protects the chronic-load ramp [Allen et al. 2019]. New weekly target is 470 TSS. IF lands inside the polarized band. No compression. The missed VO2 is gone — most lost stimuli should be skipped, not chased. TrainerRoad's Adaptive Training swaps well within an indoor block. JOIN handles partial shifts cleanly. Xert updates load continuously but defers the week-shape decision to the athlete. Wahoo SYSTM and Garmin DSW are template-rotation systems with intensity-tier swaps bolted on. None of those products are bad at what they do — they are doing the swap or the shift, not the restructure, and the marketing word does not distinguish. ## The four constraints the solver must honor Any honest restructure has to satisfy four periodization invariants simultaneously: the chronic-load ramp from the Coggan and Allen framework, the weekly intensity distribution from Seiler's polarized-pyramidal evidence, the daily-load monotony bound from Foster's overtraining work, and the macro-arc anchor from Friel's goal-event periodization. Violate any one and the restructure has solved the wrong problem. Constraint one is the chronic-load ramp. The standard guidance from Allen and Coggan [Allen et al. 2019] keeps weekly TSS increases inside roughly 4 to 7 TSS-per-day per week during a build phase. A restructure that proposes a 600-to-720 TSS jump to make up for a missed week violates this directly; a restructure that proposes 470 TSS after a partial-week disruption sits comfortably inside the band. Constraint two is the intensity distribution. Seiler's polarized-training synthesis [Seiler 2010] is the most replicated finding in endurance research: the ratio of low to high intensity across the week is what produces the adaptation, not the gross load. The restructured week has to keep roughly 75 to 85 percent of the time below the first lactate threshold. Constraint three is daily-load monotony. Foster's 25-athlete cohort study [Foster 1998] showed illness clustered in weeks where load was high and daily-load variance was low — when every day looked similar in stress. A restructure that fills the gap from a missed key session with three medium days hits the monotony trap directly. The solver has to preserve at least one genuine recovery day and one clear high-intensity day inside the rebuilt week. Constraint four is the macro arc. Friel's Base-Build-Peak-Race-Transition framework [Friel 2018] and Bosquet et al.'s 27-study taper meta-analysis [Bosquet et al. 2007] anchor the restructure to the goal date: the rebuilt week has to leave the macro phase intact, or the engine has to explicitly move the goal. Issurin's block-periodization work [Issurin 2010] sits across all four. A restructure inside a threshold block has to keep the block's concentrated stimulus coherent — skipping the week's only threshold session because Tuesday got blown by life turns the block into a maintenance week with a label. When no remaining-week shape satisfies the four invariants and delivers the block's intended adaptation, the right answer is to extend the block by a week and re-anchor the macro arc, not to compress missed work into the days that are left. ## What inputs the engine reads A restructure that gets the math right reads five inputs every time: the Strava-derived ride data for the week so far, the completion state of remaining sessions, the athlete's profile and weekly-hours budget, the calendar window with constraint days marked, and the goal-event distance. Plans that read fewer inputs end up doing swaps and shifts dressed as restructures. Strava ride data is the first input and the one most apps use well: normalized power, intensity factor, time-in-zone, and TSS from the Allen and Coggan framework [Allen et al. 2019]. A 90-minute unscheduled ride at NP 230 on a 250-watt FTP returns IF 0.92 and roughly 127 TSS — enough to know the week's high-band budget is partially spent before the solver looks at tomorrow. Completion state is the second input. A restructure on Tuesday with three sessions still on the calendar has more degrees of freedom than one on Friday with one session left. Most static plans do not read this and propose the same compression regardless of when the disruption occurred. The athlete profile is the third input — weekly hours available, FTP, experience, recovery capacity. A 6-hour-per-week rider's restructure cannot copy a 12-hour-per-week rider's even when the disruption is identical, because the constraint surface is different. The calendar window with constraint days marked — soccer tournament, work travel, wedding — is the fourth input. The fifth is goal-event distance. Eight weeks out from an A-priority event, the restructure can absorb a missed VO2 without moving the goal. Three weeks out, the same disruption forces a goal-priority decision because the taper window [Bosquet et al. 2007] cannot share days with a rebuild. This is the architecture the broader question — what adaptive cycling training plans should do when life gets in the way — depends on. The parent guide on adaptive cycling training plans makes the case that reading these five inputs and reshaping the week against the four constraints is what separates plans that survive real life from plans that just generate. A restructure that does not read what actually happened is running the loop open. ## What restructure cannot do A restructure is a solver over the days the athlete has left. It cannot violate the four periodization invariants, cannot manufacture lost adaptation, and cannot rebuild fitness faster than the underlying physiology curves allow. Athletes who expect a restructure to recover a missed peak block in two weeks are asking for an operation the engine has no legal move for. The hard physiology constraint is the detraining curve. Mujika and Padilla's short-term detraining review [Mujika & Padilla 2000] documents that VO2max decline inside the first month off is driven largely by reduced blood volume and stroke volume, with rebuild times in the 2 to 3 week range under structured training. A restructure proposing full pre-break wattages in week one of a post-illness rebuild is asking the cardiovascular system to be where it physiologically is not. The correct move is to cap intensity time at 60 to 70 percent of pre-break load and let the chronic-load ramp [Allen et al. 2019] do the rebuild work over two to three weeks, not one. The hard periodization constraint is the macro arc. A restructure cannot decide on its own that the goal event moves — that is an athlete decision the engine surfaces. What it can do is compute the math: a missed week six-plus weeks from the event usually absorbs cleanly; a missed week inside three weeks collapses the taper window and forces the goal to move, downgrade in priority, or shift to participation. The restructure surfaces the choice with numbers attached. It does not make it. The hard accounting constraint is that lost stimulus is lost. The right move when a key session is missed is almost always to skip it, accept the block ran one stimulus light, and let the next block proceed on schedule [Issurin 2010]. The wrong move — and the one static plans default to — is to compress missed work into the following weeks, which violates the chronic-load ramp and the monotony bound simultaneously and produces the overtraining signature Foster's work [Foster 1998] is built on. A restructure's most important property is that it knows when to skip. ## Common questions **Is a swap the same thing as a restructure if my app says it 'rebuilt the week'?** Usually not. The diagnostic is whether the weekly TSS target, intensity ratio, and recovery placement all changed together, or whether only the missing day changed. If Tuesday's threshold became Tuesday's endurance but every other day is identical, that is a swap. A real restructure recomputes the remaining days as one decision and the weekly target typically shifts 15 to 25 percent. **Can a restructure ever make the week harder than originally planned?** Rarely, and only when the athlete arrives at the week substantially fresher than the plan assumed — for example, when a scheduled hard ride got rained out and replaced by genuine recovery. Even then the engine should bias toward holding the planned load rather than escalating, because the chronic-load ramp [Allen et al. 2019] assumes the athlete is not starting the week over-recovered. **How often should an adaptive plan need to restructure?** For a self-coached rider with normal life variance, two to four times per twelve-week block is typical — most weeks the plan runs as written. Weekly restructures usually mean the underlying plan's volume or intensity assumptions are wrong, not that the engine is working hard. Persistent restructure pressure is a signal to revisit profile inputs, not a feature. **What happens if I keep ignoring the restructure and doing the original plan?** Two things, predictably. The chronic-load ramp drifts above the 4 to 7 TSS-per-day-per-week band [Allen et al. 2019], and the weekly intensity distribution collapses out of the polarized range [Seiler 2010]. Both load Foster's monotony predictor [Foster 1998]. Fitness gains usually still come for a few weeks; the illness or non-functional overreaching usually arrives by week four to six. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 3. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 4. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://www.velopress.com/books/the-cyclists-training-bible-5th-ed/). VeloPress. 5. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. 6. **Mujika & Padilla 2000.** [Detraining: loss of training-induced physiological and performance adaptations. Part I: short term insufficient training stimulus](https://pubmed.ncbi.nlm.nih.gov/10966148/). Sports Medicine. 7. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. --- # AI cycling coach vs human coach in 2026: who actually adapts when life breaks the plan URL: https://www.adaptcycling.com/guides/ai-cycling-coach-vs-human-coach Updated: 2026-09-03 Author: Jim Camut I raced as a pro through Bruyneel's academy from 2004 to 2009. Every coach I worked with then would have laughed at the idea that software could replace them. They would have been right — in 2009. In 2026 the question is different, and narrower than most coaches admit and wider than most AI startups admit. The honest cut is not 'AI vs human.' It is which model actually restructures your plan the Wednesday you wake up sick, and what you pay for that restructure. ## Where the line actually sits in 2026 A great human coach still wins on judgment, outside perspective, and the relational layer that drives long-term adherence. A good AI coach wins on adaptation latency, periodization invariants the model cannot violate, and structural cost. For most self-coached amateurs with a real goal and a chaotic schedule, the AI side of that line is now closer than the coaching industry markets. The parent pillar on adaptive cycling training plans argues that 'adaptive' has split into three operational tiers: intensity-adaptive (workouts get harder or easier), load-adaptive (the load target shifts with fitness), and life-adaptive (the plan itself restructures when life intrudes). The sibling spoke on adaptive versus rebranded static plans grades the major apps against those tiers; we do not repeat that comparison here. We hold the question still: at each tier, what does a human coach add that software cannot, and what does software now add that the average human coach cannot? The human coach's structural advantage is judgment under ambiguity. Research on high-level coach cognition shows that experienced endurance coaches make most in-season calls through naturalistic decision-making — fast pattern matching grounded in years of seeing athletes break and rebuild [Collins et al. 2016]. That intuition shows up exactly where data is thin: an athlete returning from a hairline fracture, a masters racer whose HRV is fine but who 'feels off', a junior whose biggest constraint is psychological. No AI in 2026 reads body language across a coffee. The AI coach's structural advantage is latency and consistency. A human coach who reviews your week every Sunday cannot restructure Wednesday at 6am when you wake up with a sinus infection. A solver-backed AI plan can — and the restructure has to respect the same periodization rules the LLM cannot wave away. That is the real comparison: a great coach reviewing weekly versus reasonable software reviewing every ride. ## What a human coach delivers that AI cannot yet Three things, in order of how much they actually move outcomes: outside-perspective accountability, judgment on intangibles, and the relational scaffold that keeps a stressed amateur showing up in week 14. The first two are real. The third is the one AI gets closest to faking and still loses. Outside-perspective accountability is the biggest. Self-coached athletes systematically over-train when fit and under-train when stressed — the dataset is your own ego. A coach who has watched fifty athletes through the same arc will tell you to back off when you are convinced you should add a fourth interval. The expert-intuition literature treats this as a domain-specific skill that compounds with rep count, not something extractable from a textbook [Collins et al. 2016]. The relational layer is the other big one. The Teixeira self-determination-theory review of 66 studies found that long-term exercise adherence is driven by autonomy, competence, and relatedness — and relatedness is the leg software struggles with [Teixeira et al. 2012]. A coach who knows your kid is sick, who texts back at 9pm, who notices when your tone shifts — that is doing relational work an LLM can simulate but not durably anchor. For an athlete whose primary block to consistency is psychological, this matters more than any plan structure. Where AdaptCycling specifically loses to a great human coach: top-5% amateurs chasing podiums where one bad call costs the season, athletes returning from major injury where conservative biomechanical judgment dominates, and athletes whose primary training problem is psychological rather than structural. We are honest about this. If you are one of those three, the math for a human coach gets clear fast. ## What an AI coach delivers that most human coaches do not Sub-day adaptation latency, periodization invariants that cannot be violated, and 24/7 availability at one-tenth the cost. The catch: only a small slice of cycling apps actually deliver tier-three life-adaptive behavior; the rest are static plans with a dynamic UI. Adaptation latency is the headline. The typical 1-on-1 cycling-coach contract reviews the calendar weekly — often Sunday night. That is a 6-day average lag between life intruding and the plan responding. A solver-backed AI coach restructures within minutes of the missed ride hitting Strava. For an amateur whose schedule breaks two or three times a month, this is not a small difference; it is the difference between a plan that is 80% executed and one that is 50% executed-plus-guilt. Invariant enforcement is the under-marketed one. The periodization rules that prevent overtraining — Foster's training-monotony bound on day-to-day load variance [Foster 1998], Mujika's taper structure of maintained intensity with cut volume [Mujika 2010], Seiler's 80/20 polarized intensity distribution [Seiler 2010] — can be encoded as solver constraints the LLM is not allowed to violate. A human coach can violate them; in fact, a tired human coach with twenty athletes routinely does, usually by stuffing a missed week into the next one. A well-designed solver cannot. This is a real, replicable edge. The cost math sharpens the picture. Established 1-on-1 online coaching from outfits like CTS runs from $207 per month at the entry tier to over $1,200 per month at elite-coach tiers [CTS 2026]. App-based AI training — TrainerRoad, JOIN, Xert, AdaptCycling — sits at $15-30. For a self-coached amateur with a $30 budget and a chaotic week, the choice is not 'AI versus human coach.' It is 'AI versus nothing,' and nothing is what most amateurs are running on. ## The hybrid most amateurs ignore Software for the daily restructure, occasional consult-only sessions with a human for the big calls. A consult-only block costs roughly $150-300 per session, runs two or three times a season, and answers the questions software is genuinely worse at: am I targeting the right race, is this training history flagging burnout, should I move up a category. Most coach-versus-app comparisons assume you pick one. You do not have to. The fact that life-adaptive software can hold the weekly schedule frees a paid coaching hour to do what humans are actually better at — outside-perspective judgment on the strategic questions, not the tactical ones. A pre-season planning call, a mid-block check-in, a race-week debrief: maybe six hours of human input across a year, used where it actually moves the needle. This is the structural answer to 'do I need a coach.' For most self-coached amateurs the answer is: not as a several-hundred-dollar monthly retainer, yes as a few consultative hours bolted onto adaptive software. The hybrid costs a fraction of a 1-on-1 contract and captures most of the human upside. Coaches who price for it (consult-only or hourly) are easy to find; most will quote $100-200 per hour. ## Common questions **Is AI cycling coaching as good as a human coach in 2026?** For most self-coached amateurs with a real goal and a chaotic schedule — yes, close enough that the cost gap is decisive. For the top 5% of amateurs chasing podiums, athletes coming back from major injury, or athletes whose primary block is psychological, a great human coach still wins on judgment and the relational layer that drives long-term adherence. The honest split favors AI being good enough for most amateurs, not a 50/50 toss-up. **What does an AI coach actually do that a human coach does not?** Three things. It restructures the plan within minutes of a missed or modified ride rather than waiting for a weekly review. It enforces periodization invariants — progression rate, intensity distribution, taper structure — as hard constraints the model cannot violate even when the athlete pushes for more. And it costs $20-30 per month instead of several hundred. The first one is the biggest in practice for amateurs whose week breaks twice a month. **What does a human coach do that no AI can do yet?** Read intangibles a Strava feed cannot see — body language, life stress, the gap between what an athlete says and what they mean. Provide outside-perspective accountability when self-coached ego pushes toward over- or under-training. Build the relational scaffold that drives long-term exercise adherence [Teixeira et al. 2012]. And exercise expert intuition on the high-stakes calls where pattern recognition from fifty prior athletes outperforms any model [Collins et al. 2016]. **Is the hybrid model (AI plan plus occasional human consults) actually viable?** Yes, and it is underused. Software handles the daily and weekly restructuring at $20-30 per month; you spend $150-300 per session on a coach two or three times a year for strategic questions — race targeting, season review, category-upgrade decisions. Total annual cost lands well below full 1-on-1 retainer pricing, and you keep the parts of human coaching that actually move outcomes for an amateur. ## References 1. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 2. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 3. **Mujika 2010.** [Intense training: the key to optimal performance before and during the taper](https://pubmed.ncbi.nlm.nih.gov/20840559/). Scandinavian Journal of Medicine & Science in Sports. 4. **Teixeira et al. 2012.** [Exercise physical activity and self-determination theory: a systematic review](https://pmc.ncbi.nlm.nih.gov/articles/PMC3441783/). International Journal of Behavioral Nutrition and Physical Activity. 5. **Collins et al. 2016.** [If It Feels Right Do It: Intuitive Decision Making in a Sample of High-Level Sport Coaches](https://pmc.ncbi.nlm.nih.gov/articles/PMC4830814/). Frontiers in Psychology. 6. **CTS 2026.** [CTS Coaching Pricing](https://trainright.com/coaching/pricing/). Carmichael Training Systems. --- # TrainerRoad vs JOIN vs AdaptCycling: which one actually adapts when life disrupts the week URL: https://www.adaptcycling.com/guides/trainerroad-vs-join-vs-adaptcycling-adaptation Updated: 2026-05-10 Author: Jim Camut I trained off TrainerRoad before I built AdaptCycling. It was the best indoor-power app I could find and Adaptive Training is real software — but it was not the thing that broke when I became a dad. The thing that broke was the week itself. Three apps now claim to adapt: TrainerRoad, JOIN, and AdaptCycling. They adapt different things at different scales. This is the head-to-head on the only dimension that decides which one survives contact with a real schedule. ## What each app actually adapts (and what it does not) TrainerRoad adapts workout difficulty inside a fixed week. JOIN restructures the week when prompted. AdaptCycling restructures continuously on observed signal. The same word — 'adaptive' — covers three different operations, and the deeper question of what happens to the macrocycle when four sessions stack up missed separates them more cleanly than any feature comparison. TrainerRoad at $21.99 per month runs an ML system trained on a large corpus of completed workouts and a Progression Level score from 1.0 to 10.0 per power zone [TrainerRoad 2026]. Plan Builder lays out the weekly skeleton; Adaptive Training swaps the workout inside that skeleton based on how the last one went. Miss a Wednesday VO2 session, the model drops your VO2 Progression Level, and Friday's workout gets softened to match. What it does not do is reshape the week — there is no Wednesday-becomes-Thursday move, and no recomputation of the block when four sessions stack up missed. That is intensity-adaptation done at high resolution, not life-adaptation. JOIN out of the Netherlands at €16.99 per month is built outdoor-first and Strava-native, with two ex-WorldTour coaches behind the methodology. The live plan rebuilds when you mark a day unavailable, and intensity adjusts based on completed work. The constraint in the field is variance: when the schedule shifts substantially three weeks running — kids sick, work trip, group ride on a recovery day — the restructure logic gets stress-tested in a way the marketing copy does not advertise. The sibling piece on adaptive versus rebranded static plans places JOIN between tiers two and three for this reason. For moderate disruption it works; for chaotic schedules it drifts. AdaptCycling sits in tier three. Where TrainerRoad and JOIN both treat the week as a unit you adjust to, AdaptCycling treats the plan as a forward-looking object recomputed against four constraints: hours available, days available, race date, current fitness. Every ride upload — completed ride or unplanned group spin — reshapes the remaining plan automatically, with block periodization invariants preserved and taper math intact [Issurin 2010, Bosquet et al. 2007]. The sibling spoke on what week-restructure actually does walks through the four-constraint solver in detail; the point here is that the unit being adapted is the plan, not the workout. ## The schedule-variance test: which app for which kind of week Use week-over-week schedule variance as the deciding question. Under 15% variance: TrainerRoad. 15-30%: JOIN. Above 30% — parents of young kids, shift work, frequent travel, unpredictable on-call: AdaptCycling. This single test predicts where each app stops working better than any feature matrix. The mechanism is straightforward and grounded in load research. Intensity-only adaptation works when the week itself is stable, because the model only has to solve one variable: how hard. Each missed session in TrainerRoad drops a Progression Level, the next workout gets softer to compensate, and inside four weeks the prescribed load drifts well below what the athlete can actually absorb. Foster's training-monotony work shows the cost: low day-to-day load variance combined with accumulated under-stimulus is its own risk profile [Foster 1998]. The sibling diagnostic on signs your training plan is not adapting names the symptom — your TSS goal slides downward every Sunday and you cannot tell why. That is the signature of a plan compensating for compliance rather than restructuring around it. Above roughly 15-20% week-over-week variance, intensity-only adaptation stops being enough. JOIN handles the middle band well. The mental model — you tell it what is happening this week, it redraws the week — fits riders with predictable disruption: a known travel cadence, a fixed weekly long-ride window. Above 30% variance, where you find out Tuesday that Wednesday is gone, you want the plan recomputing without a UI prompt. That is the case AdaptCycling is built for, and where the wider framework on adaptive cycling training plans concentrates its tier-three definition. ## Where TrainerRoad still wins, where JOIN still wins, where we win TrainerRoad has the deepest workout library, the most polished ERG-mode experience, and a large ML training corpus. JOIN has the cleanest outdoor-first UX and credible coach pedigree. AdaptCycling has continuous ride-signal restructure, no compliance scoring, and an AI chat that knows the athlete's full history. None of this is a hatchet job. TrainerRoad's workout catalog and ERG-mode polish are genuinely class-leading. If you live on a smart trainer, race short-format indoor events, or want Progression Levels per power zone as your primary feedback loop, TrainerRoad is the answer. The ML model is trained on a corpus of completed cycling workouts no competitor will match for years — that data scale is a real moat for the problem TrainerRoad chose to solve [TrainerRoad 2026]. JOIN's outdoor-first design and ex-WorldTour coach credibility matter more than the feature list suggests. The mental model is simpler than ours — fewer knobs, faster onboarding — and the Strava integration is mature. For European riders on a moderately predictable schedule who want to ride outside, JOIN is a defensible choice at €16.99 per month. AdaptCycling's wedge is narrower and sharper: life-adaptive restructure on observed ride signal, with no compliance scoring and no Wednesday-was-supposed-to-be-VO2 finger-wag. We do not have TrainerRoad's catalog depth or JOIN's UX maturity. What we have is a plan that recomputes against the four constraints — hours, days, race date, current fitness — every time the data changes, with the polarized 80/20 intensity distribution and taper-volume cuts of 41-60% over two weeks preserved across the restructure [Seiler 2010, Bosquet et al. 2007]. ## The single test that decides it Log a fake unplanned 90-minute hard outdoor ride on a Tuesday recovery day in each app. Watch what happens to Wednesday, to Saturday's long ride, and to the next two weeks. The differences are diagnostic and take 15 minutes to observe. In TrainerRoad, the extra ride lands as a manual entry. Adaptive Training will read it, possibly nudge a Progression Level, and leave the rest of the week as scheduled. The plan shape does not move — that is by design, not a bug, and it is the right design for a stable schedule. In JOIN, the ride lands in Strava, the app recognizes the load, and depending on how the day was flagged, it may reshape Wednesday and Saturday. The week-level redraw is what JOIN is built to do, and on this test it is observably different from TrainerRoad. In AdaptCycling, the same ride triggers a recomputation of the remaining plan: Wednesday becomes recovery, Saturday's long ride gets repositioned in the block, and the macrocycle invariants stay intact. If the rider does this three weeks in a row, the next block recomputes too. That is the operational definition of life-adaptive, and the easiest way to feel the difference is to do the test. ## Common questions **Is TrainerRoad's Adaptive Training the same thing as an adaptive plan?** No. Adaptive Training is intensity-adaptive — it adjusts workout difficulty inside a fixed weekly structure using Progression Levels from 1.0 to 10.0 per power zone. The plan shape itself stays put. That is the right tool for riders on stable schedules; it is not the same category as plan-restructuring. **Why is JOIN sometimes called adaptive and sometimes not?** JOIN restructures the week when you flag a change, which is more than TrainerRoad does. It sits between intensity-adaptive and life-adaptive. Where it tends to drift is high week-over-week variance — when the schedule shifts substantially three weeks running. For moderate disruption at €16.99 per month it is a credible option. **Does AdaptCycling do anything TrainerRoad does not?** Plan-level restructure on continuous ride signal, with no compliance scoring and an AI coach that retains the athlete's full history across the season. TrainerRoad's catalog depth, ERG polish, and ML training scale are areas where TrainerRoad is ahead and likely to stay ahead — different problem, different design. **When is the answer none of the three?** If you need in-person physiological testing, you are returning from a significant injury, or you are chasing top-1% podium specificity, an experienced coach with a TrainingPeaks setup is still the right call. The apps in this comparison are for self-coached riders who want structure that flexes. **Can I run TrainerRoad and AdaptCycling at the same time?** Some riders use TrainerRoad for the indoor workout library and AdaptCycling for the macrocycle. It works if you treat AdaptCycling as the planner and TrainerRoad as the workout source, and let the head-unit or intervals.icu upload close the loop. It is more setup than running one tool, but the combination is coherent. ## References 1. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 2. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. 3. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. 4. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 5. **TrainerRoad 2026.** [Adaptive Training Overview](https://support.trainerroad.com/hc/en-us/articles/4404060687387-Adaptive-Training-Overview). TrainerRoad Support. --- # Polarized training for amateurs: how to make 80/20 actually work on limited hours URL: https://www.adaptcycling.com/guides/polarized-training-for-amateurs Updated: 2026-06-01 Author: Jim Camut Polarized training is an intensity distribution where roughly 80% of your riding is genuinely easy and the remaining 20% is genuinely hard, with almost nothing in the moderate middle. The model is well-supported in trained athletes, but most amateurs who say they ride polarized are not — they have one hard day and call the rest easy while it quietly drifts into tempo. This is what the 80/20 split actually measures, why it is hard to execute on 5–7 hours a week, where the pyramidal counter-evidence pushes back, and how to run it without a lab. ## What 80/20 actually counts — and why it is not your power zones The 80/20 in polarized training refers to a three-zone physiological model anchored to your two lactate or ventilatory thresholds (VT1 and VT2), not the five-to-seven power zones on your head unit. Zone 1 is everything below VT1, Zone 3 is everything above VT2, and the moderate middle between them is the part the model deliberately keeps near-empty. The distinction is the single most common point of confusion. When Seiler and Kjerland quantified the training of elite cross-country skiers, they delineated three zones using the first and second ventilatory thresholds and found the athletes spent the large majority of sessions below VT1 and a meaningful minority above VT2, with strikingly little at the lactate-threshold intensity in between [Seiler & Kjerland 2006]. That three-zone frame is the one polarized training inherited — and it does not map cleanly onto Coggan power zones. If you want the watt targets for those Coggan bands, a power zones calculator lays out all seven at your FTP. Here is where amateurs go wrong. An effort at 88–94% of FTP is sweet spot in the power-zone world, and it feels like a hard, productive ride. But in the three-zone polarized model it is moderate Zone 2 — below VT2, above VT1, squarely in the middle the model wants empty. The deeper mechanics of why that band underperforms are covered in our glossary entries on sweet spot and polarized training; the short version is that counting a sweet-spot session as part of your hard 20% is a category error that de-polarizes the week. VT1 sits near 75% of FTP for most trained cyclists and VT2 near 95–100%. So the easy 80% has to live below roughly 75% of FTP — conversational, nose-breathing pace — and the hard 20% has to live at or above threshold, which in practice means VO2max intervals more than tempo. Anything parked between those two markers is the band that produces fatigue without the adaptation either tail delivers. ## Why polarized beats the threshold-heavy default most amateurs actually run The evidence that polarized outperforms a threshold-centered week is consistent across trained populations. The mechanism is that easy volume builds the aerobic base cheaply while a small dose of genuinely hard work supplies the high-end stimulus — and the middle buys neither efficiently. Most self-coached riders default to the threshold model without choosing it. The cleanest controlled comparison is Stöggl and Sperlich's nine-week trial across 41 well-trained endurance athletes. The polarized group trained a 68% low / 6% threshold / 26% high distribution and improved VO2peak by 11.7%, alongside gains in time-to-exhaustion and peak power. The threshold group, training 46% low / 54% threshold / 0% high — close to where many amateurs actually spend their week — showed no significant improvement in the same variables [Stöggl & Sperlich 2014]. Cyclists specifically show the same pattern. Neal and colleagues ran a six-week crossover in twelve trained cyclists comparing a polarized model (80% low / 0% moderate / 20% high) against a threshold model; peak power output rose about 8% on polarized versus 3% on threshold, with larger improvements in high-intensity exercise capacity [Neal et al. 2013]. Seiler's broader review frames why: at higher fitness, where the dose lands matters more than how big it is, and the threshold-heavy week lands almost all of it in the least productive zone [Seiler 2010]. This is the same year-two intensity-drift trap our pillar on the self-coached cyclist keeps returning to. A rider who has a passable base/build structure can still stall because every individual ride inside that structure averages out to threshold. The structure was fine; the intensity inside it never polarized. Properly run, the hard 20% is where VO2max work lives — short, hard intervals near maximum sustainable aerobic output, the kind detailed in our glossary entry on VO2max. ## The pyramidal counter-evidence: when strict 80/20 is the wrong target Polarized is not the only distribution that beats threshold, and on low volume strict 80/20 can backfire. Pyramidal training — most volume easy, a moderate threshold block, a small hard slice — is what many elite athletes actually log in base periods, and it performs well in recreational populations. The honest reading is that phase and volume matter more than dogma. The recreational evidence is more mixed than the polarized headline suggests. Filipas and colleagues held middle-aged recreational male cyclists to a roughly 60% Zone 1 / 30% Zone 2 / 10% Zone 3 pyramidal distribution for 16 weeks and produced a meaningful gain in power at the 4 mmol/L lactate marker, with Zone 2 time correlating most strongly with the improvement [Filipas 2024]. The bottom of the distribution did the heavy lifting — but the distribution that delivered it was pyramidal, not strictly polarized. Muñoz and colleagues found a similar near-tie in recreational runners: over ten weeks, a polarized group improved 10K time by 5.0% versus 3.6% for a between-thresholds group, a difference that only reached significance in the subset who actually executed the prescribed distributions most faithfully [Muñoz et al. 2014]. The signal is that adherence to whichever model you pick matters as much as the model — and on low volume, adherence to genuine easy riding is the hard part. Phase changes the right answer too. In a high-volume base period, a pyramidal lean with controlled sweet spot builds durability efficiently; as the season moves toward a goal event, training tends to polarize, concentrating intensity into VO2max-specific work while easy volume protects recovery. On 5–7 hours a week, going so polarized that there is not enough total aerobic stimulus to drive base fitness is its own failure mode — the hard sessions cannot carry a week that has too little volume underneath them. ## How to actually run polarized on 5–7 hours a week without a lab Strict 80/20 is hard to execute on limited hours because the easy portion feels almost too easy and most riders unconsciously push it into the grey zone. The fix is structural, not willpower: cap easy rides at VT1, make two or three sessions a week genuinely hard, and count by time-in-zone, not by session. Set a power or heart-rate gate on every easy ride. VT1 sits near 75% of FTP and roughly 75–78% of max heart rate for most riders; cap the easy rides there and treat going over as the actual error. A 250-watt-FTP rider sets an upper alert near 188 watts and stays within a few watts of it. Without the gate, easy rides drift up predictably — the talk test is the no-lab backstop: if you cannot hold full sentences for the whole ride, it is not Zone 1. Make the hard fraction count and keep it scarce. On 6 hours a week, 20% is roughly 70 minutes of genuine high-intensity work — that is two, at most three, sessions, and they should be true threshold or VO2max intervals, not a group ride that happens to land at tempo. Anchor the intensity to a prescribed session (4×8 at threshold, or VO2max repeats at 105–120% of FTP) and complete it fresh, with at least 48 hours of easy riding around each hard day. Count by time, not by session. The most common misapplication is calling a sweet-spot-heavy week polarized because it has one hard day, or counting a lopsided week by number of rides rather than minutes. Apps make this auditable: pull four weeks of time-in-zone from Strava, Intervals.icu, or your training log. A polarized target is roughly 80% in the easy zone and under 10% in the moderate middle; most amateurs land near 50/35/15, far too much in the band that costs recovery and returns the least. AdaptCycling's phase logic maps to these distributions rather than chasing one fixed ratio, and its post-ride analysis flags weeks whose easy rides have crept into the grey zone — the failure mode that quietly de-polarizes a self-coached rider's training. ## Common questions **Is sweet spot training part of the hard 20% in polarized training?** No. Sweet spot at 88–94% of FTP sits below your second threshold (VT2), which makes it moderate Zone 2 in the three-zone polarized model, not part of the hard 20%. Counting a sweet-spot session as a hard day is the most common way amateurs accidentally de-polarize their week. The hard 20% should be at or above threshold — true threshold or VO2max intervals. **Should a time-crunched amateur ride polarized or pyramidal?** Both beat the threshold-heavy default, and on limited hours the difference between them is smaller than the difference between either and getting it wrong. A pyramidal lean with controlled sweet spot suits high-volume base periods and recreational riders [Filipas 2024]; polarizing harder suits the run-in to a goal event. On 5–7 hours a week, the bigger risk is too little total volume to support the hard sessions, so do not over-polarize a small week. **How do I audit whether my training is actually polarized?** Pull the last four weeks of time-in-zone from Strava, Intervals.icu, or your training app and check the percentage of time below 75% of FTP versus the moderate middle. A polarized target is roughly 80% easy and under 10% in the middle band. Most amateurs land near 50% easy, 35% middle, 15% hard — too much in the tempo zone that produces fatigue without proportional adaptation. **Why does my easy riding keep drifting too hard?** Because genuine Zone 1 feels almost unproductively easy, and most riders push it up without noticing. The fix is a power or heart-rate cap at VT1 — around 75% of FTP — set as an upper-bound alert on the head unit, which converts a feel decision into a beep. The talk test is the backstop: if you cannot hold a full conversation for the entire ride, you are above Zone 1. ## References 1. **Seiler & Kjerland 2006.** [Quantifying training intensity distribution in elite endurance athletes: is there evidence for an optimal distribution?](https://pubmed.ncbi.nlm.nih.gov/16430681/). Scandinavian Journal of Medicine & Science in Sports. 2. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 3. **Stöggl & Sperlich 2014.** [Polarized training has greater impact on key endurance variables than threshold, high intensity, or high volume training](https://pmc.ncbi.nlm.nih.gov/articles/PMC3912323/). Frontiers in Physiology. 4. **Neal et al. 2013.** [Six weeks of a polarized training-intensity distribution leads to greater physiological and performance adaptations than a threshold model in trained cyclists](https://pubmed.ncbi.nlm.nih.gov/23264537/). Journal of Applied Physiology. 5. **Filipas 2024.** [Effects of a 16-Week Training Program with a Pyramidal Intensity Distribution on Recreational Male Cyclists](https://pubmed.ncbi.nlm.nih.gov/38251291/). Sports (MDPI). 6. **Muñoz et al. 2014.** [Does polarized training improve performance in recreational runners?](https://pubmed.ncbi.nlm.nih.gov/23752040/). International Journal of Sports Physiology and Performance. --- # Is TrainingPeaks worth it if you don't have a coach? URL: https://www.adaptcycling.com/guides/is-trainingpeaks-worth-it-self-coached Updated: 2026-09-01 Author: Jim Camut TrainingPeaks is the best analytics platform in endurance sport and the standard pipe a coach uses to deliver a plan. For a self-coached rider with no coach, that second job disappears and you are left paying $19.95 a month for a rear-view mirror. It will not write your plan and it will not adapt your week. The honest verdict: worth it if you work with a coach or you specifically want the deepest fitness modeling. Weaker value if what you actually need is something to build and rebuild the plan for you. ## What TrainingPeaks is actually the best at TrainingPeaks Premium ($19.95/mo or $134.99/yr) is the gold standard for two things: storing structured training and modeling fitness through the Performance Management Chart [TrainingPeaks 2026]. The PMC — CTL, ATL, and the fitness-fatigue balance it tracks — is the cleanest consumer implementation of decades of training-science modeling. As an analytics surface, nothing in the category does it better. The Performance Management Chart is not marketing. It descends directly from the Banister fitness-fatigue model, which represents performance as a positive fitness term minus a negative fatigue term, each decaying on its own time constant after a training impulse [Banister 1990]. TrainingPeaks operationalizes that model with Coggan and Allen's Training Stress Score: duration times intensity-squared, scaled to your FTP, fed into the chronic-load (CTL) and acute-load (ATL) curves that the chart plots [Allen et al. 2019]. When a coach says you are 'building fitness while shedding fatigue into race week,' they are reading this chart. It is genuinely good. TrainingPeaks is also the best place to receive a structured plan. The drag-and-drop workout builder, the structured-workout files that push targets to your head unit, the calendar a coach writes into — these are the rails the entire paid-coaching industry runs on. If a coach is delivering your training, they are almost certainly delivering it here, and the platform earns its price as the shared workspace between the two of you. For analysis depth, the separate WKO5 desktop application (a one-time ~$169 purchase, not included in Premium) adds individualized power-duration modeling that goes well beyond the web PMC. None of this is in dispute. ## Why it's a rear-view mirror for the self-coached rider Every strength above is descriptive. The PMC tells you what your training did to you; it does not decide what to do next [Mujika 2017]. TrainingPeaks does not write a periodized plan and it does not restructure your week when life breaks it. For a rider with no coach, the two jobs the platform exists to support — analyze, and deliver someone else's plan — leave a hole exactly where self-coaching is hardest. The distinction is foundational, not a feature gap. Mujika's review of training-load quantification is explicit that external-load metrics like TSS are built to estimate the biological stress an athlete already experienced, not to prescribe the stress they should impose next [Mujika 2017]. The PMC is a measurement instrument pointed backward. Reading that your CTL ramp is too steep is useful; the chart will not redesign next week to fix it, and it does not know your goal event, your base phase, or your taper. Those are prescription decisions, and TrainingPeaks deliberately leaves them to a human coach. There is a deeper caution against treating the model as an oracle. Hellard's statistical analysis of the Banister model in elite swimmers found the parameter estimates were ill-conditioned and unstable — performances correlated with modeled load, but the fitted constants carried enough uncertainty to limit how precisely the model could guide an individual's training [Hellard 2006]. The PMC inherits that property. It is an excellent dashboard and a poor autopilot. A self-coached rider who buys TrainingPeaks expecting it to tell them what to ride on Thursday has misread what the tool is for. The adaptation gap is the one that bites in real life. A self-coached week falls apart constantly — a sick kid, a work trip, an unplanned 90-minute hard group ride that banks intensity the plan was going to prescribe later. A coach sees that and rebalances the week. TrainingPeaks records all of it faithfully and changes nothing. The calendar you wrote in January is the calendar you stare at in March, now slightly wrong in a dozen small ways. ## The decision: it comes down to who writes the plan For a self-coached rider the question is not really 'is the analytics worth $19.95.' It is 'who is writing and adapting my plan, and does TrainingPeaks help them.' If the answer is a human coach, yes — buy it, it is their delivery surface. If the answer is you, the platform supports the easy half of the job (analysis) and none of the hard half (periodization and adaptation) [Mujika 2017]. Worth it, clearly, in three cases. You work with a coach and need the calendar they write into. You genuinely enjoy the analytical craft and want the deepest fitness modeling available, including WKO5's power-duration work. Or you intend to buy a static plan from the TrainingPeaks Marketplace and follow it as written — accepting that a Marketplace plan is a fixed template that will not flex when your week does. In all three the platform is doing the job it was built for, and $134.99 a year is fair for it. Weaker value if what you actually want is a plan that gets written for you and rebuilt when life disrupts it. That is the gap the parent question of self-coaching keeps returning to: the most leveraged thing a coach provides is not prettier charts, it is the periodization and recovery enforcement that keep a year-two rider improving instead of plateauing. The free TrainingPeaks Basic tier is not the answer here either — it lets you log today's ride but cannot schedule future workouts and withholds the full PMC, so it is a logbook, not a planning tool. A sibling guide covers exactly what Basic gives and withholds. This is where the category splits, and where we sit. AdaptCycling ($15/mo or $150/yr, with a free tier that gives you a real plan and the daily workout behind it — the week-restructuring is the paid half) bets on the half TrainingPeaks deliberately leaves to a coach: it reads your ride history, writes the periodized plan, and rebuilds the week when you miss a session or bank an unplanned hard ride. It does not try to out-analyze the PMC. The two tools answer different questions — TrainingPeaks asks 'what did my training do,' an adaptive coach asks 'given that, what should the next eight weeks be.' For a head-to-head on where each wins, our comparison page goes feature by feature; this guide is only here to settle whether the analytics alone justify the subscription when no coach is in the loop. ## Common questions **Do I need TrainingPeaks if I don't have a coach?** No, not as a requirement. TrainingPeaks earns its price as a coach's delivery surface and as the deepest analytics platform in the sport. Without a coach, you are paying $19.95 a month mainly for the Performance Management Chart — and free tools like Intervals.icu compute the same CTL, ATL, and fitness-fatigue curves [Allen et al. 2019]. Buy it if you specifically want the deepest modeling or you plan to follow a static Marketplace plan; skip it if you want something to write and adapt the plan for you. **Does TrainingPeaks write a training plan for you?** No. TrainingPeaks gives a coach the tools to write a plan and gives you the calendar to receive one, but it does not generate a periodized plan on its own. You either hire a coach, buy a pre-built static plan from its Marketplace, or build the plan yourself. The platform analyzes training and delivers plans; it does not author or adapt them [Mujika 2017]. **Is the free TrainingPeaks Basic tier enough for self-coaching?** Only as a logbook. Basic lets you record today's ride but cannot schedule future workouts and withholds the full Performance Management Chart — the exact features a self-coached rider would lean on. For free analytics, Intervals.icu is a stronger choice. A separate guide covers the Basic-versus-Premium split in detail. **Is WKO5 included in TrainingPeaks Premium?** No. WKO5 is a separate desktop application bought once for roughly $169, distinct from the $19.95-a-month Premium web subscription. It adds individualized power-duration modeling beyond the web Performance Management Chart and is aimed at analytically inclined athletes and coaches. Most self-coached riders do not need it; if the deep modeling is the reason you are considering TrainingPeaks at all, it is the piece that actually delivers it. **Can I trust the Performance Management Chart to tell me when to rest?** Treat it as a strong signal, not an instruction. The PMC's fitness and fatigue values come from a model whose fitted parameters can be unstable and uncertain at the individual level [Hellard 2006], and the chart measures what your training did rather than prescribing what to do next [Mujika 2017]. A rising acute-load spike or a sharply negative balance is a real cue to consider recovery — but the decision is still yours to make. ## References 1. **Banister 1990.** [Modeling human performance in running](https://pubmed.ncbi.nlm.nih.gov/2246166/). Journal of Applied Physiology. 2. **Hellard 2006.** [Assessing the limitations of the Banister model in monitoring training](https://pubmed.ncbi.nlm.nih.gov/16608765/). Journal of Sports Sciences. 3. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 4. **Mujika 2017.** [Quantification of training and competition loads in endurance sports: methods and applications](https://pubmed.ncbi.nlm.nih.gov/27918666/). International Journal of Sports Physiology and Performance. 5. **TrainingPeaks 2026.** [Premium Pricing for Athletes](https://www.trainingpeaks.com/athlete-pricing/). TrainingPeaks. --- # TrainingPeaks free vs Premium: what the free Basic tier actually gives a self-coached cyclist URL: https://www.adaptcycling.com/guides/trainingpeaks-free-tier-self-coached Updated: 2026-06-16 Author: Jim Camut TrainingPeaks free tier is a logbook, not a training tool. The free Basic account logs and displays the rides you already did, but it cannot schedule a single future workout and it does not show the full Performance Management Chart — the fitness, fatigue, and form curves that are the real reason a training-focused rider opens TrainingPeaks. For a self-coached cyclist deciding whether to pay $19.95 a month for Premium, the honest framing is narrower than the marketing: the real choice is Premium versus a genuinely free alternative that gives you the chart for nothing. ## What the free Basic tier actually does — and stops at The free Basic tier is a logbook for completed rides. It logs and displays activities and shows per-ride metrics, but it cannot schedule future workouts and it omits the full Performance Management Chart [TrainingPeaks Basic 2026]. The two features a training-focused rider opens TrainingPeaks for — forward planning and the fitness-fatigue curves — are the two the free tier withholds. The free Basic tier is a record of the past, not a plan for the future. You can connect a device, upload completed activities, and read per-ride metrics — duration, power, heart rate, the TSS for a single workout. What you cannot do is the one thing a training plan requires: place a structured workout on a future date. TrainingPeaks documents this directly — with a Basic account you can add a workout to today but not plan future workouts, and you cannot reschedule activities on the calendar the way Premium allows [TrainingPeaks Basic 2026]. The second withheld feature is the one that matters more. The Performance Management Chart — the CTL, ATL, and TSB lines that turn a pile of single-ride TSS scores into a fitness-fatigue trajectory — is locked behind Premium [TrainingPeaks Basic 2026]. Basic shows you what each ride cost. It does not show you what the accumulation is doing to your form over six weeks, which is the only view that answers should I go hard today or rest. The PMC is not a convenience bolted onto the free tier. It is the product, and it is paywalled. ## Why the paywalled chart is the part that matters The Performance Management Chart is the product, and it is paywalled. The PMC is an applied impulse-response model — fitness minus fatigue, modeled from TSS — and it is Premium-only [Calvert & Banister 1976, TrainingPeaks Basic 2026]. The free tier hands you the daily load input and hides the model that makes it useful, which is the half of self-coaching a logbook cannot do. This withholding is not arbitrary, because the chart encodes real modeling. The Performance Management Chart is an applied form of the impulse-response model Calvert and Banister formalized in 1976, which represents performance as fitness minus fatigue — two responses to the same training load that rise and decay on different timescales [Calvert & Banister 1976]. TrainingPeaks operationalizes that model with Coggan and Allen's TSS as the daily load input [Allen et al. 2019], a roughly 42-day exponential average for chronic load (CTL, the fitness proxy) and a 7-day average for acute load (ATL, the fatigue proxy), their difference reported as form (TSB). The free tier gives you the input and hides the model. That hidden model is the half of self-coaching a logbook cannot do. Refinements such as Busso's variable dose-response model, which lets the fatigue term grow nonlinearly as hard days stack [Busso 2003], and Hellard's assessment of where the Banister model breaks down in elite swimmers [Hellard et al. 2006], exist precisely because the fitness-fatigue curve is the load-bearing abstraction in periodization. A free tier that records TSS without charting CTL and TSB is the part of coaching that surfaces the uncomfortable signal — ramp climbing too fast, form not recovering before the goal — switched off. ## Premium versus a free tool that already shows the chart Premium is $19.95 a month; a free alternative gives you the same chart for nothing. Premium unlocks scheduling and the full PMC, but Intervals.icu shows the same CTL/ATL/TSB fitness-fatigue chart free [Calvert & Banister 1976, TrainingPeaks 2026]. The real decision for a self-coached rider is free-versus-free first, and only then whether TrainingPeaks Premium earns the $20. For a self-coached rider the upgrade question splits cleanly. Premium at $19.95 a month or $134.99 a year unlocks future-workout scheduling, the full PMC, and the structured-workout calendar — the planning surface the free tier withholds [TrainingPeaks 2026]. A 14-day Premium trial lets you confirm the gap before paying. The narrower question is not whether those features are valuable; they are. It is whether you must pay TrainingPeaks specifically to get them, when a free tool already shows the same fitness-fatigue chart. Intervals.icu is that tool, and it is fair to name it plainly. It is free with no trial limit and no credit card, it pulls your rides from Strava, and its free tier computes the full Fitness, Fatigue, and Form chart — the same CTL, ATL, and TSB model the PMC implements, on the same Banister-derived math [Calvert & Banister 1976]. An optional supporter tier runs about $4 a month if you want to fund the project. For a solo rider who wants the fitness-fatigue chart and power-curve analysis without paying $20 a month, the free-versus-free comparison favors Intervals.icu before the Premium decision is even reached. ## Where this fits for the self-coached rider Free Basic is fine as a record and useless as a coaching setup. The broad paid-TrainingPeaks worth-it question is its own decision [TrainingPeaks 2026]; this spoke answers the narrow one — for the fitness chart alone, free beats paying. A logbook replaces no part of a coach, which is the trap this choice hides. This is the spoke-sized version of a larger question. Whether the paid TrainingPeaks ecosystem is worth it overall — the coach-share workflow, the structured-workout library, WKO5 as a separate roughly $169 desktop purchase for deep analytics — is its own decision, and we cover the broad worth-it verdict in a companion guide [TrainingPeaks 2026]. The one-line version: for a rider with no coach, the paid case rests on planning and structured workouts, not on the fitness chart, which you can get free elsewhere. It also sits inside the broader picture of training without a coach in 2026. The self-coached rider's defining task is deciding which tools replace which parts of a coach, and a logbook replaces none of them — it records data a coach would interpret, without doing the interpreting [Allen et al. 2019]. The features TrainingPeaks paywalls are exactly the planning and load-management functions self-coaching most needs to externalize. Choosing a free logbook and calling it a coaching setup is the quiet version of the year-two plateau: plenty of data, no structure acting on it. Where AdaptCycling differs is the layer above the chart. The free TrainingPeaks tier withholds the plan; Intervals.icu gives you the fitness-fatigue chart but, by design, charts rather than prescribes. Our free tier writes an actual adaptive plan against your ride history — and rebuilds it when a week falls apart — rather than handing you curves to interpret yourself. That is a different free-tier bet: not a better logbook, but the planning the logbooks leave to you. ## Common questions **Can the free TrainingPeaks tier schedule future workouts?** No. The free Basic account lets you add a workout to today and upload completed activities, but it cannot plan workouts on future dates or reschedule activities on the calendar the way Premium does [TrainingPeaks Basic 2026]. For a self-coached rider that limit alone makes the free tier a logbook rather than a planning tool — you can record the past but not lay out the week ahead. **Does the free TrainingPeaks tier show CTL ATL and TSB?** No. The full Performance Management Chart — the fitness (CTL), fatigue (ATL), and form (TSB) curves — is a Premium feature [TrainingPeaks Basic 2026]. Basic shows the TSS for individual rides but not the chronic-load trajectory that turns those scores into a fitness-fatigue picture [Allen et al. 2019]. If the fitness chart is what you want and you do not want to pay, Intervals.icu computes it free. **Is TrainingPeaks Premium worth $19.95 a month for a self-coached rider?** It depends on what you need from it. Premium at $19.95 a month or $134.99 a year buys forward planning, the full PMC, and a structured-workout calendar [TrainingPeaks 2026]. For the fitness chart alone it is not worth it, because Intervals.icu shows the same CTL/ATL/TSB free. The paid case rests on planning and structured workouts — covered in our broader worth-it guide. **Is there a free TrainingPeaks alternative with the full fitness chart?** Yes. Intervals.icu is free, requires no credit card, pulls rides from Strava, and computes the full Fitness, Fatigue, and Form chart — the same CTL/ATL/TSB model the PMC implements [Calvert & Banister 1976]. An optional supporter tier is about $4 a month. For a solo self-coached rider it is the closest free match to TrainingPeaks Premium's chart features. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Calvert & Banister 1976.** [A Systems Model of the Effects of Training on Physical Performance](https://ieeexplore.ieee.org/document/5409179). IEEE Transactions on Systems Man and Cybernetics. 3. **Hellard et al. 2006.** [Assessing the limitations of the Banister model in monitoring training](https://pubmed.ncbi.nlm.nih.gov/16608765/). Journal of Sports Sciences. 4. **Busso 2003.** [Variable dose-response relationship between exercise training and performance](https://pubmed.ncbi.nlm.nih.gov/12840641/). Medicine & Science in Sports & Exercise. 5. **TrainingPeaks 2026.** [Pricing for Athletes](https://www.trainingpeaks.com/pricing/for-athletes/). TrainingPeaks. 6. **TrainingPeaks Basic 2026.** [Basic vs Premium Athlete Accounts](https://help.trainingpeaks.com/hc/en-us/articles/204074014-Basic-vs-Premium-Athlete-Accounts). TrainingPeaks Help Center. --- # Indoor vs outdoor FTP: why the same legs make less power on the trainer URL: https://www.adaptcycling.com/guides/indoor-vs-outdoor-ftp Updated: 2026-07-17 Author: Jim Camut Your indoor FTP is usually lower than your outdoor FTP, and for most riders that gap is real, not a measurement glitch. The dominant cause is heat: a trainer removes the 25 to 40 km/h airflow that evaporates your sweat outdoors, so core temperature climbs, heart rate drifts up, and sustainable power falls within about ten minutes. A second cause is the meter itself, when a virtual-power trainer is compared against a calibrated power meter. The gap is commonly around 5 percent, but it is individual, and you probably do not need two separate FTP numbers to manage it. ## Heat is the main reason, and it is physiology, not weakness Outdoors, riding at 30 km/h pushes air across your skin that evaporates sweat and holds core temperature down. On a trainer that airflow is gone, so heat accumulates. Endurance capacity peaks in cool-to-moderate conditions and falls off sharply as you get hotter, which is exactly the direction a still room pushes you. Galloway and Maughan had cyclists ride to exhaustion at four ambient temperatures and found an inverted-U response [Galloway and Maughan 1997]. Time to exhaustion was longest at 10.5 C (93.5 minutes) and shortest at 30.5 C (51.6 minutes) — a 45 percent collapse in endurance from the same legs, driven only by temperature. A ventilated outdoor ride sits near the top of that curve; a warm, unventilated pain cave sits well down its right-hand slope. The reason there is a ceiling at all is that fatigue in the heat tracks core temperature itself. Gonzalez-Alonso and colleagues pre-warmed and pre-cooled cyclists to different starting temperatures, yet every one of them stopped at an almost identical core temperature of 40.1 to 40.2 C [Gonzalez-Alonso et al. 1999]. Indoors you reach that ceiling faster because the heat you produce has nowhere to go, so the wattage you could hold outdoors for an hour becomes unsustainable well before it. This is why the back half of an indoor 20-minute test fades even when your legs felt fresh at the start. You are not undertrained on those days, you are overheating, and the power number you record is a heat-limited number rather than a true threshold. ## The symptom you can see is heart-rate drift At a fixed indoor wattage your heart rate keeps climbing after the first ten minutes. That is cardiovascular drift, and it is the visible fingerprint of the heat problem above. It means that to defend a target heart rate you must shed watts, so a heart-rate-anchored effort quietly downshifts your power. Cardiovascular drift is a progressive fall in stroke volume that begins after 10 to 20 minutes of steady exercise, with heart rate rising to compensate [Coyle and Gonzalez-Alonso 2001]. Each beat moves less blood, so the heart beats more often to hold cardiac output. On the road, airflow blunts the underlying heating; on a trainer it runs unchecked, so the same 250 watts that read 150 bpm early can read 165 bpm forty minutes in. Practically, this is the single biggest lever you control. A powerful fan or a cooler room restores evaporative cooling and shifts you back up the temperature curve from Galloway and Maughan, recovering much of the lost power [Galloway and Maughan 1997]. Riders who add a proper fan often watch most of their indoor deficit disappear, because they were never weaker indoors — they were 2 to 3 degrees hotter. ## You may be comparing two different rulers Not all of the gap is your body. If your outdoor number comes from a calibrated crank or pedal meter and your indoor number comes from a trainer's estimated power, part of the difference is the measurement, not the athlete. Power meters only agree when they are calibrated and compared under matched conditions. A strain-gauge power meter measures the torque you apply directly. Virtual power from a non-smart trainer is modeled from a speed-and-resistance curve, so it is an estimate rather than a measurement, and it drifts with tire pressure, roller tension, and how warm the unit is. A systematic scoping review of cycling power meters concluded that validity is conditional: accuracy shifts with temperature, cadence, rider position, and calibration state [Bouillod et al. 2022]. Two things follow. First, run a spindown or zero-offset after a few minutes of warm-up, every session, so the trainer is not reading cold. Second, where you can, record with the same meter indoors and out — a pedal or crank meter used in both places deletes the ruler problem entirely and leaves only the physiological gap to explain. Position and cadence add a little more. On a fixed trainer you cannot rock the bike or recruit stabilizers the way you do on the road, and riders tend to spin faster indoors: the professional cyclists in Lipski's study self-selected a cadence about 6 rpm higher indoors [Lipski et al. 2022]. None of this is large on its own, but it stacks on top of the heat penalty. ## Do you need two FTPs? Usually one honest estimate is better You can keep separate indoor and outdoor FTPs, and if you race seriously in both worlds that is defensible. But the gap is too individual to fix with a fixed rule, and for most self-coached riders a single estimate drawn from all your riding is simpler and more accurate than two forced tests. The individual variation is the catch. In Lipski's professional cyclists, critical power averaged 19 watts higher outdoors — roughly 5 percent for a typical rider — but the athlete-to-athlete spread was wide enough that the authors warned indoor and outdoor results cannot be used interchangeably at the individual level [Lipski et al. 2022]. So the popular advice to just subtract 5 percent indoors is a rough patch that is wrong for most people by some margin. Step back and the deeper issue is that FTP was never a single test. Andrew Coggan defined it as the highest power you can hold in a quasi-steady state for roughly an hour [Allen et al. 2019], and even that only approximates the lab's maximal lactate steady state, with a typical error near 5 percent [Borszcz et al. 2019]. This is the whole premise of estimating threshold from data you already have: rather than staging two brittle tests in two environments, you infer one threshold from the efforts you have already ridden, indoors and out. That is how we handle it at AdaptCycling. We fit a Critical Power model to your entire power curve, which pools your best efforts wherever they happened — a cool outdoor 20-minute climb and a hot indoor 5-minute block feed the same estimate. A heat-suppressed trainer day simply never becomes one of your best efforts, so it lowers nothing; it just is not selected. You get one threshold that updates as you add rides, with no test day to schedule and no 5 percent fudge factor to maintain. ## Common questions **How much lower is indoor FTP than outdoor FTP?** Commonly around 5 percent, but treat that as a loose average, not a law. In professional cyclists, outdoor critical power averaged 19 watts higher than indoor, yet the individual spread was large [Lipski et al. 2022]. Your own gap depends mostly on how well you cool the room. **Will a fan fix the difference?** It fixes most of the heat-driven part. A strong fan restores the evaporative cooling you lose without airflow, moving you back up the temperature-versus-endurance curve where capacity is highest [Galloway and Maughan 1997]. It will not fix a difference that comes from two different power meters. **Should I set two separate FTP numbers and two sets of zones?** Only if you genuinely train and race in both environments, for instance structured indoor racing plus outdoor events. Otherwise a single estimate is simpler and avoids maintaining two numbers that both drift. The wide individual variation makes any fixed offset unreliable [Lipski et al. 2022]. **Is my trainer under-reading my power?** Possibly, if it uses estimated or virtual power rather than a strain gauge. Run a spindown after warming up and, ideally, record with the same power meter indoors and out, because power-meter accuracy depends on calibration and conditions [Bouillod et al. 2022]. **Why is my heart rate higher indoors at the same watts?** Cardiovascular drift. As you heat up, stroke volume falls and heart rate rises to compensate, so the same wattage costs more beats per minute [Coyle and Gonzalez-Alonso 2001]. It is a cooling problem, not necessarily lost fitness. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Borszcz et al. 2019.** [Is the Functional Threshold Power Interchangeable With the Maximal Lactate Steady State in Trained Cyclists?](https://pubmed.ncbi.nlm.nih.gov/30676826/). International Journal of Sports Physiology and Performance. 3. **Lipski et al. 2022.** [Differences in Performance Assessments Conducted Indoors and Outdoors in Professional Cyclists](https://pubmed.ncbi.nlm.nih.gov/35361736/). International Journal of Sports Physiology and Performance. 4. **Galloway and Maughan 1997.** [Effects of ambient temperature on the capacity to perform prolonged cycle exercise in man](https://pubmed.ncbi.nlm.nih.gov/9309637/). Medicine and Science in Sports and Exercise. 5. **Gonzalez-Alonso et al. 1999.** [Influence of body temperature on the development of fatigue during prolonged exercise in the heat](https://pubmed.ncbi.nlm.nih.gov/10066720/). Journal of Applied Physiology. 6. **Coyle and Gonzalez-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. 7. **Bouillod et al. 2022.** [Caveats and Recommendations to Assess the Validity and Reliability of Cycling Power Meters: A Systematic Scoping Review](https://pmc.ncbi.nlm.nih.gov/articles/PMC8749704/). Sensors. --- # Peloton FTP test: what your number means, and where it stops being true URL: https://www.adaptcycling.com/guides/peloton-ftp-test-what-your-number-means Updated: 2026-09-03 Author: Jim Camut The Peloton FTP test is a 20-minute maximum effort, and the number you enter afterward is your average output for those 20 minutes — Peloton applies the 95% factor for you [Peloton FTP Guide 2026]. That makes the result a fine ruler inside Peloton, where every zone is a ratio of the same reading. It is not a power-meter FTP. Output on the Bike is derived from cadence and resistance rather than measured at the crank, and Peloton itself says the brake varies a little from Bike to Bike [Peloton Support 2026, Peloton Calibration 2026]. Here is what the number can and cannot carry. ## What the test measures, and what the 95% means The test is a 20-minute best effort and the Bike calculates FTP from its average output [Peloton FTP Guide 2026]. FTP itself is the output you can hold for roughly an hour, so the 20-minute average is scaled down by 5% to estimate it. The scaling is a group average; individually it can miss by tens of watts. Peloton's own guidance is precise about the input. To start using Power Zones you identify your 20-minute average output, and that number is what you enter under Preferences to have your zones calculated [Peloton Blog 2026]. The conversion to FTP happens behind the setting: multiply the 20-minute average by 0.95, or let the Bike do it [Peloton FTP Guide 2026]. So if your test averaged 200, the FTP driving your zones is 190, and a zone 4 floor at 91% sits at roughly 173 rather than 182. Riders who assume the number on screen is their FTP end up with zones about 5% too hard. The 5% is a convention, not a measurement of you. Borszcz and colleagues put 23 trained cyclists through a 20-minute test and a full 60-minute effort and found the group averages nearly identical — a bias of only 1 to 5 watts — while the individual limits of agreement ran as wide as 62 watts below to 60 above [Borszcz et al. 2018]. Time to exhaustion at the 20-minute-derived power averaged 50.9 minutes, not 60. Against the lab's maximal lactate steady state the same 95% estimate carried a bias of 1.4% with limits of about 9% either way [Borszcz et al. 2019]. Good on average, loose for one rider. The warm-up is part of the protocol, which is why Peloton pairs the test with a dedicated FTP Warmup Ride [Peloton Blog 2026]. Tramontin and colleagues had cyclists ride the same 20-minute test after different warm-ups and recorded averages from 256 to 270 watts — about 5% moved by the warm-up alone [Tramontin et al. 2022]. Skip the warm-up ride one month and take it the next, and the change in your number may be the warm-up, not your fitness. Do the test the same way every time. ## Your Peloton number is a Peloton number Output on the Bike is calculated, not measured. Hall effect sensors read flywheel speed and resistance and the console derives watts from those values [Peloton Support 2026]. The brake is calibrated at the factory with what Peloton calls minor variation from Bike to Bike [Peloton Calibration 2026]. Inside Peloton none of that matters; outside it, all of it does. A crank or pedal power meter measures the torque you apply with strain gauges. The Peloton Bike does something different: sensors capture cadence and the resistance setting and reference output values on the sensor control board [Peloton Support 2026]. That is a lookup, and it is only as good as the mapping between the resistance knob and the brake. The original Bike has a manual magnetic brake calibrated in manufacturing, and Peloton states plainly that there is some minor variation from Bike to Bike; the Bike+ has a digitally controlled brake that calibrates itself on first use [Peloton Calibration 2026]. Peloton also lists room temperature among the things that make output feel different, which is the indoor heat problem our guide on indoor versus outdoor FTP covers in full. We could not find a peer-reviewed study that tests Peloton output against a calibrated strain-gauge meter, so nobody can tell you the size of the gap with evidence. What can be said is that the reading is consistent with itself. Because Power Zones are ratios of your own test on your own Bike, a Bike that reads 5% high reads 5% high in zone 2 and in zone 5 alike, and every Power Zone class still lands at the intended intensity. The ruler is bent, but you are always measuring with the same ruler. The moment you carry the number to a different ruler, the bend shows. A Peloton FTP entered into Zwift, a head unit paired with pedals, or any app that reads outdoor rides sets zones that are off by your Bike's offset, in a direction you cannot know in advance. Different apps already disagree about FTP for reasons of method; a different sensor adds a second disagreement on top. Treat the Peloton number as valid on the Peloton Bike and as a rough starting guess anywhere else, and retest on the new equipment before you trust a zone. ## Turning the number into zones, and what each zone is for Peloton uses seven zones, each a band of FTP: zone 1 under 55%, zone 2 at 56 to 75%, zone 3 at 76 to 90%, zone 4 at 91 to 105%, zone 5 at 106 to 120%, zone 6 at 121 to 150%, and zone 7 above 151% [Peloton Blog 2026]. These are the standard power-based bands, so a Power Zone class and a coach's plan speak the same language. The bands do specific jobs. Zone 2 is where most of an endurance rider's hours belong; it builds the aerobic base and can be repeated day after day. Zone 3 is sustainable but costly, and the top of it — the sweet spot around 88 to 94% of FTP — is where time-limited riders get the most threshold progress per hour. Zone 4 straddles FTP itself and trains the ability to hold it. Zone 5 is the VO2max band, ridden in intervals of three to five minutes. Zones 6 and 7 are anaerobic and sprint work, minutes and seconds respectively. You do not need to work the percentages by hand. Once the Bike has your test, it customizes each zone's output range for you [Peloton Blog 2026], and a power zone calculator gives the same bands in watts for any FTP if you want to see them written out. What matters more than the arithmetic is remembering which FTP you are converting: the 95%-scaled value, not the raw 20-minute average. Peloton's Custom Value setting takes the raw average and does the scaling; a calculator expects the FTP. ## Retesting, and where Peloton stops Peloton's own pages give two retest intervals — every 4 to 6 weeks in its Power Zone guide and every 6 to 12 weeks in its FTP guide [Peloton Blog 2026, Peloton FTP Guide 2026]. Both are more often than most riders need. The better trigger is the workouts themselves, and the bigger gap is that a Peloton FTP feeds classes, not a plan. The retest interval is a judgement call, and Peloton's two answers show it. A 20-minute maximum effort costs real recovery, and an FTP that moves 2% in six weeks is inside the noise of the protocol — Tramontin's warm-up effect alone was 5% [Tramontin et al. 2022]. Retest when zone 4 rides have become comfortable, when zone 5 intervals are finishing with something in reserve, or after a break long enough that the old number is clearly stale. Our guide on how often to test FTP makes the case for a couple of anchored efforts a year rather than a calendar. Between tests, the honest answer to the question this article started with is that your number is already in your riding. Every hard Power Zone class leaves a best 5-, 10-, and 20-minute output behind, and a threshold estimated from that power curve tracks your fitness continuously without a test day. That is the approach our pillar on finding FTP without a test describes, and it is what AdaptCycling does for riders whose ride files reach a platform we read: we fit a Critical Power model to your mean-maximal-power curve and update the estimate as new rides arrive. We read Wahoo, Hammerhead, Polar and intervals.icu today; a Peloton-only rider can use the test number as the seed. The larger limit is structural. A Power Zone class is an instructor calling zones for a room; the number you tested personalizes the targets, but the sequence of classes is yours to choose, and nothing restructures the week when you miss Tuesday or get sick. Peloton makes the hour you show up for a hard, well-cued hour. Deciding which hours, in what order, and what changes when life intervenes is the part a plan exists to do. ## Common questions **Is the Peloton FTP 95% of my 20-minute test?** Yes. You enter your 20-minute average output and Peloton scales it by 0.95 to set your FTP and your zones [Peloton FTP Guide 2026, Peloton Blog 2026]. If you type the FTP into a separate power zone calculator, use the scaled value, not the raw average, or every zone comes out about 5% too high. **Is Peloton output the same as watts from a power meter?** No. Peloton derives output from cadence and resistance sensors rather than measuring torque, and it acknowledges minor variation between Bikes from manufacturing [Peloton Support 2026, Peloton Calibration 2026]. We found no peer-reviewed study quantifying the gap. The number is consistent on your own Bike, which is all Power Zones need, but do not assume it matches a crank or pedal meter. **How often should I retake the Peloton FTP test?** Peloton suggests every 4 to 6 weeks on one page and every 6 to 12 weeks on another [Peloton Blog 2026, Peloton FTP Guide 2026]. Retest when zone 4 and zone 5 work has clearly become easier, or after a long break. Warm-up alone can shift a 20-minute result by about 5% [Tramontin et al. 2022], so small month-to-month changes are mostly noise. **Can I use my Peloton FTP on Zwift or outdoors?** Only as a starting guess. A different sensor is a different ruler, and the direction of the offset is unknown, so zones set from the Peloton number on a smart trainer or a road power meter can miss by a meaningful margin. Retest on the new equipment, or let a model fit to your rides on that equipment estimate the threshold from the power curve. ## References 1. **Peloton Blog 2026.** [Power Zone Training with Peloton: Everything You Need to Know](https://www.onepeloton.com/blog/power-zone-training). Peloton. 2. **Peloton FTP Guide 2026.** [Functional Threshold Power (FTP): How to Test & Why It Matters](https://www.onepeloton.com/blog/ftp-cycling). The Output by Peloton. 3. **Peloton Support 2026.** [Bike and Bike+ Metrics](https://support.onepeloton.com/s/article/203325985-Bike-and-BikePlus-Metrics?language=en_US). Peloton Support. 4. **Peloton Calibration 2026.** [Calibrating Your Peloton Bike or Bike+](https://support.onepeloton.com/s/article/115005204826-Calibrating-Your-Peloton-Bike?language=en_US). Peloton Support. 5. **Borszcz et al. 2018.** [Functional Threshold Power in Cyclists: Validity of the Concept and Physiological Responses](https://pubmed.ncbi.nlm.nih.gov/29801189/). International Journal of Sports Medicine. 6. **Borszcz et al. 2019.** [Is the Functional Threshold Power Interchangeable With the Maximal Lactate Steady State in Trained Cyclists?](https://pubmed.ncbi.nlm.nih.gov/30676826/). International Journal of Sports Physiology and Performance. 7. **Tramontin et al. 2022.** [Functional Threshold Power Estimated from a 20-minute Time-trial Test is Warm-up-dependent](https://pubmed.ncbi.nlm.nih.gov/34749416/). International Journal of Sports Medicine. --- # FTP without a power meter: estimating a threshold from heart rate and effort URL: https://www.adaptcycling.com/guides/ftp-without-a-power-meter Updated: 2026-07-17 Author: Jim Camut You can estimate a working FTP-equivalent threshold without ever owning a power meter — you just accept a wider error bar. A 30-minute solo time trial gives you a lactate threshold heart rate, and percentage-of-LTHR zones then organize your training the way percentage-of-FTP zones would. Expect roughly a 10 percent margin instead of the 3 to 5 percent a power meter buys you [Achten & Jeukendrup 2003]. Here is how to anchor the number, where heart rate quietly lies to you, and the one benchmark worth borrowing a meter for. ## Anchor a threshold with your heart rate: the LTHR field test The heart-rate equivalent of FTP is lactate threshold heart rate, and you find it with a 30-minute solo time trial: your average heart rate over the final 20 minutes is your LTHR [Friel 2018]. A field study estimating threshold heart rate from a 30-minute effort reported a standard error of about 8 beats per minute against a lab criterion [McGehee et al. 2005]. From LTHR, percentage zones do the rest. The protocol is deliberately simple. Warm up, then ride 30 minutes as hard as you can sustain, alone and without drafting, on an open road or a trainer, and take the average heart rate of the last 20 minutes. The first 10 minutes are discarded on purpose: heart rate takes a minute or two to climb toward a steady state, so the early portion understates the threshold response. McGehee and colleagues validated this 30-minute time-trial approach with a standard error of roughly 8 beats per minute versus laboratory lactate testing [McGehee et al. 2005]. Turning LTHR into zones is where the method earns its keep. Friel's percentage-of-LTHR system puts endurance riding below roughly 90 percent of LTHR and threshold work around 95 to 105 percent, with sweet spot and tempo filling the gap [Friel 2018]. These are percentages of a threshold anchor, not of maximum heart rate. The alternative is the Karvonen heart-rate-reserve method, which sets each target as heart rate reserve times an intensity plus resting heart rate, but it requires a true maximum and a true resting heart rate to be accurate [Karvonen et al. 1957]. Percentage-of-LTHR zones approximate percentage-of-FTP power zones because both anchor to the same physiological boundary: the transition near lactate threshold that separates efforts you can hold for hours from efforts that spiral toward failure. For steady, moderate-length riding the two zone systems land in nearly the same place. The catch is that the mapping only holds for steady efforts; it falls apart for anything short or surging, which is the heart of heart rate's accuracy problem. ## No monitor at all? Anchor effort with the talk test and RPE If you have neither power nor heart rate, perceived effort still brackets the same zones. The talk test is the most reliable field cue: the intensity at which speaking a full sentence first becomes difficult sits almost exactly at ventilatory threshold [Persinger et al. 2004], close to the physiological boundary FTP describes. Perceived exertion covers the rest of the range. Persinger and colleagues showed that the last stage at which a rider can still recite a passage comfortably corresponds to intensity below ventilatory threshold, and the stage where speech turns choppy marks the threshold itself, on both a cycle ergometer and a treadmill [Persinger et al. 2004]. Practically, if you can hold a full-sentence conversation you are in endurance territory; if you can only force out three or four words you are at threshold or above. It costs nothing and needs no device. Rating of perceived exertion fills in the ladder. On the 1-to-10 scale, a 2 to 3 is easy endurance, 4 to 6 is tempo and sweet spot, 7 to 8 is threshold, and 9 to 10 is the VO2max and anaerobic range you can only hold for minutes. RPE is the least precise anchor in absolute terms, but it is the most robust to heat and fatigue, because it integrates everything you feel and makes you back off automatically when you are cooked, which a fixed heart-rate or power target never does. None of this is a downgrade to be embarrassed about; professionals raced on perceived effort for decades before power meters existed. The real cost of training with no device is not daily intensity control, which effort cues handle well, but the loss of a trackable number. You cannot see a 5-watt improvement in the phrase felt hard, so progression becomes something you infer rather than measure. ## Why heart rate lags, drifts, and quietly lies Heart rate is a downstream signal, not a direct measure of work, and three confounds widen its error bar to roughly 10 percent against the 3 to 5 percent a power meter delivers [Achten & Jeukendrup 2003]. Lag, cardiovascular drift, and ordinary day-to-day physiology all move your heart rate while your actual threshold sits still. Lag comes first. Heart rate needs on the order of 90 to 120 seconds to respond fully to a change in intensity, so on short intervals of 30 seconds, one minute, or even three minutes it never catches up to the true effort [Achten & Jeukendrup 2003]. That single fact is why heart rate can govern steady endurance and threshold work but is useless for prescribing VO2max or anaerobic intervals; the number you see is always describing where you were, not where you are. Cardiovascular drift comes second. During a prolonged steady effort, stroke volume falls and heart rate climbs from about 10 to 20 minutes in even when power is dead constant, as Coyle and Gonzalez-Alonso documented [Coyle & Gonzalez-Alonso 2001]. Heat and dehydration amplify the effect, so on a hot ride your heart rate can sit 10 to 20 beats per minute higher for the very same wattage. Chase a fixed heart-rate target on that ride and you will unknowingly pedal easier and easier just to hold the number down. Day-to-day physiology comes third. Sleep debt, caffeine, stress, altitude, and the early stages of illness all shift heart rate at a given effort, and Achten and Jeukendrup note that heart-rate-based predictions of aerobic capacity can deviate up to 20 percent from the true value [Achten & Jeukendrup 2003]. None of this makes heart rate useless. It means a heart-rate zone is a fuzzy band rather than a hard line, and on any day the number reads strange you should cross-check it against the talk test. ## The practical path: borrow power for one benchmark The highest-leverage move for a rider without power is a single borrowed-power benchmark. One ride on a friend's power meter, a rental, or a smart trainer's estimated watts calibrates your entire heart-rate-to-effort map, after which you can keep training by heart rate against a known reference instead of a guess. The tactic is cheap. Record one 20-to-30-minute hard effort with both power and heart rate captured on the same file, and you learn exactly what your LTHR corresponds to in watts. From then on your heart-rate zones carry a power anchor even though you never bought a meter. A smart trainer's estimated power is less accurate than a direct-force meter, but for a one-time calibration it is more than close enough to place your threshold [Allen et al. 2019]. This spoke is the branch of our FTP without a test pillar for riders who have no power at all. That pillar assumes you already own months of power data and shows how to estimate your threshold from your Strava power curve with a Critical Power model, no forced test required. The goal here is identical — estimate an FTP-equivalent threshold without a formal test — but the anchor is heart rate and effort instead of watts, and the moment real power data appears the power path becomes strictly better and you graduate to it. AdaptCycling is built for exactly that graduation. During onboarding we check whether your ride history contains power; if it does, we model FTP from the curve, and if it does not, we fall back to heart-rate zones built from your threshold and still generate a full plan, honest about the wider margin. The instant a ride with power lands — a borrowed meter, a new trainer, a rental — we re-derive your threshold from the power curve and the estimate tightens from about 10 percent toward 3 to 5 percent. You are never blocked from a plan for lacking a power meter, and never stuck with the heart-rate estimate once better data exists. ## Common questions **How accurate is an FTP estimate from heart rate alone?** Roughly a 10 percent margin, against the 3 to 5 percent a power meter delivers [Achten & Jeukendrup 2003]. That is precise enough to organize your zones and train consistently, but not precise enough to chase small watt gains week to week. The estimate tightens the moment you record even one ride with power, because heart rate stops being your only reference. **What is the difference between LTHR zones and max-heart-rate zones?** LTHR zones anchor to your lactate threshold heart rate, a repeatable number from a 30-minute time trial [Friel 2018], while percentage-of-maximum zones anchor to a maximum most riders only estimate. Because a wrong maximum corrupts every percentage-of-max zone, threshold-anchored LTHR — or the heart-rate-reserve Karvonen method, which also needs a true resting heart rate [Karvonen et al. 1957] — is usually the more reliable choice for cyclists. **Can I use my smart trainer's power number as a real benchmark?** For one calibration effort, yes. A smart trainer's estimated power is less accurate than a direct-force power meter, but a single 20-to-30-minute effort recorded with both power and heart rate is enough to map your LTHR to watts [Allen et al. 2019]. Treat that number as a reference point you calibrate against, not a precise figure to defend every day. **Why does my heart rate climb on long rides even when I am not going harder?** Cardiovascular drift. Stroke volume falls and heart rate rises from about 10 to 20 minutes into a steady effort, and heat and dehydration amplify it [Coyle & Gonzalez-Alonso 2001]. If you chase a fixed heart-rate number you will ride progressively easier to hold it. Let the zone be a soft band and cross-check with perceived effort on long or hot days. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Friel 2018.** [The Cyclist's Training Bible (5th ed.)](https://books.google.com/books/about/The_Cyclist_s_Training_Bible.html?id=VCtZDwAAQBAJ). VeloPress. 3. **McGehee et al. 2005.** [A Comparison of Methods for Estimating the Lactate Threshold](https://pubmed.ncbi.nlm.nih.gov/16095403/). Journal of Strength and Conditioning Research. 4. **Achten & Jeukendrup 2003.** [Heart Rate Monitoring: Applications and Limitations](https://pubmed.ncbi.nlm.nih.gov/12762827/). Sports Medicine. 5. **Coyle & Gonzalez-Alonso 2001.** [Cardiovascular Drift During Prolonged Exercise: New Perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. 6. **Persinger et al. 2004.** [Consistency of the Talk Test for Exercise Prescription](https://pubmed.ncbi.nlm.nih.gov/15354048/). Medicine & Science in Sports & Exercise. 7. **Karvonen et al. 1957.** [The Effects of Training on Heart Rate: A Longitudinal Study](https://pubmed.ncbi.nlm.nih.gov/13470504/). Annales Medicinae Experimentalis et Biologiae Fenniae. --- # 20-minute vs 8-minute FTP test: which one to do and what the multipliers mean URL: https://www.adaptcycling.com/guides/ftp-20-min-vs-8-min-test Updated: 2026-07-17 Author: Jim Camut Two field tests estimate the same number two different ways. The 20-minute test takes 95% of your average power over a single all-out 20-minute effort. The 8-minute test — popularized by Chris Carmichael's CTS — takes 90% of the average power from the better of two 8-minute efforts. Both approximate the power you could hold for about an hour. The different multiplier, and which test suits you, come down to one thing: how much of a short hard effort is anaerobic. Neither test is the definition of FTP, and you may not need either. ## The two protocols, side by side The 20-minute test is one maximal 20-minute effort, with FTP set at 95% of the average power [Allen et al. 2019]. The CTS 8-minute test is two 8-minute all-out efforts split by 10 minutes of easy spinning, with FTP set at 90% of the higher effort's average power [Carmichael & Rutberg 2017]. Same target: roughly your 60-minute power. The 20-minute protocol is the one most amateurs meet first, on Zwift or in TrainerRoad — though Zwift does not offer it standalone, which our guide on which Zwift FTP test to take unpacks. After a thorough warm-up you ride one 20-minute time trial as hard as you can hold evenly, then multiply the average power by 0.95. The 5% haircut exists to approximate the drop-off between 20-minute power and the hour of power that FTP is meant to represent [Allen et al. 2019]. It is a single number from a single effort, which makes it simple and also fragile. The 8-minute protocol comes from Chris Carmichael and Jim Rutberg's The Time-Crunched Cyclist [Carmichael & Rutberg 2017]. After warming up you ride two 8-minute efforts all-out, separated by about 10 minutes of easy recovery, and CTS sets training intensities from 90% of the average power of the higher of the two efforts. Some coaches average both efforts instead; either way the reduction is 10%. Two shorter efforts also reveal how well you recover and repeat hard work, which one long effort cannot. Both protocols chase the same physiological quantity: the highest power a rider can sustain in a quasi-steady state for roughly an hour, the definition Andrew Coggan gave FTP [Allen et al. 2019]. Neither test lasts an hour, because a true 60-minute time trial is brutal and rarely repeated. Both are field surrogates that lean on a fixed multiplier to bridge the gap — and the multiplier is where they part ways. ## Why the multiplier differs: 95% versus 90% A shorter maximal effort produces a higher average power relative to true threshold, so it needs a bigger discount. Eight minutes sits well into the severe-intensity domain near VO2max power, drawing more on finite anaerobic work capacity — W-prime — than a 20-minute effort does. The extra 5% subtracts that anaerobic contribution [Karsten et al. 2021]. On the power-duration curve, the shorter the effort, the further it sits above your sustainable threshold and the larger the anaerobic share of the work. Karsten and colleagues [Karsten et al. 2021] measured W-prime at 16.6 kJ in trained cyclists — worth only about 4.6 watts when spread across a full hour, but a far larger slice of an 8-minute effort. They concluded the two-parameter critical power model's validity should be limited to durations under 20 minutes, precisely because anaerobic contribution distorts shorter efforts. The 90% multiplier is the correction for that distortion; the 20-minute test needs only 5%. Field tests already read high against physiological threshold, and shorter efforts read higher still. Vinetti and colleagues [Vinetti et al. 2023] found that in junior road cyclists, even the 95%-adjusted 20-minute field FTP came in around 14% above laboratory critical power (269 versus 236 watts), on a steep outdoor climb. FTP correlated almost perfectly with aerobic power (MAP, r = 0.95) but also moderately with the anaerobic W-prime (r = 0.66). Compress the effort to 8 minutes and the anaerobic lean grows, which is why the multiplier has to work harder. This is why the two numbers are not interchangeable. Karsten and colleagues [Karsten et al. 2021] reported limits of agreement from minus 19 to plus 33 watts between critical power and 20-minute FTP — a spread of method noise before fitness even enters. Switch from the 8-minute to the 20-minute protocol mid-season and your FTP can move several percent from the change of method alone. Pick one and stay with it, so a moving number means moving fitness. ## Which test fits which rider Diesels — riders whose threshold sits close to their peak aerobic power — pace long efforts evenly and are best served by the 20-minute test. Anaerobically-gifted riders and novices, who fade or blow up over 20 minutes, get a steadier number from the 8-minute test, though its short duration over-weights their anaerobic strength. The 20-minute test rewards pacing discipline, which most amateurs lack. Tramontin and colleagues [Tramontin et al. 2022] showed that warm-up structure alone shifts the resulting 20-minute FTP by clinically meaningful amounts, and pacing errors do more damage than that: a rider who goes out too hard and cracks at minute 14 records a falsely low number. A diesel with an even power delivery holds the 20 minutes comfortably; a punchy rider fights it the whole way. The 8-minute test is easier to pace and to repeat, which is why CTS recommends it for novices and age-group racers. Eight minutes is short enough that even inexperienced riders can hold a near-even effort, and two efforts give a repeatability check. The tradeoff is the one the ramp test also suffers: because 8 minutes leans anaerobic, a sprinter-type gets a flattering number that the 90% multiplier may not fully rein in, so a ramp-test FTP that reads too high tends to read high here too. The honest caveat is that neither field test gives ground truth. Borszcz and colleagues [Borszcz et al. 2018] found that on an individual basis the 20-minute estimate and a real 60-minute effort disagreed by 40 to 60 watts — for a 250-watt rider, the gap between a productive session and a blown one. The 8-minute test stacks a second correction on top of a shorter, noisier effort. Both are starting estimates to be refined, not verdicts to be trusted. ## Why you may not need either test Both the 20-minute and 8-minute tests are protocols, not the definition of FTP — the same point the parent guide makes. If you already ride with power, the near-maximal efforts sitting in your Strava history can estimate threshold power within a few percent, without scheduling a dreaded all-out test day at all. FTP is defined as roughly 60-minute power [Allen et al. 2019]; the 20-minute and 8-minute tests are just two field routes to that estimate, each with its own multiplier and its own error. As the broader guide on estimating FTP without a test argues, the data your power meter has already collected is usually enough to model threshold from the power you actually produced. A power-duration model fit to months of riding draws on dozens of efforts of varying length instead of one anxious afternoon. AdaptCycling reads whatever efforts you did — a 20-minute climb, a hard 8-minute segment, a 12-minute time trial — from your ride history and folds them into a Critical Power fit, rather than forcing one protocol on you. If you happen to run a clean 8-minute or 20-minute test, that data simply enters the same curve; nothing is wasted and nothing is mandatory. The estimate refines itself as new rides arrive, and the free tier turns it into a real training plan. A field test still earns its place as a periodic cross-check. When the modeled FTP and a clean field effort disagree by more than 5%, one of them is wrong, and the field test usually breaks the tie. Do that roughly twice a year rather than every six weeks, and whichever protocol you choose — 20-minute or 8-minute — keep it constant so the number reflects your fitness and not your method. ## Common questions **Do I take the average or the higher of my two 8-minute efforts?** CTS's own protocol sets training intensities from the average power of the higher of the two 8-minute efforts, then takes 90% of it [Carmichael & Rutberg 2017]. Some coaches average both efforts instead. The 10% reduction is the same either way; what matters is picking one convention and applying it consistently every time you test. **Why is the 8-minute multiplier 90% when the 20-minute is 95%?** Because a shorter maximal effort sits higher above your threshold and carries a larger anaerobic contribution, so more power has to be subtracted to estimate sustainable power [Karsten et al. 2021]. An 8-minute effort runs near VO2max power; a 20-minute effort is much closer to the hour of power that FTP represents, so it needs only a 5% haircut. **Which test gives a more accurate FTP?** Neither is ground truth. At the group level the 20-minute test tracks maximal lactate steady state with about 1.4% bias [Borszcz et al. 2019], but for an individual the 20-minute estimate can miss a true 60-minute effort by 40 to 60 watts [Borszcz et al. 2018]. Accuracy depends more on your pacing and consistency than on which protocol you pick. **Can I compare an 8-minute result directly to a 20-minute result?** Not directly. The two protocols and their different multipliers produce different numbers even at identical fitness, and switching can shift FTP by several percent from method alone. Do not convert between them — if you change protocols, treat it as a fresh baseline rather than a comparison against the old number. **Do I even need to do a test at all?** If you ride with power, probably not for a starting number. A model fit to your Strava power curve can estimate FTP within roughly 3 to 5% without any forced effort. Keep a field test — 20-minute or 8-minute — for the twice-a-year sanity check against that modeled number, not as the only way to get one. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Carmichael & Rutberg 2017.** [The Time-Crunched Cyclist: Race-Winning Fitness in 6 Hours a Week (3rd ed.)](https://www.simonandschuster.com/books/The-Time-Crunched-Cyclist-Race-Winning-Fitness-in-6-Hours-a-Week-3rd-ed/Chris-Carmichael/The-Time-Crunched-Athlete/9781937715502). VeloPress. 3. **Borszcz et al. 2018.** [Functional Threshold Power in Cyclists: Validity of the Concept and Physiological Responses](https://pubmed.ncbi.nlm.nih.gov/29801189/). International Journal of Sports Medicine. 4. **Tramontin et al. 2022.** [Functional Threshold Power Estimated from a 20-minute Time-trial Test is Warm-up-dependent](https://pubmed.ncbi.nlm.nih.gov/34749416/). International Journal of Sports Medicine. 5. **Borszcz et al. 2019.** [Is the Functional Threshold Power Interchangeable With the Maximal Lactate Steady State in Trained Cyclists?](https://pubmed.ncbi.nlm.nih.gov/30676826/). International Journal of Sports Physiology and Performance. 6. **Karsten et al. 2021.** [Relationship Between the Critical Power Test and a 20-min Functional Threshold Power Test in Cycling](https://pmc.ncbi.nlm.nih.gov/articles/PMC7862708/). Frontiers in Physiology. 7. **Vinetti et al. 2023.** [Functional Threshold Power Field Test Exceeds Laboratory Performance in Junior Road Cyclists](https://pmc.ncbi.nlm.nih.gov/articles/PMC10448799/). Journal of Strength and Conditioning Research. --- # Why cycling apps show you different FTP numbers URL: https://www.adaptcycling.com/guides/why-apps-show-different-ftp Updated: 2026-07-17 Author: Jim Camut Connect the same rider to five apps and you can see five different FTPs: Strava says 250, Intervals.icu 262, TrainingPeaks 244, Xert 256, Garmin 238. Nothing is broken. Each app runs a different model on a different slice of your data, feeds it from a different power source, and reports the result with false confidence. Zwift belongs on the same list — it estimates your FTP from your power data whenever you have not tested. Here is exactly what each one computes, why the numbers diverge by 5-10%, and which estimate to trust for which decision. ## The short answer: each app runs a different model on a different slice of your data Five apps can report FTPs 10-25 watts apart on the same rider and none is malfunctioning. The spread comes from three independent choices each app makes: which model it fits, which window of your history it reads, and which power file it trusts. A 5-10% gap between the highest and lowest estimate is normal, not a bug. The apparent contradiction dissolves once you see that FTP is an estimate, not a measurement. All five apps are trying to pin the same physiological boundary — the wattage between the heavy and severe intensity domains, near the power you could hold for roughly an hour [Allen et al. 2019]. But that boundary is never observed directly; it is inferred. Two estimators handed the same rides will disagree the same way two analysts handed the same spreadsheet disagree — the raw data is identical, the assumptions are not. The size of the disagreement is predictable. For a rider whose true one-hour power is around 250 watts, the modeled estimates typically land within a 20-25 watt band — a Critical Power fit near the top, a conservative 20-minute-peak method near the bottom, and a heart-rate estimate as the wild card. That is 8-10% between the extremes, wide enough to move a sweet-spot target from a productive 218 watts to a soul-crushing 235. ## What each app actually computes Strava reads your best-efforts power curve and applies the 20-minute-minus-5% convention [Strava docs]. TrainingPeaks suggests 95% of a new best 20-minute effort [Allen et al. 2019]. Intervals.icu fits a Critical Power model to your 90-day curve and calls it eFTP [Intervals.icu docs]. Xert builds a three-parameter Fitness Signature from breakthrough rides [Xert docs]. Garmin combines a power meter with heart rate and HRV [Garmin docs]. Strava and TrainingPeaks share DNA. Strava builds a best-efforts power curve — your top average power from one second to the full ride length — and derives an estimate anchored to the 20-minute-times-0.95 convention it documents [Strava docs], smoothed across your recent bests. TrainingPeaks is more literal: it watches for a 20-minute effort that beats your previous best and proposes 95% of it as the new FTP [Allen et al. 2019]. Both are conservative by design, and both go stale the moment you stop producing fresh 20-minute peaks — a rider deep in base can hold a months-old number that no longer reflects fitness. Intervals.icu takes the modeling route. It fits a Critical Power model to your mean-maximal-power curve and reports the result as eFTP, using your efforts over a rolling 90-day window; a single long-enough maximal effort on one ride can set a new eFTP by itself [Intervals.icu docs]. Because a CP fit uses the whole curve rather than one 20-minute point, it tends to read a few watts higher and update more often than the peak-detection methods. Xert models you as a three-parameter Fitness Signature — Threshold Power, High Intensity Energy, and Peak Power — extracted from your hardest efforts and revised whenever you score a breakthrough ride the current signature cannot explain [Xert docs]. The strength is granularity; the weakness is that the signature needs breakthroughs to stay current, so a rider grinding easy zone-2 weeks watches the Threshold Power figure quietly decay. Garmin comes at it from the opposite side. Its Firstbeat-derived estimate needs a heart-rate strap and a connected power meter, plus a stable VO2max estimate, and infers threshold from the relationship between heart rate, HRV and power rather than from the power curve alone [Garmin docs]. Ramp-based apps are different again: TrainerRoad's ramp test sets FTP at 75% of your peak one-minute power, a fixed convention rather than a rider-specific ratio. Five methods, five arithmetic paths to the same target. ## Why the numbers disagree — three structural reasons The estimates disagree for three structural reasons. Different models: a Critical Power curve fit sits a few percent above a 20-minute-peak method [McGrath et al. 2021]. Different input windows: a single recent best versus a 90-day rolling curve versus live breakthroughs. Different power sources: the same legs logged through two meters can differ 5-10%. The model choice sets the systematic bias. Critical Power is the asymptote of your power-duration curve, and in trained cyclists it runs measurably above 60-minute power — McGrath and colleagues showed CP and FTP are not interchangeable, with CP the higher of the two [McGrath et al. 2021]. So a CP-based eFTP will structurally read above a TrainingPeaks suggestion built from 95% of a 20-minute effort, even on identical data. Neither is wrong; they are reporting two different points on the same curve [Poole et al. 2016]. The input window sets the lag. TrainingPeaks and Strava wait for a fresh peak effort, so their numbers can sit unchanged for two or three months if your training does not produce one. Intervals.icu's 90-day rolling curve updates as older efforts age out. Xert moves only on breakthroughs, and Garmin needs a spread of heart-rate data across intensities before it will commit. On any given Tuesday the five apps are describing five different time slices of you — last spring's peak, the last 90 days, or last week's intervals. The power source sets the noise floor. Feed one app your outdoor crank meter and another your indoor smart trainer and the same effort can differ 5-10% before any model runs, because the two devices are calibrated differently and read different parts of the drivetrain. All five apps are, in the end, attempts at the same problem our ftp-without-a-test pillar lays out — estimating threshold from data you already have, without burning a Saturday on a forced 60-minute effort. They diverge not because the physiology is unsettled but because there are many defensible ways to infer one number from a messy power file. ## Which number to trust — and where AdaptCycling lands Trust a Critical Power estimate — Intervals.icu eFTP or AdaptCycling — for day-to-day training zones, because it uses the most data per estimate and lags least [Poole et al. 2016]. Treat TrainingPeaks' threshold suggestion and Strava as conservative confirmations that update only when you go genuinely faster. Discount Garmin's estimate for prescribing intervals. Match the number to the decision. For setting the zones you ride to every day, the CP-curve estimate is the best default — it is anchored by dozens of efforts rather than one, so a single flat day cannot drag it and a single hero effort cannot inflate it. For a deliberately cautious number that only moves when you set a real personal best, TrainingPeaks' threshold suggestion is honest and useful. Garmin's estimate is fine as a trend line but too noisy to set a 4x8-minute threshold session against. When two estimates disagree by more than about 5%, one of them is wrong, and a clean 20-minute or 8-minute field effort is the tiebreaker [Allen et al. 2019]. If your eFTP reads 262 and your TrainingPeaks number reads 244, the honest move is to test once and see which the curve was closer to — usually the CP fit, if you have a recent long effort feeding it, but not always. AdaptCycling's estimate is closest to Intervals.icu's eFTP by design: we fit a Critical Power model to the mean-maximal-power curve we read from your ride history, no forced test, updated as new rides land, accurate to within roughly 3-5% for a rider with 90 days of varied data [Poole et al. 2016]. The difference is what we do with it. Intervals.icu shows you the number; we feed it straight into an adaptive plan, so when the estimate moves, your prescribed watts move with it rather than waiting for you to notice and retype a value into a settings box. ## Common questions **Which FTP number is the real one?** None of them — they are all estimates of a boundary no app measures directly. As a default, trust the Critical Power / power-curve estimate (Intervals.icu eFTP or AdaptCycling), because it is built from dozens of efforts rather than one 20-minute point. When two apps disagree by more than 5%, do a single clean 20-minute effort and let the field test break the tie [Allen et al. 2019]. **Why is my Intervals.icu eFTP higher than my TrainingPeaks FTP?** Because they compute different points on the same curve. eFTP is derived from a Critical Power fit, and CP sits a few watts above 60-minute power in trained riders [McGrath et al. 2021]. TrainingPeaks suggests 95% of your best 20-minute effort, a deliberately conservative haircut [Allen et al. 2019]. A 10-15 watt gap between the two is expected, not an error. **Why does Garmin show a different FTP than Strava or my trainer app?** Garmin's estimate comes from heart rate and HRV rather than the power curve, and it needs a heart-rate strap, a connected power meter, and a stable VO2max before it will commit [Garmin docs]. That makes it slower to update and noisier than a power-only estimate, so it commonly reads 10-20 watts off the power-curve apps. Use it as a trend, not a training target. **Should I set the same FTP in every app?** Practically, yes — pick one authoritative estimate and enter it manually everywhere else so your zones, TSS, and training load stay comparable across apps. If you let each app run its own auto-detection, one good week can show up as three different fitness stories. One number, entered everywhere, keeps the accounting honest. **Does a different FTP in each app mean one is broken?** No. A 5-10% spread across apps is the normal consequence of different models, input windows, and power sources — not a malfunction. Worry only when a single app's number jumps more than about 5% overnight with no breakthrough ride behind it, which usually points at a power-meter calibration or data-dropout problem rather than real fitness change. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Poole et al. 2016.** [Critical Power: An Important Fatigue Threshold in Exercise Physiology](https://pmc.ncbi.nlm.nih.gov/articles/PMC5070974/). Medicine & Science in Sports & Exercise. 3. **McGrath et al. 2021.** [Do Critical and Functional Threshold Powers Equate in Highly-Trained Athletes?](https://pubmed.ncbi.nlm.nih.gov/34055164/). International Journal of Exercise Science. 4. **Strava docs.** [Power — Strava Help Center](https://support.strava.com/en-us/articles/15402161-power). Strava Help Center. 5. **Intervals.icu docs.** [Power model now complete — eFTP and W'](https://forum.intervals.icu/t/power-model-now-complete/1680). Intervals.icu. 6. **Xert docs.** [Fitness Signature — Xert Breakthrough Training](https://www.baronbiosys.com/glossary/fitness-signature/). Xert / Baron Biosystems. 7. **Garmin docs.** [Getting Your FTP Estimate — Performance Measurements](https://www8.garmin.com/manuals/webhelp/GUID-EECCAC99-90D6-4AB1-9A3A-EC433D3365E2/EN-US/GUID-F5BF67CE-C94E-4842-AE96-A7A05C85B732.html). Garmin. --- # Why your ramp test FTP is too high (and how to correct it) URL: https://www.adaptcycling.com/guides/ramp-test-ftp-too-high Updated: 2026-07-17 Author: Jim Camut Your ramp test handed you an FTP, you built your intervals around it, and now every threshold and sweet-spot session falls apart before the last rep. That is not a fitness failure — it is a measurement error you can name. The ramp test estimates FTP as 75% of your peak one-minute power [Allen et al. 2019], and that single ratio can flatter anaerobically strong riders by 8 to 15 watts. This is the mechanism, the diagnostic that confirms it, and four ways to correct it. ## Why the ramp test overshoots your real threshold The ramp test never measures your hour power directly. It measures the peak one-minute power you reach as the load climbs, then multiplies by 0.75 [Allen et al. 2019]. Because that final minute is partly fueled by a finite anaerobic reserve, riders with a big reserve buy a high peak — and inherit an inflated FTP. The power-duration model explains the leak. Any effort above your critical power draws down W', a fixed quantity of work — measured in kilojoules — that you can spend above threshold before you fail [Burnley & Jones 2018, Vanhatalo et al. 2007]. The last 60 seconds of a ramp sit far above critical power, so they are paid for almost entirely out of W'. A rider with a 25 kJ anaerobic reserve reaches a higher final minute than a rider with 15 kJ at the identical aerobic ceiling, even though both hold the same sustainable power. The ramp cannot tell those two riders apart — it reads only the peak, and the peak is inflated by anaerobic capacity. This is why peak power output and sustainable power are different physiological measures, not two views of one number. Bentley and colleagues [Bentley et al. 2001] found that the peak workload from an incremental test and the power a cyclist actually held over a 90-minute time trial were correlated but distinct, tracking separate determinants across their nine well-trained subjects. The ramp collapses that distinction into a single 75% multiplier. For most of the field the collapse is harmless, because most riders cluster near the average ratio — but the further your anaerobic capacity sits from the mean, the more the multiplier lies. Worse, the peak minute is not even a stable number. Michalik and colleagues [Michalik et al. 2019] ran the same men through two ramp rates and found peak power was 51.69 watts higher — roughly 15% — on the faster ramp, because a steeper climb reaches exhaustion before the aerobic system is fully taxed. TrainerRoad's ramp and Zwift's FTP Ramp Test use fixed step sizes, so your result is partly an artifact of the protocol you happened to ride. The ramp is honest as a starting estimate; the error enters when an app treats that estimate as ground truth and prescribes threshold work against it. ## The rider types the ramp flatters (and the ones it shortchanges) The 75% ratio is a fixed protocol convention applied to every rider, not a measured per-rider constant. The true ratio of threshold to peak one-minute power varies with anaerobic capacity. Punchy sprinters and criterium riders sit below it and get inflated FTPs; aerobic diesels sit above it and get numbers that undersell them. If you sprint well, close crits from a small group, or win the last-100-meters drag race, you are the rider the ramp flatters most. Your high maximal aerobic power and large anaerobic reserve inflate the final minute, and 75% of an inflated peak overstates your true threshold — often by 8 to 15 watts on a 250-watt rider, roughly a 3 to 6% error. That is exactly enough to turn a prescribed 4x10 at 'sweet spot' into an over-threshold session you cannot finish. The mirror image is the aerobic diesel — the rider who time-trials well but gets shelled the instant the pace turns explosive. Your anaerobic reserve is modest, so your ramp peak is comparatively low, and 75% of a low peak understates a threshold you can genuinely hold. Karsten and colleagues [Borszcz et al. 2019] showed FTP sits close to maximal lactate steady state, the aerobic, sustainable boundary — precisely the quality a diesel is built around and the ramp fails to reward. These riders are frequently stronger at threshold than their ramp FTP claims, and correcting upward is as real a fix as correcting down. The scatter is not small. Borszcz and colleagues [Borszcz et al. 2018] reported that on an individual basis the agreement between a short FTP estimate and a genuine sustained effort spanned 40 to 60 watts across their cyclists, even though the group averages matched almost perfectly. A number that is accurate for the population mean and 40 watts off for you is worthless for prescribing your Tuesday intervals. The ramp's convenience is real; its per-rider precision is the part the marketing quietly omits. ## The tell: workouts that keep blowing up at your ramp FTP You do not need a lab to confirm an inflated FTP — your training log already holds the evidence. If threshold and sweet-spot sessions fail repeatedly at the prescribed watts while your easy rides feel normal, the FTP is too high, not your discipline. A pattern beats any single bad day. The clean signal is a repeated failure at one specific intensity. Two or three sweet-spot workouts — say 3x15 at 88 to 94% of FTP — that you abandon in the final interval, week after week, point at the anchor, not the legs. One failed session is noise: poor sleep, heat, under-fueling. Three failures at the same relative intensity with adequate recovery between them is signal. The giveaway is where the failure lands — inflated-FTP riders die in the last third of threshold work, when the aerobic system is finally asked to hold a wattage it never actually owned. Notice where the problem does not appear. Zone 2 and endurance rides feel entirely normal, because they sit well below the inflated number and the error never bites. Short VO2max intervals may also feel fine — 30-second and 2-minute efforts are paid for by the same anaerobic reserve that inflated your ramp in the first place. The collapse is specific to the 88 to 105% band, the sustained region where FTP is supposed to live [Borszcz et al. 2019]. If your log shows strength up top, comfort down low, and failure in the middle, that hourglass shape is the fingerprint of an FTP set too high. ## How to correct an inflated ramp FTP Four fixes, cheapest first: shave the multiplier to fit your rider type, cross-check with one sustained effort, re-anchor on a curve that pools many efforts instead of one, or let software fit your whole power history. Each moves you from a population guess toward your own sustainable power. The fastest correction costs nothing. If you are a proven sprinter or crit rider, drop the ramp multiplier from 75% toward 72% and re-run your workouts against the lower number — on a 250-watt ramp FTP that is a roughly 8-watt haircut, usually enough to make sweet-spot work completable again. Then cross-check with one clean sustained effort: a well-paced 20-minute test, or better, a genuine 30-to-40-minute solo time trial gives you real sustained data the ramp never collected [Allen et al. 2019]. Where the sustained effort and the ramp disagree by more than 5%, trust the longer effort. The deeper fix is to stop relying on any single test. The power-duration or critical-power model fits your threshold from several maximal efforts of different lengths — a 5-minute climb, a 12-minute TT, a 20-minute segment — so no one anaerobically-loaded minute can dominate the estimate [Vanhatalo et al. 2007, Burnley & Jones 2018]. This is the frame of our pillar on estimating FTP without a forced test: your threshold already lives in the data you have ridden, and the ramp is just one flawed test the curve-based approach sidesteps. A curve fit to dozens of real efforts is far harder to fool than 75% of a single peak minute. This is what AdaptCycling does with your ride history. We build your mean-maximal-power curve from every ride, fit a critical-power model, and derive FTP from your sustained efforts rather than one ramp peak — so an anaerobically gifted rider is not punished with threshold targets built on a single minute of W' spend [Bentley et al. 2001]. Because the estimate is anchored in the power you actually hold for 10, 20, and 40 minutes, it corrects the individual bias the ramp bakes in, and it re-derives every time a new hard effort lands. The plan then adapts to that corrected number instead of asking you to survive an inflated one. ## Common questions **How much can a ramp test overestimate my FTP?** For anaerobically strong riders, commonly 8 to 15 watts — roughly 3 to 6% on a 250-watt threshold — because the ramp reads 75% of an anaerobically inflated peak minute [Allen et al. 2019]. On an individual basis the gap between a short estimate and a real sustained effort has been measured as wide as 40 to 60 watts [Borszcz et al. 2018]. The direction depends on your physiology: sprinters skew high, diesels skew low. **Should I just lower my ramp FTP by a fixed amount?** Only as a stopgap. Shaving the 75% multiplier toward 72% helps a known sprinter, but a fixed haircut is still a guess. The reliable fix is to cross-check against one clean sustained effort — a paced 20-minute or 30-minute time trial — and, better, to anchor FTP on a power-duration curve that pools many efforts instead of one peak minute [Vanhatalo et al. 2007]. **Is TrainerRoad's ramp test just wrong?** No — it is a convenient starting estimate, and TrainerRoad is upfront that it is one. The ramp is the most-used indoor FTP protocol precisely because it is short and repeatable. The error appears when the number is treated as ground truth and used to prescribe threshold work, because 75% of peak one-minute power carries a per-rider bias of several percent [Bentley et al. 2001]. Use it as a benchmark, then verify it against sustained data. **Why do my VO2max intervals feel fine but threshold work blows up?** Because they draw on different systems. Short VO2max efforts are paid for largely by the same anaerobic reserve, W', that inflated your ramp peak, so they feel manageable [Burnley & Jones 2018]. Sustained threshold work asks your aerobic system to hold a wattage near maximal lactate steady state [Borszcz et al. 2019] — and if the FTP is set above that, it collapses in the final third. Strength up top with failure in the middle is the classic inflated-FTP signature. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Bentley et al. 2001.** [Peak power output, the lactate threshold, and time trial performance in cyclists](https://pubmed.ncbi.nlm.nih.gov/11740302/). Medicine & Science in Sports & Exercise. 3. **Michalik et al. 2019.** [Influence of Intensity RAMP Incremental Test on Peak Power, Post-Exercise Blood Lactate, and Heart Rate Recovery in Males: Cross-Over Study](https://pmc.ncbi.nlm.nih.gov/articles/PMC6843469/). International Journal of Environmental Research and Public Health. 4. **Borszcz et al. 2019.** [Is the Functional Threshold Power Interchangeable With the Maximal Lactate Steady State in Trained Cyclists?](https://pubmed.ncbi.nlm.nih.gov/30676826/). International Journal of Sports Physiology and Performance. 5. **Borszcz et al. 2018.** [Functional Threshold Power in Cyclists: Validity of the Concept and Physiological Responses](https://pubmed.ncbi.nlm.nih.gov/29801189/). International Journal of Sports Medicine. 6. **Vanhatalo et al. 2007.** [Determination of Critical Power Using a 3-min All-out Cycling Test](https://pubmed.ncbi.nlm.nih.gov/17473782/). Medicine & Science in Sports & Exercise. 7. **Burnley & Jones 2018.** [Power-duration relationship: Physiology, fatigue, and the limits of human performance](https://pubmed.ncbi.nlm.nih.gov/27806677/). European Journal of Sport Science. --- # How often should you test your FTP? Twice a year, not every six weeks URL: https://www.adaptcycling.com/guides/how-often-to-test-ftp Updated: 2026-07-17 Author: Jim Camut Most training apps nag you to retest FTP every four to six weeks, and Zwift recommends one before you start any training plan. For a formal field test, that cadence is roughly three times too frequent. The honest answer: run an actual 20-minute or ramp test about twice a year — once at the end of base before the build phase, and once as a sanity check two weeks out from a goal event [Allen et al. 2019]. Test off that schedule only when something specific invalidates your data. Here is the decision framework, with numbers. ## The honest cadence: about twice a year for a formal test For a dedicated field test, twice a season is enough: end of base before the build phase, and a pre-goal check about two weeks out [Allen et al. 2019]. Threshold fitness moves over weeks to months, not days, and a test's own error can exceed the change you are trying to detect in six weeks. The two-test logic is about anchoring, not measuring for its own sake. The end-of-base test resets your training zones before the hardest block of the year, when riding at the wrong intensity is most expensive; a 10-watt error on a 250-watt rider drags every threshold interval off target. The pre-goal test, run 10 to 14 days out [Allen et al. 2019], confirms the number your taper is built on. Two well-placed tests do more for a season than eight scattered ones. Testing more often mostly measures noise. Borszcz and colleagues [Borszcz et al. 2018] found that the individual limits of agreement between a 20-minute FTP estimate and a true 60-minute effort spanned 40 to 60 watts. A realistic six-week block might raise threshold by 2 to 5 percent — 5 to 12 watts for that same 250-watt rider. When the test's own error band is wider than the fitness change you are chasing, a monthly retest is measuring your warm-up and pacing, not your engine. There is also a recovery cost. A maximal 20-minute or ramp test is a genuinely hard session that needs a day of freshness going in and 24 to 48 hours to clear afterward. On a six-week cadence that is roughly eight tests a year, each one displacing a quality workout and each one vulnerable to a false low — Tramontin and colleagues [Tramontin et al. 2022] showed that warm-up structure alone shifts the resulting FTP by clinically meaningful amounts. Test under-recovered and you stamp a deflated number onto the next block. ## Why apps push a four-to-six-week retest The four-to-six-week cadence is an artifact of software design, not physiology. Static plan generators consume FTP as one fixed input, so the only way to keep your zones honest is to re-measure on a timer that matches a training block. The cadence serves the tool's need for a fresh number, and pushes the cost onto you. Consider what a conventional plan generator actually needs. It multiplies a single FTP value by fixed percentages to set every zone and every workout target. If that number drifts as you get fitter, the whole plan quietly goes stale, so the software schedules a retest every mesocycle — typically four to six weeks — to refresh the input. The logic is sound for the tool. It just assumes your only source of truth is a dedicated test. The cost lands on the athlete. At a six-week cadence you run roughly eight tests a year; each is a hard day you did not spend training, plus the taper and recovery around it. Worse, any test caught on an under-recovered day reads low, and a number that is falsely low by 15 watts then poisons every zone for the following block — your sweet-spot work drifts toward tempo, and the plan silently under-doses you for weeks. ## When to test off-cadence: the three real triggers Ignore the calendar and test when a specific event invalidates your data. Three triggers justify an off-schedule effort: a layoff long enough to start detraining, a suspected real change larger than about 5 percent, or new equipment. Each one breaks the assumption that your recent power history still reflects who you are today. A layoff is the clearest trigger, because detraining is fast at first. Coyle and colleagues [Coyle et al. 1984] tracked highly trained subjects after they stopped: VO2max fell 7 percent in the first 21 days and settled 16 percent below trained values by 56 days, with muscle oxidative enzyme activity falling on a roughly 12-day half-life. Mujika and Padilla [Mujika & Padilla 2000 Part I] attribute the earliest losses — inside the first two to four weeks — largely to a drop in blood volume. Practically: after about two weeks fully off, your zones have started to drift, and that is the point to consider a re-anchor. Longer breaks reset the curve outright. Reviewing stoppage beyond four weeks [Mujika & Padilla 2000 Part II], Mujika and Padilla report that recently acquired VO2max gains are lost completely and lactate threshold falls, even though a long-term athlete stays above untrained values. After six weeks or more off — illness, injury, a life crisis — your old FTP is fiction, and any model built on those stale rides is anchored to a rider who no longer exists. Retest before the first hard block back, not three weeks into it. The second trigger is a suspected real jump. A stable FTP should not move more than 1 to 2 percent on the strength of any single ride, so if a benchmark climb suddenly comes in 5 percent higher, or a continuously modeled estimate steps up by more than 5 percent, it is worth a clean confirmation effort. A change that large is either genuine adaptation or a data artifact — a power spike, a mis-set weight, a tailwind PR — and a controlled test is the tiebreaker. The third is new equipment. Power meters routinely disagree by 3 to 5 percent, so a new meter, a new trainer, or a swap between crank- and pedal-based measurement makes your entire historical power curve non-comparable overnight. A wattage that was FTP last week may read 10 watts high or low on the new device. The first week or two on new hardware needs a benchmark effort to rescale everything that follows; without it, every zone inherits the offset. ## The modern reframe: cross-check, not re-measure Continuous modeling changes the question entirely. When FTP is estimated from your power curve on every ride, the number updates as a background process, so cadence stops being how often you re-measure and becomes how often you cross-check the model against a real effort. For most riders that is one or two efforts a year, not eight. This is the practical payoff of the idea at the center of the parent guide — that FTP can be estimated from the ride data you already have, with no forced test. Once the estimate refreshes continuously, a scheduled monthly retest is redundant by construction: the model has already moved. The only thing a dedicated effort still buys you is verification, so the whole question collapses from 'how often should I re-measure' to 'how often should I cross-check.' The honest cross-check cadence is the same twice-a-year rhythm from the top of this article — end of base and pre-goal — except now you are confirming a number rather than generating one. AdaptCycling works this way. We read your ride history, fit a Critical Power model to your mean-maximal-power curve, and re-derive FTP after every ride, applying a small per-rider offset because Critical Power runs measurably above 60-minute power — McGrath and colleagues [McGrath et al. 2021] measured a 16-watt gap (282 versus 266 watts) in highly trained cyclists. Because the number stays current, we ask for a real test only at the two moments it earns its keep: re-anchoring after a long layoff, when the curve is built on stale rides, and a pre-goal sanity check. Cross-check, don't re-measure. ## Common questions **How often should I actually test my FTP?** For a dedicated field test, about twice a year: once at the end of your base phase before the build block, and once as a sanity check 10 to 14 days before a goal event [Allen et al. 2019]. Test outside that rhythm only when a layoff, a suspected change over 5 percent, or new equipment invalidates your recent data. The every-four-to-six-weeks cadence most apps push is roughly three times more often than the physiology requires. **Why does my app tell me to retest every four to six weeks?** Because a static plan generator uses FTP as a single fixed input and has no other way to notice you have gotten fitter — it must re-measure on a timer to keep your zones current, and four to six weeks matches a typical training block. The cadence serves the software, not you. Each test costs a hard day plus recovery, roughly eight a year, and any test caught under-recovered stamps a low number onto the next block [Tramontin et al. 2022]. **How long off the bike before I need to retest?** About two weeks fully off is when zones start to drift — VO2max drops around 7 percent in the first three weeks of no training [Coyle et al. 1984], with the earliest losses driven by falling blood volume [Mujika & Padilla 2000 Part I]. After six weeks or more off, recently gained fitness is largely gone [Mujika & Padilla 2000 Part II] and your old FTP is fiction; retest before the first hard block back rather than training weeks at the wrong intensity. **My FTP jumped 15 watts after one hard ride — is that real?** Probably not from one ride. A stable FTP should not move more than 1 to 2 percent on the strength of a single effort, so a sudden 15-watt jump — roughly 6 percent on a 250-watt rider — is more often a data artifact than an overnight adaptation: a power spike, a tailwind, a mis-entered weight. Treat it as a flag to cross-check with a clean effort, not as a new number to train against immediately. **If my FTP updates automatically, do I still need to test at all?** Only to cross-check, and only once or twice a year. A continuously modeled FTP already tracks real change within days, so a dedicated test stops being a measurement and becomes verification. The two moments still worth a real effort are re-anchoring after a long layoff, when the model is fit to stale rides, and a pre-goal check so you taper against a number you have confirmed [Allen et al. 2019]. ## References 1. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. 2. **Borszcz et al. 2018.** [Functional Threshold Power in Cyclists: Validity of the Concept and Physiological Responses](https://pubmed.ncbi.nlm.nih.gov/29801189/). International Journal of Sports Medicine. 3. **Tramontin et al. 2022.** [Functional Threshold Power Estimated from a 20-minute Time-trial Test is Warm-up-dependent](https://pubmed.ncbi.nlm.nih.gov/34749416/). International Journal of Sports Medicine. 4. **Mujika & Padilla 2000 Part I.** [Detraining: Loss of Training-Induced Physiological and Performance Adaptations. Part I](https://link.springer.com/article/10.2165/00007256-200030020-00002). Sports Medicine. 5. **Mujika & Padilla 2000 Part II.** [Detraining: Loss of Training-Induced Physiological and Performance Adaptations. Part II](https://pubmed.ncbi.nlm.nih.gov/10999420/). Sports Medicine. 6. **Coyle et al. 1984.** [Time Course of Loss of Adaptations After Stopping Prolonged Intense Endurance Training](https://pubmed.ncbi.nlm.nih.gov/6511559/). Journal of Applied Physiology. 7. **McGrath et al. 2021.** [Do Critical and Functional Threshold Powers Equate in Highly-Trained Athletes?](https://pubmed.ncbi.nlm.nih.gov/34055164/). International Journal of Exercise Science. --- # How long should your taper be? A cyclist's guide to what the evidence actually says URL: https://www.adaptcycling.com/guides/how-long-should-your-taper-be Updated: 2026-07-21 Author: Jim Camut Two weeks, cut volume 40-60 percent. That number is real, it comes from a 27-study meta-analysis, and it is the best available place to start. It is also a population average across mixed sports, and no study has ever asked whether your own optimum sits somewhere else. I raced in Europe and North America and tapered by feel and by whatever the directeur said. Here is what the literature can tell you about taper length, and — more usefully — what it cannot. ## The two-week number is a population mean, not a prescription Bosquet's meta-analysis screened 182 studies and included 27. Its headline — two weeks, volume cut 41-60 percent — is a weighted average of effect sizes across mixed sports, dominated by swimming and running. It tells you where to start. It does not tell you where your own optimum sits, and it was never built to. The numbers deserve precision. Bosquet and colleagues found the largest aggregated effects for a two-week duration (effect size 0.59), an exponential volume reduction of 41-60 percent (0.72, the biggest single effect in the paper), and leaving intensity (0.33) and frequency (0.35) unmodified — all significant at P below 0.001 [Bosquet et al. 2007]. It is the most-cited result in tapering, and a good one. The methods matter as much as the results. Effect sizes were calculated as pre-post taper standardized mean differences, weighted by within-group heterogeneity [Bosquet et al. 2007]. Pre-post means each athlete was compared against themselves before the taper, not against a matched group that did not taper. Most included studies had no non-tapering control arm, because withholding a taper from a competitive athlete before a real race is a hard study to run. So the pooled effect carries whatever regression to the mean and pre-competition motivation contribute. Tapering works; the number is softer than a lone effect size implies. Bosquet did not publish moderators either. His independent variables were taper components only — nothing on training age, chronological age, sex, or event duration. A few later studies stratify taper response by sex or career stage, but none of that reaches the two-week number in circulation. Cyclist-specific evidence is thinner still: the best-known randomized comparison ran 11 male cyclists through three fixed seven-day protocols, five or six riders per arm [Neary et al. 2003]. ## Four sources, four different volume answers Ask four respected sources how much volume to cut and you get 50 percent, 41-60 percent, 60-90 percent, and 32 percent. Ask how long and you get seven days, fourteen, four-to-more-than-twenty-eight, and thirty-five. Volume and duration are not settled parameters. Intensity and frequency are the only components where every source agrees. Neary's cyclists responded best to a 50 percent cut over seven days [Neary et al. 2003]. Bosquet's meta-analysis converges on 41-60 percent over fourteen [Bosquet et al. 2007]. Mujika and Padilla's review — a narrative review, not a meta-analysis, with no confidence intervals attached to any of its numbers — recommends reducing volume by up to 60-90 percent across a window of four to more than twenty-eight days [Mujika & Padilla 2003]. A computer simulation put the optimal linear taper at a 32 percent reduction over 35 days in non-athletes and 49 percent over 33 days in athletes [Thomas et al. 2009]. The top of one recommendation is nearly three times the bottom of another, and the durations span a factor of five. Two things survive every source. Keep intensity at pre-taper values: Bosquet found no modification optimal [Bosquet et al. 2007], Mujika and Padilla agree [Mujika & Padilla 2003], and Neary's cyclists held 85 percent of VO2max throughout [Neary et al. 2003]. And do not cut frequency by more than roughly 20 percent [Mujika & Padilla 2003]. ## Has anyone tested a personalized taper? No — and there may be a reason nobody tried No prospective study in any endurance sport has compared an individualized taper head-to-head against a population-standard one. That is absence of evidence, not evidence of absence: nobody has shown personalization fails, because nobody has run the trial. The machinery to compute a personal taper has existed for decades. Its inputs are the problem. This is the claim I most want you to leave with, so I will state it carefully. There is no controlled prospective trial, in cycling or any endurance sport, in which one group tapered by an individually derived prescription and another followed a population protocol. No study quantifies between-athlete variance in optimal taper length from measured outcomes either. That is no evidence of effect. It is not evidence of no effect. Individualized tapering has not been shown to fail; it has not been tested. Those two statements get collapsed constantly, including by people selling software. The tools are not missing. Thomas, Mujika and Busso derived non-linear model parameters for eight elite swimmers from two full seasons of each athlete's own data, then simulated the taper that would maximize each one's result [Thomas et al. 2008]. Optimal step tapers landed at 22.4 days after an overload block and 16.4 days without one — with standard deviations of 13.4 and 10.3 days around those means. If that spread is real, the between-athlete range is enormous. It is also entirely simulated. There was no prospective arm. A likely reason simulation never became a trial is identifiability. Hellard and colleagues fitted the same class of model to a season of data from nine elite swimmers, then bootstrapped the parameter estimates [Hellard et al. 2006]. The model tracked performance well. The parameters did not. The 95 percent confidence interval on the fatigue decay time constant ran 6 to 32 days; on time to peak performance after training stops — the model's own answer to taper length — 25 to 61 days. Some parameters were so correlated that the authors call interpreting them worthless, and conclude that using them to build individual training schedules from observational data is hazardous. Put those papers side by side and an explanation falls out, though no single paper states it this way. A personalized taper is computable from your own history. The computation rests on parameters that a full season of high-quality elite data cannot pin down better than a multi-week confidence interval. You cannot run a convincing trial of individualized tapering when the individualization is that unstable — you would be randomizing athletes to noise. That is my reading of why individualized tapering never got its trial. Treat it as an argument, not a citation. ## What a self-coached rider should actually do Run the population protocol, because it is the best-supported starting point available. Then record what you did and how you raced, in enough detail to compare across seasons. Your own n-of-1 record is genuinely the only taper evidence that exists about you, and nobody else is going to build it. Start with two weeks, volume cut toward the middle of the 41-60 percent band, intensity untouched, frequency within 20 percent of what you were already doing [Bosquet et al. 2007]. The execution details — how to halve interval volume without halving interval intensity, what the final three days look like, why starting four weeks out backfires — are covered operationally in our 12-week goal-event guide. Do that first. Deviating from a population mean before you have any personal data is not personalization, it is guessing with extra steps. Then write down what you did: taper length in days, the volume carried in each taper week as a percentage of your last hard week, whether you rode openers, and the part everyone skips — an honest line on how the legs felt in the first hard effort of the event. Four or five goal events across two or three seasons and you have something. It will be uncontrolled and confounded, and still the only evidence in existence about how you respond to a taper. That is the recurring bind in the broader self-coached playbook: you are replacing a coach's pattern recognition, and your own recorded history is the only raw material available. A word on our own software, since an article arguing nobody has validated personalized tapering should not imply we have. AdaptCycling does not personalize your taper. Length comes from race priority: two weeks for an A event, one for a B, none for a C. Volume retention comes from a fixed per-discipline table — criteriums hold 60 percent of build volume in the first taper week, gran fondos 70 percent. Those figures were tuned against Bosquet's 41-60 percent window [Bosquet et al. 2007], though only taper week two lands inside it — holding 60 percent of volume is a 40 percent cut, 70 percent a 30 percent cut. They descend from the same meta-analysis this article is questioning. The only athlete-facing adjustment is a per-race number between zero and five, set in chat, governing how many deep-recovery days sit immediately before that event. A population table with a small manual knob on the final days. ## Common questions **Is a two-week taper right for a short event like a criterium?** Nobody knows, and that is the honest answer. Bosquet's meta-analysis does not report taper response stratified by event duration [Bosquet et al. 2007]. Tapering shorter for shorter events is convention, not evidence. The one randomized cycling comparison used a seven-day taper and produced a 5.4 percent time-trial gain at a 50 percent volume cut, across eleven riders total [Neary et al. 2003]. **Should I taper at all for a B-priority race?** A shorter, shallower taper is the usual answer, and it is a scheduling judgment rather than an evidence-backed one. The useful question is what the B race is for. If it is a hard training day inside a build block, tapering into it costs build volume for a result that does not matter. **How do I tell whether my taper was too long or too short?** Race-morning feel is a poor signal — riders routinely feel flat and ride well, or feel sharp and fall apart. The more informative markers are your first hard effort and the power you hold in the final third. Too long tends to show as a good opening and a missing top end; too short as a strong finish after a labored start. **If I have five years of Strava data, can software compute my optimal taper?** It can compute one. Whether to trust it is a separate question. When researchers fitted this class of model to nine elite swimmers using a full season of controlled data, the confidence interval on the time to peak performance alone ran from 25 to 61 days, and some parameters were correlated to the point the authors call interpreting them worthless [Hellard et al. 2006]. Strava data is messier than that, not cleaner. ## References 1. **Bosquet et al. 2007.** [Effects of tapering on performance: a meta-analysis](https://pubmed.ncbi.nlm.nih.gov/17762369/). Medicine & Science in Sports & Exercise. 2. **Mujika & Padilla 2003.** [Scientific bases for precompetition tapering strategies](https://pubmed.ncbi.nlm.nih.gov/12840640/). Medicine & Science in Sports & Exercise. 3. **Neary et al. 2003.** [Effects of different stepwise reduction taper protocols on cycling performance](https://pubmed.ncbi.nlm.nih.gov/12904635/). Canadian Journal of Applied Physiology. 4. **Thomas et al. 2008.** [A model study of optimal training reduction during pre-event taper in elite swimmers](https://pubmed.ncbi.nlm.nih.gov/18344135/). Journal of Sports Sciences. 5. **Thomas et al. 2009.** [Computer simulations assessing the potential performance benefit of a final increase in training during pre-event taper](https://pubmed.ncbi.nlm.nih.gov/19675490/). Journal of Strength and Conditioning Research. 6. **Hellard et al. 2006.** [Assessing the limitations of the Banister model in monitoring training](https://pubmed.ncbi.nlm.nih.gov/16608765/). Journal of Sports Sciences. --- # Winter cycling training plan: come out of February faster than you went into November URL: https://www.adaptcycling.com/guides/winter-cycling-training-plan Updated: 2026-09-03 Author: Jim Camut A winter cycling training plan has two jobs: hold the fitness summer built, and raise the aerobic floor spring stands on. The structure that does both is not complicated — mostly low-intensity riding plus one or two short high-intensity sessions per week; roughly 4 to 6 hours holds fitness, 6 to 10 builds it. The harder problem is that winter is when plans break most: darkness, illness season, holidays, travel. From November to February, the plan's ability to restructure matters more than its workouts. ## What winter training is for — and the detraining math if you stop Winter training has one purpose: set the aerobic base that sizes everything spring can build. The alternative is expensive. Coyle and colleagues' detraining time-course study [Coyle et al. 1984] measured a 7% VO2max drop in the first 21 days of inactivity, stabilizing around 16% below trained baseline by day 56. A bike parked from Thanksgiving to February rides most of that curve. The decline starts faster than most riders assume. Mujika and Padilla's short-term detraining review [Mujika & Padilla 2000] documents measurable losses in VO2max and maximal aerobic power within 10 to 14 days of insufficient stimulus, driven primarily by falling blood volume and stroke volume. The cardiovascular layer moves first and fastest — in both directions. That asymmetry is the planning insight: recently acquired adaptations decay quickest, while mitochondrial and capillary adaptations built over years barely move in a month. A rider who ends the season in October and restarts in February is not resuming; they are rebuilding, and the rebuild consumes the exact weeks a spring goal needs for its build phase. One to two weeks fully off the bike after the season costs almost nothing and pays for itself in motivation — that part of off-season folklore survives the evidence. It is weeks three through twelve that decide the spring. The classical preparatory period exists because base adaptations are slow to build and slow to lose, and winter is the only uninterrupted block long enough to build them. The operational goal is not heroics. It is a floor: enough riding to hold the cardiovascular layer, enough structure that January is week ten of something rather than day one of nothing. ## Base vs intensity indoors: what the evidence actually supports Mostly low intensity, a little high intensity, not much in between. Seiler's intensity-distribution research [Seiler 2010] puts roughly 80% of training time below the first lactate threshold, and Stöggl and Sperlich's analysis of well-trained endurance athletes [Stöggl & Sperlich 2015] found elites cluster around polarized or pyramidal distributions. Winter changes the hours, not the ratio. The zone-2-only winter is folklore, not evidence. The strongest winter-specific study is Rønnestad and colleagues' transition-period trial [Rønnestad et al. 2014]: well-trained cyclists who kept one high-intensity session every 7 to 10 days across an 8-week off-season held and extended fitness, finishing with power at 4 mmol/L lactate 10.6% better than the low-intensity-only group and 40-minute time-trial power 12.4% better. The advantage persisted — after the following 16-week preparatory period the interval group was still 6.0% ahead in the 40-minute trial. The cohort was small, 13 riders, but the direction matches the broader maintenance literature: intensity is the ingredient you cannot drop. For the time-crunched rider the case for some indoor intensity is also arithmetic. A 60-minute trainer session has no coasting, no junctions, and no descents; it delivers roughly what 75 to 90 outdoor minutes deliver. Intensity compresses further still — one 4x8-minute interval session covers the week's high-intensity quota in under an hour. The trap runs the other direction: because the trainer makes hard work convenient, self-coached riders drift toward three sweet-spot sessions a week and no easy volume, inverting the distribution the evidence supports [Seiler 2010]. The ratio binds indoors exactly as it does outdoors; the trainer just makes it easier to violate. ## The minimum effective winter: hours per week that hold fitness The maintenance floor is lower than most riders fear. Spiering and colleagues' minimal-dose review [Spiering et al. 2021] found endurance performance holds for up to 15 weeks with training frequency cut to 2 sessions per week or volume cut by 33 to 66% — provided intensity is preserved. Fifteen weeks is Thanksgiving to March. Translate the review into a week. A rider who trained 8 hours per week in summer can hold most of their endurance performance on 3 to 4 winter hours if the sessions keep their intensity: one hard interval session, one steady endurance ride, nothing wasted. Intensity is the non-negotiable knob — in the studies reviewed, performance survived large frequency and volume cuts only when exercising intensity stayed [Spiering et al. 2021]. Two honesty notes: the endurance data comes mostly from general populations rather than trained cyclists, so treat the floor as a floor; and cutting the hard session while keeping the hours fails the maintenance math from the other side. Holding and building are both legitimate winters; the plan just has to know which one it is running. Four to six hours per week is a holding winter — fitness in March close to fitness in October. Six to ten hours with progressive structure is a building winter — the rider enters the spring build ahead of last year. The failure mode is neither. It is planning a 9-hour building winter in an optimistic November, missing week three to a cold and week six to holiday travel, concluding the plan is broken, and doing nothing until March. A modest plan executed beats an ambitious plan abandoned, and the gap between those two outcomes is the next section. ## Why winter is when plans break — and what an adaptive plan does about it Winter concentrates every force that breaks training plans: 5pm darkness, cold and flu season, holiday travel, year-end work crunch. A static plan demands compliance in exactly the 12 weeks compliance is hardest to give. This is the season where “adaptive” either earns its name or turns out to be a marketing word. The breakage is structural, not motivational. Weekday outdoor slots disappear with daylight. Respiratory illness clusters between November and February, and a normal cold costs 5 to 10 days. The holidays remove one to two weeks from almost everyone's December. Foster's training-monotony research [Foster 1998] adds a quieter winter risk: across his 25-athlete cohort, high load combined with low day-to-day variance was the strongest behavioral predictor of illness, and indoor winters drift naturally toward monotony — same trainer, same hour, same intervals. A static plan meets all of this with the same answer it has for everything: the next page in the schedule, plus the implication that missing it was the rider's failure. This is the mechanism question the adaptive cycling training plans pillar is built around, and the indoor cycling training season is its hardest test. When a sinus infection deletes a week of December, compressing the missed block into January is wrong and pretending the week never happened is wrong; the right answer is a restructure — the return week rebuilt at reduced load, the next deload pulled forward, the spring arc re-anchored to the new dates. We built AdaptCycling to run that restructure automatically: the plan reads every ride you upload, notices the week that actually happened, and reshapes what remains instead of grading the rider against what did not. A winter plan that cannot do this gets abandoned by February — not because the workouts were wrong, but because week one of January no longer matched reality. None of this requires a subscription platform. A trainer, a plan, and one effort signal — power, heart rate, or perceived exertion — are sufficient for every session in a 12-week winter block. The platforms are motivation layers, and good ones: Zwift makes low-intensity volume genuinely tolerable and its racing doubles as sharp interval work; TrainerRoad and Wahoo SYSTM package structured workouts with excellent execution. Use whichever keeps you on the bike in January. Just keep the roles straight: the platform makes the session happen, the plan decides what the session should be — and only the plan can adapt when the week falls apart. ## Common questions **Do I need Zwift or a smart trainer for winter training?** No. Structured winter training needs a trainer, a plan, and one effort signal — power, heart rate, or perceived exertion. A smart trainer's erg mode makes interval execution easier, and Zwift makes volume more tolerable, but both are conveniences. Riders built winter fitness on dumb trainers for decades before either existed; the intensity distribution and the consistency do the work, not the software. **How many hours a week do I need to train in winter?** To hold fitness: roughly 4 to 6 hours, with intensity preserved — the minimal-dose review [Spiering et al. 2021] supports endurance maintenance for up to 15 weeks at 2 sessions per week as long as the hard work stays hard. To build: 6 to 10 hours with progressive structure. Pick the target against your real winter schedule, not your best-case one. **When should I start building toward a spring goal?** Count backward from the event: a 2-week taper plus 8 to 12 weeks of build means a mid-May goal needs structured build from roughly mid-February, with the winter base block from November through January feeding it. The base does not need to be event-specific; it needs to hand February a rider who can absorb build-phase load at a safe progression rate. **Is it OK to take time completely off the bike after the season?** Yes — 1 to 2 weeks off costs little and helps motivation. Measurable cardiovascular decline starts around 10 to 14 days [Mujika & Padilla 2000], and the losses compound from there [Coyle et al. 1984]. Past that point, restart at reduced load rather than at the old numbers. Rønnestad's transition-period data [Rønnestad et al. 2014] argues for keeping one hard session every 7 to 10 days even through the off-block. ## References 1. **Coyle et al. 1984.** [Time course of loss of adaptations after stopping prolonged intense endurance training](https://pubmed.ncbi.nlm.nih.gov/6511559/). Journal of Applied Physiology. 2. **Mujika & Padilla 2000.** [Detraining: Loss of Training-Induced Physiological and Performance Adaptations. Part I: Short Term Insufficient Training Stimulus](https://pubmed.ncbi.nlm.nih.gov/10966148/). Sports Medicine. 3. **Rønnestad et al. 2014.** [HIT maintains performance during the transition period and improves next season performance in well-trained cyclists](https://pubmed.ncbi.nlm.nih.gov/24878691/). European Journal of Applied Physiology. 4. **Spiering et al. 2021.** [Maintaining Physical Performance: The Minimal Dose of Exercise Needed to Preserve Endurance and Strength Over Time](https://pubmed.ncbi.nlm.nih.gov/33629972/). Journal of Strength and Conditioning Research. 5. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 6. **Stöggl & Sperlich 2015.** [The training intensity distribution among well-trained and elite endurance athletes](https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2015.00295/full). Frontiers in Physiology. 7. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. --- # Why your heart rate is higher on the indoor trainer — and sometimes lower URL: https://www.adaptcycling.com/guides/heart-rate-higher-on-indoor-trainer Updated: 2026-09-03 Author: Jim Camut Heart rate on a trainer runs higher than on the road at the same watts, and lower at the same effort, and both readings come from one cause: no airflow. Riders told to hold the same perceived effort indoors and out produced 21% fewer watts in the lab and a lower heart rate to match [Mieras et al. 2014]; riders holding the same watts indoors see heart rate drift upward over the hour as stroke volume falls [Coyle & González-Alonso 2001]. This spoke of the indoor cycling training guide explains both directions, what a fixed heart-rate target quietly does indoors, and the two fixes that actually work. ## It goes both ways, and both are heat The forum claim that heart rate is always higher on the trainer is half right. At matched effort it is lower, because the watts are lower; at matched watts it climbs, because nothing carries heat away. Twelve cyclists riding 40 km at the same perceived effort averaged 208 W outdoors and 163 W in the lab, with heart rate 152 versus 143 bpm [Mieras et al. 2014]. The Mieras study is the cleanest test of the question because it removed everything but the room. Same riders, same distance, same instruction to hold the same effort. Core temperature finished the same in both conditions; skin temperature was 1.6 °C warmer indoors; the only environmental difference was wind, 2.5 m/s outdoors against zero in the lab [Mieras et al. 2014]. Perceived exertion did not differ, and neither did where the riders reported their attention going. The body simply produced fewer watts for the same sensation, and heart rate followed the watts down. Most riders never see that reading, because they ride the trainer by power, and at a fixed power the picture inverts. The 180 W that sat at 135 bpm in the fifth minute reads 150 bpm by the fiftieth. Nothing about the muscle changed; the cooling did. Both observations are the same physiology viewed from two sides — the trainer raises the heart-rate cost of every watt, and a rider who holds effort constant pays that cost in watts, while a rider who holds watts constant pays it in heart rate. The practical consequence is that an indoor heart-rate reading is only comparable with an outdoor one at the same point in the ride, in the same state of cooling. A 20-minute reading on a well-fanned trainer is close to road conditions. A 60-minute reading in a shut room is not a fitness number at all. With no coasting to let heart rate recover, the drift described next has an unbroken hour to build. ## What cardiovascular drift does to a heart-rate target After 10 to 20 minutes of steady work, stroke volume begins to fall and heart rate rises to hold cardiac output — cardiovascular drift [Coyle & González-Alonso 2001]. On a trainer the drift is larger and starts sooner, and it changes what a heart-rate zone means: the same beats per minute now sit at a higher fraction of a reduced VO2max [Wingo et al. 2020]. The mechanism was revised in 2001. Older explanations blamed blood pooling in the skin; Coyle and González-Alonso showed the progressive fall in stroke volume is driven primarily by the rising heart rate itself, with skin blood flow a smaller contributor [Coyle & González-Alonso 2001]. Heat is still the trigger, and the loop tightens across the hour. Dehydration accelerates it. The trainer supplies all three conditions at once: no convective cooling, sweat that stays on the skin, and a bottle that is easy to forget. The size of the effect is not small. In seven active men cycling at 60% of VO2max in 35 °C, heart rate rose 17% and stroke volume fell 15% between minute 15 and minute 45 — and the drift corresponded to a proportional fall in VO2max measured immediately afterward [Wingo et al. 2020]. The authors' conclusion is the one to hold onto: the upward drift in heart rate reflects an increased relative metabolic intensity. Heart rate is not misreporting. The session genuinely got harder. That is what a fixed heart-rate target does indoors. A rider told to hold 140 bpm for 90 minutes will, as drift progresses, back the power off to keep the number — and end the ride training a lower and lower fraction of the watts the session was meant to deliver. A rider told to hold 180 W will keep the training stimulus and watch heart rate leave the zone. The mistake is reading either as a statement about fitness rather than about cooling, which is why the fix is equipment before it is physiology. ## Airflow and temperature are the cause, so they are the fix Time to exhaustion at 70% of VO2max was 93.5 minutes at 10.5 °C and 51.6 minutes at 30.5 °C in near-still air [Galloway & Maughan 1997]. Airflow of about 33 km/h removed the heat-storage and perceived-effort penalty of two hours of cycling, and 50 km/h added nothing [Saunders et al. 2005]. One strong fan, running from the first minute, is the whole intervention. The temperature curve is an inverted U with its optimum in the cold, and a spare room is closer to the hot end than it looks. Galloway and Maughan tested 3.6, 10.5, 20.6 and 30.5 °C at an air velocity around 0.7 m/s — a room with no fan — and 10.5 °C won [Galloway & Maughan 1997]. Twenty-one degrees with the door shut feels fine at minute one, but the air 20 cm in front of a sweating body warms and saturates within minutes, and the effective condition drifts toward the 30 °C trial. Airflow substitutes for the road. Saunders and colleagues cycled nine subjects for two hours at 33 °C with facing air at 0.2, 9.9, 33.3 and 50.1 km/h; heat storage, body temperature and perceived exertion were all higher at the two low speeds than at the two high ones, and there was no difference between 33 and 50 km/h [Saunders et al. 2005]. Those are hot-room numbers, not a fan wattage for a 20 °C room. The principle transfers: moving air across the torso at something like road speed closes most of the gap, and past that point more fan buys nothing. Timing matters as much as size. Wingo and colleagues let drift develop over 45 minutes in the heat and then aimed a 4.5 m/s fan at the rider shortly before a maximal test; skin temperature dropped about 1 °C and VO2max still fell 15%, statistically indistinguishable from the 17% fall with no fan at all [Wingo et al. 2019]. Once drift is established, cooling does not reverse it. The fan goes on before the warm-up, and the window opens before the fan, so the air being moved is not already at skin temperature. ## What to do with a heart rate that reads high Fix the cooling first, train to power where power exists, treat indoor heart-rate zones as ceilings rather than targets, and read the gap between heart rate and power across a steady ride — aerobic decoupling — as a diagnostic of the setup before reading it as a statement about fitness. Power is the indoor signal of choice because it does not drift. A smart trainer reports watts regardless of room temperature, and a workout prescribed in watts delivers the intended stimulus whether heart rate cooperates or not. Heart rate then becomes the second channel — the one that says how much the session cost — rather than the one that decides how hard it is. Riders with a strap and no power meter should hold indoor heart-rate ceilings and let effort come down as drift builds, rather than chasing the number. Aerobic decoupling is the number that ties the two channels together: how far heart rate rises relative to power between the first and second halves of a steady effort. Some decoupling is physiology and unavoidable, and a lot of it on a trainer is a cooling problem. A few percent over a well-fanned endurance ride is ordinary; a much larger gap usually says the fan is too small, the room is too warm, or the bottle is empty. Fix those before concluding anything about aerobic fitness. AdaptCycling reports decoupling on every ride long enough to measure it, from Wahoo, Hammerhead, Polar and intervals.icu rides alike, and treats a high indoor value as a setup flag first. The checklist is short. Fan on the torso at road-like speed, running before the warm-up. Door or window open before the fan. A bottle per hour, started early. Intervals prescribed in watts and executed in ERG where the trainer supports it; endurance rides prescribed with a heart-rate ceiling that wins when drift pushes past it. And a plan that reads the ride that happened: a heart rate 10 bpm above last week's at the same watts is information about the room, recovery, or an incoming cold, and a coach should notice first. ## Common questions **Is heart rate higher or lower on an indoor trainer?** Both, depending on what is held constant. At the same perceived effort riders produce fewer watts indoors and heart rate reads lower, because the work is lower [Mieras et al. 2014]. At the same watts, the missing airflow drives cardiovascular drift and heart rate reads higher as the ride goes on [Coyle & González-Alonso 2001]. Both come from heat. **Does a fan really lower heart rate on the trainer?** Yes, if it is on from the start. Facing airflow around 33 km/h removed the heat-storage and perceived-effort penalty of two hours of cycling in one study, and raising it to 50 km/h added nothing [Saunders et al. 2005]. A fan switched on after drift has developed did not reverse the loss of VO2max in another [Wingo et al. 2019]. Run it before the warm-up. **Should I use different heart-rate zones indoors?** Use the same zones as ceilings rather than targets. Drift means a fixed heart rate corresponds to a rising fraction of VO2max as the ride progresses [Wingo et al. 2020], so holding a number indoors quietly makes the session harder. Where power is available, prescribe the session in watts and let heart rate report the cost. **Is a high heart rate on the trainer a sign of poor fitness?** Usually not. The first suspects are the room, the fan, hydration and sleep, in that order. Aerobic decoupling over a well-cooled steady ride is the useful check: a few percent is ordinary, and a large gap on a trainer is far more often a cooling problem than an aerobic one. Change the setup first and compare the same session a week later. ## References 1. **Mieras et al. 2014.** [Physiological and psychological responses to outdoor vs. laboratory cycling](https://pubmed.ncbi.nlm.nih.gov/24476776/). Journal of Strength and Conditioning Research. 2. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. 3. **Wingo et al. 2020.** [Cardiovascular drift and maximal oxygen uptake during running and cycling in the heat](https://pubmed.ncbi.nlm.nih.gov/32102057/). Medicine & Science in Sports & Exercise. 4. **Galloway & Maughan 1997.** [Effects of ambient temperature on the capacity to perform prolonged cycle exercise in man](https://pubmed.ncbi.nlm.nih.gov/9309637/). Medicine & Science in Sports & Exercise. 5. **Saunders et al. 2005.** [The effects of different air velocities on heat storage and body temperature in humans cycling in a hot, humid environment](https://pubmed.ncbi.nlm.nih.gov/15743384/). Acta Physiologica Scandinavica. 6. **Wingo et al. 2019.** [Fan cooling after cardiovascular drift does not reverse decrements in maximal oxygen uptake during heat stress](https://pmc.ncbi.nlm.nih.gov/articles/PMC6773240/). Temperature. --- # How long should indoor trainer rides be? Session lengths that earn their place URL: https://www.adaptcycling.com/guides/how-long-should-indoor-trainer-rides-be Updated: 2026-09-03 Author: Jim Camut An interval session on the trainer needs 45 to 75 minutes including warm-up and cool-down. An endurance ride earns its keep between 60 and 120 minutes, and past 90 the heat problem starts to eat the stimulus unless the fan and the room are very good. Twenty-six minutes a day, at full intensity, held VO2max for 15 weeks in Hickson's reduced-duration study [Hickson et al. 1982]. The right length depends on the job of the session, not on the road ride it replaces, because a trainer hour is not a road hour. ## Why a trainer hour is not a road hour A trainer removes coasting, airflow and distraction, and each one shortens the useful session. Sixty minutes of pedalling on a trainer carries roughly the work of 75 to 90 minutes on a road with junctions and descents — a rule of thumb, not a measurement, but a fair one. It also stores more heat, which is why the ceiling on indoor duration is lower than the ceiling outdoors [Galloway & Maughan 1997]. Coasting is the obvious difference. A road ride with descents, corners and traffic lights spends a meaningful share of its clock at zero watts; a trainer in ERG mode spends none. That is why a 90-minute outdoor endurance ride moved indoors unchanged usually feels long in its last half hour: the pedalling time was already there in a shorter session. Treat the 75-to-90 conversion as a planning rule and let the ride's power data correct it. Heat sets the ceiling. In near-still air — roughly what a spare room with the door shut and no fan provides — cyclists riding at 70% of VO2max lasted 93.5 minutes at 10.5 °C and 51.6 minutes at 30.5 °C, with 20.6 °C also shorter than the cold condition [Galloway & Maughan 1997]. Time to exhaustion is not the target of a training ride, but the shape of that curve is the point: as the room warms, the minutes a rider can hold a steady endurance power shrink, and the last minutes of a long indoor ride fight temperature rather than train the aerobic system. Monotony is the third and least discussed factor. Foster's index of training monotony — daily mean load divided by its standard deviation — feeds training strain, and strain was the index that best explained when 25 athletes got sick [Foster 1998]. Indoor sessions are structurally monotonous: same room, same hour, same view. A week of identical 90-minute trainer rides scores badly on that index even at modest volume. Shorter sessions with different jobs keep the week's variance up. ## Session length by job: the hard day fits in an hour The interval literature was written on ergometers, and its best-supported formats fit inside 60 minutes. Four by eight minutes at 90% of peak heart rate — 32 minutes of work — beat both shorter and longer intervals on VO2peak and power in 35 trained cyclists [Seiler et al. 2013]. Add a 15-minute warm-up and a 10-minute cool-down and the session is 60 to 65 minutes long. Seiler's design answers the length question directly. Three groups trained two interval sessions a week for seven weeks on about six hours of total riding: 4×4 minutes at 94% of peak heart rate, 4×8 at 90%, or 4×16 at 88%, each at the hardest tolerable intensity. The 4×8 group improved VO2peak, power at VO2peak and power at 4 mmol lactate by 11.4%; the other two groups improved by 4 to 6% [Seiler et al. 2013]. The middle length won, and the whole session sits inside an hour. Short intervals shorten it further. Ten weeks of 30-second efforts with 15-second recoveries, effort-matched against 4×5 minutes and performed twice a week, raised VO2max 8.7% against 2.6%, with moderate-to-large effects across 30-second, 5-minute and 40-minute power [Rønnestad et al. 2015]. Three sets of 13 repetitions of 30/15 is under 30 minutes of interval time. Neither format needs a longer session to work, and both are far easier to execute on a trainer in ERG mode than on an open road, which is the trainer's real advantage. The practical rule: the hard session is 45 to 75 minutes, its length set by the interval format plus a proper warm-up, never padded with easy riding to reach a round number. Extra easy volume belongs in a separate endurance ride on another day, not on top of fatigued legs in a warming room. ## The endurance ride indoors: where the return runs out Sixty to 90 minutes is the productive range for an indoor endurance ride; up to two hours is defensible with a strong fan and a cool room; past that, cardiovascular drift and heat have usually replaced the aerobic stimulus with a heat-tolerance one. Stroke volume starts falling after 10 to 20 minutes of steady work, and heart rate rises to cover it [Coyle & González-Alonso 2001]. Drift is the mechanism that puts a ceiling on indoor duration. After 10 to 20 minutes, stroke volume begins a progressive decline and heart rate rises to hold cardiac output, driven primarily by the rising heart rate itself rather than by blood diverted to the skin [Coyle & González-Alonso 2001]. On the road, the airflow that comes free with speed limits the drift. On a trainer, only the fan does. The 180 W that sat at 135 bpm in minute five reads 150 bpm in minute fifty. The consequence for length is not that long indoor rides are wrong, but that they pay less per minute the longer they run. The first 60 to 90 minutes of a well-cooled endurance ride deliver the steady aerobic work the session exists for. The minutes after that, in most home setups, deliver a rising heart rate at a falling power — a different session from the one prescribed. If the ride must be long, the fan goes on before the warm-up, the door opens before the session, and the target is held by power with a heart-rate ceiling that wins when the two disagree. A longer indoor ride pays when the calendar has removed every outdoor option, the cooling setup holds decoupling to a few percent over two hours, and the goal event needs long steady riding. A rider preparing for a spring gran fondo in a January with no rideable roads has a case for one two-hour trainer ride a week. A rider with a four-hour winter budget does not; that rider's best long ride is 90 minutes, with the remaining time spent on a second short session. ## The minimum that still counts, and how to fit the week Twenty-six minutes a day held VO2max and short-term endurance for 15 weeks in Hickson's study; 13 minutes a day held VO2max too but lost 10% of long-term endurance [Hickson et al. 1982]. Intensity was held constant in both. A 30-minute trainer session is a real session. A 15-minute one keeps the engine but not the range. The reduced-duration study is the clearest answer to the question of the shortest session worth doing. Thirteen subjects trained 40 minutes a day, six days a week, for ten weeks, raising VO2max 10 to 20%. They then cut each session to either 26 or 13 minutes for a further 15 weeks at the same intensity and frequency. VO2max stayed at trained levels in both groups. Long-term endurance, tested at two hours or more, held in the 26-minute group and fell 10% — from 139 to 123 minutes — in the 13-minute group [Hickson et al. 1982]. A two-thirds cut in duration kept almost everything, provided the intensity did not move. That reframes a busy winter week. The question is not how to find 90 minutes but how to protect the hard 30. A 45-minute session with a 10-minute warm-up and 30 minutes of intervals is well inside the range the evidence supports; the indoor cycling training pillar puts the same point as a tiered week, where every hour budget keeps the hard session and scales the easy volume. The companion piece on how many trainer sessions a week maintain fitness works through the frequency half of the same studies. Fitting sessions into the week means not stacking them. Two 45-minute sessions on separate days beat one 90-minute session indoors: each gets a full warm-up, each finishes before drift dominates, and the week's variance — Foster's protection against strain — is higher. The indoor cycling training plan for four hours a week lays out the concrete week: one hard session of about an hour, one 90-minute endurance ride, one shorter session, and a day off between the hard days. We build the same skeleton when the plan reads a winter week from a Wahoo, a Hammerhead, a Polar or intervals.icu, and rebuild it when the calendar takes a day away. ## Common questions **Is a 30-minute trainer ride worth doing?** Yes, if it carries intensity. Sessions cut to 26 minutes a day at full intensity held VO2max and short-term endurance for 15 weeks, and only the 13-minute group lost long-term endurance [Hickson et al. 1982]. A 30-minute session with a short warm-up and 20 minutes of hard work is a real training day. A 30-minute easy spin is recovery, not the week's aerobic work. **How long should an indoor endurance ride be?** Sixty to 90 minutes for most riders, up to two hours with a strong fan and a cool room. Stroke volume starts falling after 10 to 20 minutes of steady work and heart rate rises to compensate [Coyle & González-Alonso 2001]; in a warm room without airflow that drift turns the back half of a long ride into heat tolerance rather than aerobic training. A trainer hour carries roughly the pedalling of 75 to 90 road minutes — a rule of thumb, not a measurement. **How long should an interval session on the trainer be?** Forty-five to 75 minutes including warm-up and cool-down. The best-supported formats are short: 4×8 minutes at 90% of peak heart rate outperformed both 4×4 and 4×16 in trained cyclists [Seiler et al. 2013], and 30-second on, 15-second off intervals beat 4×5 minutes on an effort-matched basis [Rønnestad et al. 2015]. Neither needs more than an hour. Do not pad the session with easy riding to reach a round number. **Is one long trainer ride better than two short ones?** Usually not, indoors. Two sessions on separate days each get a full warm-up, each finishes before cardiovascular drift dominates, and the week's day-to-day variance stays higher, which matters because training strain — load multiplied by monotony — was the index that best predicted illness in Foster's 25 athletes [Foster 1998]. The exception is a rider with a long-event goal, a cool room and no outdoor option. ## References 1. **Hickson et al. 1982.** [Reduced training duration effects on aerobic power, endurance, and cardiac growth](https://pubmed.ncbi.nlm.nih.gov/6214534/). Journal of Applied Physiology. 2. **Galloway & Maughan 1997.** [Effects of ambient temperature on the capacity to perform prolonged cycle exercise in man](https://pubmed.ncbi.nlm.nih.gov/9309637/). Medicine & Science in Sports & Exercise. 3. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 4. **Seiler et al. 2013.** [Adaptations to aerobic interval training: interactive effects of exercise intensity and total work duration](https://pubmed.ncbi.nlm.nih.gov/21812820/). Scandinavian Journal of Medicine & Science in Sports. 5. **Rønnestad et al. 2015.** [Short intervals induce superior training adaptations compared with long intervals in cyclists — an effort-matched approach](https://pubmed.ncbi.nlm.nih.gov/24382021/). Scandinavian Journal of Medicine & Science in Sports. 6. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. --- # Sweet spot vs zone 2 for the indoor winter: what the trials found and where each belongs URL: https://www.adaptcycling.com/guides/sweet-spot-vs-zone-2-indoor-winter Updated: 2026-09-03 Author: Jim Camut Mostly zone 2, with one properly hard session, and sweet spot as a capped third lever — that is the distribution the trials support. In the two studies that pitted a polarized week against a threshold-heavy one in trained cyclists, the polarized group gained more on equal or fewer hours [Neal et al. 2013, Stöggl & Sperlich 2014]. Sweet spot did not fail in those trials; it lost. The distinction matters, because a rider with four hours a week has a real use for it. Under the indoor cycling training guide, this is the one choice that decides most winters. ## What zone 2 and sweet spot are each for Zone 2 builds the aerobic engine that everything else runs on; sweet spot buys a threshold stimulus in less time. Professional endurance athletes oxidise far more fat at a given power and produce less lactate than moderately active people, and the two measures track each other almost perfectly [San-Millán & Brooks 2018]. That capacity is what easy volume trains. Zone 2 is 56 to 75% of FTP, and its job is metabolic rather than mechanical. In San-Millán and Brooks' comparison of professional cyclists with moderately active and metabolically unhealthy adults, fat oxidation and blood lactate were inversely correlated in every group, at r = −0.97 in the professionals [San-Millán & Brooks 2018]. That capacity comes from hours below the first threshold, not from riding at it. It is slow to build and slow to lose, which is why winter is where it gets built. Sweet spot is 88 to 94% of FTP: hard enough to load the threshold system, easy enough to repeat. Its case is arithmetic: a 2×20 at 90% of FTP delivers roughly the training load of 90 minutes of endurance riding in one 60-minute slot, and a rider with four hours a week cannot spend two of them on one easy ride. That is the honest use of sweet spot — a compromise for the time-limited. The trainer tilts the choice toward sweet spot for a reason unrelated to physiology. At matched intensity, riders rate the same work as harder indoors — by 1.4 to 1.6 RPE units for breathing and 2.5 to 2.7 for the legs [Olsson et al. 2024] — and at matched perceived effort they produce 21% fewer watts on the trainer than outside [Mieras et al. 2014]. Zone 2 feels like nothing on a road and like work on a trainer, so the target creeps up to make the hour pass. That is how a zone 2 winter becomes a sweet spot winter without anyone deciding it should. ## What happened when polarized met threshold on equal hours Twelve trained cyclists rode six weeks polarized — 80% low, 20% high, nothing in the middle — and six weeks threshold-heavy, 57% low and 43% moderate, in crossover. Polarized won on peak power (8% vs 3%), lactate threshold (9% vs 2%) and high-intensity capacity (85% vs 37%), on 6.4 hours a week against the threshold model's 7.5 [Neal et al. 2013]. The threshold model improved every measure; it just improved them less, and on more hours. Muscle enzyme activity did not change in either period, so the difference was systemic [Neal et al. 2013]. The threshold group's 43% of time between the thresholds is, in practice, a sweet spot winter — three or four sessions a week at 85 to 95% of FTP with easy riding around them. It is exactly the week the trainer makes convenient. The second trial widened the field. Stöggl and Sperlich put 48 well-trained runners, cyclists, triathletes and skiers through nine weeks of one of four distributions. The polarized group gained 11.7% in VO2peak and 17.4% in time to exhaustion; the threshold group's VO2peak did not change and its time to exhaustion rose a non-significant 6.2%; the high-volume group's VO2peak did not change either [Stöggl & Sperlich 2014]. In runners, the same pattern held over five months when the easy share was pushed from 67% to 80% with high-intensity work held constant — the low-intensity group improved a 10.4 km race by 157 seconds against 121 [Esteve-Lanao et al. 2007]. None of this says sweet spot does not work; every threshold group got fitter. What the trials say is narrower: on the same hours, easy volume plus a small dose of genuinely hard work beats a week built around the middle. The middle is the less efficient place to spend a limited hour, and on a trainer it is the place riders drift to without noticing. ## The honest case for sweet spot on four hours a week Below roughly five hours a week the polarized arithmetic gets tight, because 80% easy of four hours is three hours of zone 2 that the rider may not have in one or two sittings. One sweet spot session covers the threshold stimulus and part of the load in a single hour. The cap is what makes it honest: one, at most two, never the whole week. The reason the cap exists is Hickson's intensity study. Twelve subjects trained 40 minutes a day, six days a week, for ten weeks, then kept the frequency and duration but cut the work rate by a third or two-thirds for 15 more weeks. VO2max was not maintained at either reduction, and long endurance fell 21% in the one-third group and 30% in the two-thirds group [Hickson et al. 1985]. A winter of nothing but sweet spot is reduced intensity with the frequency kept — the protocol that lost fitness in that trial. So the four-hour week keeps its hard day hard and lets sweet spot replace part of the easy volume, not the intensity. A workable shape: one 60-minute session carrying the week's high-intensity work — 4×8 minutes at 90% of peak heart rate, or 30/15s — one 60-minute 2×20 at 88 to 94% of FTP, and one 90 to 120 minute endurance ride at 60 to 70%. The endurance ride is the one riders cut when the week gets short, and the one that should survive: it is the only session training the capacity San-Millán and Brooks measured. The failure mode is a third sweet spot session because it is convenient in ERG. Two sessions at 90% of FTP plus one at 94% of peak heart rate and no easy riding is the threshold model from Neal's trial with the easy volume removed [Neal et al. 2013]. The rider gets fitter for six to eight weeks, plateaus around the new year, and reads the plateau as a need for more intensity. The week that holds through February keeps the hardest session very hard, the sweet spot session capped, and the easy ride genuinely easy. ## A winter week that uses both Mostly easy, one hard, sweet spot as the third lever that scales with how few hours there are. Four hours: hard, sweet spot, endurance. Six hours: hard, sweet spot, two endurance rides. Eight and up: two hard, one long ride, the rest easy, and sweet spot drops out because the volume now does its job. The tiers follow the trials. Above six hours the polarized shape is affordable — 80% easy of eight hours is more than six hours of zone 2, enough to build the base without borrowing from the middle — and the second hard session adds more than a sweet spot session would [Stöggl & Sperlich 2014]. Below six, the sweet spot session is the compromise that keeps the threshold system loaded while the easy volume shrinks. At four it is doing real work; at three the rider is in maintenance and the hard session alone holds VO2max [Hickson & Rosenkoetter 1981]. Two indoor rules keep the easy rides easy. The endurance session is prescribed by power and a heart-rate ceiling, and the ceiling wins: when drift pushes heart rate out of zone 2 in minute fifty, the watts come down. And the fan runs from the first minute, because a fan switched on after drift has set in does not reverse it — the indoor cycling training guide covers the heat mechanism and the airflow numbers, and the sibling spoke on session length covers how long each of these rides should be on a trainer. Zone 2 on a trainer is a cooling problem before it is a discipline problem. This is also the week that survives winter's interruptions best, and it is the shape we build around. When a cold takes ten days out of December, the sweet spot session is the first thing to drop on the return week and the last thing to add back; the hard session comes back first at reduced volume, the easy ride never really left. AdaptCycling reads the rides that actually happened — from a Wahoo, a Hammerhead, a Polar or intervals.icu — and rebuilds the week that way instead of scheduling the next 2×20 on the next free Tuesday. The distribution is easy to write down. Keeping it intact through an indoor winter needs a plan that reads the calendar. ## Common questions **Is sweet spot or zone 2 better for winter base training?** Zone 2 for the base, with one hard session a week, and sweet spot only where hours are short. Polarized distributions beat threshold-heavy ones in trained cyclists on equal or fewer hours [Neal et al. 2013, Stöggl & Sperlich 2014]. Sweet spot still improved every measure in those trials; it just improved them less per hour. **How many sweet spot sessions a week is too many?** More than two, and for most riders more than one. Three sweet spot sessions with no easy riding is the threshold model that lost in both cycling trials, and a winter of only sub-maximal work is the reduced-intensity protocol that lost VO2max [Hickson et al. 1985]. Keep one session genuinely hard and cap sweet spot at one, two at most. **Why is zone 2 so hard to hold on the trainer?** Heat and perception. Without airflow, heart rate drifts upward at the same watts, and riders rate matched work as harder indoors than outside [Olsson et al. 2024] while producing 21% fewer watts at the same perceived effort [Mieras et al. 2014]. Prescribe by power with a heart-rate ceiling, run a strong fan from minute one, and let the watts drop when the ceiling is reached. **Can I do a whole winter of zone 2 and skip intensity?** Not without losing the top end. Cutting training intensity while keeping frequency and duration failed to maintain VO2max over 15 weeks and cut long endurance by 21 to 30% [Hickson et al. 1985]. One hard session a week is the non-negotiable; the easy volume is what scales with the hours. ## References 1. **Neal et al. 2013.** [Six weeks of a polarized training-intensity distribution leads to greater physiological and performance adaptations than a threshold model in trained cyclists](https://pubmed.ncbi.nlm.nih.gov/23264537/). Journal of Applied Physiology. 2. **Stöggl & Sperlich 2014.** [Polarized training has greater impact on key endurance variables than threshold, high intensity, or high volume training](https://pmc.ncbi.nlm.nih.gov/articles/PMC3912323/). Frontiers in Physiology. 3. **Esteve-Lanao et al. 2007.** [Impact of training intensity distribution on performance in endurance athletes](https://pubmed.ncbi.nlm.nih.gov/17685689/). Journal of Strength and Conditioning Research. 4. **Hickson & Rosenkoetter 1981.** [Reduced training frequencies and maintenance of increased aerobic power](https://pubmed.ncbi.nlm.nih.gov/7219129/). Medicine & Science in Sports & Exercise. 5. **Hickson et al. 1985.** [Reduced training intensities and loss of aerobic power, endurance, and cardiac growth](https://pubmed.ncbi.nlm.nih.gov/3156841/). Journal of Applied Physiology. 6. **San-Millán & Brooks 2018.** [Assessment of metabolic flexibility by means of measuring blood lactate, fat, and carbohydrate oxidation responses to exercise in professional endurance athletes and less-fit individuals](https://pubmed.ncbi.nlm.nih.gov/28623613/). Sports Medicine. 7. **Olsson et al. 2024.** [Perceived exertion can be lower when exercising in field versus indoors](https://pmc.ncbi.nlm.nih.gov/articles/PMC11135770/). PLoS One. 8. **Mieras et al. 2014.** [Physiological and psychological responses to outdoor vs. laboratory cycling](https://pubmed.ncbi.nlm.nih.gov/24476776/). Journal of Strength and Conditioning Research. --- # An indoor cycling training plan for 4 hours a week: three sessions, one of them hard URL: https://www.adaptcycling.com/guides/indoor-cycling-training-plan-4-hours-week Updated: 2026-09-03 Author: Jim Camut Four hours a week on a trainer is enough to hold fitness through a winter and, spent right, enough to add a little. The evidence is specific about the spending: two hard sessions a week held a full VO2max gain for 15 weeks when the volume was cut by two-thirds, and the same schedule with the intensity cut by a third lost it [Hickson & Rosenkoetter 1981, Hickson et al. 1985]. So the four-hour week is built around one session that is genuinely hard, one that is long and genuinely easy, and one that flexes. This is the week, the numbers, and what to drop when December removes a day. ## What four hours can do, and what it cannot Four hours holds an aerobic base and keeps the top end alive. It does not build a big spring on its own. Hickson's riders trained 40 minutes a day, six days a week for ten weeks, then held every bit of the VO2max gain for another 15 weeks on two sessions a week — as long as those two sessions stayed at the trained intensity [Hickson & Rosenkoetter 1981]. The maintenance dose is smaller than most riders fear. After the ten-week build, one group cut to four sessions a week and one to two, with intensity and duration unchanged; VO2max stayed at trained levels in both for 15 weeks [Hickson & Rosenkoetter 1981]. A second study cut the duration instead, to 26 or 13 minutes a day; VO2max held again, and only the 13-minute group lost long endurance, by about 10% [Hickson et al. 1982]. Frequency and duration are both negotiable. Intensity is not. When the same six-day schedule was kept but the work rate dropped by a third, VO2max was not maintained and the time riders could ride long fell 21%; at two-thirds off, the endurance loss was 30% [Hickson et al. 1985]. That is the exact shape of the winter most self-coached riders run — every session kept, every session softened because it is cold and dark — and it is the version that loses fitness. The four-hour week works because it protects one hard session before it protects anything else. The ceiling is real. The cyclists in the interval studies this plan borrows from rode about six hours a week [Seiler et al. 2013], and the polarized model that beat a threshold-heavy one did so on 6.4 hours [Neal et al. 2013]. Expect to come out of February with what October built, plus a sharper top end, not with a new FTP. The winter cycling training plan spoke covers what six to ten hours can add; this page is about four. ## The week: three sessions, four hours Three sessions, not four. One 60-minute interval session that carries all of the week's intensity, one two-hour endurance ride at the bottom of zone 2, and one 60-minute session that is sweet spot when the week is normal and easy when it is not. Everything below is in ranges of FTP because a trainer target is a band, not a number. Session one, 60 minutes, hard. Fifteen minutes of warm-up with a couple of 30-second openers, then 4×8 minutes at 100 to 106% of FTP with 3 minutes easy between, then 10 minutes down. In the trial that built it, 4×8 at about 90% of peak heart rate produced an 11.4% gain in VO2peak and power, roughly double the 4×4 at 94% and the 4×16 at 88%, in cyclists on about six hours a week [Seiler et al. 2013]. Every third week, swap in short intervals — three blocks of 30 seconds hard and 15 seconds easy at around 115 to 125% of FTP, about 9 to 10 minutes per block — which beat 4×5 minutes on VO2max 8.7% to 2.6% when effort was matched [Rønnestad et al. 2015]. Session two, 120 minutes, easy. Sixty to 72% of FTP with a heart-rate ceiling at the top of zone 2, and the ceiling wins: when drift lifts heart rate over the line, the watts come down. A trainer hour has no coasting — a rule of thumb, not a measurement, is that 60 trainer minutes carry the pedalling of 75 to 90 road minutes — so two hours indoors is a long ride. The fan runs from the first minute; how long indoor trainer rides should be, and why this one stops at two hours, has its own spoke. Session three, 60 minutes, the flex. In a normal week it is sweet spot: 2×20 at 88 to 94% of FTP with 5 minutes between, a threshold stimulus and a load stimulus in one hour. The trials that favoured polarized training beat a threshold model with three or more of these a week and no easy riding around them [Neal et al. 2013]; one, next to a hard day and a long easy day, is a defensible use of a four-hour rider's third slot, and the sweet spot vs zone 2 spoke argues that case in full. In the first three weeks, or after a bad hard session, it is a second easy hour at 60 to 70% instead. Laid across a week: Tuesday hard, Thursday flex, Saturday or Sunday long, at least one full day off, never two hard days in a row. Four hours on three days beats four hours on five, because the recovery between the hard session and the long one is where the adaptation happens. ## When the week breaks: what survives Winter removes days. The order to drop them in is fixed: the flex session goes first, the long ride second, the hard session last — unless the reason is illness, in which case the hard session goes first and nothing replaces it. A plan that cannot restructure to that order is a calendar. Two sessions is still a real week. The maintenance data says the hard session and one more hold VO2max [Hickson & Rosenkoetter 1981], so a week that loses Thursday keeps Tuesday and the long ride and loses nothing that matters. A week that loses the weekend keeps Tuesday and moves the flex session to Saturday as an easy 60 to 90 minutes. What never happens is compressing the missed session into the next week; five hours stacked into a four-hour plan is the monotony and strain pattern the pillar warns about, not a catch-up. Illness reverses the order. A normal cold costs five to ten days between November and February, and the hard session is the first thing to go: the return is an easy 45 minutes at 55 to 65% of FTP, then the long ride shortened, then the intervals back in the second week at the low end of their range. A lost week in a four-hour plan costs almost nothing in fitness and costs a great deal if it is paid back with intensity in the first days out of bed. This is the part of indoor cycling training where the plan, not the workouts, does the work. A static plan meets a lost Thursday with next Thursday's page. We built AdaptCycling to meet it with a restructure: connect a Wahoo, a Hammerhead, a Polar or an intervals.icu account, and the plan reads the week that actually happened, notices the ride that did not, and reshapes what is left in the order above. That runs on the 14-day trial with no card; after it, the Free tier is read-only and the rebuilding plan is $15 a month. Strava seats are limited right now, so Strava riders may see a short waitlist. ## Running it without a subscription Every session on this page is a set of power targets and durations, which means a head unit can run it. A Wahoo ELEMNT or a Hammerhead Karoo drives a smart trainer in ERG from a planned workout with nothing on a laptop or a phone. That buys the structure on hardware already owned. It does not buy a world. The mechanics are documented by both vendors and covered in the head-unit spokes on this site. The workout reaches the device in the morning, the device pairs the trainer over ANT+ FE-C, and the trainer holds each target while the rider holds the cadence. A session with no power targets gives the trainer nothing to hold, which surprises riders the first time; every session on this plan has watts, so all three run. What is missing is the world: no group ride under the two-hour session, no race on a Thursday, no map. For a January long ride it is probably not enough for most riders — distraction lowers the perceived cost of easy volume in a way a blank wall does not, and Zwift is honestly the best indoor riding experience available. Keeping the plan on the head unit and the world on the screen is the coherent position, and we ship a free Zwift workout generator for riders who want these sessions inside it. ## Common questions **Is 4 hours a week enough for cycling training in winter?** Enough to hold fitness and sharpen the top end, provided one session a week is genuinely hard. Riders who cut to two intense sessions a week held their VO2max gains for 15 weeks; riders who kept every session and cut the intensity by a third did not [Hickson & Rosenkoetter 1981, Hickson et al. 1985]. **How many trainer sessions a week should I do on 4 hours?** Three. One 60-minute interval session, one two-hour easy ride, one 60-minute session that is sweet spot in a normal week and easy in a hard one. Three sessions on three days beat four short ones because the day off between the hard session and the long ride is where the adaptation lands. **Which session do I drop when I lose a day?** The flex session first, the long ride second, the hard session last. Two sessions — the intervals plus one more — still hold VO2max in the maintenance data [Hickson & Rosenkoetter 1981]. Illness reverses the order: the hard session goes first, the return is an easy 45 minutes, and the intervals come back in the second week at the low end of their range. **What intervals work best in one hard trainer session a week?** Accumulated time near threshold beats short maximal work: 4×8 minutes at about 90% of peak heart rate, roughly 100 to 106% of FTP, produced double the gains of 4×4 or 4×16 in cyclists on six hours a week [Seiler et al. 2013]. ## References 1. **Hickson & Rosenkoetter 1981.** [Reduced training frequencies and maintenance of increased aerobic power](https://pubmed.ncbi.nlm.nih.gov/7219129/). Medicine & Science in Sports & Exercise. 2. **Hickson et al. 1982.** [Reduced training duration effects on aerobic power, endurance, and cardiac growth](https://pubmed.ncbi.nlm.nih.gov/6214534/). Journal of Applied Physiology. 3. **Hickson et al. 1985.** [Reduced training intensities and loss of aerobic power, endurance, and cardiac growth](https://pubmed.ncbi.nlm.nih.gov/3156841/). Journal of Applied Physiology. 4. **Seiler et al. 2013.** [Adaptations to aerobic interval training: interactive effects of exercise intensity and total work duration](https://pubmed.ncbi.nlm.nih.gov/21812820/). Scandinavian Journal of Medicine & Science in Sports. 5. **Rønnestad et al. 2015.** [Short intervals induce superior training adaptations compared with long intervals in cyclists — an effort-matched approach](https://pubmed.ncbi.nlm.nih.gov/24382021/). Scandinavian Journal of Medicine & Science in Sports. 6. **Neal et al. 2013.** [Six weeks of a polarized training-intensity distribution leads to greater physiological and performance adaptations than a threshold model in trained cyclists](https://pubmed.ncbi.nlm.nih.gov/23264537/). Journal of Applied Physiology. --- # The fan is equipment: how much airflow indoor cycling actually needs URL: https://www.adaptcycling.com/guides/best-fan-for-indoor-cycling-airflow Updated: 2026-09-03 Author: Jim Camut The best fan for indoor cycling is the one that puts riding-speed air across your torso and face from the first minute of the warm-up. That is the whole buying rule, and the evidence behind it is specific: about 33 km/h of facing air cut heat storage, body temperature and perceived exertion during two hours of cycling, and raising it to 50 km/h added nothing [Saunders et al. 2005]. A fan switched on after drift has set in does not undo it [Wingo et al. 2019]. Brand matters far less than speed, placement and timing, so this guide covers those three. ## Why the fan is equipment, not comfort On a trainer, evaporation is the only cooling channel that scales with effort, and evaporation needs moving air. In near-still air, cyclists at 70% of VO2max lasted 93.5 minutes at 10.5 °C and 51.6 minutes at 30.5 °C [Galloway & Maughan 1997]. Remove the airflow that road speed provides and a temperate room behaves like a warm one. The body sheds heat four ways — radiation, conduction, convection and evaporation — and the absolute rate of each is set by the skin and the environment, not by effort [Cramer & Jay 2016]. Convection and evaporation both depend on air moving across the skin, and evaporation is the only one that can carry away the 700 to 1,000 W of heat a hard session produces. Sweat that sits on the skin cools nothing. Sweat that evaporates does, at a rate set by air velocity and humidity at the skin surface. Indoors, that surface layer stops moving. Twelve cyclists riding 40 km in a laboratory versus outdoors had the same core temperature in both, but skin temperature was 33.0 °C in the lab against 31.4 °C outside, with air movement of 0.0 m/s versus 2.5 m/s [Mieras et al. 2014]. A warmer skin means a smaller gradient between core and surface, so the same heat takes more blood flow to move. That is the mechanism behind the higher heart rate at the same watts that riders notice on the trainer, and the sibling spoke on why heart rate runs higher indoors walks through the numbers in detail. Temperature then sets the ceiling. Galloway and Maughan's inverted-U result — best at 10.5 °C, worse at 3.6 °C, 20.6 °C and much worse at 30.5 °C — was measured with air velocity around 0.7 m/s, which is roughly what a spare room with the door shut and no fan provides [Galloway & Maughan 1997]. At 250 W in still air, the 20 cm of air in front of the rider warms and saturates within minutes, and the room is no longer 21 °C where it counts. ## How much airflow is enough Riding-speed air, and no more. Saunders and colleagues cycled nine subjects for two hours at 33 °C and 59% humidity with facing air at 0.2, 9.9, 33.3 and 50.1 km/h; heat storage, body temperature and effort were higher at the two slow speeds than at the two fast ones, and 33 and 50 km/h were indistinguishable [Saunders et al. 2005]. A fan you can feel across your whole torso and face at a brisk riding pace is the target. The shape of that result is the useful part. Ten km/h — a gentle breeze, what a small desk fan delivers at a metre — performed like no fan at all. The step from 10 to 33 km/h did all the work; the step from 33 to 50 did none. Nothing in the study translates into a fan wattage or an airflow rating, and the conditions were hot and humid rather than a heated house in January, so the honest reading is a principle: air that feels like a steady road pace on the face, not a flutter. If a sheet of paper taped to the handlebar barely moves, the fan is at the 10 km/h end of the curve. The same study answered the hydration question by accident. At the high airflow settings, drinking more fluid did not change heat storage [Saunders et al. 2005]. A bottle replaces what the sweat cost; it does not cool the rider, because the cooling was never in the water. Drink to the session, cool with air. Size follows from that. A fan that only moves air at the head leaves the largest sweating surfaces — chest, back, thighs — in still air, so a floor or pedestal fan with a wide face outperforms a small high-velocity unit aimed at the eyes. The two cheapest ways to get coverage are a bigger fan and a shorter distance. Inside those constraints, fan choice is taste, noise and floor space; none of the studies here would separate two fans that both put riding-speed air across the torso. ## Placement, count and timing Front and slightly below face height, aimed at the torso, running before the warm-up starts. Wingo and colleagues let cardiovascular drift develop over 45 minutes of cycling at 35 °C, then switched on a 4.5 m/s fan; skin temperature dropped about 1 °C, and VO2max still fell 15% against 17% with no fan at all [Wingo et al. 2019]. Once drift has happened, cooling does not undo it. Drift is why timing matters. After 10 to 20 minutes of steady work, stroke volume falls and heart rate climbs to hold cardiac output, driven primarily by the rising heart rate itself rather than by blood moving to the skin [Coyle & González-Alonso 2001]. A fan that starts when the sweat starts is starting after the drift, and the Wingo result says the drift stays. So the fan switch is part of the setup, next to the trainer pairing, and it goes on before the first pedal stroke. The same goes for the room: a window opened before the session means the fan is not moving air already at skin temperature. One fan placed well beats two placed badly. In front of the bike, low enough to hit the chest rather than the forehead, angled slightly upward so the stream crosses the torso and reaches the face, at a distance where the rider can feel it across the whole upper body. In a warm room, or for sessions past 90 minutes, a second fan behind or to the side covers the back and legs, which are large surfaces the front fan misses. Two moderate fans covering the body beat one strong fan covering the head. A plan makes this easier to keep. A structured indoor cycling training session on a Wahoo ELEMNT, a Hammerhead Karoo or any trainer app runs to a script — warm-up, work, recovery — and the fan belongs at the top of that script. Across the 40 to 60 sessions of a November-to-February season, a fan that runs for every minute is a bigger intervention than any single workout choice, and the cheapest one available. The sibling spoke on how long trainer rides should be explains why heat dominates past 90 minutes even with good airflow. ## What the fan changes in your numbers Heart rate at a given power comes down, the gap between the two stops widening across the hour, and the session's numbers start to mean what they would mean outdoors. Cardiovascular drift is the mechanism [Coyle & González-Alonso 2001]; the fan does not remove it, but a well-cooled rider shows a fraction of the drift a hot one does. Decoupling is the signal that the setup is right. Decoupling is the percentage by which power-to-heart-rate slips between the first and second halves of a steady ride. Some of it is physiology; on a trainer a lot of it is cooling. A few percent across a well-cooled 90-minute endurance ride is ordinary. Double digits usually says the fan is too small, the room is too warm, or both, before it says anything about aerobic fitness. No study attaches a precise threshold to fitness, so read the number as a diagnostic of the environment first and the rider second. We report it on every ride long enough to measure, for that reason. The second change is in prescription. A heart-rate-anchored session in still air silently gets harder as drift progresses, because the same heart rate corresponds to a rising fraction of a falling maximum [Wingo et al. 2019]. With riding-speed air on the torso, a power target and a heart-rate ceiling agree with each other for the length of the session, which is the only condition under which either is worth prescribing. A rider who sees the ceiling hit at minute 40 of an endurance ride has, most winters, found the fan problem before finding a fitness problem. The last change is the one nobody measures: the session gets done. Perceived exertion was higher at the slow air speeds [Saunders et al. 2005], and perceived exertion is what ends indoor rides early. Buy the airflow, place it on the body, switch it on first. The brand is the least important decision in the room. ## Common questions **Do I really need a fan for indoor cycling?** Yes, for any session longer than a warm-up. In still air, riders at a moderate intensity lasted 51.6 minutes at 30.5 °C against 93.5 minutes at 10.5 °C [Galloway & Maughan 1997], and a trainer in a closed room drifts toward the warm end of that range within minutes because there is no airflow to carry heat away. The fan replaces the wind that road speed provides for free. **How powerful should a fan for the bike trainer be?** Strong enough that riding-speed air reaches the whole torso and face. Around 33 km/h of facing air matched 50 km/h for heat storage and effort in a two-hour cycling study, while 10 km/h performed like no fan at all [Saunders et al. 2005]. No study converts that into a wattage or airflow rating; judge it by feel, and prefer a wide fan close to the bike over a small one far away. **Where should the fan point when cycling indoors?** At the chest and face from the front, slightly below face height and angled up, close enough to feel across the upper body. In a warm room or for rides past 90 minutes, add a second fan behind or to the side for the back and legs. Coverage of sweating skin matters more than raw speed at one spot. **Should I turn the fan on before I start sweating?** Before the warm-up, every time. When a fan was switched on after 45 minutes of cycling in the heat, VO2max still fell 15% against 17% with no fan at all [Wingo et al. 2019]. Cardiovascular drift begins within 10 to 20 minutes [Coyle & González-Alonso 2001], and cooling applied after it has developed does not reverse it. ## References 1. **Saunders et al. 2005.** [The effects of different air velocities on heat storage and body temperature in humans cycling in a hot, humid environment](https://pubmed.ncbi.nlm.nih.gov/15743384/). Acta Physiologica Scandinavica. 2. **Galloway & Maughan 1997.** [Effects of ambient temperature on the capacity to perform prolonged cycle exercise in man](https://pubmed.ncbi.nlm.nih.gov/9309637/). Medicine & Science in Sports & Exercise. 3. **Mieras et al. 2014.** [Physiological and psychological responses to outdoor vs. laboratory cycling](https://pubmed.ncbi.nlm.nih.gov/24476776/). Journal of Strength and Conditioning Research. 4. **Cramer & Jay 2016.** [Biophysical aspects of human thermoregulation during heat stress](https://pubmed.ncbi.nlm.nih.gov/26971392/). Autonomic Neuroscience. 5. **Wingo et al. 2019.** [Fan cooling after cardiovascular drift does not reverse decrements in maximal oxygen uptake during heat stress](https://pmc.ncbi.nlm.nih.gov/articles/PMC6773240/). Temperature. 6. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. --- # How many trainer sessions a week does it take to maintain fitness through winter? URL: https://www.adaptcycling.com/guides/minimum-trainer-sessions-to-maintain-fitness Updated: 2026-09-03 Author: Jim Camut Two, if both are hard. That is the answer the maintenance studies give, and it has held since 1981: riders who trained six days a week for ten weeks kept their entire VO2max gain for a further 15 weeks on two sessions a week, provided intensity and duration did not change [Hickson & Rosenkoetter 1981]. The catch is in the word maintained. It applies to aerobic power. It does not apply to the ability to ride long, which slipped in every reduced-training study that measured it. This spoke of the indoor cycling training guide covers what the studies did, what maintenance hides, and what the two sessions should be. ## What the maintenance studies actually did Hickson's group ran the same experiment three times, each time cutting one training variable and holding the others. Frequency cut from six days to two or four held VO2max for 15 weeks [Hickson & Rosenkoetter 1981]. Duration cut from 40 minutes to 26 or 13 held it too [Hickson et al. 1982]. Intensity cut by a third did not [Hickson et al. 1985]. The build phase was identical across the three studies: 40 minutes a day, six days a week, ten weeks, alternating cycling and running at a work rate that raised VO2max by 10 to 25%. Then the reduced phase ran for 15 weeks, roughly the length of an indoor season. In the frequency study, the two-day and four-day groups both measured the same VO2max at weeks 5, 10 and 15 of reduced training as they had at the end of the build [Hickson & Rosenkoetter 1981]. Twelve sessions a fortnight had become four, and the number did not move. The duration study cut each session to 26 minutes or 13 minutes while keeping six days a week at the trained work rate. VO2max held in both groups, and so did short-term endurance of about five minutes. Even the 13-minute group, training a total of 78 minutes a week, kept its aerobic power for 15 weeks [Hickson et al. 1982]. Cardiac mass, which had grown 15 to 20% with training, stayed elevated in both groups. The intensity study is the one that matters for winter. Frequency and duration stayed at six days and 40 minutes; only the work rate dropped, by one third or two thirds. VO2max fell in both groups, further in the two-thirds group, and the training-induced growth in left-ventricular mass returned to baseline in both [Hickson et al. 1985]. The caveats are real: 12 or 13 subjects per study, average age 23, a cycling-and-running mix, laboratory ergometers. But three studies pointing the same way is a pattern, and the pattern is that intensity is the variable a reduced winter cannot give up. ## What maintained hides: aerobic power held, endurance slipped Every study that held VO2max also measured how long subjects could ride, and that number fell. Long-term endurance dropped 21% in the one-third intensity group (184 to 145 minutes) and 30% in the two-thirds group (202 to 141 minutes) [Hickson et al. 1985]. Even the 13-minute duration group, with VO2max intact, lost 10% of its ride-to-exhaustion time [Hickson et al. 1982]. So maintenance is a statement about one number. The engine's ceiling stays where the summer put it; the ability to hold a fraction of that ceiling for two hours does not. Hickson's own conclusion from the duration study was that the endurance-trained state is not regulated uniformly under reduced training — VO2max and five-minute power held while two-hour endurance decreased [Hickson et al. 1982]. A rider who tests well in March and fades at 90 minutes in the first spring group ride has run exactly this experiment. The modern name for the number that slipped is durability: the time of onset and magnitude of deterioration in physiological characteristics over prolonged exercise [Maunder et al. 2021]. Threshold, efficiency and the power-to-heart-rate relationship all measured fresh are not the same values measured in hour three, and the gap between them is trained by riding long. Two hard sessions a week do nothing for it. That is not a flaw in the two-session week; it is the boundary of what it maintains. The practical read is that a winter of minimum sessions protects the wrong number if the spring goal is long. For a rider whose events are 60 to 90 minutes, two hard sessions hold what matters. For a rider aiming at a 100-mile sportive or a five-hour gravel race, the long ride is the session that cannot be dropped for 15 weeks, and the pillar's endurance tier — one ride of 90 to 120 minutes on the trainer, longer when the weather allows — is the floor, not the ceiling. ## The two-session week that holds fitness One interval session and one endurance ride. The interval session accumulates 30 minutes or so at roughly 90% of peak heart rate: 4×8 minutes produced more than double the VO2peak gain of 4×4 or 4×16 in cyclists training six hours a week [Seiler et al. 2013]. The endurance ride is 90 minutes at 60 to 70% of FTP with the fan on from minute one. The 4×8 result deserves its detail because the intuition runs the other way. Thirty-five trained recreational cyclists did two interval sessions a week for seven weeks at 4×4, 4×8 or 4×16 minutes, each at the hardest sustainable intensity: 94%, 90% and 88% of peak heart rate respectively. The 4×8 group improved VO2peak, power at VO2peak and power at 4 mmol lactate by 11.4%, against 4.2 to 5.6% in the other groups [Seiler et al. 2013]. Accumulating 32 minutes at 90% beat 16 minutes at 95%, and it hurt less. On a trainer in ERG mode, 4×8 with two-minute recoveries fits inside 60 minutes with warm-up and cool-down. Short intervals are the alternative for riders who cannot face eight-minute blocks. Ten weeks of 30-second efforts with 15-second recoveries, effort-matched against 4×5 minutes and done twice a week, raised VO2max 8.7% against 2.6%, with moderate-to-large effects across 30-second, 5-minute and 40-minute power [Rønnestad et al. 2015]. Three sets of 13 repetitions with three minutes between sets is the published format. Alternate the two across the winter; the variety also lowers the monotony that the pillar flags as the indoor season's real risk. The endurance ride is where most two-session weeks go wrong, because on a trainer 90 minutes at 65% of FTP feels like work. Hold it anyway: this is the session that slows the durability loss above. A trainer hour carries roughly the pedalling of 75 to 90 road minutes — a rule of thumb, not a measurement — so 90 minutes indoors is a fair stand-in for a two-hour road ride. When a third hour appears in the week, spend it on a second endurance ride before a second interval session; the indoor cycling training plan for four hours a week lays out that version. ## Maintenance is not building Two sessions hold. They do not raise anything. Hickson's subjects kept the VO2max the ten-week build had given them and gained nothing further in 15 weeks [Hickson & Rosenkoetter 1981]. A rider whose spring goal is to be faster than last spring needs the winter to build, and the winter cycling training plan is the version of this cluster that does that. The distinction sets the honest expectation. Maintenance is the right winter for a rider with a new baby, a project that ends in February, or a season that finished with genuine fatigue; two hard hours a week and a clear conscience beat five half-hearted ones, and the intensity study says exactly why. It is the wrong winter for a rider who was 20 watts short of the front group in September and wants to close the gap. Holding the gap is what maintenance does. Building costs hours, and it costs them in a particular shape: mostly easy, some very hard, little in between. The self-coached indoor season the pillar describes keeps the same two sessions as its skeleton and adds easy volume around them — six hours holds and adds a little, eight builds. The mistake to avoid is adding the hours in the middle zone, which is the threshold model that lost to polarized training on the same or fewer hours in two separate trials. The two-session week scales up by adding zone 2, not by adding a third sweet-spot session. Whichever winter it is, the plan has to survive the weeks that go missing, because between November and February they will. We built AdaptCycling to treat a lost week as a restructure rather than a failure: connect the rides from a Wahoo, a Hammerhead, a Polar or intervals.icu, and the plan reads the week that happened, keeps the hard session where the evidence says it belongs, and rebuilds the rest around the calendar that now exists. Two sessions a week is a floor the evidence supports. Keeping them on the calendar is the part that takes a plan. ## Common questions **Can I maintain cycling fitness with two trainer sessions a week?** Yes, for aerobic power. Riders who dropped from six sessions a week to two held their full VO2max gain for 15 weeks as long as the sessions stayed as hard and as long as before [Hickson & Rosenkoetter 1981]. Long-ride endurance is the number that slips on that schedule, so add a longer easy ride if spring events run past 90 minutes. **Is it better to ride easy every day or hard twice a week in winter?** Hard twice a week. Keeping six days a week but cutting intensity by a third lost VO2max and 21% of long-term endurance in 15 weeks [Hickson et al. 1985]; cutting frequency to two days at full intensity lost nothing. Easy daily riding is a fine addition around two hard sessions. It is a poor substitute for them. **How short can a trainer session be and still count?** Shorter than most riders think, if the intensity holds. Sessions cut to 26 minutes a day maintained VO2max and five-minute endurance for 15 weeks; even 13 minutes held VO2max, though two-hour endurance fell 10% [Hickson et al. 1982]. A 40-minute trainer session with 4×8 minutes at 90% of peak heart rate is a complete hard day. **Which interval session is best for maintaining VO2max on the trainer?** Either 4×8 minutes at about 90% of peak heart rate, which outperformed both 4×4 and 4×16 in cyclists training six hours a week [Seiler et al. 2013], or 30-second efforts with 15-second recoveries in three sets of 13, which beat 4×5 minutes on effort-matched volume [Rønnestad et al. 2015]. Both run cleanly in ERG mode. Alternating them across the winter also breaks up monotony. ## References 1. **Hickson & Rosenkoetter 1981.** [Reduced training frequencies and maintenance of increased aerobic power](https://pubmed.ncbi.nlm.nih.gov/7219129/). Medicine & Science in Sports & Exercise. 2. **Hickson et al. 1985.** [Reduced training intensities and loss of aerobic power, endurance, and cardiac growth](https://pubmed.ncbi.nlm.nih.gov/3156841/). Journal of Applied Physiology. 3. **Hickson et al. 1982.** [Reduced training duration effects on aerobic power, endurance, and cardiac growth](https://pubmed.ncbi.nlm.nih.gov/6214534/). Journal of Applied Physiology. 4. **Maunder et al. 2021.** [The importance of 'durability' in the physiological profiling of endurance athletes](https://pubmed.ncbi.nlm.nih.gov/33886100/). Sports Medicine. 5. **Seiler et al. 2013.** [Adaptations to aerobic interval training: interactive effects of exercise intensity and total work duration](https://pubmed.ncbi.nlm.nih.gov/21812820/). Scandinavian Journal of Medicine & Science in Sports. 6. **Rønnestad et al. 2015.** [Short intervals induce superior training adaptations compared with long intervals in cyclists — an effort-matched approach](https://pubmed.ncbi.nlm.nih.gov/24382021/). Scandinavian Journal of Medicine & Science in Sports. --- # Garmin bought TrainingPeaks: what it actually means for your training URL: https://www.adaptcycling.com/guides/garmin-bought-trainingpeaks-what-it-means Updated: 2026-08-13 Author: Jim Camut On July 22, 2026, Garmin announced it had acquired TrainingPeaks and TrainHeroic from Peaksware, bringing roughly 120 staff into the company; financial terms were not disclosed. Nothing in your TrainingPeaks account changed that day — pricing, coach connections, and device sync all work as before. What changed is who owns the most load-bearing neutral layer in endurance training: a device company that competes directly with Wahoo, Polar, and Hammerhead. Here is what is known, what is genuinely unknown, and what a self-coached rider should do about it. ## What actually happened on July 22 Garmin acquired TrainingPeaks and TrainHeroic — the endurance and strength coaching platforms — from Peaksware Holdings on July 22, 2026. About 120 combined staff join Garmin; financial terms were not disclosed [Garmin 2026]. Garmin says both platforms continue to function as they do today [Triathlon Today 2026]. So far this is an ownership change, not a product change. The mechanics are simple. Peaksware Holdings, the Colorado parent company, sold its two sports platforms to Garmin: TrainingPeaks, the calendar-and-analytics surface most paid endurance coaching runs through, and TrainHeroic, its strength-and-conditioning counterpart. Roughly 120 employees move to Garmin, and neither side published a price [Garmin 2026]. For scale, Garmin makes the Edge head units and Forerunner watches a large share of TrainingPeaks users already own. What did not change at announcement: everything user-facing. Accounts, subscriptions, coach relationships, Marketplace plans, and sync from non-Garmin devices all continue — Garmin's stated position is that both platforms will continue to function as they currently do [Triathlon Today 2026]. That continuity is rational. TrainingPeaks' commercial asset is a network of working coaches billing athletes through the platform; breaking their delivery rail on day one would burn the thing Garmin just bought. The reason this lands differently from routine SaaS consolidation is neutrality. TrainingPeaks became the standard pipe of paid endurance coaching by treating every device identically — it integrates with rival hardware makers including Wahoo, Hammerhead, and Coros [SGI Europe 2026]. A coach prescribes one workout and it lands on whatever head unit each athlete owns. As of July 22, that neutral layer belongs to one of the hardware companies it was neutral between. ## The open question: does vendor neutrality survive? Nobody outside Garmin knows, and Garmin has not committed. Asked what changes, TrainingPeaks' managing director said athletes and coaches would keep using devices and services 'that best suit their needs' — no operational detail, no binding promise [SGI Europe 2026]. Today a Wahoo file syncs exactly as it did on July 21. The question is the next product cycle, not this one. Neutrality is not a nice-to-have for this platform; it is the business model. A coach's roster might span 30 athletes on a mix of Garmin, Wahoo, Polar, and Hammerhead hardware, and TrainingPeaks' job is to make that mix invisible. The concrete worry, raised across industry coverage within days of the deal: whether workout export and activity sync for competing devices stay on equal footing, or whether Garmin hardware quietly becomes the first-class citizen [SGI Europe 2026, Triathlon Today 2026]. No restriction has been announced. No guarantee has been given either. Both halves of that sentence matter. The precedent everyone cites is Firstbeat. On June 30, 2020, Garmin acquired Firstbeat Analytics, the Finnish physiology engine whose metrics — VO2 max estimates, training load, recovery time — were licensed across the wearable industry [Garmin 2020]. Coverage of the TrainingPeaks deal explicitly echoes the question asked then: whether Garmin would keep licensing those metrics to competitors [SGI Europe 2026]. Six years later the pattern is visible on spec sheets: Firstbeat's newest work ships on Garmin's own watches, and the broad licensing business that defined the company before 2020 faded. The precedent is suggestive, not conclusive. Firstbeat was a licensing engine Garmin could absorb without breaking anyone's daily workflow; TrainingPeaks is a two-sided marketplace where degrading a rival device's sync would push working coaches — and their recurring revenue — toward competing platforms the day it happened. The incentives point in different directions, which is exactly why honest analysis stops at 'unknown.' Treat anyone claiming certainty in either direction as selling something. ## The developer-program pause nobody noticed Months before the deal, Garmin stopped admitting new partners to its Connect Developer Program — the API third-party apps use to read Garmin data. The official line, relayed on Garmin's own forums, is that it has 'temporarily paused the review and approval of new API access requests,' with no timeline given [Garmin Forums 2026]. The timeline matters. The pause was in force through spring 2026 — one developer who reapplied in April 2026 was still receiving the same no-timeline response in August [Garmin Forums 2026]. The freeze applies to new applications, not established integrations. Roughly four months after it took hold, Garmin announced the TrainingPeaks acquisition. A pause is not a strategy announcement, and connecting the two is inference — but a platform's posture toward third-party access is data, and the current posture is a closed door with no reopening date. The strategic logic explains why this question will not resolve quickly. Hardware commoditizes; analysts frame the deal as Garmin assembling 'a full-stack platform spanning devices, software, and coaching,' where insight rather than hardware is the differentiator [Fitt Insider 2026]. One trade analysis is blunter: the recurring subscription revenue 'actually lives' in TrainingPeaks and TrainHeroic, not in the devices [SGI Europe Strategy 2026]. TrainingPeaks also holds decades of coach-written plans and athlete compliance records — if Garmin builds the AI coaching analysts expect, that archive is the obvious training material. None of that logic depends on staying neutral. ## What a self-coached rider should actually do about it Nothing drastic. Migrating a working setup over an unannounced risk is a worse trade than watching two concrete signals for the next 12 months: whether workout export to non-Garmin head units stays feature-equal, and whether the developer program reopens. Meanwhile, keep the parts of your stack that carry switching costs portable. The two signals are cheap to monitor and hard to fake. First, feature parity: if structured-workout export to a Wahoo or Hammerhead unit ever lags what a Garmin Edge gets — new fields, new workout types, faster sync — that is the neutrality answer, delivered by product instead of press release. Second, the Connect Developer Program: reopening it to new partners would signal an open platform louder than any statement, and another year closed says the opposite. Watch behavior, not communications. The durable lesson belongs to the larger project of self-coaching. Our self-coached cyclist playbook frames the whole discipline as deciding which tools replace which parts of a coach — and this acquisition is the reminder that every tool in that stack has an owner whose incentives can change in one morning. The defensive moves cost nothing: know where your ride history lives (five seasons of data should never be hostage to one platform), and know which layer owns your plan, because the plan is where switching costs actually accumulate. Where we sit in this: AdaptCycling is not a TrainingPeaks replacement and does not try to be. TrainingPeaks remains the best analytics and coach-delivery surface in the sport, now under new ownership. We occupy the layer it deliberately leaves to a coach — reading your ride history, writing the periodized plan, and rebuilding the week when life disrupts it, at $15 a month and free to start. We compare the two head-to-head elsewhere on the site. The acquisition does not change that division of labor; it concentrates ownership of one side of it. ## Common questions **Did anything change in my TrainingPeaks account when Garmin bought it?** No. Garmin's stated position is that TrainingPeaks and TrainHeroic continue to function as they currently do [Triathlon Today 2026]. Pricing, coach relationships, Marketplace plans, and sync from non-Garmin devices were all unchanged at the July 22 announcement. The things worth watching are longer-horizon: export parity for competing head units and whether Garmin's paused developer program reopens. **Will Wahoo and Polar devices keep syncing with TrainingPeaks?** Today, yes — nothing has been restricted. Long term, nobody outside Garmin knows. Asked directly, TrainingPeaks offered only that athletes and coaches would keep using the devices 'that best suit their needs,' with no operational commitment attached [SGI Europe 2026]. The practical tell will be feature parity: whether workout export to a Wahoo or Hammerhead unit keeps pace with what Garmin's own Edge units get. **How much did Garmin pay for TrainingPeaks?** Not disclosed. Garmin's announcement confirms the acquisition of TrainingPeaks and TrainHeroic from Peaksware and the transfer of roughly 120 combined staff, but neither company published financial terms [Garmin 2026]. Garmin is a public company, so a deal material to its financials would have required disclosure; the silence suggests a price that is small by Garmin's standards. **Should I leave TrainingPeaks because of the acquisition?** Not on this news alone. Nothing has degraded, and migrating a working coach relationship over an unannounced risk is a bad trade. If TrainingPeaks is your coach's delivery surface, stay and watch the signals. The honest caveat predates July 22: if what you actually need is a plan written and adapted for you, TrainingPeaks never did that job — under Peaksware or under Garmin. ## References 1. **Garmin 2026.** [Garmin acquires TrainingPeaks and TrainHeroic, leading endurance and strength training platforms for athletes and coaches](https://www.garmin.com/en-US/newsroom/press-release/corporate/garmin-acquires-trainingpeaks-and-trainheroic-leading-endurance-and-strength-training-platforms-for-athletes-and-coaches/). Garmin Newsroom. 2. **SGI Europe 2026.** [Garmin adds coaching to its wearables business](https://www.sgieurope.com/mergers-and-acquisitions/garmin-adds-coaching-to-its-wearables-business/122408.article). Sporting Goods Intelligence Europe. 3. **SGI Europe Strategy 2026.** [Garmin CIRQA: the real story might be elsewhere](https://www.sgieurope.com/strategies/garmin-cirqa-the-real-story-might-be-elsewhere/122414.article). Sporting Goods Intelligence Europe. 4. **Triathlon Today 2026.** [Major acquisition in the world of endurance sports: Garmin acquires TrainingPeaks and TrainHeroic](https://tri-today.com/2026/07/major-acquisition-in-the-world-of-endurance-sports-garmin-acquires-trainingpeaks-and-trainheroic/). Triathlon Today. 5. **Fitt Insider 2026.** [Garmin Acquires TrainingPeaks and TrainHeroic to Turn Wearable Data Into Coaching](https://insider.fitt.co/garmin-acquires-trainingpeaks-trainheroic/). Fitt Insider. 6. **Garmin 2020.** [Garmin acquires Firstbeat Analytics, a leading provider of physiological analytics for health, fitness and performance](https://www.garmin.com/en-US/newsroom/press-release/corporate/2020-garmin-acquires-firstbeat-analytics-a-leading-provider-of-physiological-analytics-for-health-fitness-and-performance/). Garmin Newsroom. 7. **Garmin Forums 2026.** [Garmin Connect API access paused for months — what are startups supposed to do?](https://forums.garmin.com/apps-software/mobile-apps-web/f/garmin-connect-mobile-andriod/441607/garmin-connect-api-access-paused-for-months-what-are-startups-supposed-to-do). Garmin Forums. --- # Critical Power vs FTP: why your two numbers disagree, and which one should anchor your training URL: https://www.adaptcycling.com/guides/critical-power-vs-ftp-explained Updated: 2026-07-21 Author: Jim Camut Your software shows you two numbers: a Critical Power fitted from your power curve, and an FTP from a 20-minute test. They are not the same construct, and for you specifically the gap can run up to about 36 watts in either direction. Across three studies of trained cyclists the average gap was +7 W, +16 W, and -3 W. Small, and inconsistent even in sign. That is exactly why no correction factor exists. Here is what the agreement data shows, and which number should anchor your prescriptions. ## CP and FTP are not the same construct Critical Power is the asymptote of your power-duration curve — a modelled boundary above which the physiological response stops being steady-state [Jones & Vanhatalo 2017]. FTP is an operational convention: roughly the power you could hold for an hour, usually estimated as 95% of a 20-minute effort. One is a fitted parameter. The other is prescription shorthand. The distinction is not academic hair-splitting, because the two have very different evidence behind them. Jamnick and colleagues reviewed the methods used to mark the boundary between the heavy and severe intensity domains and concluded there is little evidence supporting the validity of most commonly used approaches — with critical power and critical speed named as the exceptions [Jamnick et al. 2020]. FTP is not among the validated markers. It is a field convention that has proven extremely useful for organising training, which is a different claim from being a measured physiological threshold. FTP is also not one-hour power, despite the definition. Wong and colleagues had 13 cyclists ride to exhaustion at their measured FTP and got 33.7 plus or minus 7.6 minutes — barely half the implied hour. Add 15 watts and time to exhaustion collapsed to 22.0 plus or minus 5.7 minutes [Wong et al. 2022]. The authors concluded FTP is not a valid marker of the maximal metabolic steady state. So when your software puts CP and FTP side by side, it is comparing a modelled domain boundary against a number whose own definition does not survive a stopwatch. A gap between them is the expected output of that comparison, not a fault in either estimator. ## What the agreement studies actually found Three agreement studies in trained cyclists, all reaching the same verdict: not interchangeable. Karsten found CP 7 W above FTP [Karsten et al. 2021]. McGrath found it 16 W above [McGrath et al. 2021]. Morgan found it 3 W below [Morgan et al. 2019]. Every one of them reported individual limits of agreement spanning 40 watts or more. Karsten and colleagues tested 17 trained cyclists and triathletes: CP 256 plus or minus 50 W against FTP 249 plus or minus 44 W. Mean bias +7 W, about 2.8%. Correlation r = 0.969. And 95% limits of agreement of -19 to +33 W [Karsten et al. 2021]. Read those last two numbers together. The correlation is near-perfect, and the interval inside which any individual rider's gap will fall is 52 watts wide. Both facts come from the same 17 riders. McGrath and colleagues, working with highly-trained athletes, found CP 282 plus or minus 53 W against FTP 266 plus or minus 55 W — a 16 W bias, roughly 6%, at p < 0.001, exceeding the agreement thresholds they had set in advance [McGrath et al. 2021]. Morgan and colleagues went the other way: 12 competitive male cyclists, CP 275 plus or minus 40 W against FTP 278 plus or minus 42 W, a non-significant -3 W bias, limits of agreement -36 to +30 W [Morgan et al. 2019]. That non-significant result is worth stating carefully. It is no evidence of a difference in the group mean. It is not evidence that the two numbers are the same, and the 66-watt agreement band in that same study is the proof. This is the association-versus-prediction distinction, and it is the whole article. A correlation of 0.969 answers a ranking question: among these riders, does a high CP travel with a high FTP? Yes, almost perfectly. The limits of agreement answer a prediction question: given one rider's FTP, how tightly can I pin their CP? Within roughly 40 to 66 watts, depending on which cohort you ask. High correlation paired with wide limits of agreement is the signature of a metric that describes a cohort well and prescribes to an individual badly. On a 250-watt rider, a 30-watt error moves a sweet-spot target from a productive 218 W to a session-ending 244 W. ## Why there is no correction factor The group-mean gap is small and its sign flips between studies: +7 W, +16 W, -3 W. A constant that adds 6% in one cohort subtracts 1% in another, while individual limits of agreement run 40 to 66 watts wide in all three. There is nothing stable enough to correct with. McGrath and colleagues did propose a correction factor, derived from the bias they measured in their own athletes [McGrath et al. 2021]. That is a defensible thing to publish and a dangerous thing to generalise. Apply their +16 W adjustment to Morgan's cyclists, where CP actually sat 3 W below FTP, and you have made the estimate 19 watts worse on average [Morgan et al. 2019]. A correction factor that flips sign between two cohorts of trained cyclists is not describing a physiological offset. It is describing the two protocols each lab happened to run. Which is the second reason to distrust a tidy conversion: the CP your software reports is partly a protocol artifact. The reference protocol is 3 to 5 fresh maximal trials on separate days, each exhausting the rider in roughly 2 to 15 minutes [Jones & Vanhatalo 2017]. Fitting CP from a whole-season best-effort power curve — what nearly every consumer tool does, including us — violates every clause of that. Season data mixes fresh efforts with fatigued ones, race-day motivation with Tuesday-night motivation, and durations chosen by the terrain rather than by the protocol. Borszcz and colleagues, meta-regressing 36 studies, found agreement between critical power and the maximal lactate steady state shifted with the longest predictive trial duration, the type of predictive trials, and the model-fitting parameters [Borszcz et al. 2024]. Change the protocol and you change the number. ## Which number should anchor your training Anchor to one number and know which one it is. If your intervals are prescribed as a percentage of FTP, CP is a cross-check, not a substitute — and vice versa. Do not spend a season trying to reconcile the two into a single true watt figure. The disagreement is structural, and no amount of extra data closes it. The failure mode is specific and easy to miss. Take a 4x8 prescribed at 105% of a 249 W FTP: 261 watts. Recompute the same session off a 256 W CP and you get 269 watts. Eight watts of silent drift, entirely because both numbers were sitting in the app under the word threshold. The mistake is never having two numbers on screen. The mistake is letting a target computed against one of them get applied to the other, which is what happens every time a rider copies an interval target out of one platform and into another. This sits inside the larger argument our ftp-without-a-test pillar makes — that estimating threshold from the data you already have is good enough for nearly all self-coached training, and a forced test is rarely the best use of a Saturday. The caveat this spoke adds is about what that estimate is. Whatever the model reports, CP or FTP, is a parameter estimate carrying a confidence interval, not a measurement. Treat your anchor as stable rather than exact, and let it move slowly, on trend rather than on any single ride. Here is our own position, stated plainly. We fit CP from a season-best mean-maximal-power curve using efforts between 2 and 20 minutes, drop the longest effort when leaving it out shifts CP by more than 5% — the signature of a 20-minute personal best that was not actually maximal — floor the result at 85% of a demonstrated 5-minute maximum, and then convert CP to an FTP-equivalent using a fixed 0.97 multiplier. That multiplier is precisely the correction factor this article has just argued does not exist for an individual. We use it because it is the best available population approximation, and because a field-fitted number that actually moves your training beats a lab number you will never collect. It is a pragmatic field approximation, not the Jones and Vanhatalo protocol [Jones & Vanhatalo 2017], and we are not going to pretend otherwise. Intervals.icu, Xert, TrainingPeaks and Garmin all make some version of the same compromise. The part we will commit to is naming it. ## Common questions **Is Critical Power the same thing as FTP?** No. Critical Power is the asymptote of your power-duration curve, fitted from maximal efforts. FTP is a convention approximating one-hour power, usually taken as 95% of a 20-minute test. They correlate strongly in trained cyclists — r = 0.969 in one cohort — but the 95% limits of agreement in that same study ran -19 to +33 W [Karsten et al. 2021]. Strong relationship, unreliable individual conversion. **My CP reads 15 watts higher than my FTP. Which one is wrong?** Probably neither. A 15-watt gap sits comfortably inside the individual limits of agreement reported in every published comparison [Karsten et al. 2021, McGrath et al. 2021, Morgan et al. 2019]. It tells you the two estimators disagree about you by a normal amount. Chasing a third test to break the tie usually just adds a third number. **Can I convert CP to FTP with a multiplier?** Not reliably for one rider. The published mean biases are +7 W, +16 W and -3 W across three cohorts of trained cyclists — a multiplier fitted to one of them is wrong in the others, sometimes in the opposite direction [Karsten et al. 2021, McGrath et al. 2021, Morgan et al. 2019]. Population conversions are fine for describing a group. They are not accurate enough to price an individual interval. **Should I train to CP or to FTP?** Train to whichever number your plan is anchored to, and keep it consistent. If your workouts arrive as percentages of FTP, switching the anchor to CP without recomputing every target shifts your whole intensity distribution upward. Consistency matters more here than picking the physiologically purer construct, because the prescription error from mixing anchors is larger than the difference between them. ## References 1. **Karsten et al. 2021.** [Relationship Between the Critical Power Test and a 20-min Functional Threshold Power Test in Cycling](https://pubmed.ncbi.nlm.nih.gov/33551839/). Frontiers in Physiology. 2. **McGrath et al. 2021.** [Do Critical and Functional Threshold Powers Equate in Highly-Trained Athletes?](https://pubmed.ncbi.nlm.nih.gov/34055164/). International Journal of Exercise Science. 3. **Morgan et al. 2019.** [Road Cycle TT Performance: Relationship to the Power-Duration Model and Association with FTP](https://pubmed.ncbi.nlm.nih.gov/30387374/). Journal of Sports Sciences. 4. **Jones & Vanhatalo 2017.** [The Critical Power Concept: Applications to Sports Performance with a Focus on Intermittent High-Intensity Exercise](https://pmc.ncbi.nlm.nih.gov/articles/PMC5371646/). Sports Medicine. 5. **Jamnick et al. 2020.** [An Examination and Critique of Current Methods to Determine Exercise Intensity](https://pubmed.ncbi.nlm.nih.gov/32729096/). Sports Medicine. 6. **Wong et al. 2022.** [Functional Threshold Power is Not a Valid Marker of the Maximal Metabolic Steady State](https://pubmed.ncbi.nlm.nih.gov/36803419/). Journal of Sports Sciences. 7. **Borszcz et al. 2024.** [Agreement Between Maximal Lactate Steady State and Critical Power in Different Sports: A Systematic Review and Bayesian Meta-Regression](https://pubmed.ncbi.nlm.nih.gov/38781475/). Journal of Strength and Conditioning Research. --- # Can you hold your FTP for an hour? What the research actually shows URL: https://www.adaptcycling.com/guides/can-you-hold-ftp-for-an-hour Updated: 2026-07-21 Author: Jim Camut You were told FTP is your one-hour power, you went out to prove it, and you came apart at 35 minutes. That is a normal result, not a failure. When researchers rode trained cyclists to exhaustion at their own measured FTP, the mean was 33.7 minutes in one study [Wong et al. 2022] and 50.9 minutes in another [Borszcz et al. 2018] — with individual spreads of 8 to 16 minutes around each mean. Neither reached the hour. Here is what your FTP number actually contains, and what it never did. ## Where the one-hour idea came from The hour is a definition of convenience, not a measurement anyone took on you. Coggan and Allen defined FTP as the highest power sustainable in a quasi-steady state for approximately one hour [Allen et al. 2019]. Almost no rider's FTP was ever established by riding for an hour — it came from a 20-minute effort multiplied by 0.95. The definition was pointing at a physiological boundary, not a stopwatch. FTP was meant to mark the divide between the heavy and severe intensity domains — the wattage below which lactate and oxygen uptake settle, and above which they climb until you stop. That boundary is a real construct, though in the literature it is critical power that marks it, not FTP [Poole et al. 2016]. The hour was shorthand for a long time near that boundary, chosen because it is a duration cyclists understand. It was never the thing being measured. The number you actually own came from a protocol. The 20-minute test takes 95% of your best 20-minute average [Allen et al. 2019]. TrainerRoad's ramp test takes 75% of peak one-minute power. TrainingPeaks watches for a new 20-minute best and suggests a threshold increase at the same 0.95, which you accept or ignore. Zwift and TrainerRoad both ship ramp and 20-minute variants. Not one of these protocols involves riding for 60 minutes, so testing the result against a 60-minute effort holds the number to a claim it never made. The boundary itself is real; the duration bolted onto it was always loose. Poole and colleagues note that critical power was originally defined as the output sustainable indefinitely, or for a very long time without fatigue, then state plainly that this definition should be considered theoretical, since no exercise can ever be undertaken indefinitely [Poole et al. 2016]. What the boundary separates is power outputs for which exercise tolerance is predictably limited from those for which it is not. ## What time-to-exhaustion at FTP actually measures Two studies rode cyclists to failure at their own FTP. Wong and colleagues measured 33.7 plus or minus 7.6 minutes in 13 cyclists [Wong et al. 2022]. Borszcz and colleagues measured 50.9 plus or minus 15.7 minutes in 23 trained cyclists [Borszcz et al. 2018]. Neither mean reaches 60. Blowing up at 35 minutes puts you inside the published distribution. The Wong result is the sharper one. Thirteen cyclists rode to task failure at FTP, then again at FTP plus 15 watts. Tolerance fell from 33.7 minutes to 22.0 plus or minus 5.7 minutes — a 15-watt step cut the sustainable duration by a third. End-test blood lactate differed significantly between the two intensities, roughly 6.7 against 9.2 millimolar. The authors concluded that FTP should not be considered a threshold marker between the heavy and severe intensity domains [Wong et al. 2022]. The gap between 33.7 and 50.9 minutes is not a contradiction to resolve — it is the finding. Different cohorts, different FTP determination protocols and different laboratories produced means 17 minutes apart. If time-to-exhaustion at FTP were a stable property of the metric, two studies of trained cyclists would not land that far from each other. Both figures are group means with wide standard deviations, which describes a population rather than predicting any individual inside it. The rest of the threshold literature sits in the same band. Jones and colleagues, reviewing critical power as the maximal metabolic steady state, describe exercise at CP as sustainable for no more than approximately 20 to 30 minutes [Jones et al. 2019]. Poole and colleagues put the ceiling near 30 minutes for efforts above the boundary [Poole et al. 2016]. Whether you anchor to FTP or to CP, measured tolerance clusters between 20 and 50 minutes. The hour sits outside the range in every dataset here. ## Why hold-duration is individual and your FTP number doesn't contain it Two riders with identical 300-watt FTPs can differ by 15 minutes or more in how long they hold it. The standard deviations say so: plus or minus 7.6 minutes in one cohort [Wong et al. 2022], plus or minus 15.7 in the other [Borszcz et al. 2018]. Duration is a second, independent property, and your FTP does not encode it. The mechanism is where your estimate lands relative to your true boundary. If your FTP sits a few watts below your critical power, you are riding in the heavy domain and tolerance is long. If it sits a few watts above, you are in the severe domain from the first pedal stroke, spending a finite anaerobic work capacity that empties on a predictable schedule [Poole et al. 2016]. McGrath and colleagues showed CP and FTP are not interchangeable in highly trained athletes, and the offset between them is rider-specific [McGrath et al. 2021]. Same 300 watts on the screen, two different intensity domains. Measurement error stacks on top of that. Borszcz and colleagues compared FTP derived from a 20-minute test against mean power from an actual 60-minute time trial in 23 cyclists: the bias was only 1 to 5 watts, but the limits of agreement ran from 40 watts under to 32 watts over [Borszcz et al. 2018]. That is the difference between group agreement and individual prediction. The two methods agree on average and disagree on you, which is why the authors concluded they should not be used interchangeably on an individual basis. We should be straight about our own limits. AdaptCycling does not estimate your individual time-to-exhaustion at FTP, and we do not prescribe against one. We fit a critical-power model to the power curve in your ride history to estimate threshold, and that pipeline outputs a wattage, not a duration. Some platforms do publish a duration: TrainingPeaks lists Time to Exhaustion alongside its modelled FTP, defined as the predicted time an athlete can hold that power, and Xert ships a live time-to-exhaustion field. Intervals.icu reports eFTP, and Garmin's FTP estimate needs a paired power meter as well as a heart-rate strap. So this gap is ours, not the industry's — and where a platform does hand you a clock, it is a model's prediction, not a measurement of you. ## A better validity check than the hour test Stop using can I hold it for an hour as a validity test. Judge the number by whether prescriptions built on it land correctly — threshold work should feel like threshold work, and sweet spot should be finishable. That check costs one session. An hour-long time trial costs several days of recovery and settles nothing about the estimate. The functional test is 2x20 minutes at 95 to 100% of your FTP with 5 minutes of easy spinning between. If you complete both intervals at the same power, with the second hard but controlled, the number is working. If rep two disintegrates before minute 12, the estimate is high — most often because a ramp test flattered a large anaerobic capacity. If you finish comfortably and could plausibly have started a third, the estimate is low. That is a decision you can act on inside a single Tuesday. Cross-check against your own data rather than a fresh test. This is the broader argument our ftp-without-a-test pillar makes — estimating threshold power from the rides you have already done, without a formal test — and it carries more force once you accept that an hour-long time trial validates nothing. Ninety days of riding already holds dozens of near-maximal efforts across durations. If your modeled FTP and your best real 20 to 40 minute efforts disagree by more than about 5%, one of them is wrong and that is worth investigating. For pacing a long effort, treat the hour question as an open one about you specifically. If a 40-kilometre time trial or a 50-minute climb is the goal, ride the first attempt 5 to 10 watts under FTP and record where you actually came apart. Do that twice and you own a personal number no model gave you. The failure mode to retire is not a wrong FTP — it is the assumption that the number arrived with a duration attached. ## Common questions **Is my FTP wrong if I can only hold it for 35 minutes?** Probably not. In the Wong study, 13 cyclists riding at their measured FTP lasted 33.7 plus or minus 7.6 minutes before task failure [Wong et al. 2022]. Thirty-five minutes sits at the centre of that distribution. Your FTP is wrong only if the workouts prescribed from it stop landing — threshold intervals that are uncompletable, or sweet spot that feels easy. **How long should I actually be able to hold my FTP?** There is no honest single answer. Published means for trained cyclists run from 33.7 minutes [Wong et al. 2022] to 50.9 minutes [Borszcz et al. 2018], with individual standard deviations of 8 and 16 minutes respectively. Somewhere between 30 and 50 minutes is unremarkable. The only way to learn your own figure is to ride to failure at it, which is a costly session with little training benefit. **Why does 15 watts above FTP make such a large difference?** Because it moves you across an intensity boundary rather than up a gradient. Wong and colleagues measured time to task failure at 33.7 minutes at FTP and 22.0 minutes at FTP plus 15 watts, with end-test blood lactate significantly higher at the harder intensity [Wong et al. 2022]. Above the boundary you are drawing on a finite anaerobic work capacity that empties on a schedule [Poole et al. 2016]. **Does AdaptCycling estimate how long I can hold my FTP?** No. We fit a critical-power model to your power curve to estimate threshold power, and that produces a wattage, not a time to exhaustion. We do not model your individual hold duration and we do not prescribe against one. Others do — TrainingPeaks publishes a modelled Time to Exhaustion and Xert ships a live one — so treat this as a gap in our product rather than a limit of the field. ## References 1. **Wong et al. 2022.** [Functional threshold power is not a valid marker of the maximal metabolic steady state](https://pubmed.ncbi.nlm.nih.gov/36803419/). Journal of Sports Sciences. 2. **Borszcz et al. 2018.** [Functional Threshold Power in Cyclists: Validity of the Concept and Physiological Responses](https://pubmed.ncbi.nlm.nih.gov/29801189/). International Journal of Sports Medicine. 3. **Jones et al. 2019.** [The maximal metabolic steady state: redefining the gold standard](https://pubmed.ncbi.nlm.nih.gov/31124324/). Physiological Reports. 4. **Poole et al. 2016.** [Critical Power: An Important Fatigue Threshold in Exercise Physiology](https://pmc.ncbi.nlm.nih.gov/articles/PMC5070974/). Medicine & Science in Sports & Exercise. 5. **McGrath et al. 2021.** [Do Critical and Functional Threshold Powers Equate in Highly-Trained Athletes?](https://pubmed.ncbi.nlm.nih.gov/34055164/). International Journal of Exercise Science. 6. **Allen et al. 2019.** [Training and Racing with a Power Meter (3rd ed.)](https://www.velopress.com/books/training-and-racing-with-a-power-meter-3rd-ed/). VeloPress. --- # FTP stopped improving: are you a non-responder? URL: https://www.adaptcycling.com/guides/ftp-stopped-improving-non-responder Updated: 2026-07-21 Author: Jim Camut You have trained consistently for months and the number has not moved. Somewhere online you found the word non-responder, and it read like a verdict. It is not one. When one dataset was run through several accepted statistical methods for classifying responders, only 11 of 20 subjects kept the same label across all of them [Hecksteden et al. 2018]. Real between-person variation in training response almost certainly exists. A defensible way to tell one specific rider they are a non-responder does not. ## What the non-responder literature actually shows The idea comes from large training studies reporting wide spreads in VO2max gain. The critique is methodological, not physiological: most of those studies ran no control group, so within-subject noise was never separated from true response [Williamson et al. 2017]. That is no evidence of effect, which is a different claim from evidence of no effect. The canonical source is the HERITAGE Family Study, whose published gain distributions — some participants improving enormously, some barely at all — became the origin of the responder and non-responder vocabulary. Williamson and colleagues searched more than 180 HERITAGE publications and could not find a comparator arm in any of them [Williamson et al. 2017]. Without a control condition, a spread of observed changes cannot be decomposed into real individual response and ordinary test-retest variation. Their conclusion is blunt: true inter-individual differences in response cannot be quantified, let alone appraised for clinical relevance. The people who built HERITAGE agree with the design point. The 2019 precision exercise medicine consensus statement in the British Journal of Sports Medicine, co-authored by Claude Bouchard and James Skinner, states that designs without a control group cannot isolate changes due to treatment from changes that would have occurred without it [Ross et al. 2019]. That is the field conceding the methodology, in print, under the names of the investigators who generated the original data. The same paper carries the other half, and honesty requires carrying it too. Monozygotic twin resemblance and familial aggregation in the response data show that response variance is not randomly distributed — a familial component accounts for roughly 30% to 60% of it depending on the trait [Ross et al. 2019]. A genuine individual-response component is therefore almost certainly non-zero. But that finding is an association measured across a population. It is not a prediction about any one person, and no published method converts it into one. ## Why the label flips depending on who does the maths Hecksteden and colleagues took a single training dataset and applied several accepted analytical approaches for classifying responders. Only 11 of 20 subjects were consistently classified [Hecksteden et al. 2018]. Nine of twenty changed label on nothing but the statistic chosen. A classification that unstable cannot carry a verdict about a person. The detail that matters is that the data never changed. One year-long training study, one set of VO2max measurements, several defensible ways to draw the responder line — and the labels moved. The authors describe the disagreement between approaches as remarkable [Hecksteden et al. 2018]. If a rider is a responder or a non-responder depending on which reviewer's preferred method gets applied to identical numbers, the label is describing the analyst, not the athlete. Measurement error is the reason. In a pooled analysis of three randomised trials — 251 intervention participants against 87 controls — only 45% of participants exceeded twice the technical error of measurement for absolute VO2peak [Brennan et al. 2022]. In supervised, funded, laboratory-measured training, most participants did not produce a change large enough to separate from the equipment and the day — even though the group mean improved significantly. Brennan calls those participants uncertain, not unresponsive. Undetectable and absent are different things, and most published training data cannot tell them apart. Hecksteden's proposed remedy is worth knowing for what it implies. The fix is not more subjects; it is repeated testing during the training phase, so each individual carries their own error estimate [Hecksteden et al. 2018]. It was published as a proof of concept, and where others have applied it the picture did not improve: measuring fitness five times across 24 weeks, Bonafiglia and colleagues could confidently classify only 55 of 109 participants as responders [Bonafiglia et al. 2019]. No protocol specifies how many repeat tests one person needs for a stated misclassification rate, which means nobody currently holds the evidence that would settle it. ## The four things far more likely than being a non-responder Before reaching for genetics, check four things: the two FTP numbers came from different protocols, the stimulus never actually changed, fatigue was never resolved, or the window was too short. Each is more common than true non-response, and each is fixable. A January ramp test against a June 20-minute test is not a comparison. Measurement first, because it is the cheapest to rule out and the most often guilty. Warm-up structure alone shifts the FTP a 20-minute test reports [Tramontin et al. 2022], and the ramp test's assumption that FTP equals 75% of peak one-minute power is a protocol convention rather than a constant — the true ratio varies rider to rider. Two tests, two protocols, two states of readiness, and most of the difference between them is method. This is exactly the problem our ftp-without-a-test pillar addresses: an FTP modelled from ride history you already have is anchored to dozens of efforts across months rather than two isolated days, so a flat trend line means closer to what you think it means. Second, the stimulus. Eight months of the same sweet-spot template at the same three weekly durations is not eight months of progressive overload; it is one month repeated eight times. Adaptation follows a change in demand. If volume, session structure, and intensity distribution have all been stable since January, the honest reading is that the plan stopped asking for anything new — not that the body stopped answering. Third, unresolved fatigue. A threshold estimate taken while carrying accumulated load reads low, and a rider stacking hard weeks on five hours of sleep can hold a suppressed number for a full season without ever seeing the fitness underneath it. The diagnostic is cheap: take a genuine easy week, then re-measure. If the number jumps, the problem was recovery, not response. Fourth, the window. A change smaller than roughly twice the measurement error of your protocol cannot be told apart from noise [Brennan et al. 2022], and for most amateur field tests that band is wide enough to swallow a real gain. Six weeks is a training block, not an evaluation period. Twelve or more weeks, measured the same way each time, is the minimum before flat means anything — and switching apps or protocols mid-window resets the clock. ## What an honest answer looks like No protocol establishes how many repeat tests are needed to classify one person as a responder at a stated error rate. Until that exists, no coach, laboratory, or app — AdaptCycling included — can tell you that you are a non-responder. What can honestly be said is narrower: this stimulus, measured this way, did not move this number. The narrower statement is also the more useful one, because every term in it is something you control. The stimulus can change. The measurement method can be held constant. The window can be extended. The genetic verdict, by contrast, is unfalsifiable in practice — there is no test you can run on yourself that returns it, which is precisely why it is a comfortable thing to believe and a poor thing to act on. Response is also trait-specific, which makes single-number verdicts worse than they look. In the same pooled trial analysis, the share of participants exceeding twice the technical error ranged from 37% for glucose disposal to between 51% and 77% for body-composition measures [Brennan et al. 2022]. Same people, different traits, very different response rates. A rider whose FTP is flat may have added durability late in long rides, repeatability across intervals, or five-minute power, and would still be labelled a non-responder by an assessment that only ever looked at threshold. AdaptCycling does not classify riders as responders or non-responders, and will not until the literature supports it. What we do is remove two of the four confounders above by default: the FTP estimate is modelled continuously from your ride history rather than sampled from occasional tests, which holds the measurement method constant, and the plan rotates its stimulus on a periodized cadence rather than repeating one block indefinitely. It does not detect that your numbers have stalled — that read is still yours to make. That is not a genetic answer. It is the part of the question that is answerable. ## Common questions **Am I a non-responder to training?** There is currently no defensible way for anyone to answer that about you specifically. Classification of individuals is unstable — in one dataset run through several accepted analytical approaches, only 11 of 20 subjects were consistently classified [Hecksteden et al. 2018] — and no protocol establishes how many repeat tests would settle it for one person at a stated error rate. A flat FTP is far more likely to be a measurement, stimulus, or recovery problem. **What percentage of people are non-responders?** Every figure you will see quoted is method-dependent, which is the whole problem. Nine of twenty subjects in one analysis changed classification based purely on which accepted statistic was applied to the same data [Hecksteden et al. 2018]. Separately, only 45% of participants in a pooled trial analysis exceeded twice the technical error for absolute VO2peak [Brennan et al. 2022] — but that measures whether a change was detectable, not whether a person is incapable of adapting. **Does genetics affect how much I improve?** Almost certainly yes, at the population level. Twin and family data show response variance is not randomly distributed, with a familial component accounting for roughly 30% to 60% of it depending on the trait [Ross et al. 2019]. That is an association across groups, not a prediction for an individual. Knowing a familial component contributes to the spread tells you nothing about where you personally sit within it. **What should I change first if my FTP will not move?** Change one variable and hold the rest still. Start with a real recovery week and re-measure, because a suppressed number from unresolved fatigue is the fastest thing to rule out. If it does not move, change the stimulus rather than the volume of the same stimulus — different session structure, different intensity distribution — and give it twelve weeks measured the same way before judging. ## References 1. **Hecksteden et al. 2018.** [Repeated testing for the assessment of individual response to exercise training](https://pubmed.ncbi.nlm.nih.gov/29357481/). Journal of Applied Physiology. 2. **Williamson et al. 2017.** [Inter-Individual Responses of Maximal Oxygen Uptake to Exercise Training: A Critical Review](https://pubmed.ncbi.nlm.nih.gov/28097487/). Sports Medicine. 3. **Ross et al. 2019.** [Precision exercise medicine: understanding exercise response variability](https://pubmed.ncbi.nlm.nih.gov/30862704/). British Journal of Sports Medicine. 4. **Brennan et al. 2022.** [Toward Personalized Exercise Medicine: A Cautionary Tale](https://pubmed.ncbi.nlm.nih.gov/36007156/). Medicine & Science in Sports & Exercise. 5. **Bonafiglia et al. 2019.** [The application of repeated testing and monoexponential regressions to classify individual cardiorespiratory fitness responses to exercise training](https://pubmed.ncbi.nlm.nih.gov/30666410/). European Journal of Applied Physiology. 6. **Tramontin et al. 2022.** [Functional Threshold Power Estimated from a 20-minute Time-trial Test is Warm-up-dependent](https://pubmed.ncbi.nlm.nih.gov/34749416/). International Journal of Sports Medicine. --- # Durability: why your power fades late in a long ride even when your FTP is fine URL: https://www.adaptcycling.com/guides/durability-late-ride-power-fade Updated: 2026-07-21 Author: Jim Camut Your FTP is 280 and you can hold it in a 20-minute effort on a Tuesday. Four hours into a Saturday group ride the same wattage feels like 330. That gap has a name — durability — and since 2021 it has become the most-hyped number in cycling analytics [Maunder et al. 2021]. It is real, it is measurable from ride files you already own, and almost everything sold around it runs well ahead of the evidence. Here is what the research establishes and what it does not. ## What durability actually means and where the term came from Ed Maunder and colleagues coined the term in a 2021 Sports Medicine paper, defining durability as the time of onset and magnitude of deterioration in physiological-profiling characteristics during prolonged exercise [Maunder et al. 2021]. The definition is stated in time terms. It specifies no kilojoule criterion at all — that part was added later, by other people. The observation behind the term is mundane once stated. Everything in a standard physiological profile — VO2max, lactate threshold, gross efficiency, critical power — is measured on a fresh rider in a rested state. Maunder and colleagues argued those characteristics are not static, and that a profile ignoring their deterioration describes an athlete who no longer exists after hour two [Maunder et al. 2021]. Valenzuela and colleagues put a number on it: twelve male professional cyclists lost a mean 2.9% of time-trial power after roughly four hours of submaximal riding [Valenzuela et al. 2023]. The mean hides the interesting part. Individual responses in that same group ranged from an 8.5% loss to a 1.1% gain [Valenzuela et al. 2023]. One rider was measurably worse late; another was fractionally better. Both had elite fresh numbers. That spread is the entire case for measuring durability at all — if every trained rider decayed by the same 3%, the metric would be a constant and you could safely ignore it. It also explains a mismatch most self-coached riders feel before they have a word for it. A single FTP describes a rider at minute zero. The power-duration curve our ftp-without-a-test pillar shows you how to fit from ride history you already have — the estimate that needs no formal test — is the curve you carry into a ride, not the one you carry out of it. Hour four has its own curve, and it sits lower. ## Why the kJ number your app uses is arbitrary Every app reporting durability picks a work threshold — 1000 kJ, 2000 kJ, 40 kJ/kg — and calls everything past it fatigued. A systematic review catalogued the thresholds used across the literature and concluded that kilojoules alone are insufficient, because the metric ignores the intensity at which that work was done [Sanchez-Jimenez et al. 2025]. The number is an investigator's choice, not a physiological boundary. List them and the arbitrariness is hard to miss. Published work has used 2.5, 5 and 7.5 kJ/kg; 15, 25, 35 and 45 kJ/kg; a 0 to 50 kJ/kg sweep; roughly 40 kJ/kg; and a flat absolute 2000 kJ [Sanchez-Jimenez et al. 2025]. The most-quoted of them, 40 kJ/kg, traces back to a study with exactly two conditions: 0 and 40 [Valenzuela et al. 2023]. That study was not designed to locate a threshold and could not have, because you cannot find a breakpoint from two points. When the range is swept, the decline starts earlier. Mateo-March and colleagues tracked mean-maximal power across accumulated work from 0 to 40 kJ/kg in professional cyclists and found progressive decline detectable after 20 kJ/kg [Mateo-March et al. 2025]. The threshold in wide circulation is roughly double the point at which the effect becomes measurable, which means a durability score gated at 40 kJ/kg quietly discards the first half of the decline it claims to describe. The deeper problem is that total work is the wrong axis. The same review found efforts above critical power produced 10 to 20% power declines at only 2.5 to 15 kJ/kg, while similar or greater volumes at lower intensities produced smaller decrements [Sanchez-Jimenez et al. 2025]. A 2000 kJ steady endurance ride and a 2000 kJ road race leave you in completely different states. Our own gate is 20 kJ/kg of body mass — closer to Mateo-March than the 40 kJ/kg default, still a field heuristic, and still blind to how those kilojoules were earned. ## Is durability independent of FTP? The honest answer The marketing claim is that durability is a fourth axis beside FTP, VO2max and anaerobic capacity. The evidence is split. One study found change in critical power correlated 0.891 with relative peak power, 0.835 with relative VO2max and 0.869 with gross efficiency at 300 W [Spragg physiology 2023]. Another, in a similar cohort, found no association with any laboratory endurance measure [Valenzuela et al. 2023]. That evidence comes from ten under-23 professionals. The headline was an R-squared of 0.96 to 0.98 for predicting durability from physiological measures, but it came from backwards stepwise selection fitting three predictors with six residual degrees of freedom and no out-of-sample validation [Spragg physiology 2023]. A model at that ratio fits nearly anything you hand it. And it does not replicate. Valenzuela and colleagues asked the same question of twelve professionals and found no significant association between power decay and ventilatory threshold, peak power output or VO2max [Valenzuela et al. 2023]. Two small studies, opposite answers — that is what an open question looks like, not a settled one. The study usually cited as proof durability wins races is more careful than its reputation. Van Erp and colleagues examined 26 male professionals across 85 rider-seasons and found a smaller decline in power after accumulated work was associated with season-level competitive success, with the discriminating durations differing by specialization — 20-minute and 5-minute power for climbers, 10-second and 1-minute for sprinters [Van Erp et al. 2021]. The design was retrospective, exposure and outcome came from the same season, success was a ranking-points proxy, and the study never measured FTP or critical power at all. An association observed within a season, in a sample where nobody's threshold was measured, cannot establish prediction, and it cannot establish independence from a variable that was never collected. None of this makes durability fake. It makes its independence an open question rather than the settled fourth axis it is sold as. AdaptCycling computes a durability index — the median ratio of your best late-ride 20-minute power to a matched fresh 20-minute effort earlier in the same ride, across long rides in the past year — and reports it as descriptive, never as a verdict. One limit worth stating plainly: we score everyone on the 20-minute and 5-minute windows Van Erp found discriminating for climbers, so a sprinter or crit racer is being graded on the wrong pair. ## What we know about training it — and what to do in the meantime No controlled training trial in trained cyclists has measured a change in durability. The nearest work is uncontrolled dose-response, and the one randomised trial to move a durability measure ran in runners. That is no evidence of a trainability effect size, not evidence that durability cannot be trained. Every durability protocol sold to cyclists rests on observational data. The strongest observational signal comes from Spragg and colleagues tracking 30 professional cyclists across a competitive season: time spent below the first ventilatory threshold correlated with improvement in fatigued 2-minute power at r = 0.43 (p = 0.018) [Spragg training 2023]. That is the empirical basis for the standard advice to ride more easy volume. The caveats are load-bearing. The analysis was observational and uncorrected for multiple comparisons, and both the training exposure and the fatigued-power outcome were extracted from the same pooled training and race files — riders accumulating more easy hours were also riders racing and training more overall. We act ahead of that evidence, and it is worth saying so. When a learned durability index lands below 88% of the matched fresh effort, that deficit can pull a fatigue-resistance long ride into a build week — quality efforts placed after an aerobic pre-load rather than at the fresh start of the session. The reasoning is specificity: if you fade at hour three, practice the hard part at hour three. That is a defensible coaching argument. It is not a proven dose-response, and no controlled trial has shown that session moves the number. In the meantime, the intensity finding is the more actionable one. If work above critical power costs 10 to 20% at 2.5 to 15 kJ/kg while steady riding at the same or greater volume costs measurably less [Sanchez-Jimenez et al. 2025], then how you rode the first 90 minutes may matter more for your hour-four watts than how far you went. The rider who contests every roller in the opening hour of a five-hour ride has spent their durability before the ride is a third done. Audit pacing and fueling discipline before buying a training block aimed at a score nobody has yet shown is trainable. ## Common questions **Why does my power fade at hour three when my FTP is fine?** Because FTP describes you fresh. Professional cyclists lost a mean 2.9% of time-trial power after roughly four hours of submaximal riding, with individual responses ranging from an 8.5% loss to a 1.1% gain [Valenzuela et al. 2023]. Some fade is normal physiology. An unusually large fade is worth measuring against your own ride history rather than a population average. **How many kilojoules before durability starts to matter?** Nobody knows. The thresholds apps use vary widely across the literature, and a systematic review concluded kilojoules alone are insufficient because they ignore intensity [Sanchez-Jimenez et al. 2025]. The widely quoted 40 kJ/kg comes from a study with only two conditions, 0 and 40 [Valenzuela et al. 2023]. The one study that actually swept the range found progressive decline detectable after 20 kJ/kg [Mateo-March et al. 2025]. **Is durability a fourth training axis alongside FTP and VO2max?** Unresolved. One study of ten under-23 professionals found change in critical power correlated 0.891 with relative peak power and 0.835 with relative VO2max [Spragg physiology 2023]. Another found no association between power decay and any laboratory endurance measure [Valenzuela et al. 2023]. Two small samples, opposite results. Treat a durability score as informative, not as an established independent axis. **Can I train durability with long slow rides?** Possibly. One season-long observational study found time below the first ventilatory threshold correlated with improvement in fatigued 2-minute power at r = 0.43 [Spragg training 2023]. No controlled trial in trained cyclists has measured a change in durability, so this is no evidence of an effect size rather than evidence of no effect. ## References 1. **Maunder et al. 2021.** [The Importance of 'Durability' in the Physiological Profiling of Endurance Athletes](https://pubmed.ncbi.nlm.nih.gov/33886100/). Sports Medicine. 2. **Van Erp et al. 2021.** [Maintaining Power Output with Accumulating Levels of Work Done Is a Key Determinant for Success in Professional Cycling](https://pubmed.ncbi.nlm.nih.gov/33731651/). Medicine & Science in Sports & Exercise. 3. **Spragg physiology 2023.** [The Relationship between Physiological Characteristics and Durability in Male Professional Cyclists](https://pubmed.ncbi.nlm.nih.gov/35977108/). Medicine & Science in Sports & Exercise. 4. **Spragg training 2023.** [The relationship between training characteristics and durability in professional cyclists across a competitive season](https://pubmed.ncbi.nlm.nih.gov/35239466/). European Journal of Sport Science. 5. **Sanchez-Jimenez et al. 2025.** [Is intensity the most important factor in determining the amount of prior work accumulated that affects cyclists' acute durability? A systematic review](https://pubmed.ncbi.nlm.nih.gov/40613880/). European Journal of Applied Physiology. 6. **Mateo-March et al. 2025.** [Reliability of the durability concept in professional cyclists: a field-based study](https://pubmed.ncbi.nlm.nih.gov/40373793/). International Journal of Sports Medicine. 7. **Valenzuela et al. 2023.** [Durability in Professional Cyclists: A Field Study](https://pubmed.ncbi.nlm.nih.gov/36521188/). International Journal of Sports Physiology and Performance. --- # How to get a structured workout on an ELEMNT without TrainingPeaks URL: https://www.adaptcycling.com/guides/elemnt-structured-workout-without-trainingpeaks Updated: 2026-08-01 Author: Jim Camut The premise of this question is half wrong, and the half that is wrong is the expensive half. TrainingPeaks is not required to run a structured workout on a Wahoo ELEMNT — Wahoo's own requirements list names SYSTM, TrainingPeaks, TrainerRoad, or another third-party provider [Wahoo ELEMNT Support]. What a TrainingPeaks subscription buys is not the workout. It is the week: how many days of your plan the computer will hold at once. ## The four routes Wahoo actually supports Wahoo names four sources of planned workouts for an ELEMNT: SYSTM, TrainingPeaks, TrainerRoad, or another third-party provider [Wahoo ELEMNT Support]. Three of the four do not involve TrainingPeaks at all, and the computer ships with a small set of workouts before you connect anything. Start with what you already own. Wahoo includes a starter set with ELEMNT computers — two fitness tests (a 20-minute and an 8-minute FTP test) and seven workouts: 12x1min MAP, 2x20min FTP, 3 sets of 40/20s, 5x5min FTP, plus climbing, mixed, and sprinting sessions [Wahoo ELEMNT Support]. If your question is really "can I ride intervals off my head unit tonight," the answer needs no account anywhere. What the built-in set cannot do is progress: it is nine fixed sessions, identical in week one and week twelve. Linking a provider is where riders get stuck, because the direction of the handshake differs. For SYSTM, you use the same email address and password on the SYSTM and Wahoo apps. For TrainingPeaks, you enter your credentials in the Wahoo app under Authorized Apps. For any other third-party provider you authorize from the provider's side, signing into your Wahoo account in their app or web interface rather than looking for them in Wahoo's [Wahoo ELEMNT Support]. Riders who hunt for a missing provider inside the Wahoo app conclude the integration does not exist, when the connect button was on the other side the whole time. TrainerRoad is the one case Wahoo's own documentation disagrees on: the older ELEMNT article lists it in the app beside TrainingPeaks, while the ACE, BOLT 3 and ROAM 3 article groups it with the providers you connect from their side [Wahoo ELEMNT Support, Wahoo ELEMNT ACE Support]. The routes are not equivalent in what they will do for you, which is the actual decision. SYSTM brings a large structured library and the 4DP profile, and is built around indoor riding. TrainerRoad brings adaptive indoor plans. TrainingPeaks brings the deepest workout builder in the business and is what most coached athletes are already on. A third-party provider brings whatever that provider is good at — which, if it reads your completed rides, can include a plan that changes when your week does. ## What the subscription actually buys: the week, not the workout How many days of your plan reach the device is a property of your provider, not your computer. A free TrainingPeaks account syncs today's workout and nothing else. TrainerRoad syncs today plus three days. Everything else syncs today plus six [Wahoo ELEMNT Support]. That is the whole mechanism behind the folklore. TrainingPeaks Basic will put today's session on your ELEMNT, and after the first thirty days of a trial, a Premium subscription is required to sync a week of planned workouts [Wahoo ELEMNT Support]. So the rider who concludes they must pay TrainingPeaks to train off their head unit is half right: they must pay to see the week. The workout itself was never the gated thing. The same table applies on current hardware — ELEMNT ACE, BOLT 3, and ROAM 3 follow identical rules [Wahoo ELEMNT ACE Support]. Seven days is also the ceiling, and it belongs to Wahoo rather than to any provider. Today plus six is the most an ELEMNT will hold, so no service can put week nine of your plan on the handlebars in advance, and any that claims to is describing its phone app. In practice this matters less than it sounds: a plan worth following changes inside seven days anyway, and a week that was queued a month ago is a week that has not accounted for the ride you did on Saturday. Read the window as a filter on providers rather than a limitation. If a provider only ever gives you today, you are managing the plan on your phone and using the computer as a screen. If it gives you the full week and updates it, the computer becomes the thing you ride from — which is the difference between owning a plan and consulting one. ## Getting it onto the device, and why it sometimes doesn't Two transports, and they are not interchangeable. Wi-Fi pulls multiple days directly from the Wahoo Cloud to the computer; Bluetooth through the phone app moves exactly one workout at a time [Wahoo ELEMNT Support]. Most sync complaints are one of three specific failures. The first: only today's session appears despite a full week on the calendar. The usual cause is that the computer itself — not just the phone — never joined the Wi-Fi network. An ELEMNT shows a solid Wi-Fi symbol in its top bar when it is actually connected, and without that it can only receive the single workout Bluetooth carries [Wahoo ELEMNT Support]. This one failure explains most of the "my provider says it synced" reports. The second: the targets look wrong. Power targets arrive keyed to the zones set in your provider's app, and those can differ from the zones stored in your Wahoo account [Wahoo ELEMNT Support]. Update your threshold in one place and not the other and the computer will faithfully render targets computed from a number you no longer believe. The fix is to reconcile the two rather than scaling every session on the road. The third catches watch owners specifically. The ELEMNT RIVAL is a different pipeline: it takes planned workouts from TrainingPeaks only, syncs through the companion app rather than over Wi-Fi, and holds today plus the next five days [Wahoo RIVAL Support]. Every route in this article reaches an ELEMNT computer. On a RIVAL, the TrainingPeaks requirement is real. ## Picking a source when you don't want another subscription Three honest cases. If a coach already writes your plan, stay on TrainingPeaks and treat Premium as the cost of the week. If you train indoors and want structure with production value, SYSTM. If you want the plan to come from the riding you actually do, use a provider that reads your rides back. The case for staying put is stronger than the internet suggests. A rider working with a coach has the plan written where the coach works, and moving that to save a subscription creates a second place for the truth to live. The Premium gate is annoying, not irrational — it is the price of the seven-day window on the device, and if someone else is authoring your week it is money well spent. The case for changing is about where the plan comes from. SYSTM is indoor-first by design and does not rebuild a plan around your outdoor ride history; TrainerRoad's adaptation happens within indoor power workouts and syncs a four-day window. Neither is a flaw — they are what those products are for. But if your training is mostly outdoors, or if what breaks your plan is a sick kid and a work trip rather than a missed interval, the thing you need is not a bigger workout library. It is a plan that rewrites itself around the week you had, which is the subject of our broader guide to training with Wahoo. That is the gap AdaptCycling was built for. Connect a Wahoo account and we read your ride history, estimate your threshold from your own power curve rather than a maximal test — the critical-power relationship makes that possible from efforts you have already ridden [Jones et al. 2010] — and push today plus the next six days to your Wahoo account as structured workouts. When the plan changes, the revised sessions replace the old ones on the device. There is no TrainingPeaks subscription in the chain, and the device push runs on the free tier. ## Common questions **Can I use TrainerRoad workouts on my ELEMNT instead of TrainingPeaks?** Yes. TrainerRoad is one of the providers Wahoo names, with one caveat worth knowing: a TrainerRoad account syncs today plus the next three days, rather than the full seven-day window other providers get [Wahoo ELEMNT Support]. Where you connect it is the one thing Wahoo's documentation is inconsistent about — the older ELEMNT article lists TrainerRoad inside the Wahoo app, the ACE, BOLT 3 and ROAM 3 article puts it with the providers you authorize from their own side [Wahoo ELEMNT ACE Support]. Check both. **Do I need a power meter for structured workouts on a Wahoo?** No. Interval targets can be power, heart rate, rate of perceived exertion, or speed [Wahoo ELEMNT Support]. With heart-rate targets, expect the later intervals of a long session to feel easier to satisfy — cardiovascular drift raises heart rate over prolonged work even at constant power [Coyle & González-Alonso 2001], so the target effectively comes down to meet you. **How many days of workouts can an ELEMNT hold at once?** Seven at most — today plus the next six — and only if your provider supplies them. TrainingPeaks free gives you today; TrainerRoad gives you four days; SYSTM, paid TrainingPeaks, and other third-party providers give the full seven [Wahoo ELEMNT Support]. The ceiling is Wahoo's, so no service can queue a month of sessions on the device. **Why did my workout sync to the app but not to the computer?** Almost always Wi-Fi. Bluetooth moves a single workout from the phone to the computer; syncing more than one requires the computer itself to be on a saved Wi-Fi network, which it confirms with a solid Wi-Fi symbol in the top bar [Wahoo ELEMNT Support]. Having the phone on Wi-Fi is not the same thing, and it is the most common version of this problem. ## References 1. **Wahoo ELEMNT Support.** [Planned Workouts [ELEMNT]](https://support.wahoofitness.com/hc/en-us/articles/115001223770-Planned-Workouts-ELEMNT). Wahoo Fitness Support. 2. **Wahoo ELEMNT ACE Support.** [Planned workouts with ELEMNT ACE, BOLT 3 and ROAM 3](https://support.wahoofitness.com/hc/en-us/articles/22710662013458-Planned-workouts-with-ELEMNT-ACE-BOLT-3-and-ROAM-3). Wahoo Fitness Support. 3. **Wahoo RIVAL Support.** [Planned Workouts on ELEMNT RIVAL](https://support.wahoofitness.com/hc/en-us/articles/360021974640-Planned-Workouts-on-ELEMNT-RIVAL). Wahoo Fitness Support. 4. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. 5. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. --- # Wahoo SYSTM alternatives for riders who train mostly outdoors URL: https://www.adaptcycling.com/guides/wahoo-systm-alternative-outdoor-riders Updated: 2026-08-01 Author: Jim Camut SYSTM is better outdoors than its reputation suggests. Most of its workouts have a proper outdoor variant, they sync to an ELEMNT over Wi-Fi without you doing anything, and a completed outdoor session pairs back to your plan [Wahoo SYSTM Outdoor]. The gap is narrower and more specific than "it's an indoor app": the assessment that personalizes everything has no outdoor version, and outdoor load is scored the same for every athlete. Whether that matters depends entirely on where your training actually happens. ## What SYSTM actually gives an outdoor rider More than most comparisons admit. The majority of SYSTM workouts have a corresponding outdoor version, converted to target ranges rather than exact wattages, and they appear on an ELEMNT automatically once both apps share an email address [Wahoo SYSTM Outdoor]. The conversion is thoughtful rather than mechanical. Outdoor variants smooth the intervals and replace exact target values with ranges, on the honest reasoning that nobody holds a precise wattage on an open road the way ERG enforces it indoors. Zones are widened deliberately: zone 4 becomes 92-101% of the target value, MAP efforts get plus or minus 5%, anaerobic capacity 8%, neuromuscular 10% [Wahoo SYSTM Outdoor]. Some sessions are rebuilt entirely to reproduce the training stimulus rather than the shape. Delivery is genuinely automatic. Sync the computer over Wi-Fi and today's plus the coming days' sessions arrive; reschedule or delete something on the SYSTM calendar and the change propagates on the next sync, including over cell data through the phone app [Wahoo SYSTM Outdoor]. On the device you get the power target range, the time left in the interval, how many intervals remain, and the workout profile along the bottom of the screen. Completion counts, too. If the outdoor session belonged to a training plan, it pairs back to the planned workout and turns it compliant once you finish 80% of it [Wahoo SYSTM Outdoor]. That is a real closed loop, and it is why the honest verdict on device delivery is a tie rather than a win for anyone. If you are shopping purely on "can I ride SYSTM sessions outside," the answer is yes and you can stop reading. ## The three places the indoor-first design shows Not in delivery — in personalization. The fitness tests have no outdoor variant, outdoor targets are computed from your indoor profile, and the load score for an outdoor session is identical for every athlete who rides it [Wahoo SYSTM Outdoor, Wahoo SYSTM Assessments]. First, the tests. Full Frontal and Half Monty have not been optimized for outdoors, so only the indoor icon appears beside them on the calendar [Wahoo SYSTM Outdoor]. Since workout targets and the plan itself are tailored to the 4DP profile and Rider Type those tests produce [Wahoo SYSTM Assessments], an outdoor-only rider cannot personalize the system without going indoors. Wahoo's guidance on repeating the test also varies by which of its own pages you read — the SYSTM support docs say optimally every 10-12 weeks, while its coaching blog advises against more often than every 12-16 [Wahoo SYSTM Assessments, Wahoo Sports Science]. Either way, you owe the trainer a maximal hour a few times a year. Second, the derivation. Wahoo is explicit that there is no separate indoor and outdoor 4DP profile: outdoor ranges are calculated from your indoor value with an allowance for the expected variance between the two [Wahoo SYSTM Outdoor]. That allowance is doing real work, because the difference is real — in a controlled comparison of 40-km efforts at matched perceived exertion, riders produced higher power and higher heart rate outdoors, with cooler skin temperature and a wider core-to-skin gradient [Mieras et al. 2014]. A modelled offset from an indoor number is a reasonable estimate. It is not the same as measuring you where you ride. Third, the accounting. The TSS and intensity factor for an outdoor variant are based on the workout rather than your profile, so they are the same for every athlete who completes it [Wahoo SYSTM Outdoor]. For a rider whose training is mostly indoors with occasional outdoor sessions, that rounding is harmless. For a rider whose season is built outdoors — long weekend rides, a hilly commute, an unplanned four hours with the fast group — the load ledger stops being about them specifically. ## What an outdoor-first setup has to do instead Three things, none of which require a trainer. Estimate your threshold from efforts you already produce, count every ride toward the same load account, and rebuild the plan when the week breaks rather than when the calendar says so. Threshold first, because everything downstream is scaled to it. The critical-power relationship derives sustainable power from the hyperbolic curve between power output and time to exhaustion [Jones et al. 2010], and a rider doing climbs, group rides, and intervals is generating efforts across exactly the durations that model needs. A power curve built from real riding updates continuously and never asks you to schedule a maximal hour — which for an outdoor rider is the difference between a current number and a quarterly one. Then the ledger. The reason to insist that every ride lands in one account is that intensity distribution, not any single session, is what decides whether a training year works [Seiler 2010]. Split the record — structured sessions in one system, the Saturday group ride somewhere else — and nothing is checking whether the easy rides stayed easy. This is the failure mode that quietly produces a year of tempo. Finally, reaction. An outdoor season generates disruptions an indoor season does not: weather, daylight, a route that turned into a race, an event moved by a month. A plan that only re-sequences when you complete or skip a scheduled workout cannot respond to a four-hour ride that was never on the calendar. This is the broader argument in our guide to training with Wahoo — the head unit is the delivery layer, and what matters is the quality of the decision upstream of it. ## The honest options Four, and the right one depends on where your center of gravity is. Stay on SYSTM if it is indoors. Pair it with something if it is split. Change the plan source if it is outdoors. And if you have no ELEMNT computer, your options narrow considerably. Stay on SYSTM if most of your quality work happens on the trainer and outdoor riding is the reward rather than the training. The video library and the 4DP profile are genuinely strong, the outdoor variants are well built, and none of the friction above is worth losing them over. If you want the head-to-head rather than the outdoor-specific question, our SYSTM comparison page covers the full matrix including price. Know this one before you subscribe: without a Wahoo ELEMNT computer you cannot upload outdoor structured workouts at all. Wahoo's recommendation in that case is to pick a No-Vid session and follow its written interval breakdown by feel — duration, target RPE, target power as a percentage of a 4DP metric, target heart rate as a percentage of LTHR [Wahoo SYSTM Outdoor]. That is a workable manual fallback, not an integration. Change the plan source if outdoors is where your season actually happens. AdaptCycling reads every ride from your Wahoo account — structured or not, indoor or out — estimates your threshold from your own power curve rather than a test you have to schedule, rebuilds the plan when the week breaks, and sends the next seven days back to the head unit as structured workouts. There is no indoor assessment to ride first, and the device push runs on the free tier. TrainerRoad and TrainingPeaks are the other serious answers: the first for indoor-anchored adaptive plans, the second when a coach is authoring your week. ## Common questions **Can I do SYSTM workouts outdoors?** Yes. The majority of SYSTM sessions have a corresponding outdoor version with smoothed intervals and target ranges instead of exact wattages, and they sync to a Wahoo ELEMNT automatically over Wi-Fi once both apps use the same email address [Wahoo SYSTM Outdoor]. The outdoor version deliberately will not play in the SYSTM app — it is meant for the head unit. **Do I have to do Full Frontal indoors?** Yes. Full Frontal and Half Monty have not been optimized for outdoor use, so no outdoor variant exists for either [Wahoo SYSTM Outdoor]. Since your workout targets and plan are tailored to the 4DP profile those tests produce [Wahoo SYSTM Assessments], personalizing SYSTM requires trainer time even if you never otherwise ride indoors. **Does my outdoor ride count toward my SYSTM training plan?** A planned outdoor structured workout does — it pairs back to the calendar entry and marks it compliant once you complete 80% of it [Wahoo SYSTM Outdoor]. Note that the TSS and intensity factor recorded for an outdoor variant come from the workout rather than your profile, so the load is scored identically for every athlete who rides it. **Is my outdoor FTP different from my indoor one?** Usually, and measurably. Riders produced higher power and heart rate outdoors than in a laboratory at matched perceived exertion over 40 km [Mieras et al. 2014]. Wahoo does not maintain separate indoor and outdoor 4DP profiles; it calculates outdoor target ranges from your indoor value with an allowance for the expected variance [Wahoo SYSTM Outdoor]. **What if I don't have a Wahoo ELEMNT computer?** Then outdoor structured workouts cannot be uploaded to a computer at all. Wahoo suggests choosing a No-Vid workout and following its written interval breakdown — duration, RPE, power as a percentage of a 4DP metric, and heart rate as a percentage of LTHR — by feel [Wahoo SYSTM Outdoor]. ## References 1. **Wahoo SYSTM Outdoor.** [Wahoo SYSTM - Outdoor Structured Workouts](https://support.wahoofitness.com/hc/en-us/articles/7671514838674-Wahoo-SYSTM-Outdoor-Structured-Workouts). Wahoo Fitness Support. 2. **Wahoo SYSTM Assessments.** [Fitness Assessments in the SYSTM app](https://support.wahoofitness.com/hc/en-us/articles/4403358637586-Fitness-Assessments-in-the-SYSTM-app). Wahoo Fitness Support. 3. **Wahoo Sports Science.** [Power Tests: What, When and Why Do Them?](https://www.wahoofitness.com/blog/power-testing-what-when-and-why-do-them/). Wahoo Fitness. 4. **Mieras et al. 2014.** [Physiological and psychological responses to outdoor vs. laboratory cycling](https://pubmed.ncbi.nlm.nih.gov/24476776/). Journal of Strength and Conditioning Research. 5. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. 6. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. --- # Training without Strava: connecting a Wahoo account directly URL: https://www.adaptcycling.com/guides/train-without-strava-connect-wahoo Updated: 2026-08-02 Author: Jim Camut Most training apps never talk to your bike computer. They read finished rides out of Strava, which makes Strava a required link in a chain it is not really part of — and a one-way one. Wahoo's own app-partner matrix has Strava checked for importing routes, segments and completed activity summaries, and for sharing summaries back. It is not checked for importing plans [Wahoo App Partners]. Eleven other partners are. If you ride a Wahoo, the direct route already exists, and it runs in both directions. ## What Strava is actually doing in your training stack Acting as the universal adapter, and doing it well. What it cannot do is carry anything back. Strava moves routes, segments and finished rides in Wahoo's partner matrix, and never planned workouts [Wahoo App Partners] — so a Strava-fed training app can score last Saturday but cannot hand you Wednesday. The usual stack has three parties in it for one ride. Your computer records, it uploads to Strava, and a training platform reads that activity through Strava's API. Nobody designed this arrangement; it accumulated, because Strava arrived first and built the one integration every device maker already supported. The cost is a dependency nobody consciously signed up for. Your training app's relationship with your riding is mediated by a third company's account, and cancelling or disconnecting that account quietly ends it. For a rider whose training app is the thing they actually pay for, that is a strange place for the single point of failure to sit. The direction is the part most riders have never had reason to notice. Strava's own ELEMNT documentation describes recording with the computer and automatically uploading activities to Strava, and adds that anything uploaded before you connected will not sync [Strava ELEMNT Support]. No return leg is described, because there isn't one. Wahoo's matrix says the same thing from the other side: import plans is unchecked for Strava, while eleven other partners have it [Wahoo App Partners]. Whatever your training app concludes on Tuesday, Strava has no route to put it on your handlebars. So a Strava-based training platform is structurally read-only at the point where it matters most. The result is a setup most self-coached riders will recognise: the plan lives on a phone, the computer is a stopwatch that happens to record power, and when the plan changes mid-week nothing on the bike finds out. That is not a flaw in any of those apps. It is the shape of the pipe they chose. ## The direct route, and who already rides it Wahoo runs its own cloud, and services connect to it with no Strava in the middle. Eleven of Wahoo's listed app partners can import plans onto an ELEMNT [Wahoo App Partners]. Most are newer connections made through the Wahoo Cloud API, which you authorize from the provider's side rather than Wahoo's. That direction of authorization is the single most common source of confusion. Some partners are authorized inside the Wahoo app, under Today, Settings, Authorized apps. Others — mostly the more recently added ones — are connected by signing into your Wahoo account in the partner's own app or web interface [Wahoo App Partners]. A rider who searches the Wahoo app for a provider that isn't listed there reasonably concludes the integration does not exist, when the button was simply on the other side. However a link was made, all of them end up in one place: Wahoo's profile page lists every connected partner, what it is permitted to do, and a control to revoke it [Wahoo App Partners]. The list of services that can put a plan on your handlebars is worth reading in full, because it is longer and more varied than the folklore suggests: TrainingPeaks, TrainerRoad, Intervals.icu, JOIN, Xert, TrainerDay, Humango, FasCat, Enduco, Nolio and MATS [Wahoo App Partners]. Those products disagree about nearly everything — price, philosophy, whether a human writes your week. They share exactly one property. Not one of them needs a Strava account to reach an ELEMNT. History is the other half of going direct. Connecting Strava to a Wahoo pulls nothing backwards — Strava states plainly that activities uploaded before you connect will not sync [Strava ELEMNT Support]. A provider reading the Wahoo Cloud directly can instead work from the rides already sitting in your Wahoo account, which for most ELEMNT owners is everything the computer has recorded since they set it up. That difference decides whether a new training app starts with months of your real riding or with an empty ledger and a questionnaire. ## What you actually give up One thing, and it is real. Strava is the only partner in either of Wahoo's tables checked for importing segments [Wahoo App Partners], so Live Segments leave your handlebars with it. Route sync and your ride history are far more portable than they look. Take the segment loss seriously rather than arguing around it. Import segments is checked for Strava and for nobody else in either table [Wahoo App Partners]. If racing the climb near your house against everyone who has ever ridden it is part of why you go out, that is a feature with no substitute, and no argument about training quality should talk you out of it. None of this requires cancelling anything, either: keeping Strava for the social layer while taking your coaching from elsewhere is a perfectly coherent setup, and a common one. Navigation survives the change easily. Importing routes to an ELEMNT is checked for Komoot, RideWithGPS, MapMyFitness, Bikemap and Trailforks as well as Strava [Wahoo App Partners]. If what keeps you subscribed is that your Saturday route appears on the computer, that particular job has five other applicants. And your riding history is not hostage. Strava provides an option to export an archive of your account, and in many cases the original activity file comes back in FIT format [Strava Data Export] — the format your computer wrote in the first place. That matters, because the ride file is the substrate every analysis is built from: power, heart rate and cadence second by second, not a table of averages. Any system that reads FIT can be handed years of riding regardless of which platform it travelled through. Leaving is an export problem, not a loss. ## Building the loop without Strava in it Three requirements, and none of them is a subscription: estimate your threshold from riding you already do, keep every ride in one ledger, and get the revised plan back onto the device. The third is the one a Strava-shaped pipeline structurally cannot satisfy. Threshold first, because everything downstream is scaled to it. The critical-power relationship derives sustainable power from the hyperbolic curve between power output and time to exhaustion [Jones et al. 2010], and a rider doing climbs, group rides and intervals is already producing near-maximal efforts across the durations that model needs. A power curve built from real riding updates continuously and never asks you to book a Saturday for a maximal test — which also means it cannot quietly go stale for a quarter the way a tested number does. Then the ledger, which is where a split setup does its quiet damage. Intensity distribution, not any individual session, is what decides whether a training year works [Seiler 2010]. If structured sessions live in one system and the Sunday group ride lands in another, nothing holds the whole picture, and nothing is checking whether the easy rides stayed easy. The requirement is not a particular platform. It is that every ride counts toward the same account. The third requirement is what the direct route buys, and it is the argument running through our broader guide to training with Wahoo: the head unit is the delivery layer, and its worth is set by the decision upstream of it. AdaptCycling connects to a Wahoo account as its own sign-in, so there is no Strava step to complete first. We read the ride history already in your Wahoo account, estimate threshold from your power curve rather than a test, and send the current day plus the next six back to the ELEMNT as structured workouts [Wahoo ELEMNT Support], replacing them when the plan changes. That push runs on the free tier. ## Common questions **Can I use a training app without a Strava account?** Yes. Eleven of Wahoo's listed app partners can import plans onto an ELEMNT, and Strava is not one of them [Wahoo App Partners]. Most of those connect through the Wahoo Cloud API, which means you authorize your Wahoo account inside the provider's app or website rather than hunting for it in Wahoo's. **Can Strava send a structured workout to my Wahoo?** No. In Wahoo's partner matrix, Strava is checked for importing routes, segments and completed summaries, and for sharing summaries back — not for importing plans [Wahoo App Partners]. Strava's own ELEMNT documentation describes the link in one direction only: the computer records, and the activity uploads to Strava [Strava ELEMNT Support]. **Does my ELEMNT need Strava to upload my rides?** No. Strava is one destination among many that you can authorize, and Wahoo lists more than a dozen others for sharing completed activity summaries [Wahoo App Partners]. The computer's own upload path is to the Wahoo Cloud; where the ride goes after that is a permissions question you control from your Wahoo profile. **What happens to my training history if I stop using Strava?** You can take it with you. Strava provides an option to export an archive of your account, and in many cases the original file comes back in FIT format [Strava Data Export]. Any system that reads FIT can be handed that history directly, so switching costs an export rather than the data. **I ride a Garmin, or train indoors on Zwift. Does any of this apply?** Partly. The Wahoo Cloud route needs a Wahoo device, and Garmin does not appear on Wahoo's app-partner list at all [Wahoo App Partners]. What does travel is the file: essentially every cycling computer and most indoor platforms export FIT, so uploading the ride puts it in the same ledger as everything else without needing either company's integration. ## References 1. **Wahoo App Partners.** [Wahoo App Partners](https://support.wahoofitness.com/hc/en-us/articles/35014623885714-Wahoo-App-Partners). Wahoo Fitness Support. 2. **Wahoo ELEMNT Support.** [Planned Workouts [ELEMNT]](https://support.wahoofitness.com/hc/en-us/articles/115001223770-Planned-Workouts-ELEMNT). Wahoo Fitness Support. 3. **Strava ELEMNT Support.** [Wahoo ELEMNT and Strava](https://support.strava.com/en-us/articles/15401928-wahoo-elemnt-and-strava). Strava Support. 4. **Strava Data Export.** [Exporting your Data and Bulk Export](https://support.strava.com/hc/en-us/articles/216918437-Exporting-your-Data-and-Bulk-Export). Strava Support. 5. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. 6. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. --- # Wahoo planned workouts not syncing: finding the actual failure URL: https://www.adaptcycling.com/guides/wahoo-planned-workouts-not-syncing Updated: 2026-08-02 Author: Jim Camut Four unrelated problems produce the same sentence. The workout sits outside the window your provider syncs, the file was never a structured workout to begin with, the connection between your provider and the Wahoo Cloud has quietly broken, or the computer itself never joined the Wi-Fi. Wahoo's own troubleshooting list is ordered to separate them, and its first two steps are both checks most riders skip: whether the workout reached the app at all, and whether it was ever inside the window your provider syncs [Wahoo ELEMNT Support]. ## First, prove the workout was supposed to be there A large share of these reports are the sync window working as designed. A free TrainingPeaks account syncs today's workout only, TrainerRoad syncs today plus three days, and everything else syncs today plus six [Wahoo ELEMNT Support]. A session scheduled eight days out is correctly absent, and re-syncing will never produce it. The window is not a setting you can widen. Wahoo answers "how do I select which planned workouts are on my ELEMNT" with "you don't, really" — the computer holds today plus whatever future days your provider supplies, and each Wi-Fi sync rebuilds that list against the same rule [Wahoo ELEMNT Support]. So the first check is arithmetic: count the days to the workout you are hunting for, and compare that against your provider's allowance. The table is identical on ELEMNT ACE, BOLT 3 and ROAM 3 [Wahoo ELEMNT ACE Support]. The second check is whether the entry on your calendar is a structured workout at all. Wahoo names this explicitly for TrainingPeaks: the workout has to contain blocks built in the Workout Builder, not just a description of the targets [Wahoo ELEMNT Support]. A session typed as "3x10min at 95%" into a description field is a note to yourself — no intervals to render, nothing to send, and it fails silently rather than erroring. The principle holds whatever your plan source: what syncs is a structured file, not the text on the calendar. ## Bisect the pipe at the app There are two halves — provider to Wahoo Cloud, and cloud to computer — and one look tells you which failed. If the workout is in the app but not on the computer, it is a transport problem. If it is missing from the app too, the fault sits upstream of the device entirely [Wahoo ELEMNT Support]. That is step one of Wahoo's own troubleshooting order, and it halves the search space before you change anything [Wahoo ELEMNT Support]. On older computers the list lives in the ELEMNT app under the Workout tab, View Planned Workouts; on ACE, BOLT 3 and ROAM 3 it is the My plan panel on the Today tab of the Wahoo app [Wahoo ELEMNT ACE Support]. Pull down on that page to refresh it before drawing any conclusion [Wahoo ELEMNT Support]. When the workout is missing from the app as well, the link between your provider and your Wahoo account is the suspect, and the documented repair is blunt: deauthorize the provider and authorize it again [Wahoo ELEMNT Support]. Wahoo prescribes the same reset for route and upload failures, and there names the reason: the Cloud connection to a third-party app can be damaged and need re-establishing, done from the Profile tab under Authorized apps [Wahoo Authorized Apps]. A connection that still appears in the list is not evidence that it works; the listing can outlive the token behind it. Reset the correct side, though, because the direction differs by provider. SYSTM and TrainingPeaks are linked inside Wahoo's app; other providers are linked from their own app or web interface using your Wahoo account [Wahoo ELEMNT Support]. TrainerRoad is the case Wahoo's documentation disagrees with itself about: the older ELEMNT article lists it alongside TrainingPeaks inside the app, and the ACE, BOLT 3 and ROAM 3 article groups it with the providers you connect from their side [Wahoo ELEMNT Support, Wahoo ELEMNT ACE Support]. Check both before concluding the integration is gone. However a link was made, all of them are listed and revocable in one place — your Wahoo profile [Wahoo App Partners]. ## The device half is Wi-Fi, and sometimes it is the band Bluetooth moves exactly one workout; only Wi-Fi moves the week, and the computer itself has to be on the network rather than the phone [Wahoo ELEMNT Support]. One specification decides it: older ELEMNT, BOLT and ROAM units need a 2.4 GHz network, while ACE, BOLT 3 and ROAM 3 accept 2.4 or 5 [Wahoo ELEMNT ACE Support]. That band split produces a failure that looks like nothing else. A mesh router presenting one merged SSID, or a household that has moved to 5 GHz, can leave an older ELEMNT unable to join while every other device in the house is fine — and the computer never reports a failure. Confirm the join on the computer itself, which shows a solid Wi-Fi symbol in its top bar when connected [Wahoo ELEMNT Support]. Then force a sync rather than waiting for one. On an ELEMNT, BOLT or ROAM: tap the left power button for the Menu, scroll to Planned Workouts, select the "Synced X ago" line and choose Sync — it reads SYNCING with your provider's name, then Synced All, and selecting Sync more than once is expected rather than a symptom [Wahoo ELEMNT Support]. On ACE, BOLT 3 and ROAM 3, Workouts on the Ready to Ride dashboard carries a SYNC control, and on the ACE and ROAM 3 touchscreens you can pull down instead [Wahoo ELEMNT ACE Support]. It typically runs in under a minute [Wahoo ELEMNT Support]. If Wi-Fi is genuinely unavailable, know the ceiling. Exactly one workout transfers from the phone over Bluetooth, chosen in the app, and that is a limit rather than a slower path to the same result [Wahoo ELEMNT Support]. A rider travelling for a week on Bluetooth alone gets today, every day, and reasonably concludes the provider stopped sending the rest. Two prerequisites sit underneath all of this and are easy to skip: a current app version and current computer firmware [Wahoo Authorized Apps]. ## When it synced and the workout is still wrong Three residual cases get filed as sync failures and are not: targets computed from the wrong zones, completed rides that never travel back to your provider, and the RIVAL watch, which is a different pipeline altogether [Wahoo ELEMNT Support, Wahoo RIVAL Support]. The common one is targets that look impossible, or oddly easy. Power targets in a synced workout are keyed to the zones held in your provider's app, and those can differ from the zones stored in your Wahoo account [Wahoo ELEMNT Support]. The sync did its job exactly; the two accounts disagree about your threshold. Reconcile them rather than scaling every session on the road. The deeper version is a threshold that is simply old: a number set in March and still prescribing intervals in August is not a configuration error, and no amount of re-authorizing touches it. The second is the return leg, which has a switch of its own. A completed ride only reaches a third-party app if Automatic upload is on and that specific app is toggled on beneath it [Wahoo Authorized Apps]. Nothing in the outbound path hints at this, so workouts can arrive on the computer all week while nothing goes back the other way. The third catches watch owners. The ELEMNT RIVAL takes planned workouts from TrainingPeaks alone, syncs them through the companion app rather than over Wi-Fi, and holds today plus the next five days [Wahoo RIVAL Support]. No third-party provider reaches it. The pattern behind all of these is the argument running through our broader guide to training with Wahoo: the head unit is a delivery layer, and it renders faithfully whatever the layer above it committed to. It is also why AdaptCycling re-sends the current day plus the next six to your Wahoo account each morning after the plan rolls over [Wahoo ELEMNT Support] — a window that slides daily has to be re-pushed daily — and why we estimate your threshold from your own power curve rather than a test you have to remember to schedule [Jones et al. 2010]. The device push runs on the free tier. ## Common questions **Why does only today's workout show up on my ELEMNT?** Either your provider only supplies one day, or the computer is not on Wi-Fi. Bluetooth carries a single workout at a time; multiple days require the computer itself — not just the phone — to be joined to a saved network, which it confirms with a solid Wi-Fi symbol in the top bar [Wahoo ELEMNT Support]. On ELEMNT, BOLT and ROAM that network has to be 2.4 GHz [Wahoo ELEMNT Support]. **How do I force my Wahoo to sync planned workouts?** On an ELEMNT, BOLT or ROAM, tap the left power button, scroll to Planned Workouts, select the line reading Synced X ago and choose Sync; repeat if it does not report Synced All [Wahoo ELEMNT Support]. On ACE, BOLT 3 or ROAM 3, use SYNC under Workouts on the Ready to Ride dashboard, or pull down on the touchscreen [Wahoo ELEMNT ACE Support]. **My workout is on my TrainingPeaks calendar but not on my ELEMNT. Why?** Two likely causes. A free TrainingPeaks account syncs today's workout only, so anything further out is correctly missing [Wahoo ELEMNT Support]. The other is format: the workout must contain blocks created in the TrainingPeaks Workout Builder rather than target text typed into a description, or there are no intervals to send [Wahoo ELEMNT Support]. **How do I delete a planned workout from my ELEMNT?** You don't do it on the computer. The device mirrors a rolling window of your provider's calendar, so removing a workout means moving it out of that window or deleting it from the calendar, then syncing again [Wahoo ELEMNT Support]. The same sync that adds tomorrow's session drops the one you removed. **Does re-authorizing my provider actually fix anything?** Often, yes. Wahoo lists deauthorizing and reauthorizing your provider as a standard step for missing workouts [Wahoo ELEMNT Support], and describes the same reset elsewhere as repairing a damaged Cloud connection [Wahoo Authorized Apps]. Do it wherever the link was made — inside the Wahoo app for SYSTM and TrainingPeaks, in the provider's own app for most others [Wahoo App Partners]. ## References 1. **Wahoo ELEMNT Support.** [Planned Workouts [ELEMNT]](https://support.wahoofitness.com/hc/en-us/articles/115001223770-Planned-Workouts-ELEMNT). Wahoo Fitness Support. 2. **Wahoo ELEMNT ACE Support.** [Planned workouts with ELEMNT ACE, BOLT 3 and ROAM 3](https://support.wahoofitness.com/hc/en-us/articles/22710662013458-Planned-workouts-with-ELEMNT-ACE-BOLT-3-and-ROAM-3). Wahoo Fitness Support. 3. **Wahoo Authorized Apps.** [Authorize a 3rd party app for route / planned workout sync or completed ride upload [ELEMNT]](https://support.wahoofitness.com/hc/en-us/articles/115000504304-Authorize-a-3rd-party-app-for-route-planned-workout-sync-or-completed-ride-upload-ELEMNT). Wahoo Fitness Support. 4. **Wahoo App Partners.** [Wahoo App Partners](https://support.wahoofitness.com/hc/en-us/articles/35014623885714-Wahoo-App-Partners). Wahoo Fitness Support. 5. **Wahoo RIVAL Support.** [Planned Workouts on ELEMNT RIVAL](https://support.wahoofitness.com/hc/en-us/articles/360021974640-Planned-Workouts-on-ELEMNT-RIVAL). Wahoo Fitness Support. 6. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. --- # Getting a Wahoo FTP without riding Full Frontal URL: https://www.adaptcycling.com/guides/wahoo-ftp-without-4dp-test Updated: 2026-08-02 Author: Jim Camut Full Frontal is an hour of maximal work, and Wahoo asks for it every 10 to 12 weeks [Wahoo SYSTM Assessments]. Plenty of riders never do it. What is less known is that Wahoo publishes its own ranked ladder of alternatives — the Half Monty ramp test as second best, an in-app questionnaire as third [Wahoo Fitness Metrics]. Every rung down costs something specific. Which of the four numbers stay measured, and which quietly become estimates, is the whole decision. ## Why the test takes an hour Full Frontal is four maximal efforts in one session: two 5-second sprints, a 5-minute effort, a 20-minute effort, and a closing 1-minute effort, separated by 5-minute recoveries [Wahoo Sports Science]. The order is not incidental. Each number is measured in a deliberately fatigued state, which is also why the protocol cannot be shortened. Start with the part most riders get wrong. The familiar 20-minute test takes 95% of the result, on the reasoning that a fresh 20-minute effort overstates threshold. Full Frontal takes 100% of its 20-minute value instead — because two sprints and a 5-minute maximal effort came first and have already drained anaerobic capacity, Wahoo treats that number as a truer reading rather than an inflated one [Wahoo 4DP Profile]. That is defensible protocol design, and it explains the length. The fatigue is the instrument, not padding. The other three numbers come out of the same logic. Neuromuscular power is your best 5-second power across both sprints. Maximal aerobic power comes from the 5-minute effort and acts as a ceiling on FTP. Anaerobic capacity comes from the 1-minute effort at the very end, and Wahoo is explicit that a fresh 1-minute number would be higher and less useful — measured last, it captures how quickly you recover and go again rather than raw peak [Wahoo 4DP Profile]. Four numbers describing four systems is more information than one. The fair criticism is that the number is only as current as the last time you rode it, and Wahoo's own guidance lets it age. The support documentation asks for Full Frontal before each new plan, optimally every 10 to 12 weeks [Wahoo SYSTM Assessments]; the coaching blog advises against repeating it more often than every 12 to 16 [Wahoo Sports Science]. Either way the profile behind every target on your screen is allowed to be a quarter old. And a threshold produced by a 20-minute protocol estimates maximal lactate steady state well on average — a bias of 1.4% — while its 95% limits of agreement span roughly 9%, so it is a sound group-level number carrying real individual uncertainty [Borszcz et al. 2019]. ## Wahoo's own ladder, in Wahoo's own words Wahoo ranks the routes to an Athlete Profile explicitly: Full Frontal is best, Half Monty is better, the in-app questionnaire is ok [Wahoo Fitness Metrics]. All three produce a working profile. What separates them is how many of the four numbers are measured rather than inferred. Half Monty is a ramp test followed by a 20-minute sub-maximal effort held to a heart-rate range rather than a power target [Wahoo Half Monty]. The second half exists for a specific reason. A conventional ramp takes 75% of peak 1-minute power as FTP, and Wahoo's analysis of its own Full Frontal database found that formula inaccurate for 17% of athletes — the more anaerobically fit ones, who ramp well and threshold worse [Wahoo Half Monty]. Reading the relationship between heart rate and power across a steady 20 minutes is what corrects it. Ridden correctly it updates FTP, maximal aerobic power, and lactate threshold heart rate. If you have never completed Full Frontal, it also supplies estimated anaerobic capacity and neuromuscular values, which Wahoo says beat estimating them from FTP alone [Wahoo 4DP Profile]. What it will not do is set your Rider Type or Weakness [Wahoo SYSTM Assessments]. Since plans are tailored to that profile rather than to the raw watts, that is the true gap between the two assessments — not precision, but which plan you get handed. Two constraints decide whether the shorter route is open to you. Half Monty's second half is constrained by heart rate, so without a monitor you cannot complete it — you get FTP and MAP, less accurately [Wahoo Half Monty]. And with no power trainer at all, the questionnaire in the Wahoo app produces an estimated profile from self-report; Wahoo calls it a great start and says an assessment will always be more accurate [Wahoo Fitness Metrics]. It is the placeholder that stops the app prescribing from nothing. ## What the profile is quietly driving It is not a display value. Workout targets scale off it, plan selection keys off Rider Type and Weakness, and Wahoo's newer Fitness Metrics score every ride you upload against it [Wahoo Fitness Metrics]. An estimated or stale profile does not sit still — it propagates into everything downstream. The Fitness Metrics are the clearest case. They analyze every cycling activity in your account carrying power and cadence, scoring each ride's impact on threshold, maximal aerobic power and repeatability separately [Wahoo Fitness Metrics]. They need a paid subscription, at least 90 days of history to behave, and a current Athlete Profile — because the method is comparing what you did against what the profile says you can do. Feed it a questionnaire estimate and every ride is scored against a guess about you. One default deserves naming because nothing surfaces it. Preferred cadence is part of the profile, and the metrics use cadence as a modifier on power to work out how you produced it. If you have not completed Full Frontal, that field is simply set to 85 RPM [Wahoo Fitness Metrics]. A rider who habitually turns 75, or 95, has an analysis layer reading their power wrong on every ride — and the fix is a manual field in the Athlete Profile, not another hour on the trainer. Staleness is hard to see from the inside, which is why Wahoo names a symptom rather than a date: if your key sessions — not the recovery rides — are consistently too easy, take a Half Monty about two-thirds of the way through the plan, at the end of a recovery week when you are carrying less fatigue [Wahoo SYSTM Assessments]. That is good advice, and it concedes the design — the profile drifts between assessments, and the athlete is the drift detector. ## The route Wahoo doesn't offer: the rides you already did There is a third way to a threshold number that involves no assessment at all. The critical-power model derives sustainable power from the hyperbolic relationship between power output and time to exhaustion [Jones et al. 2010] — fitted to efforts already sitting in your ride history. The durations line up with normal riding better than most people expect. A rider doing group rides, climbs and interval sessions produces near-maximal efforts across seconds, minutes and tens of minutes without scheduling one: the sprint for a town sign, two minutes over a rise, twenty minutes of a real climb. Fitted continuously, that curve moves every week instead of every quarter, and it never costs a Saturday. Be precise about what transfers, because the honest answer is not all four. Threshold and maximal aerobic power estimate well from ride data, since most riders generate efforts near those durations whether they meant to or not. Neuromuscular power does not — Wahoo's own note is that the sprint value has to come from indoors, because indoor sprint power is what sets your indoor sprint targets [Wahoo 4DP Profile] — and a rider who never sprints outdoors gives the curve nothing to fit. Rider Type and Weakness have no ride-history equivalent at all. That is the trade we make deliberately. AdaptCycling estimates threshold from your rolling power curve rather than a scheduled test, rebuilds the plan when your week changes, and pushes today plus the next six days to your Wahoo account as structured workouts, with targets sent as ranges rather than a single number — an estimate is an interval, and a target that pretends otherwise is false precision. We produce no Rider Type, and if a 4DP profile is what you want, ride Full Frontal, because it is the only thing that yields one. The larger argument is in our guide to training with Wahoo: the head unit renders faithfully whatever it is handed, and the decision upstream is the part that matters. ## Common questions **Can I get personalized SYSTM targets without riding Full Frontal?** Yes. Wahoo's own ranking puts Half Monty second and the in-app Athlete Profile questionnaire third, behind Full Frontal [Wahoo Fitness Metrics]. Half Monty measures FTP and maximal aerobic power directly and estimates the other two dimensions if you have never tested [Wahoo 4DP Profile]. Targets are personalized either way — the question is how much of the profile was measured. **What does Half Monty update, and what does it leave alone?** Ridden correctly it updates FTP, maximal aerobic power and lactate threshold heart rate, plus estimated anaerobic capacity and neuromuscular values if you have no prior Full Frontal [Wahoo 4DP Profile]. It does not update Rider Type or Weakness [Wahoo SYSTM Assessments] — and since plans are tailored to those, the plan you are handed does not change. **Do I need a heart-rate monitor for Half Monty?** For the full result, yes. The second half is an effort constrained to a heart-rate range rather than a power target, so without a monitor you cannot complete it and will see only updated FTP and MAP, less accurate than they would otherwise be [Wahoo Half Monty]. The ramp on its own still returns numbers. **How often does Wahoo want Full Frontal repeated?** Its two answers differ slightly. The SYSTM support documentation recommends it before each new training plan, optimally every 10 to 12 weeks [Wahoo SYSTM Assessments]; the coaching blog advises against repeating it more often than every 12 to 16 [Wahoo Sports Science]. Quarterly is the practical read of both. **Can I just type my own numbers into my Wahoo profile?** You can, through the Athlete Profile questionnaire, which Wahoo ranks as legitimate but least accurate [Wahoo Fitness Metrics]. While you are in there, check preferred cadence: it defaults to 85 RPM if you have never ridden Full Frontal, and the Fitness Metrics use cadence to interpret how you produced your power. ## References 1. **Wahoo 4DP Profile.** [4DP (Four-Dimensional Power) Cycling Metrics / Athlete Profile](https://support.wahoofitness.com/hc/en-us/articles/360021387420-4DP-Four-Dimensional-Power-Cycling-Metrics-Athlete-Profile). Wahoo Fitness Support. 2. **Wahoo Half Monty.** [The Half Monty fitness assessment: Everything you need to know](https://support.wahoofitness.com/hc/en-us/articles/4404067414418-The-Half-Monty-fitness-assessment-Everything-you-need-to-know). Wahoo Fitness Support. 3. **Wahoo SYSTM Assessments.** [Fitness Assessments in the SYSTM app](https://support.wahoofitness.com/hc/en-us/articles/4403358637586-Fitness-Assessments-in-the-SYSTM-app). Wahoo Fitness Support. 4. **Wahoo Fitness Metrics.** [Set up your Wahoo account for Fitness Metrics](https://support.wahoofitness.com/hc/en-us/articles/30285425990290-Set-up-your-Wahoo-account-for-Fitness-Metrics). Wahoo Fitness Support. 5. **Wahoo Sports Science.** [Power Tests: What, When and Why Do Them?](https://www.wahoofitness.com/blog/power-testing-what-when-and-why-do-them/). Wahoo Fitness. 6. **Borszcz et al. 2019.** [Is the functional threshold power interchangeable with the maximal lactate steady state in trained cyclists?](https://pubmed.ncbi.nlm.nih.gov/30676826/). International Journal of Sports Physiology and Performance. 7. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. --- # When ERG mode is too hard: the KICKR lockdown and what causes it URL: https://www.adaptcycling.com/guides/kickr-erg-mode-target-too-hard Updated: 2026-08-02 Author: Jim Camut ERG holds the prescribed watts no matter what you do with the pedals. That is the feature and the failure mode. Outdoors you would shed a few watts on the second rep without ever deciding to; ERG removes that option, so a target pitched slightly high does not fail early, when you still have the composure to change it. It fails at rep five, after you have already paid for the first four. Wahoo's own documentation names the endpoint: ERG Lockdown [Wahoo ERG Guide]. ## Why the session fails late instead of early Above critical power you are spending a finite work capacity [Jones et al. 2010] that only refills when you drop back below it [Skiba et al. 2012]. A target 5% too high does not cost 5% more on the first rep. It empties the reservoir faster on every rep while the recoveries keep refilling the same amount. Skiba's balance model is the useful way to picture it: work capacity above critical power is expended during hard efforts and reconstituted during easy ones, and the point of exhaustion is predictable from the pattern of the two [Skiba et al. 2012]. Run that forward with a target a few percent high and the arithmetic is unforgiving — each rep takes a little more than the recovery gives back, the deficit compounds, and nothing feels wrong until the balance approaches zero. That is why the rider's report is almost always the same shape: reps one through four were fine, rep five was impossible. Self-selection is the safety valve ERG removes. On the road or in Level mode you shed a handful of watts when a rep starts to bite, usually without noticing you did it, and the session degrades gracefully instead of collapsing. ERG will not let you. It is also worth knowing that perfectly constant power is not a neutral choice: variable and constant power output at matched averages produce measurably different physiological responses [Kolsung et al. 2020]. So "the target is too hard" is a symptom with three different causes, and they need separating before anything gets changed. The trainer's setup can make a correct target unrideable. The workout's structure can be wrong while its target is right. And the number itself can be wrong. This article works through them in that order. ## The lockdown is a control loop, and you can break it Wahoo describes the mechanism plainly: if the power target is too high and you cannot hold the cadence, the trainer keeps adding resistance as your cadence decreases until it locks down and you can no longer turn the pedals [Wahoo ERG Guide]. The escape is counterintuitive — stop pedalling and let the trainer release. Resistance is a function of flywheel speed, so the loop is self-reinforcing. Cadence falls, the trainer adds resistance to defend the power target, the pedals get heavier, cadence falls further. Wahoo's own framing of the rider's job is the sentence worth internalizing: in ERG mode you meet the cadence target, and the trainer meets the power target for you [Wahoo ERG Guide]. Riders who spend an ERG session watching the power number are watching the one variable they do not control. The instinctive response makes it worse. A KICKR takes roughly 3 to 5 seconds to settle on a new target and other trainers 5 to 10, and with a fast cadence increase the trainer bleeds resistance to avoid overshooting — so spinning up to "help" at the start of an interval produces the sensation that ERG has stopped working entirely [Wahoo ERG Guide]. Shifting gears has the same problem from the other direction: it changes flywheel speed, which makes the trainer re-adjust, which slows the response you were trying to speed up. Hold the cadence target because the control loop needs a stable flywheel, not because a particular number is physiologically correct. The evidence there is genuinely equivocal — a systematic review found that riders self-select from a narrow band of the cadences available to them, and concluded it remains unclear whether cadence itself plays a role in the onset of fatigue [Mater et al. 2021]. Steadiness is what the machine is asking for. The specific rpm is mostly yours. ## Two setup faults that impersonate a bad target Before revising anything about your fitness, rule out gear selection and control interference. Both produce sessions that feel wrong in ways indistinguishable from a bad prescription, and both are free to check. Every gear has a basement. No matter what number the app sends, the trainer cannot produce less power than the gear you are in allows, so if the recovery valleys never actually feel like recovery you are geared too high [Wahoo ERG Guide]. Wahoo's starting point is the small chainring in front and the middle of the cassette at the back; from there, shift easier until the lowest power the trainer will give you sits below the lowest target in the workout. The mirrored fault is real too — too easy a gear cannot reach high targets on low-cadence intervals, and that is one of the few moments when shifting inside ERG is the right call. Interference is the other free check. Lockdown also happens when something else is talking to your trainer at the same time — a head unit, a smartwatch, a second app left running in the background [Wahoo ERG Lockup]. Wahoo's repair is to power off every other device and app that connects to the trainer, unplug the trainer for two minutes, restart the app, and reconnect. A session that locks up at a target you rode comfortably last month is far more likely to be this than a sudden loss of fitness. Some sessions should not be in ERG at all. Wahoo names several that are better ridden in Level mode, along with a hard rule: Full Frontal should never be done in ERG, and the second half of Half Monty is Level mode only [Wahoo ERG Guide]. The general test it offers is a good one to carry — any workout where your trainer visibly struggles to add or remove resistance during the interval belongs in Level mode. ## When the number really is wrong Wahoo's troubleshooting reaches this cause from both directions: struggling to hold power and cadence on new test result numbers means lowering the intensity, and resistance that feels too low means checking whether the profile numbers were ever right [Wahoo ERG Lockup, Wahoo ERG Guide]. Use the scale control rather than quitting. A planned workout can be scaled from the workout page before or during the session in 10% or 1% steps, intervals can be skipped or replayed, and the workout can be paused while the ride keeps recording [Wahoo ELEMNT Support]. This matters more than it sounds. A session finished at 95% is training that happened; a session abandoned at rep five is a data point about your FTP setting and nothing else. Then read which branch you are in, because the fix differs. If every quality session is too hard, the threshold driving the targets is too high, and that is a profile problem rather than a workout problem — our guide to getting a Wahoo FTP without Full Frontal covers the routes to a better number. If only the last reps fail, the target is probably fine and the structure is not: recoveries too short to reconstitute what the work efforts spend [Skiba et al. 2012]. And if rep one is already brutal, you are under-recovered today and nothing about the prescription is at fault. The durable fix is that a failed session should change the next one. AdaptCycling estimates threshold from your rolling power curve rather than a scheduled maximal test, reads what you actually completed including the reps you did not, and rebuilds the following week around it, pushing the revised sessions to your Wahoo account as structured workouts with targets sent as ranges rather than a single midpoint. That is the argument our guide to training with Wahoo makes at length: the KICKR and the head unit render faithfully whatever they are handed, and nothing in that chain notices that Tuesday's prescription was wrong. ## Common questions **How do I get out of an ERG death spiral mid-interval?** Stop pedalling and let the trainer release the resistance [Wahoo ERG Guide]. The loop is self-reinforcing — falling cadence makes the trainer add resistance, which makes cadence fall further — so pedalling harder cannot break it. If it keeps happening, Wahoo's advice is to reduce the target intensity rather than keep riding into the lockdown. **Should I shift gears in ERG mode?** Usually not. Shifting changes flywheel speed, which makes the trainer readjust and slows its response [Wahoo ERG Guide]. Set up before the session instead: small chainring, middle of the cassette, then shift easier until the trainer's floor power sits below the workout's lowest target. The exception is low-cadence intervals, where a harder gear may be needed to reach the target. **Why does it feel like there's no resistance when an interval starts?** Because the trainer is still catching up, and because you are probably helping. A KICKR settles on a new target in about 3 to 5 seconds and other trainers take 5 to 10; a fast cadence increase makes the trainer bleed resistance to avoid overshooting the target [Wahoo ERG Guide]. Hold your cadence steady and let it arrive. **Does a failed ERG session mean my FTP is set too high?** Only if it keeps happening across sessions. Wahoo's own troubleshooting names new test numbers as a cause of resistance that feels too high, and recommends lowering the workout intensity [Wahoo ERG Lockup]. One bad night is fatigue; every quality session collapsing in the last reps is a threshold problem, and the profile is what needs revisiting. **Which workouts shouldn't be ridden in ERG mode?** Wahoo lists several SYSTM sessions as Level mode work, and is unambiguous about the assessments: Full Frontal should never be done in ERG, and the second half of Half Monty is Level mode only [Wahoo ERG Guide]. The general rule is that any workout where your trainer visibly struggles to add or shed resistance mid-interval belongs in Level mode. ## References 1. **Wahoo ERG Guide.** [A Guide to using ERG mode](https://support.wahoofitness.com/hc/en-us/articles/4402565516946-A-Guide-to-using-ERG-mode). Wahoo Fitness Support. 2. **Wahoo ERG Lockup.** [Trainer resistance is too high or the trainer locks up](https://support.wahoofitness.com/hc/en-us/articles/4402741190802-Trainer-resistance-is-too-high-or-the-trainer-locks-up). Wahoo Fitness Support. 3. **Wahoo ELEMNT Support.** [Planned Workouts [ELEMNT]](https://support.wahoofitness.com/hc/en-us/articles/115001223770-Planned-Workouts-ELEMNT). Wahoo Fitness Support. 4. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. 5. **Skiba et al. 2012.** [Modeling the expenditure and reconstitution of work capacity above critical power](https://pubmed.ncbi.nlm.nih.gov/22382171/). Medicine & Science in Sports & Exercise. 6. **Kolsung et al. 2020.** [Physiological Response to Cycling With Variable Versus Constant Power Output](https://pmc.ncbi.nlm.nih.gov/articles/PMC7481374/). Frontiers in Physiology. 7. **Mater et al. 2021.** [Effect of Cycling Cadence on Neuromuscular Function: A Systematic Review of Acute and Chronic Alterations](https://pubmed.ncbi.nlm.nih.gov/34360206/). International Journal of Environmental Research and Public Health. --- # KICKR CORE 2 with the Zwift Cog: what one gear means for structured workouts URL: https://www.adaptcycling.com/guides/kickr-core-2-zwift-cog-structured-workouts Updated: 2026-08-07 Author: Jim Camut The Zwift Cog replaces your trainer's cassette with a single 14-tooth sprocket, and the Zwift Click puts 24 virtual gears on the handlebar [Zwift Cog Specs]. For structured training the honest answer is that this changes almost nothing. Zwift's own documentation is explicit: virtual shifting is available inside a workout only during free-ride segments, because in ERG mode Zwift is already controlling resistance to match the targets [Zwift Virtual Shifting FAQ]. The gear you are in stops mattering the moment ERG takes over. What the Cog changes is everything you ride outside it. ## What the Zwift Cog actually replaces One cassette, and nothing else. The Cog is a 14-tooth sprocket on a Shimano Hyperglide freehub body, compatible with almost any 8-13-speed bike [Zwift Cog Specs]. It contains no electronics. The 24 gears people associate with it live in software on the trainer and in Zwift, not in the part bolted to the freehub. The mechanics are simpler than the marketing. You pull the rear wheel, mount the bike, and park the rear derailleur roughly mid-range so the chain lines up with the single cog; a numbered ring then trims the cog laterally, and Zwift ships 10 of those positions for exactly that job [Zwift Cog Specs, DC Rainmaker CORE 2 Review]. The compatibility payoff is the reason the configuration exists — a 12-speed road bike and a 13-speed gravel bike go on the same trainer without a cassette swap or a tool [DC Rainmaker CORE 2 Review]. Zwift's published exclusions are mostly freehub-side — Micro Spline, XD/XDR and Campagnolo bodies are out — which is moot on a CORE 2 that ships with the Cog fitted [Zwift Cog Specs, Zwift Cog Compatibility]. The bike-side exclusion that does apply is a 1/8 inch fixed-gear or single-speed chain [Zwift Cog Specs]. And the decision reverses in both directions: Wahoo sells the CORE 2 three ways — 11-speed cassette, Zwift Cog and Click, or cassette-ready [Wahoo CORE 2 Product Page] — and independent testing of both variants found the hardware and firmware identical, the trainer unable to tell what is bolted to its freehub [DC Rainmaker CORE 2 Review]. For training purposes the change is narrow and worth stating in one sentence: you have given up the rear half of your gear range and kept the front. On a 2x bike that leaves two ratios. On a 1x gravel or mountain bike it leaves one. Everything below follows from that. ## Why ERG mode makes the single cog a non-issue Because ERG had already made your gear irrelevant. Zwift's FAQ says virtual shifting works inside a workout only during free-ride segments, since in ERG mode Zwift controls resistance to match the workout targets [Zwift Virtual Shifting FAQ]. During an ERG interval neither a Cog rider nor a cassette rider is choosing their effort with the drivetrain — the trainer is. ERG holds the prescribed watts by varying resistance against whatever gear and cadence you present it with, so a 4x8 at threshold executes identically on one cog and on eleven. An independent review that tested the CORE 2 across platforms states it without qualification: for structured workouts in ERG mode no shifting is required, so TrainerRoad and the rest run on the Zwift Cog version without issue [DC Rainmaker CORE 2 Review]. That is a hardware-behaviour observation rather than a vendor claim, and it matches what Zwift documents. The control protocols survive the drivetrain choice intact. Wahoo lists Wi-Fi, Bluetooth with up to three concurrent connections, ANT+ and ANT+ FE-C on the spec sheet for the Zwift Cog and Click configuration itself [Wahoo CORE 2 Product Page], and independent testing confirms ANT+ FE-C control, Bluetooth FTMS and Wi-Fi direct connect are all present on that variant [DC Rainmaker CORE 2 Review]. So TrainerRoad, the Wahoo app, Rouvy, TrainingPeaks Virtual and an ELEMNT head unit can each put the trainer into ERG and hold a target with the Cog fitted — the ELEMNT with no app in the room at all. That locates the Cog question correctly, which is the broader point our guide to training with Wahoo argues: the hardware is a delivery layer, and it renders faithfully whatever it is handed. The Cog changes the last few inches of that chain and nothing about what the week should contain. One ERG-adjacent detail does change — Wahoo's standard setup advice is small chainring plus the middle of the cassette, and with the Cog only the front half of that is available to you. If the recovery valleys stop feeling like recovery, our guide to ERG targets that feel too hard is the one to read next. ## Where one gear does bite: outside Zwift, outside ERG Simulation and level modes, in apps that do not carry Zwift's shifting protocol. There the trainer sets resistance from the gradient and your gear decides the cadence — and with the Cog your only remaining lever is the front chainring. Two ratios on a 2x bike, one on a 1x. What that feels like is specific. The gradient rises in simulation mode and you cannot answer it, so cadence falls until you are grinding or you concede the power. A sprint has no top gear to reach for; deliberate low-cadence torque work has no way to get low. The same independent review's guidance is well calibrated: if you ride Zwift and do structured work in something like TrainerRoad, the Cog version is fine; if your free riding happens in Rouvy or similar, take the cassette [DC Rainmaker CORE 2 Review]. The cadence cost is narrower than it sounds, and overclaiming it would be easy. A systematic review of cycling cadence reported that heterogeneity in the cadences studied, the assessment methods and the populations made comparison across studies difficult, and left unclear whether cadence itself contributes to the onset of fatigue [Mater et al. 2021]. So this is not a departure from some correct rpm — no such number is established. It is the loss of self-selection, which matters mainly because self-selection is how a rider normally sheds a little strain without ever deciding to. Riding those sessions in ERG instead is a good option and it is what most riders should do, but it is not a null substitution. Variable and constant power output at matched averages produce measurably different physiological responses [Kolsung et al. 2020], so a tempo hour held at a metronomic 240 W is not the same stimulus as the same hour over rolling terrain. Which is the real test: the drivetrain only becomes a training question if your plan deliberately contains non-ERG work — big-gear torque sets, race-simulation efforts, sprints off a free ride. ## What to check before your first structured week Four things, none of them slow. Pair over Bluetooth or Wi-Fi rather than ANT+ if you want the Click to work at all, trim the chainline, choose the front ring you will live in, and confirm which platform your non-Zwift riding actually happens on. Start with the pairing, because it fails silently. Zwift's FAQ states that ANT+ does not support virtual shifting, and that supported methods vary by trainer and may include Bluetooth, Zwift Companion bridging, Ethernet, Wi-Fi or USB [Zwift Virtual Shifting FAQ]. Wahoo says the same from the other side on the product page: Bluetooth is required for Zwift virtual shift [Wahoo CORE 2 Product Page]. A rider who reaches for an ANT+ dongle out of habit gets a perfectly working trainer and a dead Click, with nothing on screen explaining why. On whether the Click works anywhere but Zwift, two credible sources disagree and we will not pick between them. DC Rainmaker's November 2025 review describes Rouvy's virtual shifting as an unofficial workaround a Zwift update could break, and recommends a cassette on that basis [DC Rainmaker CORE 2 Review]. Zwift's current documentation states support plainly in two places: virtual shifting is supported in ROUVY with compatible Zwift hardware, and the Click is compatible with Zwift and ROUVY [Zwift Virtual Shifting FAQ, Zwift Cog Specs]. The safe reading is that this is one sanctioned exception rather than a general capability — Zwift names no other platform. The setup itself is small: trim the chainline, pick a chainring, stop thinking about gears — which is close to the whole point of the thing. What none of it decides is what Tuesday should contain. AdaptCycling reads your ride history, estimates threshold from your rolling power curve rather than a scheduled test, and sends today plus the next six days to your Wahoo account as structured workouts with targets as ranges rather than a single midpoint, reaching the KICKR through exactly the ERG path the Cog leaves untouched. That runs on the free tier. ## Common questions **Can I do structured workouts on a KICKR CORE 2 with the Zwift Cog?** Yes, with no compromise in ERG mode. Zwift's own FAQ notes that virtual shifting is unavailable during ERG segments because Zwift is controlling resistance to match the targets [Zwift Virtual Shifting FAQ], and independent testing confirms no shifting is required for structured work on the Cog version [DC Rainmaker CORE 2 Review]. **Will TrainerRoad or another app control the trainer with the Cog fitted?** Yes. Wahoo lists Wi-Fi, Bluetooth, ANT+ and ANT+ FE-C on the Zwift Cog and Click configuration's own spec sheet [Wahoo CORE 2 Product Page], and independent testing confirms FE-C control, Bluetooth FTMS and Wi-Fi direct connect on that variant [DC Rainmaker CORE 2 Review]. The trainer cannot tell what is bolted to its freehub. **Does the Zwift Click work outside Zwift?** Sources disagree, so treat it as one exception rather than a capability. A November 2025 review calls Rouvy's support an unofficial workaround Zwift could break [DC Rainmaker CORE 2 Review]; Zwift's current documentation says virtual shifting is supported in ROUVY with compatible Zwift hardware [Zwift Virtual Shifting FAQ]. No other platform is named by either. **Can I put a normal cassette back on later?** Yes, and Wahoo sells the CORE 2 three ways to begin with — with an 11-speed cassette, with Zwift Cog and Click, or cassette-ready [Wahoo CORE 2 Product Page]. The hardware and firmware are identical across configurations, so switching either direction is a cassette swap, not a platform change [DC Rainmaker CORE 2 Review]. **Do I lose sprint or low-cadence work with one gear?** Only outside ERG. In ERG the trainer sets resistance to the target whatever gear you are in. In simulation or level mode the front chainring is your only lever, so big-gear torque sets and sprint efforts lose their range — and constant and variable power at matched averages are not physiologically interchangeable [Kolsung et al. 2020]. ## References 1. **Zwift Virtual Shifting FAQ.** [Virtual Shifting FAQ](https://support.zwift.com/virtual-shifting-faq-r16UiRFlT). Zwift Support. 2. **Zwift Cog Compatibility.** [Zwift Ready and Zwift Cog Compatible Trainers](https://support.zwift.com/en_us/zwift-ready-and-zwift-cog-compatible-trainers-r1x1sNas0). Zwift Support. 3. **Zwift Cog Specs.** [Zwift Cog and Click Upgrade Kit](https://us.zwift.com/products/zwift-cog-and-click-upgrade-kit). Zwift. 4. **Wahoo CORE 2 Product Page.** [KICKR CORE 2 with Zwift Cog and Click](https://www.wahoofitness.com/devices/indoor-cycling/bike-trainers/kickr-core-2-zwift-cog-and-click-buy). Wahoo Fitness. 5. **DC Rainmaker CORE 2 Review.** [Wahoo KICKR CORE 2 In-Depth Review](https://www.dcrainmaker.com/2025/11/wahoo-kickr-core-2-in-depth-review.html). DC Rainmaker — hardware-behaviour observation only. 6. **Mater et al. 2021.** [Effect of Cycling Cadence on Neuromuscular Function: A Systematic Review of Acute and Chronic Alterations](https://pubmed.ncbi.nlm.nih.gov/34360206/). International Journal of Environmental Research and Public Health. 7. **Kolsung et al. 2020.** [Physiological Response to Cycling With Variable Versus Constant Power Output](https://pmc.ncbi.nlm.nih.gov/articles/PMC7481374/). Frontiers in Physiology. --- # Which Zwift FTP test to take: Standard, Ramp, Lite, or Elite Verification URL: https://www.adaptcycling.com/guides/which-zwift-ftp-test-to-take Updated: 2026-08-07 Author: Jim Camut Zwift ships four FTP tests and the menu gives you no basis for choosing. Two facts settle most of it. FTP Ramp Test (Lite) is not a shorter or easier test — it is a body-weight split at 60 kg. And ERG mode does not run an FTP test: Zwift holds resistance neutral and you shift gears yourself [Zwift Support 2026]. Having ridden the ramp myself, it is the one I would put a nervous rider on — you are only really in pain for the last few minutes. ## The four tests, and the only real decision among them You are choosing between two. FTP Test (Standard) stacks a warmup, a ramp effort, a 5-minute effort and a 20-minute effort, and derives FTP from that final 20 minutes. The two ramp tests are the same progressive ramp, split by rider weight. Elite Racing Verification is a race-categorisation tool, not a training one [Zwift Support 2026]. Zwift's options table is precise. FTP Test (Standard) is listed for general FTP assessment, its format given as warmup, ramp effort, 5-minute effort and 20-minute effort, with the note that FTP is based on the final 20-minute segment. Both ramp tests are described identically as a progressive ramp test, carrying the same note: FTP may display as 100 at the start, and the final FTP is calculated during the workout [Zwift Support 2026]. The difference sits in the intended-for column, not the format column. The Elite Racing Verification Test is the odd one out; ignore it unless you have been asked for it. Zwift lists it for elite racing pre-verification, describes the format as multiple timed efforts, and notes that ERG mode is not used during this test [Zwift Support 2026]. It exists to substantiate a performance claim for race categorisation, not to set training zones. Whichever you pick, note that Zwift uses FTP only during workouts, so the number moves your workout targets and nothing else. ## Lite is a weight split at 60 kg, not a short cut Zwift describes both ramp tests in identical language. The standard FTP Ramp Test is intended for riders 60 kg or over and more experienced riders; FTP Ramp Test (Lite) is intended for riders under 60 kg and newer riders [Zwift Support 2026]. Lite is not shorter, not easier, and not less valid. The name misleads almost everyone. Riders read Lite as the quick version and pick it to save an evening, or avoid it assuming a lightweight test yields a lightweight number. The two rows differ only in who each test is for, and Zwift gives two criteria on each side: a body-weight threshold at 60 kg, and a rider-experience judgement [Zwift Support 2026]. Splitting a ramp on body weight is defensible, though Zwift does not publish its reasoning. A ramp raises load in absolute watts, so a fixed step is a far larger relative demand on a 55 kg rider than on an 85 kg one — the lighter rider is effectively riding a steeper ramp. Ramp rate is not a neutral choice. Michalik and colleagues [Michalik et al. 2019] put the same men through two ramp rates and recorded peak power of 341 watts on the slower protocol against 393 watts on the faster one, a gap of 51.69 watts, or roughly 15%. Because a ramp derives FTP from that peak, a steeper relative ramp inflates the number you take home. The practical instruction is dull and load-bearing: pick by Zwift's stated criteria, then stay on that test. A rider who takes the standard ramp in March and Lite in June has changed method, not fitness, and any movement in the number is uninterpretable. That matters most near the 60 kg line. ## ERG does not run the test, and the 20-minute segment is not fresh Two corrections that change how you ride. ERG does not hold your target during an FTP test — Zwift keeps resistance neutral and you shift gears manually, which makes the on-or-off question close to moot [Zwift Support 2026]. And Zwift has no standalone 20-minute test: that effort arrives after a ramp effort and a 5-minute effort. Zwift is explicit. In ERG mode during an FTP test you still shift gears manually to adjust resistance, because Zwift does not automatically adjust resistance during FTP test intervals — resistance stays neutral so you control cadence and power output throughout [Zwift Support 2026]. Riders hunting for the correct ERG setting are hoping the trainer will pace the effort for them. It will not. Since cadence is yours too, do not go hunting for an optimal rpm. Mater and colleagues [Mater et al. 2021] reviewed cycling cadence and neuromuscular function, reporting that heterogeneity in the cadences selected, the assessment methods and the populations studied made comparison across studies difficult. The literature does not hand you a single best cadence. Ride what you habitually ride, and keep it consistent between tests. The second correction matters more. The only route to a 20-minute test on Zwift is FTP Test (Standard), and its 20-minute segment is preceded inside the same workout by a ramp effort and a 5-minute effort [Zwift Support 2026]. That is not a warm-up; it is prior severe-intensity work. Chorley and colleagues [Chorley et al. 2021] modelled the recovery of W-prime, the finite work capacity above critical power, and found it reconstitutes bi-exponentially with a slow-component time constant of 388 seconds — rising to 716 seconds after a second maximal effort, and at three minutes the second recovery had restored about 9% less than the first. Capacity spent earlier is only partly back when the 20 minutes begins. The consequence is comparability. Warm-up structure alone shifts the FTP a 20-minute test reports [Tramontin et al. 2022], and a preceding ramp plus 5-minute effort is a much larger difference than a warm-up. Compare a Zwift Standard result only to previous Zwift Standard results. ## Choosing tonight, and when not to test at all If dread is what has stopped you, take the ramp matching your weight — it front-loads easy and concentrates the hard part into the last few minutes. If you want a number anchored in sustained work, take Standard and pace it evenly. If you already ride with power, you may need neither [Zwift Support 2026]. Every page ranking for this question compares protocols on accuracy and stops there. Tolerability is what riders actually decide on, and it is a legitimate criterion. I have taken the ramp, and it is not bad at all if you want a quicker and somewhat accurate number — you are only really in pain for the last few minutes rather than facing a full 20-minute effort. An abandoned or badly paced 20-minute effort produces a falsely low FTP, and short estimates are noisy against genuine sustained efforts: individually the two have disagreed by 40 to 60 watts [Borszcz et al. 2018]. The best test is the one you finish. State the trade honestly. The ramp is also the protocol most likely to hand you a number you cannot hold for an hour, particularly if you sprint well — our guide on why a ramp test FTP comes out too high covers that mechanism. A ramp result is partly an artifact of the step size you happened to ride [Michalik et al. 2019], which is why somewhat accurate is the right ceiling. Zwift's own troubleshooting agrees: if your workouts feel too difficult, your FTP may be set too high. There is also the option of not testing. Zwift states that without a completed FTP test it estimates your FTP from your weight and the fitness level profile you selected, and that once you have started riding it estimates from your power data [Zwift Support 2026]. That is the broader frame our pillar on estimating FTP without a forced test works in — your threshold is already implied by the efforts sitting in your ride history. The limit of Zwift's version is its window: it sees Zwift rides, and uses the result only inside Zwift workouts. AdaptCycling fits a Critical Power model to the mean-maximal-power curve built from your whole ride history, indoor and outdoor, so the estimate rests on dozens of efforts across months rather than one evening's willingness to suffer. A Zwift test simply joins that curve. Whichever protocol you pick, run it roughly twice a year rather than every six weeks — our guide on how often to test FTP makes that case. ## Common questions **Which Zwift FTP test is best?** For most riders, the ramp matching your weight — the standard FTP Ramp Test at 60 kg or over, Lite under 60 kg [Zwift Support 2026]. It is far easier to face than a 20-minute effort, at the cost of being the protocol most likely to read high. Choose FTP Test (Standard) if you want the number anchored in sustained work and can pace 20 minutes evenly. **Should ERG mode be on or off for a Zwift FTP test?** It barely matters, because ERG does not control an FTP test. Zwift keeps resistance neutral and you shift gears manually, so you control cadence and power output either way [Zwift Support 2026]. Zwift separately documents that ERG is not used at all during the Elite Racing Verification Test. Plan to pace the effort yourself. **How long is the Zwift FTP test?** It depends which one. FTP Test (Standard) is longest by construction — a warmup, a ramp effort, a 5-minute effort and a 20-minute effort in one workout [Zwift Support 2026]. Zwift publishes no fixed duration for the ramp tests, and a ramp is self-terminating: it ends when you can no longer hold the step, so a fitter rider stays on longer. **What is the difference between the Zwift FTP Ramp Test and Ramp Test Lite?** Intended rider, not length or difficulty. Zwift describes both as a progressive ramp test with identical notes, splitting them by audience: the standard ramp for riders 60 kg or over and more experienced riders, Lite for riders under 60 kg and newer riders [Zwift Support 2026]. Because a ramp raises load in absolute watts, a fixed step is a steeper relative demand on a lighter rider, and ramp rate measurably changes peak power [Michalik et al. 2019]. **Do I need to take a Zwift FTP test at all?** Not necessarily. Zwift estimates FTP from your weight and chosen fitness level profile before you test, then from your power data once you have been riding [Zwift Support 2026]. With months of power files, a model fit to your power curve gets you close without a forced effort. Test when you want a clean anchor before a block, and expect twice a year to be enough. ## References 1. **Zwift Support 2026.** [FTP Test](https://support.zwift.com/en_us/ftp-test-HypJnI_NH). Zwift Support. 2. **Michalik et al. 2019.** [Influence of Intensity RAMP Incremental Test on Peak Power, Post-Exercise Blood Lactate, and Heart Rate Recovery in Males: Cross-Over Study](https://pmc.ncbi.nlm.nih.gov/articles/PMC6843469/). International Journal of Environmental Research and Public Health. 3. **Chorley et al. 2021.** [Bi-exponential modelling of W-prime reconstitution kinetics in trained cyclists](https://pmc.ncbi.nlm.nih.gov/articles/PMC8854279/). European Journal of Applied Physiology. 4. **Mater et al. 2021.** [Effect of Cycling Cadence on Neuromuscular Function: A Systematic Review of Acute and Chronic Alterations](https://pubmed.ncbi.nlm.nih.gov/34360206/). International Journal of Environmental Research and Public Health. 5. **Tramontin et al. 2022.** [Functional Threshold Power Estimated from a 20-minute Time-trial Test is Warm-up-dependent](https://pubmed.ncbi.nlm.nih.gov/34749416/). International Journal of Sports Medicine. 6. **Borszcz et al. 2018.** [Functional Threshold Power in Cyclists: Validity of the Concept and Physiological Responses](https://pubmed.ncbi.nlm.nih.gov/29801189/). International Journal of Sports Medicine. --- # How to do a structured indoor workout without a Zwift subscription URL: https://www.adaptcycling.com/guides/indoor-workout-without-zwift-subscription Updated: 2026-08-07 Author: Jim Camut This one came out of our own August. Riding outside all summer, our founder had paused his Zwift subscription until October, then wanted a single structured session on a KICKR CORE 2 and did not want to reactivate a subscription for one ride. AdaptCycling pushed the planned workout to his ELEMNT BOLT, the BOLT drove the trainer in ERG, and the session got done. Wahoo documents this path directly: planned workouts can be completed indoors, with or without the computer controlling the trainer [Wahoo ELEMNT Support]. ## What the head unit can actually do on its own Wahoo's answer is published and unambiguous. Planned workouts can be completed indoors, with or without your ELEMNT, BOLT or ROAM controlling your KICKR or another ANT+ FE-C compatible smart trainer [Wahoo ELEMNT Support]. Nothing has to be running on a phone or a laptop while you ride. The control surface is one page with six modes on it: Passive, Level, ERG, Resistance, Route and Planned Workout, with Passive the default when you enter trainer control [Wahoo Trainer Control]. Planned Workout is the one that matters here — it follows an uploaded workout according to the resistance design of that workout, so the computer handles the interval timing, the target display and the resistance together. Passive is the inverse, worth knowing for the weeks your subscription is active again: the computer becomes a mounted display, recording metrics while another app drives the trainer [Wahoo Trainer Control]. Getting there takes four steps. Pair the smart trainer to the computer exactly like any other sensor, go to the Menu page and set Location to Indoor, then start pedalling — the computer switches Location to KICKR on its own. If the control page does not appear, confirm Location is KICKR and page through the stack to find it [Wahoo Trainer Control]. One gap is worth naming: that article covers ELEMNT, BOLT 1/2 and ROAM 1/2, and there is no equivalent for ACE, BOLT 3 or ROAM 3. The capability is documented for the current generation all the same, inside its planned-workouts article [Wahoo ELEMNT ACE Support]. Both generations do this; only the older one has the step-by-step written down. The compatibility is broader than the brand on the box: Wahoo says these computers can control most ANT+ FE-C enabled third-party trainers, with the hedge that you should check FE-C compatibility with your trainer's manufacturer [Wahoo Trainer Control]. What none of this requires is a subscription to a riding platform. Wahoo lists a planned-workout provider among the requirements, then footnotes an exception — a short set of 2 fitness tests and 7 workouts ships on ELEMNT computers to test out the feature functionality [Wahoo ELEMNT Support]. The list is specific: 20- and 8-minute FTP tests, 12x1min MAP, 2x20min FTP, 40/20s, 5x5min FTP, plus climbing, mixed and sprinting sessions. That is enough to ride intervals tonight with nothing linked to anything. One caveat, because the difference is real: that footnote appears only in the legacy ELEMNT article — the ACE, BOLT 3 and ROAM 3 article lists a provider as a flat requirement [Wahoo ELEMNT ACE Support], so do not assume the starter set ships on current hardware. ## The one setting that decides whether the trainer resists at all Planned Workout mode follows the resistance design of the workout, which in most cases is ERG — but if there are no power targets, the trainer will not provide resistance [Wahoo Trainer Control]. A session built on heart rate or perceived exertion renders its intervals perfectly and leaves the flywheel free. This is the failure that reads as a broken trainer. Wahoo's planned workouts accept power, heart rate, rate of perceived exertion or speed targets [Wahoo ELEMNT Support], which is genuinely useful — a rider without a power meter is not locked out of structured training on an ELEMNT. Indoors, though, only one of those four target types gives the trainer a number to hold. Load an RPE-based session onto a KICKR, start it, and the intervals count down on schedule, the alerts fire, the graph draws itself, and the resistance sits exactly where it was when you started. The fix is not to abandon the session. Switch the control page to Level or Resistance mode [Wahoo Trainer Control] and ride the interval structure the computer is still displaying. The head unit keeps doing the timing and the prompting; you supply the effort ERG was going to enforce. That is a worse session than a power-targeted one, and it is a long way from no session. The better answer is upstream. Whether an indoor workout carries power targets is a property of the plan, decided before anything reaches the handlebars, and it is the single thing to check when you build or choose a session you intend to ride in ERG. Where the plan comes from is a separate subject, and our companion guide on getting a structured workout onto an ELEMNT without TrainingPeaks covers it properly. The one-line version: the size of the window is a property of your provider, not your computer, and today plus six days is the ceiling [Wahoo ELEMNT Support]. ## What you give up when you leave the virtual world Everything that makes Zwift worth paying for. No virtual world, no group rides, no racing, no gamification — just a rider, a set of numbers and the intervals. For a shoulder-season session that is fine. For January motivation it may well not be. State the trade before the workaround, because it is the whole decision. Zwift is the best indoor riding experience available, and pausing it for the summer is normal rider behaviour rather than a defection — most riders who pause in June want it back when the weather turns, and they are right to. Nothing here is an argument against that. The claim is narrow and it is the only one this article makes: a paused subscription should not cost you your training structure for the two or three indoor sessions a summer actually needs. The research is clearer about why the loss matters than about how large it is. A positive affective response during moderate-intensity exercise predicts future physical activity, while how you feel once it is over does not [Rhodes & Kates 2015]. Boredom during the session is what costs you the next one. So the cost of a head-unit session is not physiological. The intervals are the intervals, the trainer holds the watts, and the ride file is the same file. What you lose is the part that makes you want to start, and that part is not decoration. ## When this is the right call — and when it isn't It is a volume question, not a principle. If indoors is where your winter happens, the subscription is buying adherence and you should keep it. If indoors is a weather contingency you hit a handful of times between May and September, the head unit answers it completely. Three or four trainer sessions a week for four months is a different problem from one wet Tuesday in July. At winter volume you are asking the room to hold your attention across roughly sixty sessions, and a platform built to do that will beat a numbers screen every time. At summer volume you are asking it once, then you go back outside. Reactivating a subscription for that is a reasonable thing to decide against, and it is the only case this article is about. The corollary matters as much: if your subscription is active, use it. We ship a free Zwift workout generator for exactly that reason. The head-unit path is what you reach for when the subscription is paused, not a replacement for it. This sits inside the broader question our guide to training with Wahoo takes on, which is that the head unit is an excellent delivery layer and decides nothing at all. What reaches it still has to come from somewhere. AdaptCycling reads your ride history, estimates threshold from your own power curve rather than a scheduled test, and pushes today plus the next six days to your Wahoo account as structured workouts carrying power targets — which is precisely what makes the ERG path above work rather than coast. When the week breaks, the revised sessions replace the old ones on the device. The push runs on the free tier, and there is no riding-platform subscription anywhere in the chain. ## Common questions **Can a Wahoo ELEMNT run a structured workout on a smart trainer with no app running?** Yes. Wahoo states that planned workouts can be completed indoors, with or without your ELEMNT, BOLT or ROAM controlling your KICKR or another ANT+ FE-C compatible smart trainer [Wahoo ELEMNT Support]. The same answer appears for ELEMNT ACE, BOLT 3 and ROAM 3 [Wahoo ELEMNT ACE Support]. You pair the trainer as a sensor, set Location to Indoor, and select Planned Workout on the trainer control page [Wahoo Trainer Control]. **Do I need any subscription to ride a planned workout indoors on an ELEMNT?** Wahoo lists a planned-workout provider as a requirement, then footnotes a short set of 2 fitness tests and 7 workouts shipped on ELEMNT computers [Wahoo ELEMNT Support] — enough to ride intervals with nothing linked. That footnote appears only in the legacy article, not the ACE, BOLT 3 and ROAM 3 one [Wahoo ELEMNT ACE Support]. Beyond the starter set you need a provider, and what it costs is its decision, not Wahoo's. **Why is my trainer providing no resistance during a planned workout?** Almost certainly because the workout has no power targets. Planned Workout mode follows the resistance design of the session, but with no power targets the trainer will not provide resistance [Wahoo Trainer Control]. Interval targets can be power, heart rate, RPE or speed [Wahoo ELEMNT Support], and only power gives the trainer something to hold. Switch to Level or Resistance mode and ride the structure manually. **Does this work with a trainer that isn't a Wahoo?** Wahoo says the ELEMNT, BOLT and ROAM can control most ANT+ FE-C enabled third-party trainers, and asks you to check FE-C compatibility with your trainer's manufacturer [Wahoo Trainer Control]. Both planned-workout articles use the same wording — your KICKR or another ANT+ FE-C compatible smart trainer [Wahoo ELEMNT Support]. **Is riding workouts off a head unit a replacement for Zwift?** No, and it is not meant to be. You get the intervals, the targets and the ride file, and you lose the world, the group rides, the racing and the gamification. Enjoyment during exercise predicts whether you come back to it [Rhodes & Kates 2015], so for a winter built indoors that loss is expensive. For a handful of sessions in a summer spent outside, it is not. ## References 1. **Wahoo ELEMNT Support.** [Planned Workouts [ELEMNT]](https://support.wahoofitness.com/hc/en-us/articles/115001223770-Planned-Workouts-ELEMNT). Wahoo Fitness Support. 2. **Wahoo Trainer Control.** [Control a smart trainer (KICKR, CORE, SNAP, ROLLR, or BIKE) with ELEMNT, BOLT 1/2, or ROAM 1/2](https://support.wahoofitness.com/hc/en-us/articles/115000129570-Control-a-smart-trainer-KICKR-CORE-SNAP-ROLLR-or-BIKE-with-ELEMNT-BOLT-1-2-or-ROAM-1-2). Wahoo Fitness Support. 3. **Wahoo ELEMNT ACE Support.** [Planned workouts with ELEMNT ACE, BOLT 3 and ROAM 3](https://support.wahoofitness.com/hc/en-us/articles/22710662013458-Planned-workouts-with-ELEMNT-ACE-BOLT-3-and-ROAM-3). Wahoo Fitness Support. 4. **Rhodes & Kates 2015.** [Can the Affective Response to Exercise Predict Future Motives and Physical Activity Behavior? A Systematic Review of Published Evidence](https://pubmed.ncbi.nlm.nih.gov/25921307/). Annals of Behavioral Medicine. --- # How to get structured workouts on a Karoo without TrainingPeaks URL: https://www.adaptcycling.com/guides/karoo-structured-workouts-without-trainingpeaks Updated: 2026-08-04 Author: Jim Camut The common answer to this question is that you cannot, and that answer is out of date. Part of the reason it persists is that Hammerhead's own smart-trainer article still says only workouts synced from TrainingPeaks or uploaded to the Dashboard will work. Its Workouts article, updated more recently, is broader: workouts import from your connected third-party accounts or through the Dashboard. Both are Hammerhead's. The second one is current, and it opens two routes that cost nothing. ## The free route almost nobody mentions: upload the file The Hammerhead Dashboard imports .ZWO and .FIT workout files directly. Add the workout, and the next time your Karoo reaches the internet it syncs across [Hammerhead Workout Import]. There is no subscription anywhere in that chain, and no partner list to appear on. This is the answer to the literal question, and it has been available the whole time. If a platform can export a structured workout as a file — and most can, including free ones — that workout can reach your Karoo. Open the Dashboard, choose Workouts in the sidebar, click Add Workout, and drop the file in [Hammerhead Workout Import]. The forum answer that circulates for this question is that you have to recreate your sessions by hand in TrainingPeaks. You do not. The rules the importer enforces are stricter than most exporters assume, which is why this route sometimes fails in a way that looks like a bug. For a .ZWO file, every step must be a warmup, cooldown, free ride, repeating interval, or steady state — maximum-effort and ramp steps are unsupported — and every step must carry a power or cadence target, with zone-based power targets rejected outright [Hammerhead Workout Import]. A file that breaks any of those rules is rejected whole rather than partially imported. If your export contains a ramp or an all-out effort, that single step is what stopped the whole session loading. One rule decides more than the others. A .FIT workout's steps can target speed, heart rate, cadence, or power — heart rate and power as absolute values or percentages, speed and cadence as absolute values — and heart rate is not on the .ZWO list at all [Hammerhead Workout Import]. The two formats are therefore not interchangeable. If you train to power, export .ZWO or .FIT and it makes little difference. If you train to heart rate, only .FIT will carry your targets, and choosing the wrong export is the difference between a usable session and a set of numbers you cannot follow. The honest limit of this route is that a hand-uploaded file is a snapshot. It does not change when your week does, and doing it every evening is a chore that quietly stops after a fortnight. It proves the device is not locked down. It does not solve the problem of keeping a plan current. ## Connecting a provider instead, and which ones qualify The Dashboard links ten third-party accounts: AXS, Intervals.icu, Komoot, Ride with GPS, Sentiero, Strava, Suunto, TrainerRoad, TrainingPeaks, and Xert [Hammerhead Account Linking]. Not all of them send workouts — several are route or activity destinations — but the list has not been a single name for some time. The change worth knowing about is Intervals.icu, which now syncs with the Karoo directly. It is the option most likely to end this search for a rider who does not want another subscription, because its analysis is genuinely strong and its planning is free to use. For several years the only workaround was a community-built script that pushed workouts in unofficially, with a warning not to press Sync afterwards or they would be removed. That era is over, and pages describing it are describing 2022. Xert is the other one worth naming, because it does something structurally different: it models your fatigue continuously and reshapes what it recommends rather than serving a fixed calendar. TrainerRoad appears on the account list and does real within-plan adaptation for indoor power sessions. TrainingPeaks remains the deepest structured-workout builder available and is where most coached athletes already work — if a coach writes your plan, moving it elsewhere to save a subscription just creates a second place for the truth to live. What none of them is built to do is rebuild your next fortnight around the week you actually had. That is a different job from delivering a workout, and it is the one the Karoo ecosystem leaves open — Hammerhead ships no training platform of its own, which is the subject of our broader guide to training with a Hammerhead Karoo. The device renders a session extremely well. Deciding which session it should be was never any part of the chain's job. ## Choosing a route when you don't want another subscription Three cases, and the right answer depends on where your plan is authored rather than on price. Upload files by hand if your sessions are fixed. Connect a free provider if you want a calendar. Use something that reads your rides back if what breaks your plan is your life rather than your legs. If your training is a repeating block you already know — the same four sessions on a six-week rotation — hand-uploading is entirely reasonable. Build them once, upload them once, and the Karoo holds them. The chore only bites when the plan is supposed to change. If you want a calendar without paying for one, connect Intervals.icu and stop there. It costs nothing, it syncs, and it will do a better job of showing you what you have done than most paid tools. This is a real answer, not a consolation prize, and any page that tells you TrainingPeaks Premium is the only route to a week of workouts on a Karoo is out of date. The third case is the one worth thinking about properly. Most amateur plans do not fail because the intervals were wrong. They fail because a work trip erased Wednesday, a chest infection took ten days, or a four-hour group ride on Saturday quietly made Tuesday's session a bad idea. A calendar cannot notice any of that, and neither can a file you uploaded last Sunday. That is where AdaptCycling sits. We read your ride history from the source you already use, estimate your threshold from your own power curve rather than asking you to schedule a maximal test — the critical-power relationship makes that possible from efforts you have already ridden [Jones et al. 2010] — and build a plan from what you have actually been doing. When the week goes sideways, the plan is rewritten around it rather than left to rot. Rides recorded on a Karoo help here more than most, because the device opens a new lap at every interval boundary [Hammerhead Karoo Workouts], so the analysis reads what you held rep by rep instead of guessing where the reps were. ## Common questions **Do I need a TrainingPeaks subscription for structured workouts on a Karoo?** No. The Hammerhead Dashboard imports .ZWO and .FIT workout files directly with no subscription involved [Hammerhead Workout Import], and the Dashboard links ten third-party accounts including Intervals.icu, Xert, and TrainerRoad [Hammerhead Account Linking]. The claim that TrainingPeaks is the only route appears in Hammerhead's older smart-trainer article [Hammerhead Trainer Support] and is contradicted by its own more recent Workouts documentation [Hammerhead Karoo Workouts]. **How do I upload my own workout to a Karoo?** In the Hammerhead Dashboard, select Workouts in the sidebar, click Add Workout, and choose or drag in a .ZWO or .FIT file. The workout syncs to the Karoo the next time the device connects to the internet [Hammerhead Workout Import]. Any file type other than those two is rejected. **Why did my .ZWO file get rejected?** Because one step broke a rule and the file is rejected whole. Steps must be warmup, cooldown, free ride, repeating interval, or steady state — maximum-effort and ramp steps are unsupported — and every step needs a power or cadence target, with zone-based power targets not accepted [Hammerhead Workout Import]. A ramp or an all-out sprint in the export is the usual culprit. **Can I use heart-rate targets instead of power?** Yes, but the format decides it. .FIT workout steps accept heart rate as an absolute value or a percentage; the .ZWO rules list only power and cadence [Hammerhead Workout Import]. So a heart-rate session has to arrive as a .FIT workout. Note also that RPE-based workouts are not supported on the Karoo in any format [Hammerhead Karoo Workouts]. **Is the free Intervals.icu route as good as a paid one?** For getting a planned workout onto the device, yes — it syncs with the Karoo and costs nothing. What no calendar-based tool does, paid or free, is revise the plan when your week changes. That is a different capability from delivery, and it is worth separating the two when you compare options. ## References 1. **Hammerhead Karoo Workouts.** [Karoo OS - Workouts](https://support.hammerhead.io/hc/en-us/articles/18001019006363-Karoo-OS-Workouts). Hammerhead Support. 2. **Hammerhead Workout Import.** [Dashboard - Importing ZWO & FIT based Workouts](https://support.hammerhead.io/hc/en-us/articles/4416694669851-Dashboard-Importing-ZWO-FIT-based-Workouts). Hammerhead Support. 3. **Hammerhead Trainer Support.** [Karoo OS - Smart Trainer Integration](https://support.hammerhead.io/hc/en-us/articles/25559624673179-Karoo-OS-Smart-Trainer-Integration). Hammerhead Support. 4. **Hammerhead Account Linking.** [Dashboard - Linking Third-Party Accounts](https://support.hammerhead.io/hc/en-us/articles/360006755614-Dashboard-Linking-Third-Party-Accounts). Hammerhead Support. 5. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. --- # Which training plan app actually works with a Hammerhead Karoo URL: https://www.adaptcycling.com/guides/karoo-training-plan-app Updated: 2026-08-04 Author: Jim Camut There is no Hammerhead training app, and that absence is the whole reason this question gets asked. The brands the Karoo competes with all ship something that decides what you should ride. Hammerhead ships a slot. Whatever you connect to it becomes your plan, right down to the FTP and training zones the device holds you to, which the Karoo syncs in from whichever platform you nominate rather than working out for itself. So the real question is not which app works. It is which job you need the app to do. ## Why the Karoo arrives without a plan Hammerhead builds a head unit and a dashboard, not a training platform. The Karoo does not generate workouts, and it does not set your own numbers either: FTP, power zones, heart-rate zones, and rider weight all sync in from a third-party platform you choose [Hammerhead Training Metrics]. The contrast with the obvious rival is stark. A Garmin Edge recommends a daily suggested workout based on your previous activities saved to Garmin Connect, once you have ridden with heart rate and power for a week [Garmin Edge 1050 Manual]. The computer has an opinion about your training. A Karoo never will, and no firmware update is going to change that, because Hammerhead competes on being open rather than on owning the decision. The training-metrics setting is the sharpest illustration of where the line sits. You nominate Strava, TrainingPeaks, intervals.icu, or anything else connecting through Hammerhead's public API as your sync partner, and the Karoo's FTP and zones follow it. You cannot change that partner from the device — only from the Dashboard — and if you would rather not sync at all, you edit the numbers by hand [Hammerhead Training Metrics]. Even the threshold your zones are built from is imported. Which means the search that brought you here resolves to a narrower question than it looks. No app turns a Karoo into a coached device, because Hammerhead never built the first-party training software that Garmin and Wahoo both ship. There is a slot, four or five candidates for it, and they behave far less alike than the comparison articles suggest. ## What each app actually does once it is connected The Dashboard links ten third-party accounts [Hammerhead Account Linking], but linking is not the same as delivering workouts. Of the names people actually shortlist, two push structured sessions to the device, one runs its own app on it instead, and one route needs no account at all. TrainingPeaks is the long-standing default and still the deepest structured-workout builder available. It is also the answer nearly every page on this question gives, which has more to do with how long it was the only one than with a comparison anybody ran recently. If a coach already writes your plan there, leave it there — moving it to save a subscription only creates a second place for the truth to live. TrainerRoad has a real integration with one constraint that belongs in every comparison and appears in almost none. You authorize it from the Accounts section of the Hammerhead Dashboard, and a scheduled workout only reaches the device if you set it to be completed outside. TrainerRoad then states plainly that outside workouts are made available to Hammerhead on the day they are scheduled, and that you will not see them in advance of that day [TrainerRoad Hammerhead Sync]. The Ramp Test cannot be pushed outside at all. Completed rides return to your TrainerRoad calendar automatically. Xert behaves differently on a Karoo than it does on a Wahoo, by its own account. Xert describes its ELEMNT and RIVAL integration as pushing adaptive workouts directly to the device; it describes the Karoo integration as launching the Xert EBC app on the Karoo mid-ride for real-time feedback and activity sync [Xert Integrations]. That is a meaningful difference. On a Karoo, Xert is an app you ride inside rather than a session sitting in the Karoo's own workout list, so if you expected today's Xert workout to appear alongside the others, it will not. Intervals.icu is the one that ends this search for most riders who do not want another subscription. It syncs rides and planned workouts, it costs nothing, and it can double as the training-metrics partner above [Hammerhead Training Metrics]. And underneath all of them sits the route that needs no account on any list: the Dashboard imports .ZWO and .FIT workout files directly, and the Karoo collects them the next time it reaches the internet [Hammerhead Workout Import]. ## The delivery window nobody quotes in the comparison Every one of these options is a rolling window, not a training plan parked on your bars. TrainerRoad publishes its window explicitly — the day of, and no earlier [TrainerRoad Hammerhead Sync]. That one line reframes the shortlist: you are not choosing where your plan lives, you are choosing what arrives tomorrow. Riders compare these platforms on plan quality, buy one, and then discover the device is showing a day or a few days rather than the block they were sold. That is not a fault in any of them. No provider can place week nine on a head unit today, and a tool implying otherwise is describing its own calendar rather than the device. It is worth reading every integration claim with that distinction in hand. The useful consequence is that authoring a plan and delivering the next session are separable jobs, and most people asking this question are actually shopping for the first one. Delivery is already solved and free — a hand-uploaded file reaches the Karoo perfectly well [Hammerhead Workout Import]. What a subscription buys you is the part that decides what the file should contain, which is also the part that is hardest to evaluate from a feature list. So evaluate on the job. This is the same argument our fuller guide to training with a Hammerhead Karoo makes about the hardware: the device renders a session extremely well and decides nothing, and the ecosystem's real gap sits one layer above it. ## Choosing by the failure you actually have Three failure modes, three answers, and price is the wrong axis to sort them on. If your sessions are fixed, upload files. If you want a calendar without paying, connect Intervals.icu. If your plan keeps going stale because your week moves, no calendar-based tool solves that — paid or free. If your training is a block you already know, hand-uploading is entirely reasonable and permanently free. If a coach writes your plan, TrainingPeaks. If your work is mostly indoor structured power and you want the outdoor version on the bars, TrainerRoad — as long as you have made peace with seeing it on the morning of. If you want continuous fatigue modelling, Xert, understanding that on a Karoo you will be riding inside its app. The third failure mode is the one worth thinking about properly, because it is the one that actually ends most amateur training years. Plans rarely fail because the intervals were wrong. They fail because a work trip erased Wednesday, a chest infection took ten days, or an unplanned four hours on Saturday quietly made Tuesday's session a bad idea. A calendar cannot notice any of that, and neither can the file you uploaded last Sunday. That is the job AdaptCycling was built for. We read your ride history from the source you already use, estimate your threshold from your own power curve rather than asking you to schedule a maximal test — the critical-power relationship makes that possible from efforts you have already ridden [Jones et al. 2010] — and rebuild the plan around the week you actually had. The Karoo's openness is what makes that reachable at all: Hammerhead documents a public API alongside its named partners [Hammerhead Training Metrics], so the slot is genuinely open rather than reserved. ## Common questions **Does the Hammerhead Karoo have its own training plans?** No. Hammerhead ships no training platform, and the Karoo does not generate workouts. It does not even calculate your own numbers — FTP, power zones, heart-rate zones, and weight sync in from a third-party platform you nominate [Hammerhead Training Metrics]. This is the main functional difference from a Garmin Edge, which suggests a daily workout from your previous activities in Garmin Connect [Garmin Edge 1050 Manual]. **Does TrainerRoad work with a Hammerhead Karoo?** Yes, with two conditions. The workout must be set to be completed outside, and TrainerRoad states that outside workouts are made available to Hammerhead on the day they are scheduled — you will not see them in advance [TrainerRoad Hammerhead Sync]. The Ramp Test cannot be pushed outside. Completed Karoo rides sync back to your TrainerRoad calendar automatically. **Can I sync Xert workouts to my Karoo's workout list?** Not the way you can on a Wahoo. Xert describes its ELEMNT and RIVAL integration as pushing adaptive workouts to the device, but describes the Karoo integration as launching the Xert EBC app on the Karoo mid-ride, with rides syncing back afterwards [Xert Integrations]. On a Karoo, Xert is an app you ride inside rather than a session in the Karoo's own workout list. **What is the cheapest way to get a training plan onto a Karoo?** Two routes cost nothing. The Dashboard imports .ZWO and .FIT workout files directly with no subscription anywhere in the chain [Hammerhead Workout Import], and Intervals.icu syncs planned workouts and rides for free. The first suits a fixed block of sessions you already know; the second gives you a calendar as well. **Why is my Karoo showing an FTP I did not set on the device?** Because it is synced. Third-party sync pulls FTP, power zones, heart-rate zones, and rider weight from whichever platform you selected as your sync partner [Hammerhead Training Metrics]. You cannot change the partner from the Karoo — that has to be done on the Dashboard — and you can switch to editing the numbers manually there instead. ## References 1. **Hammerhead Training Metrics.** [Dashboard - Training Metrics](https://support.hammerhead.io/hc/en-us/articles/25219578603291-Dashboard-Training-Metrics). Hammerhead Support. 2. **Hammerhead Account Linking.** [Dashboard - Linking Third-Party Accounts](https://support.hammerhead.io/hc/en-us/articles/360006755614-Dashboard-Linking-Third-Party-Accounts). Hammerhead Support. 3. **Hammerhead Workout Import.** [Dashboard - Importing ZWO & FIT based Workouts](https://support.hammerhead.io/hc/en-us/articles/4416694669851-Dashboard-Importing-ZWO-FIT-based-Workouts). Hammerhead Support. 4. **TrainerRoad Hammerhead Sync.** [Pushing an Outside Workout to Hammerhead](https://support.trainerroad.com/hc/en-us/articles/24512658012955-Pushing-an-Outside-Workout-to-Hammerhead). TrainerRoad Support. 5. **Xert Integrations.** [Partner Integrations](https://www.baronbiosys.com/integrations/). Xert. 6. **Garmin Edge 1050 Manual.** [Edge 1050 Bike Computer Owner's Manual - Daily Suggested Workouts](https://www8.garmin.com/manuals/webhelp/GUID-08ACA9FC-DEE6-4C8D-8A95-F62181C512E9/EN-US/GUID-AFECC098-0B34-4913-9944-66E8DAF64039.html). Garmin. 7. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. --- # Syncing intervals.icu workouts to a Hammerhead Karoo URL: https://www.adaptcycling.com/guides/intervals-icu-karoo-workout-sync Updated: 2026-08-04 Author: Jim Camut This one sat as an open feature request from July 2021 until it closed in January 2025. Intervals.icu now syncs both directions with the Karoo — planned workouts out to the device, completed rides back — and it can also be the source your Karoo takes its FTP and zones from. It is the single best answer for a rider who wants a training calendar on a Karoo without paying for one. The setup has one non-obvious step, and the workout format has one trap. ## What the integration does, and where you turn it on Intervals.icu appears on Hammerhead's list of linkable third-party accounts [Hammerhead Account Linking], and the sync runs both ways: planned workouts reach the Karoo, and rides recorded on the Karoo land back in intervals.icu. The non-obvious part is where the connection gets made. Intervals.icu's developer announced the integration on 24 January 2025, describing it as connecting to intervals.icu to sync your rides and planned workouts, and noting that although there is a Hammerhead box on the intervals.icu settings page, the connection has to be created from your Karoo [Intervals.icu Karoo Sync]. That is the step people miss. If you have been hunting for an intervals.icu button on the Hammerhead Dashboard and coming up empty, you were looking in the wrong place — start on the device, and make sure it has been updated first. There is a second capability here that almost nobody uses, and it is arguably the more useful half. Hammerhead's third-party sync pulls training metrics — FTP, power zones, heart-rate zones, and rider weight — from a platform you nominate, and intervals.icu is on that list alongside Strava and TrainingPeaks [Hammerhead Training Metrics]. Set it there and the numbers your Karoo trains you by follow your intervals.icu profile instead of drifting out of date on the device. The sync partner can only be changed from the Dashboard, not from the Karoo. Once a workout is on the Dashboard, delivery is automatic: the Karoo collects it the next time it connects to the internet [Hammerhead Workout Import]. On the device, a workout scheduled for today carries a Today tag in the workouts list and raises a notification in the Control Center [Hammerhead Karoo Workouts]. Nothing else needs pressing. ## The format trap, and the steps the importer rejects Intervals.icu will happily author a workout the Karoo refuses, and the failure is all-or-nothing rather than partial. Hammerhead publishes exactly what it accepts, and two of the rules catch intervals.icu users specifically [Hammerhead Workout Import]. The first is ramps. A .ZWO step must be a warmup, cooldown, free ride, repeating interval, or steady state — maximum-effort and ramp steps are explicitly unsupported [Hammerhead Workout Import]. Intervals.icu makes ramped efforts easy to write, so a single ramp inside an otherwise ordinary session is a common cause of a workout that simply never appears. The whole file is rejected, not the offending step, which is why the failure looks so much like a sync bug. The second is targets. Every .ZWO step must carry a power or cadence target, and zone-based power targets are not supported at all [Hammerhead Workout Import]. A session written as ride zone 4 for ten minutes rather than as a percentage of FTP will not import. Where a step has both power and cadence, power wins. Distance-based workouts are converted to time-based ones on the way in. The trap proper is heart rate, and it decides which export you should reach for. A .FIT workout's steps can target speed, heart rate, cadence, or power — heart rate and power as absolute values or percentages, speed and cadence as absolute values — and heart rate does not appear on the .ZWO list at all [Hammerhead Workout Import]. The two formats are not interchangeable. If you train to power, either export works. If you train to heart rate, only .FIT carries your targets, and exporting the same session as .ZWO silently strips the thing you were training by. Note too that RPE-based workouts are not supported on the Karoo in any format [Hammerhead Karoo Workouts]. ## Where this route stops For getting a planned session onto a Karoo, this is a genuinely good answer and it costs nothing. What it does not do — what no calendar does — is revise the plan when the week that plan assumed stops happening. Be clear about how good the free option is, because plenty of pages have an incentive to be vague about it. Intervals.icu's analysis is stronger than most paid tools', the sync is official rather than a workaround, and pairing it with the training-metrics sync gives you one consistent set of numbers across the site and the device. If that is what you came for, you are done, and you can ignore the rest of this page. The limit is structural rather than a missing feature. A calendar holds the plan you made on Sunday. It does not know that Wednesday was erased by a work trip, that a chest infection took ten days out of the block, or that Saturday's unplanned four hours changed what Tuesday should be. Those are the events that actually end amateur training years, and they arrive after the plan is written. That gap is the subject of our broader guide to training with a Hammerhead Karoo, and it is where AdaptCycling sits. We read your rides from the source you already use, estimate threshold from your own power curve rather than asking you to schedule a maximal test [Jones et al. 2010], and rebuild the coming fortnight around the week you actually had. Karoo-recorded rides help here more than most, because the device opens a new lap at every workout interval [Hammerhead Karoo Workouts], so the analysis reads what you held rep by rep instead of guessing where the reps were. ## Common questions **How do I connect intervals.icu to my Hammerhead Karoo?** From the Karoo itself, not the Dashboard. Intervals.icu's announcement notes that although a Hammerhead box exists on the intervals.icu settings page, the connection has to be created from your Karoo, and the device needs updating first [Intervals.icu Karoo Sync]. Intervals.icu then appears among the Dashboard's linked third-party accounts [Hammerhead Account Linking]. **When did intervals.icu start syncing with the Karoo?** 24 January 2025, announced by intervals.icu's developer as syncing rides and planned workouts with the Karoo [Intervals.icu Karoo Sync]. It closed a feature request that had been open on both sites since 2021, so any page describing community scripts as the only route is describing the era before that date. **Why did my intervals.icu workout not appear on my Karoo?** Usually one unsupported step, and the file is rejected whole rather than partially. Ramp and maximum-effort steps are not supported, every step needs a power or cadence target, and zone-based power targets are rejected [Hammerhead Workout Import]. A single ramp inside an otherwise ordinary session is the most common cause. **Can I sync heart-rate workouts from intervals.icu to a Karoo?** Only as .FIT. A .FIT workout's steps accept heart rate as an absolute value or a percentage; the .ZWO rules list power and cadence only [Hammerhead Workout Import]. Exporting a heart-rate session as .ZWO drops the targets you were training by. RPE-based workouts are not supported in any format [Hammerhead Karoo Workouts]. **Can intervals.icu set the FTP and zones on my Karoo?** Yes. Hammerhead's third-party sync pulls FTP, power zones, heart-rate zones, and rider weight from a nominated platform, and intervals.icu is one of the options [Hammerhead Training Metrics]. The sync partner has to be selected on the Dashboard — you cannot change it from the Karoo. ## References 1. **Intervals.icu Karoo Sync.** [Hammerhead Karoo now syncs with Intervals.icu](https://forum.intervals.icu/t/hammerhead-karoo-now-syncs-with-intervals-icu/88156). Intervals.icu. 2. **Hammerhead Account Linking.** [Dashboard - Linking Third-Party Accounts](https://support.hammerhead.io/hc/en-us/articles/360006755614-Dashboard-Linking-Third-Party-Accounts). Hammerhead Support. 3. **Hammerhead Workout Import.** [Dashboard - Importing ZWO & FIT based Workouts](https://support.hammerhead.io/hc/en-us/articles/4416694669851-Dashboard-Importing-ZWO-FIT-based-Workouts). Hammerhead Support. 4. **Hammerhead Training Metrics.** [Dashboard - Training Metrics](https://support.hammerhead.io/hc/en-us/articles/25219578603291-Dashboard-Training-Metrics). Hammerhead Support. 5. **Hammerhead Karoo Workouts.** [Karoo OS - Workouts](https://support.hammerhead.io/hc/en-us/articles/18001019006363-Karoo-OS-Workouts). Hammerhead Support. 6. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. --- # Why your workouts are not reaching your Hammerhead Karoo URL: https://www.adaptcycling.com/guides/karoo-workouts-not-syncing Updated: 2026-08-04 Author: Jim Camut One look at the Hammerhead Dashboard tells you which half of the chain broke, and it is worth doing before you touch anything. If the workout is sitting in the Dashboard's workout list, the fault is between the Dashboard and the device. If it is not there, nothing you do to the Karoo will help, and the cause is almost always a documented provider rule rather than a bug. A surprising share of these reports turn out to be a feature behaving exactly as its vendor describes it. ## Start at the Dashboard, because that one look halves the problem Open the Dashboard and check the Workouts list. Anything imported there syncs to the Karoo the next time the device connects to the internet [Hammerhead Workout Import], so the workout's presence or absence in that list splits every possible cause into two piles that share no fixes. Know what arrival looks like on the device before you decide it has not arrived. A workout scheduled for today carries a Today tag in the Karoo's workouts list and raises a notification in the Control Center [Hammerhead Karoo Workouts]. If the Dashboard has the workout and the Karoo shows neither, you have a transport problem and the rest of this section is the fix. If the Dashboard has it and the Karoo does too, but the session looks wrong, skip to the last section — that is not a sync failure at all. The device pulls rather than being pushed to, which is why an unconnected Karoo looks broken. To force it, connect the Karoo Companion App and press Sync on the Karoo — the phone needs an active internet connection for this to do anything [Hammerhead Companion Sync]. The same article documents a genuinely useful side door: share a workout file to the Companion App from your phone's share menu and it becomes a workout on the Karoo, immediately if the device is connected, otherwise the next time it reaches Wi-Fi or you press Sync. If the Dashboard does not have the workout, stop working on the Karoo. Re-pairing the device, reinstalling the app, and factory resets all address the half of the chain that is working. And if you uploaded a file yourself and it never appeared in the Dashboard, that is a rejection rather than a sync failure — a different problem, with its own published rules. ## When it never reached the Dashboard: the provider rules that block it Linking an account is not the same as that account sending workouts. Several of the services the Dashboard connects to are route or activity destinations that never send a session at all, and three of the most common failure reports here are documented behaviour rather than faults. TrainerRoad has two conditions and both catch people. A scheduled workout only reaches Hammerhead if you set it to be completed outside, and TrainerRoad states that outside workouts are made available on the day they are scheduled, and that you will not see them in advance of that day [TrainerRoad Hammerhead Sync]. So Thursday's session being absent on Tuesday is the integration working. The Ramp Test is a separate case — it cannot be pushed outside at all, because it is not designed as an outdoor assessment. Intervals.icu fails for a different reason: people look for the connection in the wrong place. Its developer's announcement notes that although a Hammerhead box appears on the intervals.icu settings page, the connection has to be created from your Karoo, and the device needs updating first [Intervals.icu Karoo Sync]. Riders who hunt for an intervals.icu button on the Dashboard and find nothing usually conclude the integration does not exist. Xert produces the report that is not a fault at all. Xert describes its Wahoo integration as pushing adaptive workouts to the device, but describes the Karoo integration as launching the Xert EBC app on the Karoo mid-ride [Xert Integrations]. Xert sessions are not meant to appear in the Karoo's own workout list, so no amount of unlinking and relinking will put one there. If that is what you were waiting for, you were waiting for something the integration never promised. One rule spans all of them. What syncs is a structured workout, not a line of text on a calendar. A session written as a description of the targets rather than as built intervals has nothing to send, and it fails quietly rather than erroring — which is exactly what makes it hard to spot. ## In the Dashboard but not on the Karoo This half is short, because there are only a few moving parts: an internet connection, a current device, and a pull the Karoo has not made yet. The Companion App's Sync button is the manual override [Hammerhead Companion Sync]. Wi-Fi is the usual answer and the usual blind spot, because the phone being online is not the same as the Karoo being online. The device syncs when it connects to the internet [Hammerhead Workout Import], so a Karoo that has sat in a kit bag since Sunday holds Sunday's list no matter what the Dashboard shows. Wake it, let it join a network, and check again before diagnosing anything further. The Companion App covers the case where Wi-Fi is not available — connect it and press Sync on the Karoo, with the phone online [Hammerhead Companion Sync]. Note that the app's ride-upload feature is limited to the latest generation of Karoo, so an older unit still depends on Wi-Fi for the return leg even when the outbound one works. Firmware matters more here than it does on most devices, because integrations have arrived over time rather than all at once. The intervals.icu connection in particular requires an updated Karoo before it can be created at all [Intervals.icu Karoo Sync]. A device that has not been updated in a year is not missing a fix; it is missing the feature. ## It synced, and the targets are still wrong This gets filed as a sync failure and is a different sync entirely. Your Karoo's FTP, power zones, heart-rate zones, and rider weight come from a third-party platform you nominate [Hammerhead Training Metrics] — so a workout can arrive perfectly and still prescribe the wrong watts. The mismatch is easy to create without noticing. Your plan lives on one platform, your training-metrics sync partner is set to another, and a workout expressed as a percentage of FTP resolves against whichever number the Karoo holds. The sync did its job; the two accounts disagree about your threshold. The partner can only be changed on the Dashboard, not from the Karoo, and Hammerhead's own advice when recent changes are not reflected on the device is to trigger a manual sync from that page [Hammerhead Training Metrics]. The deeper version is a threshold that is simply stale. A number set in March and still prescribing intervals in August is not a configuration error, and nothing in the sync chain will flag it. If every session has felt slightly wrong for a month rather than one session feeling wrong today, suspect the number rather than the pipe. All of this is the argument our broader guide to training with a Hammerhead Karoo makes in another form: the device is a delivery layer, and it renders faithfully whatever the layer above it committed to. AdaptCycling sits in that upper layer. We keep one set of numbers, estimate your threshold from your own power curve rather than a test you have to remember to schedule, and re-send the coming days after the plan rolls over — because a rolling window has to be re-pushed, not set once and trusted. ## Common questions **How do I force my Hammerhead Karoo to sync workouts?** Connect the Karoo Companion App and press Sync on the Karoo, with your phone on an active internet connection [Hammerhead Companion Sync]. Otherwise the device pulls on its own the next time it reaches the internet [Hammerhead Workout Import]. If the workout is not in the Dashboard's workout list, syncing will never produce it — the fault is upstream. **Why is my TrainerRoad workout not showing on my Karoo?** Two documented rules. The workout has to be set to be completed outside, and TrainerRoad states outside workouts are made available to Hammerhead on the day they are scheduled, not in advance [TrainerRoad Hammerhead Sync]. A session later in the week is correctly absent today. The Ramp Test cannot be pushed outside at all. **Why can I not find intervals.icu on my Hammerhead Dashboard?** Because the connection is made from the device, not the Dashboard. Intervals.icu's announcement notes that although a Hammerhead box exists on the intervals.icu settings page, you have to create the connection from your Karoo, and the Karoo needs updating first [Intervals.icu Karoo Sync]. Riders who search the Dashboard for it and find nothing usually conclude the integration does not exist. **Why do my Xert workouts never appear in my Karoo's workout list?** They are not supposed to. Xert describes its ELEMNT and RIVAL integration as pushing workouts to the device, but the Karoo integration as launching the Xert EBC app on the Karoo mid-ride [Xert Integrations]. On a Karoo, Xert is an app you ride inside, so relinking accounts will not put a session in the workout list. **The workout synced but the power targets are wrong. What happened?** Almost certainly your training-metrics sync partner. FTP, power zones, heart-rate zones, and rider weight sync from a platform you nominate [Hammerhead Training Metrics], so a percentage-of-FTP target resolves against that number rather than the one on the platform that wrote your plan. Change the partner on the Dashboard — it cannot be changed from the Karoo. ## References 1. **Hammerhead Workout Import.** [Dashboard - Importing ZWO & FIT based Workouts](https://support.hammerhead.io/hc/en-us/articles/4416694669851-Dashboard-Importing-ZWO-FIT-based-Workouts). Hammerhead Support. 2. **Hammerhead Karoo Workouts.** [Karoo OS - Workouts](https://support.hammerhead.io/hc/en-us/articles/18001019006363-Karoo-OS-Workouts). Hammerhead Support. 3. **Hammerhead Companion Sync.** [Companion App - Uploads and Syncing](https://support.hammerhead.io/hc/en-us/articles/25476943333275-Companion-App-Uploads-and-Syncing). Hammerhead Support. 4. **Hammerhead Training Metrics.** [Dashboard - Training Metrics](https://support.hammerhead.io/hc/en-us/articles/25219578603291-Dashboard-Training-Metrics). Hammerhead Support. 5. **TrainerRoad Hammerhead Sync.** [Pushing an Outside Workout to Hammerhead](https://support.trainerroad.com/hc/en-us/articles/24512658012955-Pushing-an-Outside-Workout-to-Hammerhead). TrainerRoad Support. 6. **Intervals.icu Karoo Sync.** [Hammerhead Karoo now syncs with Intervals.icu](https://forum.intervals.icu/t/hammerhead-karoo-now-syncs-with-intervals-icu/88156). Intervals.icu. 7. **Xert Integrations.** [Partner Integrations](https://www.baronbiosys.com/integrations/). Xert. --- # Why your ZWO file will not import to a Hammerhead Karoo URL: https://www.adaptcycling.com/guides/zwo-file-not-importing-karoo Updated: 2026-08-04 Author: Jim Camut Hammerhead publishes exactly what its importer accepts, and the rules are stricter than the ones your exporter was almost certainly written against. The consequence that catches everyone is that a file is rejected whole rather than partially: one unsupported step in an otherwise ordinary session and nothing imports at all. There are five permitted step types, one target rule, and a second failure mode where the file imports and quietly drops something instead. ## The rejection is all-or-nothing, and five step types is the whole list Every step in a .ZWO file must be a warmup, cooldown, free ride, repeating interval, or steady state. MaxEffort is unsupported, and Ramp is listed as unsupported with coming soon beside it [Hammerhead Workout Import]. Anything else and Hammerhead rejects the file with an error rather than importing what it can. The reason this bites so often is that the .ZWO format was designed for Zwift, where ramps and maximum-effort blocks are entirely ordinary. Most exporters target Zwift's tolerance rather than Hammerhead's, so a workout that opens with a ramped warmup or closes with an all-out sprint is a perfectly valid file that this particular importer will not take. Hammerhead states that unsupported interval types will be rejected and that you will see an error message to that effect [Hammerhead Workout Import]. That makes the diagnosis mechanical rather than mysterious. Open the file in a text editor and read the element names. If you find a MaxEffort or Ramp step anywhere in it, you have your answer, and it does not matter that the other fourteen steps are fine. The usual repair is to replace a ramp with a short staircase of steady-state blocks — three or four steps stepping up to the target approximate it closely enough for a warmup, and every one of them is a permitted type. Treat the Ramp status as a moving target rather than a fixed rule. Hammerhead's own documentation marks it coming soon, which means a file that fails today may import later without anything on your side changing. Worth re-checking before rebuilding a library of workouts around the limitation. Durations, meanwhile, are always time in seconds in a .ZWO file, and distance-based workouts are converted to time-based ones on import. ## The target rule, and one contradiction nobody can resolve Every step must include at least one target: Power, a PowerHigh and PowerLow pair, Cadence, or a CadenceHigh and CadenceLow pair. Power takes preference when both are supplied, and zone-based power targets are not supported at all [Hammerhead Workout Import]. The zone rule is the sneaky one, because a zone-based workout is not malformed — it is just expressed in the wrong currency. A session written as ride zone 4 for ten minutes will not import; the same session written as a percentage of FTP will. If you author in a tool that thinks in zones, the export setting that switches to percentages is the fix, and it is usually one checkbox away from where you gave up. Those percentages then resolve against whichever FTP your Karoo holds, which is synced from a platform you nominate rather than set on the device [Hammerhead Training Metrics]. There is one contradiction in the published rules that we have not been able to resolve, and it is worth stating rather than papering over. Free ride is listed as a supported step type, and a free ride is self-paced by definition, yet the same document says every step must include at least one target. Both cannot hold literally. In our own exporter we treat the explicit support listing as the more specific rule and let free-ride steps through without a target — but that is an inference from two lines of documentation, not something we have tested against the live importer, and if your file consists mainly of free rides it is the first thing we would suspect. The distinction that saves the most time is between rejected and ignored. Unsupported step types are rejected and take the file with them. Unsupported step fields are not: Hammerhead supports fields relating to cadence, power, duration, and repeat, and states that all others are ignored while the workout remains usable [Hammerhead Workout Import]. So a file that never appears and a file that appears with something missing are different diagnoses. The first is a step type. The second is a field you can stop hunting for. ## When it is not a ZWO problem at all Two more causes sit outside the .ZWO rules. The file may not be a workout, and the format may be incapable of carrying the targets you are trying to send. Only files with a .ZWO or .FIT extension are accepted in the first place [Hammerhead Workout Import]. The .FIT rules trip people who switch formats to escape the .ZWO ones. An activity or a route exported as .FIT is not a workout and is rejected, which is an easy mistake because all three share an extension and an analysis platform will happily offer you all three. The workout's sport type must be one of generic, cycling, e-bicycle, all, or invalid, and zone targets are rejected in .FIT just as they are in .ZWO [Hammerhead Workout Import]. There is also a quiet one here: an open-duration step with no target is silently ignored rather than rejected, so a file can import looking shorter than you built it. The format decision that matters most is heart rate. A .FIT workout's steps can target speed, heart rate, cadence, or power — heart rate and power as absolute values or percentages, speed and cadence as absolute values — and heart rate does not appear among the .ZWO targets at all [Hammerhead Workout Import]. If you have been fighting a .ZWO export to carry heart-rate intervals, you have been fighting the wrong format, and no amount of editing will fix it. Note too that the Karoo does not support RPE-based workouts in any format [Hammerhead Karoo Workouts]. The rules follow the file rather than the route it took. Besides the Dashboard, you can share a workout file to the Karoo Companion App from your phone's share menu and have it turned into a workout on the device [Hammerhead Companion Sync] — convenient, but not a way around any of the constraints above. A step type the importer rejects is rejected whichever door the file came through. We ended up encoding these rules in software rather than trusting an exporter to respect them. Every workout AdaptCycling pushes to a Karoo is checked against the permitted step types and the target rule before it leaves us, and a workout that fails is rejected with the offending step named — because Hammerhead rejects the file whole and gives no usable reason back, so a single bad step otherwise costs an athlete that entire day on the device with no explanation anywhere. That is the same principle our broader guide to training with a Hammerhead Karoo applies to the hardware: the device is faithful, and everything upstream of it has to earn that. ## Common questions **Why was my ZWO file rejected by the Hammerhead Dashboard?** Almost always one unsupported step, and the file is rejected whole rather than partially. Steps must be warmup, cooldown, free ride, repeating interval, or steady state; MaxEffort is unsupported and Ramp is listed as unsupported with coming soon beside it [Hammerhead Workout Import]. Open the file in a text editor and look for those two element names first. **Which step types does the Karoo accept in a ZWO file?** Warmup, cooldown, free ride, repeating intervals, and steady state [Hammerhead Workout Import]. Durations are always time in seconds, and distance-based workouts are converted to time-based ones on import. Replacing a ramp with three or four stepped steady-state blocks is the usual way to keep the session's shape. **Why will my zone-based workout not import?** Zone-based power targets are not supported [Hammerhead Workout Import]. Express the same targets as percentages instead, which most authoring tools offer as an export option. Those percentages resolve against the FTP held on your Karoo, which syncs from a platform you nominate rather than being set on the device [Hammerhead Training Metrics]. **My file imported but the workout looks wrong. Is that the same problem?** No, and the difference is worth knowing. Unsupported step types are rejected and take the file with them; unsupported step fields are ignored while the workout stays usable, since only fields relating to cadence, power, duration, and repeat are supported [Hammerhead Workout Import]. An open-duration step with no target in a .FIT file is also silently ignored rather than rejected. **Can a ZWO file carry heart-rate targets to a Karoo?** No. Heart rate is a .FIT target only — .FIT steps accept speed, heart rate, cadence, or power, while the .ZWO rules list power and cadence [Hammerhead Workout Import]. A heart-rate session has to be exported as .FIT. RPE-based workouts are not supported in any format [Hammerhead Karoo Workouts]. ## References 1. **Hammerhead Workout Import.** [Dashboard - Importing ZWO & FIT based Workouts](https://support.hammerhead.io/hc/en-us/articles/4416694669851-Dashboard-Importing-ZWO-FIT-based-Workouts). Hammerhead Support. 2. **Hammerhead Karoo Workouts.** [Karoo OS - Workouts](https://support.hammerhead.io/hc/en-us/articles/18001019006363-Karoo-OS-Workouts). Hammerhead Support. 3. **Hammerhead Training Metrics.** [Dashboard - Training Metrics](https://support.hammerhead.io/hc/en-us/articles/25219578603291-Dashboard-Training-Metrics). Hammerhead Support. 4. **Hammerhead Companion Sync.** [Companion App - Uploads and Syncing](https://support.hammerhead.io/hc/en-us/articles/25476943333275-Companion-App-Uploads-and-Syncing). Hammerhead Support. --- # Structured workouts on a Karoo without a power meter URL: https://www.adaptcycling.com/guides/karoo-workouts-without-power-meter Updated: 2026-08-04 Author: Jim Camut You can follow a structured workout on a Karoo without a power meter, because a heart-rate strap satisfies the device's requirement that something report your effort against the target. But one detail decides whether it works, and it lives in the file rather than in the hardware. Heart-rate targets exist in the FIT workout format and not in ZWO. Export the same session in the wrong one and what arrives on your bars is a list of watt numbers you have no way to follow. ## The strap is enough; the file format is the real question Hammerhead states that following a workout needs a smart trainer or sensors that can report the rider's effort against the selected workout [Hammerhead Karoo Workouts]. A heart-rate strap is one of those sensors. The same page closes the other door: RPE-based workouts are not supported at all. The format split is the whole story. A FIT workout's steps can target speed, heart rate, cadence, or power, with heart rate accepted as an absolute value or a percentage. The ZWO target list is power and cadence only [Hammerhead Workout Import]. The two are not interchangeable, and nothing about the device or your strap changes that — a heart-rate session simply has to arrive as a FIT workout. That reframes the question you should be asking a platform. Not whether it works with a Karoo, but whether it exports a FIT workout carrying heart-rate targets. Plenty of tools that sync happily to the device produce ZWO by default, which is the right answer for a rider with a power meter and the wrong one for you. Worth checking before you subscribe rather than after. One more constraint applies to both formats: zone-based targets are rejected, for heart rate as well as power [Hammerhead Workout Import]. A session written as ride zone 3 for twenty minutes will not import. It has to be expressed as beats per minute or as a percentage. And because the Karoo offers no RPE fallback, a rider with neither power meter nor strap has no route to a structured session on this device — some head units allow one, and this is a genuine narrowness rather than an oversight worth defending. ## What heart rate can target, and what it cannot Heart rate is a good target for long efforts and a poor one for short. Because interval training built from brief bursts produces a lag in heart rate and oxygen uptake, the rise during each interval is not representative of the absolute exercise intensity [Astorino et al. 2025]. A forty-second rep cannot be steered by a number that has not arrived yet. The practical line falls somewhere around a few minutes. Efforts long enough for heart rate to settle — tempo blocks, threshold intervals, long endurance riding — are exactly where heart-rate targets earn their keep, and a plan built mostly from those is not a compromised plan. Short, sharp repetitions are a different matter: by the time the number reflects the effort, the rep is over, and chasing it during the rep makes you ride the first half too hard. Two corrections belong on any heart-rate target you follow. The first is cardiovascular drift: heart rate rises over prolonged exercise even when work rate is held constant [Coyle & González-Alonso 2001], so a target satisfied at hour three represents less mechanical work than the same target at minute twenty. Hitting the number late in a long ride is not the same achievement as hitting it early, and reading it that way inflates what you did. The second is that the relationship itself moves. Dehydration and ambient temperature have a profound effect on the link between heart rate and oxygen uptake, and there is measurable day-to-day variability in heart rate [Achten & Jeukendrup 2003]. On a hot day the same number means less work; on a cold morning after a good night it may mean more. Treat the target as a band and the session as the unit you judge, rather than grading yourself against a single figure. ## Building a session the Karoo will actually accept Two supported targets do the work between them. Heart rate carries the long blocks, delivered as FIT. Cadence carries the short ones — it is a first-class target in its own right in both formats [Hammerhead Workout Import], which gives a rider without power something legal to put on reps too short for heart rate. The cadence route solves a real problem. Every step in the file has to carry a target, and Cadence or a CadenceHigh and CadenceLow pair satisfies that rule with no power anywhere in the workout [Hammerhead Workout Import]. For the short hard efforts you were always going to ride by feel, a cadence target makes the step valid, gives the device something to display, and quietly enforces the gear selection that usually decides whether those reps are any good. The device itself is unaffected by any of this, which is worth saying because it is the encouraging half. Before you start, the Karoo shows which zones the workout targets and how long you will spend in each [Hammerhead Karoo Workouts]. During it you get the workout drawer, controls for skipping, scaling and pausing, and a new lap opened automatically at every interval boundary. Your ride file comes back with its structure already marked, whether the targets in it were watts or beats. Check the numbers underneath before blaming a session that felt wrong. Your Karoo's heart-rate zones sync in from a platform you nominate, alongside FTP and rider weight [Hammerhead Training Metrics], so the zones the device shows you are only as current as that profile. A maximum heart rate estimated from a formula years ago, or a threshold set in a different season, will produce targets that feel wrong in a way no amount of re-syncing fixes. ## Where this leaves you, and where it leaves us A limitation in our own product belongs here, because the honest version of this page has to include it. AdaptCycling's Karoo push emits ZWO, and ZWO cannot carry heart-rate targets. Until we ship a FIT workout encoder, a rider training on heart rate is better served reaching the device another way. The route that works today costs nothing. Export or author the session as a FIT workout with heart-rate targets and upload it to the Hammerhead Dashboard, or connect a provider that produces one — the Dashboard accepts FIT workout files directly, with no subscription in the chain [Hammerhead Workout Import]. That is the same free route that serves power-meter riders, and the file rules are the only thing standing between you and it. What we do handle is everything upstream of the device. We read heart-rate rides, build and revise plans for riders training without power, and judge sessions on what your heart rate actually did rather than assuming a power file exists. The gap is specifically the last hop onto a Karoo, and it is a format gap rather than a modelling one — which is why it is worth naming plainly instead of leaving you to discover it after signing up. The pattern is the one running through our broader guide to training with a Hammerhead Karoo. The device is a faithful delivery layer that decides nothing and renders whatever it is handed. Here the constraint is one level lower still, in the file format itself, and it is the sort of thing that is invisible until it costs you a session. Confirm the export before you build a training block around it. ## Common questions **Can I do structured workouts on a Karoo without a power meter?** Yes. Hammerhead requires a smart trainer or sensors that can report your effort against the selected workout, and a heart-rate strap qualifies [Hammerhead Karoo Workouts]. The catch is the file: heart-rate targets exist only in the FIT workout format, not in ZWO [Hammerhead Workout Import]. RPE-based workouts are not supported in any format, so a strap is effectively required. **Why did my heart-rate targets vanish when I exported the workout as ZWO?** Because ZWO has no heart-rate target. The ZWO rules list power and cadence; the FIT rules add speed and heart rate, with heart rate accepted as an absolute value or a percentage [Hammerhead Workout Import]. Exporting a heart-rate session as ZWO strips the thing you were training by and leaves targets you cannot follow. Re-export as a FIT workout. **Can I use heart-rate zones as workout targets on a Karoo?** No. Zone-based targets are rejected for heart rate and power alike [Hammerhead Workout Import]. Targets have to be absolute values or percentages. Your Karoo's heart-rate zones themselves sync in from whichever platform you nominated [Hammerhead Training Metrics], so it is worth confirming that profile is current before blaming the file. **Are short intervals worth doing on heart rate?** Not as heart-rate targets. Interval training built from brief bursts produces a lag in heart rate and oxygen uptake, so the rise during each rep does not represent the actual exercise intensity [Astorino et al. 2025]. Give short reps a cadence target instead — cadence is a supported target in its own right [Hammerhead Workout Import] — and save heart-rate targets for efforts long enough to settle. **Does AdaptCycling push heart-rate workouts to a Karoo?** Not yet. Our Karoo push emits ZWO, which carries power and cadence targets only, so a heart-rate rider gets targets they cannot use. Until we add a FIT workout encoder, upload a FIT workout to the Hammerhead Dashboard instead — it is free and accepts the format directly [Hammerhead Workout Import]. We do build and revise plans for riders training on heart rate; the gap is the last hop onto the device. ## References 1. **Hammerhead Karoo Workouts.** [Karoo OS - Workouts](https://support.hammerhead.io/hc/en-us/articles/18001019006363-Karoo-OS-Workouts). Hammerhead Support. 2. **Hammerhead Workout Import.** [Dashboard - Importing ZWO & FIT based Workouts](https://support.hammerhead.io/hc/en-us/articles/4416694669851-Dashboard-Importing-ZWO-FIT-based-Workouts). Hammerhead Support. 3. **Hammerhead Training Metrics.** [Dashboard - Training Metrics](https://support.hammerhead.io/hc/en-us/articles/25219578603291-Dashboard-Training-Metrics). Hammerhead Support. 4. **Astorino et al. 2025.** [Heart Rate and Oxygen Uptake During Recovery from High-Intensity Interval Training: A Retrospective Analysis](https://pmc.ncbi.nlm.nih.gov/articles/PMC12294763/). International Journal of Environmental Research and Public Health. 5. **Achten & Jeukendrup 2003.** [Heart rate monitoring: applications and limitations](https://pubmed.ncbi.nlm.nih.gov/12762827/). Sports Medicine. 6. **Coyle & González-Alonso 2001.** [Cardiovascular drift during prolonged exercise: new perspectives](https://pubmed.ncbi.nlm.nih.gov/11337829/). Exercise and Sport Sciences Reviews. --- # Does Polar have a cycling training plan? URL: https://www.adaptcycling.com/guides/polar-cycling-training-plan Updated: 2026-08-27 Author: Jim Camut No, and Polar is unusually direct about it. The company publishes a page enumerating every training plan it offers, there are five of them, and not one is built for cycling [Polar Training Plans]. Asked in its own FAQ why the programs cover running only, Polar answers that in the future training programs for other sports like cycling are possible [Polar Running Program FAQ]. Here is what each of the five actually does on a bike, and what you can assemble instead. ## The five options, and what each one is really for Polar's own menu lists Flow manual planning, the Running Program, TrainingPeaks integration, the Fitness Program and FitSpark [Polar Training Plans]. Two of those are plans in the ordinary sense. Neither is for cyclists. The other three are a workout builder, someone else's calendar, and a daily suggestion engine. The Running Program is the one that proves Polar can do this. It is free, it runs on Flow web, it takes a race distance and a target date, and it returns a personalized plan of two to five sessions a week spread across nine to twenty weeks [Polar Training Plans]. That is a periodized training plan with a goal and an arc. It exists for 5K, 10K, half and marathon runners, and for nobody else. Flow manual planning is a workout builder, not a plan. You can construct any session you like as a phased training target and the watch will guide you through it [Polar Training Targets], which is genuinely useful and completely unopinionated. It answers what does this interval look like. It has no view on which intervals you should be doing in week six. The TrainingPeaks integration is real and it is worth knowing about, because it is the only route by which a structured cycling workout reaches a Polar watch from outside. Once linked, every compatible TrainingPeaks training plan and all target workouts sync to Flow automatically once a day [Polar TrainingPeaks Sync]. What arrives is whatever plan you or your coach put into TrainingPeaks. Polar is the delivery layer; the training decisions were made elsewhere. FitSpark is the fifth, and it is the one most often mistaken for a plan. It offers two to four ready-made workouts a day, chosen against your recovery, fitness level and training history, and its cardio sessions can be completed with any sport profile on your watch [Polar FitSpark]. So it does work on the bike. It also has no goal date and no memory of what it suggested three weeks ago, which is the difference between a suggestion and a plan. ## The Fitness Program does adapt, and it is not cycling training This is the claim people get wrong in both directions. Polar does ship an adaptive training program that accepts cycling: the Fitness Program, 9,99 euro a month, running on four-week periods, promoting or demoting you a level after each one [Polar Fitness Program]. It is also general-fitness cardio with no power targets. The adaptation is documented and genuine. After a finished period you are suggested a new level if the program detects you have progressed enough, or if the level has been too demanding for you [Polar Fitness Program]. You can choose any sport except Multisport, so cycling is in scope. Anyone writing that Polar has no adaptive program has not read this page. What the page also says is the decisive part: the Polar Fitness Program is based on training in the right heart rate zone [Polar Fitness Program]. That is the whole prescription model. There are no power targets, no cycling-specific interval structures, and no build toward a dated goal. A four-week cardio ladder that repeats is a fitness product, and Polar markets it as one, aimed at busy people looking for lasting results [Polar Training Plans]. Set that against what a cyclist preparing for something actually needs: a base phase, a build phase with intensity that escalates in a controlled way, a taper, and a plan that changes shape when a week is lost. The Fitness Program is not attempting any of that. So the precise, defensible statement is that Polar has no cycling-specific adaptive training plan — not that Polar cannot adapt. ## What a Polar cyclist can assemble instead Three routes, in ascending order of effort and descending order of independence: build the sessions yourself in Flow, import someone else's calendar through TrainingPeaks, or use a service that reads your Polar rides and writes the plan. Each solves a different part of the problem. Building it yourself is free and the tooling is better than its reputation. Phased training targets accept intensity by heart rate, speed or power [Polar Training Targets], phases can advance automatically or on a button press, and the watch guides you through the session. The cost is that you are now the coach: you decide the week's structure, the progression, and what to do after a missed session. That is the work, and no amount of workout-builder polish reduces it. TrainingPeaks solves the delivery problem completely and leaves the decision problem open. It pushes structured workouts to the watch, it works on a basic free account as well as a premium one [Polar TrainingPeaks Sync], and if you have a coach it is almost certainly where they already work. What it does not do is notice that you rode four hours on Saturday instead of Tuesday's threshold session and rebuild the fortnight accordingly. A calendar is a calendar. The third route is a service that reads the rides your Polar records and generates the plan from them. This is the layer the whole question of training with a Polar keeps arriving at, and it is where AdaptCycling sits: we read your rides, estimate threshold from your power curve rather than sending you to test, build a plan from what you have actually been doing, and rewrite it when the week collapses. We cannot push the session to your watch, because Polar's public API has no endpoint that writes one — so the honest arrangement is that the plan lives in the app and you rebuild today's session as a phased target if you want it on the wrist. ## Common questions **Is there a free Polar cycling training plan?** Not as a plan. Flow's phased training targets let you build any cycling session for free and the watch will guide you through it [Polar Training Targets], but you write the schedule. The only free plan Polar generates is the Running Program, which covers 5K to marathon and no other sport [Polar Training Plans]. **Will Polar ever add a cycling program?** Polar's Running Program FAQ says it is possible, in those words: we have had the most requests for a running program, so we started with that, and in the future training programs for other sports like cycling are possible [Polar Running Program FAQ]. That is a roadmap statement with no date attached, and it has stood for several product generations. **Does the Polar Fitness Program work for cycling?** It runs on a bike, and it is not cycling training. You can choose any sport except Multisport, it costs 9,99 euro a month, and it adapts across four-week periods [Polar Fitness Program]. But it is based on training in the right heart rate zone, with no power targets and no build toward a dated goal. Useful for general fitness; not a plan for an event. **Can I put a cycling plan from another app onto my Polar watch?** Through TrainingPeaks, yes. Every compatible training plan and target workout syncs to Flow automatically once a day after you link the accounts, and it works with a basic TrainingPeaks account as well as premium [Polar TrainingPeaks Sync]. No other service can, because Polar's public API is read-only for training and exposes no endpoint that creates a workout. ## References 1. **Polar Training Plans.** [Which Polar Training Plan Is Right for You?](https://www.polar.com/en/guide/polar-training-plans). Polar. 2. **Polar Running Program FAQ.** [Polar Running Program FAQ](https://support.polar.com/en/support/polar_running_program_faqs). Polar Support. 3. **Polar Fitness Program.** [Polar Fitness Program FAQ](https://support.polar.com/en/fitness-program-faqs). Polar Support. 4. **Polar FitSpark.** [FitSpark daily training guide](https://support.polar.com/us-en/fitspark-daily-training-guide). Polar Support. 5. **Polar Training Targets.** [How do I create a training target in Polar Flow web service?](https://support.polar.com/en/support/how_do_i_create_a_training_target_in_polar_flow). Polar Support. 6. **Polar TrainingPeaks Sync.** [Using Polar Flow with TrainingPeaks](https://support.polar.com/en/support/how_can_i_automatically_sync_my_data_from_polar_flow_to_trainingpeaks). Polar Support. --- # TrainingPeaks workout not syncing to Polar Flow: the documented causes URL: https://www.adaptcycling.com/guides/polar-flow-structured-workout-not-syncing Updated: 2026-08-27 Author: Jim Camut Four different failures wear the same symptom, and Polar publishes all four. The workout can be waiting on a once-a-day sync clock. It can exceed the twenty-phase cap and be refused outright. It can land in a sport profile your watch does not have in favourites. Or it can arrive correctly and look wrong, because Flow remapped your zones. Work through them in that order; the first one costs nothing to check and explains a large share of cases. ## Start with the clock, not the workout TrainingPeaks plans sync to Flow automatically once a day [Polar TrainingPeaks Sync]. That is the single most common reason a workout you scheduled twenty minutes ago is not in your diary. There is a manual Refresh button in the Flow calendar view that triggers the sync immediately, and it is easy to miss. Once you link the accounts on Flow's Partners page, every compatible TrainingPeaks training plan and all target workouts you make will sync to Polar Flow automatically once a day [Polar TrainingPeaks Sync]. Once a day is the contract. A coach who adds tomorrow's session at 9pm and a rider who checks the watch at 6am the next morning may simply be inside the same sync window, with nothing wrong at either end. The second clock-shaped cause is a stale link. When Polar extended the integration to carry structured workouts and training plans in the TrainingPeaks-to-Flow direction, existing users had to reconnect their accounts for the new direction to work [Polar TrainingPeaks Update]. Before that change the sync only ran the other way, from Flow to TrainingPeaks. An account linked before the change and never re-linked will keep uploading your completed rides perfectly while never delivering a single planned workout, which is a confusing failure because half of it works. So the first two things to do are press Refresh in the Flow calendar, and unlink then relink TrainingPeaks on the Partners page. If the workout appears, nothing was broken. If it does not, the cause is one of the content rules below. ## The rules that reject a workout outright One rule accounts for most hard rejections and it is a number: in Flow the number of unique phases allowed in a training target is limited to 20, so a TrainingPeaks target containing more than 20 phases cannot be synced to Flow at all [Polar TrainingPeaks Sync]. Twenty sounds generous until you count the way a workout builder counts. Every warm-up block, every work interval, every recovery interval and every cool-down is a phase. A session written as ten by three minutes with three minutes recovery, plus a warm-up and a cool-down, is twenty-two phases before you add anything else. Sessions with short reps hit the cap fastest, which is why the workouts that vanish tend to be the VO2max and anaerobic ones rather than the endurance rides. The word unique is doing work in Polar's phrasing, so a repeating block expressed as a repeat rather than as expanded steps will generally survive where the expanded version will not. If a session you need is over the limit, the practical fix is to rebuild it in TrainingPeaks using a repeat structure, or to collapse the warm-up from four graded steps into one. The other outright rejection is narrower and worth knowing if you cross-train: Polar supports only the Old Strength Builder workout type, not the New Strength Builder [Polar TrainingPeaks Sync]. That does not touch cycling sessions, but a mixed plan can appear to sync partially, with the bike work arriving and the gym work silently absent. ## The rules that change a workout on the way through Two documented transformations catch riders who assume a synced workout is an identical copy. Flow always uses five training zones while TrainingPeaks does not restrict the number, so zones are remapped to be compatible with Flow on sync [Polar TrainingPeaks Sync]. And TrainingPeaks does not separate indoor cycling from outdoor cycling the way Flow does. Take the zone remap first, because it produces the report that the workout synced but the targets are wrong. If your TrainingPeaks plan is written against a seven-zone or eight-zone model, Flow has to compress it into five on arrival. The session structure survives; the intensity boundaries are Flow's, not TrainingPeaks'. A target written as the top of zone 3 in an eight-zone model does not land at the top of zone 3 in a five-zone one, and no error is raised because nothing failed. The sport-profile mismatch is the subtler one and Polar gives the fix directly. Because TrainingPeaks has fewer workout types than Flow and does not separate indoor from outdoor cycling, Polar recommends that if you only do indoor cycling in Flow you should also keep the regular cycling sport profile in your favourites, with custom zones turned on [Polar TrainingPeaks Sync]. Without that profile on the watch, the target has nowhere to land and the structured guidance does not appear even though the session is in your diary. Two smaller transformations round it out. Flow imposes a 45-character limit on the title and a 500-character limit on training notes, so long workout names and detailed coach instructions are truncated. And once a workout is in your Flow diary you cannot edit it in Flow, only in TrainingPeaks [Polar TrainingPeaks Sync] — which means an attempted fix made on the Polar side has no effect and will be overwritten on the next daily sync. ## When the sync is fine and the plan is still wrong Worth separating from the troubleshooting above, because it is the more common complaint underneath it. A workout that arrives exactly as written can still be the wrong workout, and no amount of sync diagnosis will surface that. This is the gap that training with a Polar keeps running into. The TrainingPeaks-to-Flow pipe is a delivery mechanism. It moves a decision that something upstream already made. If that decision was made ten days ago against a week you then did not have — you missed Tuesday, you rode four hours unplanned on Saturday, you caught a cold on Thursday — the sync will deliver Thursday's threshold session on Thursday with complete fidelity, and it will be wrong. Polar's side of this is documented and finite. Its public AccessLink API exposes read endpoints for exercises and activity and nothing that creates a training target [Polar AccessLink API], which is exactly why TrainingPeaks' partner integration is the only door and why no other service can push a session to your watch. Polar built the delivery layer and the measurement layer and left the decision layer to whoever you connect. That decision layer is where AdaptCycling works. We read the rides your Polar records, build the plan from what you have actually been doing, and rewrite it when the week goes sideways instead of re-delivering a stale calendar. We cannot push the result to your watch, for the API reason above, so the honest arrangement is that the plan lives in the app and you rebuild today's session as a phased training target in Flow if you want it on the wrist [Polar Training Targets]. ## Common questions **How long should a TrainingPeaks workout take to reach Polar Flow?** Up to a day. Polar states that plans and target workouts sync automatically once a day after the accounts are linked [Polar TrainingPeaks Sync]. If you need it sooner, use the Refresh button in the Flow calendar view, which triggers the sync immediately. **Why did only some of my plan sync?** Most often the twenty-phase cap. Flow limits a training target to 20 unique phases and refuses any TrainingPeaks target that exceeds it [Polar TrainingPeaks Sync], so the short-interval sessions in a plan disappear while the endurance rides arrive. Strength sessions built with the New Strength Builder are also unsupported, so a mixed plan can lose those too. **The workout synced but the power targets look wrong. Why?** Because Flow always uses five training zones and TrainingPeaks does not restrict the count, so the zones are remapped to be compatible with Flow when the workout syncs [Polar TrainingPeaks Sync]. The structure survives the trip; the zone boundaries become Flow's. Nothing failed, which is why there is no error. **My indoor cycling workout has no guidance on the watch. What is missing?** Probably the outdoor cycling sport profile. TrainingPeaks does not separate indoor from outdoor cycling the way Flow does, so Polar recommends keeping the regular cycling sport profile in your watch favourites with custom zones turned on, which is what lets the indoor target carry structured guidance [Polar TrainingPeaks Sync]. **Can I fix a synced workout inside Polar Flow?** No. Once a workout is in your Flow diary you can only edit it in TrainingPeaks [Polar TrainingPeaks Sync]. Anything you change on the Polar side is not saved and would be replaced at the next daily sync anyway. Edit at the source. ## References 1. **Polar TrainingPeaks Sync.** [Using Polar Flow with TrainingPeaks](https://support.polar.com/en/support/how_can_i_automatically_sync_my_data_from_polar_flow_to_trainingpeaks). Polar Support. 2. **Polar TrainingPeaks Update.** [Polar Flow update - Further integration with TrainingPeaks](https://support.polar.com/en/updates/trainingpeaks-integration). Polar Support. 3. **Polar Training Targets.** [How do I create a training target in Polar Flow web service?](https://support.polar.com/en/support/how_do_i_create_a_training_target_in_polar_flow). Polar Support. 4. **Polar AccessLink API.** [Polar AccessLink API](https://www.polar.com/accesslink-api/). Polar Developers. --- # Polar FTP without a power meter: what actually still works URL: https://www.adaptcycling.com/guides/polar-ftp-without-power-meter Updated: 2026-08-27 Author: Jim Camut Polar's Cycling Performance Test is a real FTP protocol and it has one prerequisite you cannot work around: a compatible cycling power sensor [Polar Cycling Performance Test]. No sensor, no test, and no power zones on the watch either. That is an awkward wall for a brand whose centre of gravity is the heart-rate strap. Here is exactly where the wall sits, which watches are behind it, and what a heart-rate rider can anchor training to instead. ## The test is good, and its prerequisite is absolute Polar states it in one line: to take the Cycling Performance Test you need a compatible cycling power sensor [Polar Cycling Performance Test]. The test itself is well designed — 20, 30, 40 or 60-minute options, with the shorter ones estimating the 60-minute result — and it returns FTP in watts, maximum heart rate and VO2max in a single session. It runs on eight watches: Grit X Pro, Pacer Pro, Vantage V2, Vantage V3, Grit X2 Pro, Vantage M3, Grit X2 and Street X [Polar Cycling Performance Test]. Results sync to the Flow app and update your individual power zones, which is the part that makes the test worth doing rather than merely interesting. If you own one of those eight and a power meter, take it. Twenty minutes is enough for a usable number. Two groups are excluded and they are excluded differently. The first is anyone on a watch outside that list of eight — the Ignite line and the base Pacer among them — where the test simply is not present. The second is the far larger group who own a listed watch and no power meter. For them the test is visible and unusable, which is the more frustrating version. There is also a quiet asymmetry between running the test and using the result. Power-based phased training targets are supported only on the Grit X Pro, Grit X2 Pro, Street X, Vantage V2 and Vantage V3 [Polar Training Targets]. That is five models, and the Vantage M3 and Grit X2 are on the test list but not this one. Both can measure your FTP and neither can then guide you through an interval expressed in watts. On those two, build the same session with heart-rate phases instead. ## Even with a power meter, Polar is Bluetooth-only Before buying a sensor to unlock the test, check the protocol. Polar's compatibility page annotates the PowerTap G3 entry directly: Bluetooth only and Dual models are supported, ANT+ only model is not supported [Polar Power Sensors]. An ANT+-only power meter does not reach a Polar watch at all. The supported list is short and specific: PowerTap C1, PowerTap G3, PowerTap P1 and P1S, single-sided Stages, left-side 4iiii PRECISION, Rotor 2INpower, Quarq DZero, Power2Max NG, Favero Assioma, Wattbike Atom and the Wahoo KICKR [Polar Power Sensors]. That list covers Grit X, Grit X Pro, Grit X2, Grit X2 Pro, Pacer, Pacer Pro, Street X, Vantage M, Vantage M2, Vantage M3, Vantage V2 and Vantage V3. The Ignite family is absent from it entirely. This explains something that otherwise looks like user error. A rider buys a Polar watch, already owns a power meter, pairs nothing, and concludes the watch is broken. It is not: their meter is ANT+-only, which is common on older cranks and on plenty of trainers, and Polar's radio never sees it. The fix is a dual-protocol sensor, not a firmware update. It is also why a meaningful share of Polar riders who do own a power meter still record heart-rate-only rides. The population that cannot run the Cycling Performance Test is therefore larger than the population without a power meter, and that is worth knowing before you assume the HR-only path is a minority case in this ecosystem. It is closer to the main road. ## What to anchor training to when there is no watts number Heart rate, properly configured, and a threshold estimate that arrives from riding rather than testing. Polar's own load model already runs on heart rate: Cardio Load is TRIMP, calculated after every workout from your heart rate and session duration [Polar Training Load Pro]. The infrastructure for an HR-anchored training year is already on the watch. The first job is to stop using an age formula for maximum heart rate. Tanaka's meta-analysis put the population relationship at 208 minus 0.7 times age and was explicit that individual variation around that line is wide [Tanaka et al. 2001], so an estimate can sit well away from your real number. Every heart-rate zone on the watch is a percentage of it, which means one wrong anchor mis-scales the entire prescription and every Cardio Load score computed from it. The highest heart rate you have genuinely produced in a hard effort in the last few months is a better anchor than any equation. The second job is to read heart-rate targets with drift in mind. On a long session heart rate climbs at constant work rate, so an interval that asks for a number late in a ride is asking for less power than the same number asked for early. That is physiology rather than fading fitness, and it is the single largest reason HR targets feel inconsistent across a three-hour ride. Short reps are where heart rate is least useful, because the response lags the effort; that is exactly where a phased target built on speed, or simply on time and perceived effort, does a better job. The third route only opens once some power data exists, and it is worth naming because it changes the calculus on buying a meter. A threshold estimate fitted to your mean maximal power across durations targets the same physiological boundary the Cycling Performance Test is aiming at [Jones et al. 2010], and it updates continuously from ordinary riding rather than from a scheduled maximal effort. A rider who buys a Bluetooth meter to unlock Polar's test may find they never need to take it. Whichever anchor you end up with, hold its precision loosely. Measured against maximal lactate steady state, FTP tracks the group mean closely while its 95% limits of agreement run to roughly 9% [Borszcz et al. 2019]. A tested number and a modelled one that differ by fifteen watts are not in disagreement, and a heart-rate-anchored plan is not a degraded version of a power-anchored one so much as a plan with different error bars. Training with a Polar and no power meter is a legitimate configuration, not a workaround. ## Common questions **Can I run Polar's Cycling Performance Test without a power meter?** No. Polar's requirement is unconditional: to take the Cycling Performance Test you need a compatible cycling power sensor [Polar Cycling Performance Test]. There is no heart-rate variant of the test, and without a paired sensor the watch cannot produce the FTP figure the test exists to give you. **Which Polar watches have the Cycling Performance Test?** Eight: Grit X Pro, Pacer Pro, Vantage V2, Vantage V3, Grit X2 Pro, Vantage M3, Grit X2 and Street X [Polar Cycling Performance Test]. You can choose a 20, 30, 40 or 60-minute test, and the result updates your power zones in Flow. The Ignite line does not have it and does not support power sensors either [Polar Power Sensors]. **Will my ANT+ power meter work with a Polar watch?** Not if it is ANT+-only. Polar's compatibility page states of the PowerTap G3 that Bluetooth only and Dual models are supported while the ANT+ only model is not [Polar Power Sensors]. You need a Bluetooth or dual-protocol sensor. This catches out a lot of riders who own a perfectly good meter. **My watch ran the FTP test but will not follow a power interval. Why?** Because those are two different capability lists. Power-based phased training targets are supported only on Grit X Pro, Grit X2 Pro, Street X, Vantage V2 and Vantage V3 [Polar Training Targets], while the Cycling Performance Test also runs on the Vantage M3 and Grit X2 [Polar Cycling Performance Test]. On those two, rebuild the session with heart-rate phases, which are supported far more widely. **Is a heart-rate-only training plan actually worse?** It has wider error bars, not a lower ceiling. Polar's own Cardio Load is heart-rate-based by design [Polar Training Load Pro], so the load accounting works. The two things to get right are a maximum heart rate taken from a real effort rather than an age formula [Tanaka et al. 2001], and an awareness that heart rate lags on short reps and drifts upward on long ones. Bear in mind FTP itself carries roughly 9% limits of agreement against maximal lactate steady state [Borszcz et al. 2019], so power is not the precise instrument it is often assumed to be. ## References 1. **Polar Cycling Performance Test.** [Cycling Performance Test](https://support.polar.com/en/cycling-performance-test). Polar Support. 2. **Polar Power Sensors.** [Which third-party power sensors are compatible with Grit X/Pacer/Street X/Vantage?](https://support.polar.com/en/which-third-party-sensors-are-compatible-with-polar-vantage). Polar Support. 3. **Polar Training Targets.** [How do I create a training target in Polar Flow web service?](https://support.polar.com/en/support/how_do_i_create_a_training_target_in_polar_flow). Polar Support. 4. **Polar Training Load Pro.** [Training Load Pro](https://support.polar.com/en/training-load-pro). Polar Support. 5. **Tanaka et al. 2001.** [Age-predicted maximal heart rate revisited](https://pubmed.ncbi.nlm.nih.gov/11153730/). Journal of the American College of Cardiology. 6. **Jones et al. 2010.** [Critical power: implications for determination of VO2max and exercise tolerance](https://pubmed.ncbi.nlm.nih.gov/20195180/). Medicine & Science in Sports & Exercise. 7. **Borszcz et al. 2019.** [Is the functional threshold power interchangeable with the maximal lactate steady state in trained cyclists?](https://pubmed.ncbi.nlm.nih.gov/30676826/). International Journal of Sports Physiology and Performance. --- # Polar Flow to Strava sync: what it carries, and what it quietly drops URL: https://www.adaptcycling.com/guides/polar-flow-strava-sync-cycling Updated: 2026-08-27 Author: Jim Camut The Flow-to-Strava connection is one-way and mostly reliable, and it has four documented exclusions that surprise people months later. Rides recorded before you linked the accounts never travel. Manually added sessions never travel. Sessions with no sample data never travel. Calorie data is not transferred at all [Polar Strava Sync]. None of these produce an error, which is why the gaps are usually discovered by noticing a hole in a chart rather than by a failure. ## How the connection actually works It runs in one direction, Flow to Strava, and it fires when Flow itself syncs rather than on a schedule of its own. You connect it from Flow's Settings under the Partners tab, click Connect beside Strava, and authorize through a Strava login [Polar Strava Sync]. From then on sessions are synced to Strava when you sync them to Polar Flow. That coupling matters more than it looks. The trigger is your Flow sync, not a background job at Strava's end, so a ride sitting on a watch you have not synced is not late — it has not been handed over yet. Riders who leave the watch un-synced for a few days and then wonder why Strava is behind are seeing the design rather than a fault. It is also worth being precise about the direction, because the reverse assumption causes real confusion. Nothing travels from Strava back into Polar Flow. A ride recorded on a different head unit and uploaded to Strava will never appear in your Flow diary, and no Strava data feeds Polar's Training Load Pro or its zone calculations. Flow's picture of your training is built only from what Polar recorded. For anything the automatic sync refuses, Polar's answer is manual: those sessions can be transferred to Strava with Polar Flow's export function [Polar Strava Sync]. That works, and it is a per-session chore rather than a backfill, which is why almost nobody does it for more than a handful of rides. ## The four things that never arrive Polar lists the exclusions plainly. Previously recorded sessions require manual export rather than syncing automatically, manually added training sessions do not sync, sessions without sample data do not sync, and sessions synced from Polar group solutions or third-party applications do not sync [Polar Strava Sync]. The pre-link exclusion is the one that costs most. Connecting Flow to Strava does not backfill your history — the connection starts working from the moment you make it, and everything before it stays in Flow unless you export it by hand. A rider with three years of Polar riding who links Strava today has a Strava account that starts today. If both accounts matter to you, link them on day one rather than when you first need the data somewhere else. The sample-data rule is more specific than it sounds and is the one that catches indoor riders. To sync, a session has to contain sample data needed to draw curves, from GPS, heart rate, altitude, cadence or power [Polar Strava Sync]. Any one of those is enough, so a trainer session with a heart-rate strap qualifies. A session with a duration and nothing else does not, which is exactly the shape of a ride logged from memory. That connects directly to the third exclusion, because a manually added session is usually a session with no samples. If you add a ride to your Flow diary by typing in the time and the sport, Strava will not receive it. The same applies to anything that arrived in Flow from a Polar group product or another service rather than from your own device. The fourth item is not an exclusion so much as a silent omission: calorie data is not transferred from Polar Flow to Strava, although Strava calculates its own figure for cycling and running [Polar Strava Sync]. So the number is not missing on Strava, it is simply a different number, computed by a different model from a different input. Comparing the two is meaningless, and neither should be treated as a measurement. ## Which one should a training app read? Both, if you can, and for different reasons. Strava holds your archive and can be bulk-exported [Strava Data Export]. Polar holds the live feed and the sensor detail. The relevant asymmetry is that Polar's public API hands over no history at all when you first connect it [Polar AccessLink API]. That last point is worth stating exactly, because it is the mirror image of the Strava exclusion above. Polar's API returns only exercises uploaded to Flow in the last 30 days, and only those uploaded after your account was registered with the connecting client [Polar AccessLink API]. Both conditions apply together. At the moment you connect, nothing satisfies them, so a Polar-only connection starts from zero and fills forward. The 30 days is how long the live feed keeps a ride available, not a window into the past. So the two platforms fail in opposite directions and the fix is the same in both cases: connect early. A rider who links Strava to Flow on day one has an archive that any training service can read. A rider who connects Polar to a training service today gets an excellent feed of everything from today, and should connect Strava alongside it if their history matters. There is a real cost to reading only one of them, and it is not tidiness. Whether a training year works is largely decided by whether the easy rides stayed easy and the hard ones stayed hard [Seiler 2010], and that is an audit you can only run over a complete record. A ledger missing every pre-link ride, every manually logged session and every sample-less trainer effort will systematically under-count the easy end, because those are exactly the sessions people log by hand. Polar's own metrics inherit the same problem from the other side. Cardio Load is calculated from heart rate and session duration [Polar Training Load Pro], so it scores only what the watch recorded with a strap. Training with a Polar means holding two partial records in your head unless something reconciles them. That reconciliation is the job we take on: we read the rides, deduplicate what arrives twice, and build the plan from the merged picture rather than from whichever feed happened to be complete. ## Common questions **Why did my old Polar rides not appear on Strava after I connected?** Because the connection is not a backfill. Polar states that previously recorded sessions require manual export using Flow's export function, while future sessions sync automatically once the accounts are linked [Polar Strava Sync]. Everything before the link stays in Flow until you export it yourself. **Does Strava data sync back into Polar Flow?** No. The connection is one-way, Flow to Strava [Polar Strava Sync]. A ride recorded on another device and uploaded to Strava will not appear in your Flow diary, and it will not contribute to Training Load Pro or to any Polar zone calculation. **Why is one of my rides missing from Strava?** Check for sample data first. A session must contain sample data from GPS, heart rate, altitude, cadence or power to sync [Polar Strava Sync]. Any one of them is enough. Sessions added manually to the Flow diary, and sessions that came into Flow from a Polar group product or another app, are also excluded from the automatic sync. **Why do my calories differ between Polar Flow and Strava?** Because they are two independent estimates. Calorie data is not transferred from Polar Flow to Strava, and Strava calculates its own figure for cycling and running [Polar Strava Sync]. Different models, different inputs, different numbers. Neither is a measurement, so do not reconcile them. **Should I connect a training app to Polar or to Strava?** Both, and connect Strava first if you have history. Polar's API returns only exercises from the last 30 days that were also uploaded after your account was registered with the connecting client [Polar AccessLink API], so a Polar-only connection begins with nothing and fills forward. Strava holds the archive. Polar gives the cleaner live feed from the watch. ## References 1. **Polar Strava Sync.** [How can I automatically sync my data from Polar Flow to Strava?](https://support.polar.com/en/support/how_can_i_automatically_sync_my_data_from_polar_flow_to_strava). Polar Support. 2. **Polar AccessLink API.** [Polar AccessLink API](https://www.polar.com/accesslink-api/). Polar Developers. 3. **Polar Training Load Pro.** [Training Load Pro](https://support.polar.com/en/training-load-pro). Polar Support. 4. **Strava Data Export.** [Exporting your Data and Bulk Export](https://support.strava.com/hc/en-us/articles/216918437-Exporting-your-Data-and-Bulk-Export). Strava Support. 5. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. --- # Training with just a Polar H10 and no Polar watch URL: https://www.adaptcycling.com/guides/polar-h10-only-cycling-training Updated: 2026-08-27 Author: Jim Camut A great many H10 owners have never owned a Polar watch. The strap sells to runners, gym users and heart-rate-variability trackers, and a lot of cyclists bought one on reputation alone. The good news is that an H10 and a phone is a complete, free, genuinely usable training stack: Polar Beat records the session, Flow stores it, and any training service that reads Polar sees a proper ride. Three specific things are missing, and it is worth knowing which before you buy a watch to fix them. ## What an H10 and a phone actually give you More than most people expect, and it costs nothing. Polar Beat is free, gives live heart rate and real-time voice guidance, tracks your route and distance through the phone's location services, and syncs your sessions to Polar Flow [Polar Beat]. That is a recorded ride with heart rate, GPS, distance and duration. The consequence for Polar's own analytics is the important part. Cardio Load is TRIMP, calculated after every workout from your heart rate and session duration, and affected by your physical settings, resting and maximal heart rate, and gender [Polar Training Load Pro]. Every one of those inputs is present in an H10-plus-Beat ride. So the strap-only rider gets the full Cardio Load and the full Cardio Load Status ratio — Detraining below 0.8, Maintaining to 1.0, Productive to 1.3, Overreaching above it — with no watch involved at all. The rides also travel. Flow's automatic Strava sync requires a session to contain sample data from GPS, heart rate, altitude, cadence or power [Polar Strava Sync], and a Beat-recorded ride carries at least two of those. So the H10 route feeds Strava as cleanly as a watch does, which matters because Strava is where most training services find your history. One practical note that is easy to get wrong: set your maximum heart rate in Flow from a real effort rather than accepting a default. Every zone and every Cardio Load figure is computed against it, so it is the single setting that scales everything else. A strap-only rider has no watch to prompt them, which makes it the setting most often left at whatever the app assumed. ## The strap records without the phone, with a catch at each end The H10 has an internal memory that can store heart rate data of one training session [Polar H10 Memory], so you can leave the phone behind. The catch is that Polar Beat has to be there to start the recording and to collect it afterwards, which makes it a phone-free ride rather than a phone-free workflow. The procedure is specific. Open Polar Beat on the phone the sensor is paired with, select your sport profile, tap the three dots at the top right, choose Save HR with sensor, and tap Start [Polar H10 Memory]. Then go and ride. When you come back into Bluetooth range, Beat automatically offers to download the finished session from the sensor memory and syncs your Flow diary. The capacity is generous for a strap and hard-limited in one dimension: the H10 records heart rate data once per second and a stored session can run beyond 30 hours [Polar H10 Duration], but the memory holds one session at a time [Polar H10 Memory]. So it covers a long ride comfortably and will not cover a weekend of two rides without a download in between. What comes back is heart rate and nothing else. There is no GPS in the strap, so a memory-recorded ride has no route, no distance and no speed — those come from the phone in the normal Beat workflow. For a training service the session is still perfectly usable, because heart rate and duration are what Cardio Load and any heart-rate-anchored plan actually need. It just will not look like a ride on a map. ## The three things a watch would add, and which of them matter Muscle Load, Recovery Pro, and on-wrist workout guidance. Only one of the three is really about the strap; the other two are watch features. Deciding whether to buy a watch is mostly a question of which of them you would actually use. Muscle Load is the clean gap and it is a power gap rather than a watch gap. It is calculated by multiplying average power by session duration [Polar Training Load Pro], so it needs a power meter, and a Polar watch without one produces no Muscle Load either. If you have no power meter, buying a watch does not close this. Cardio Load will remain your only load axis, which is a coherent way to train and simply means half of Training Load Pro stays empty. Recovery Pro is the genuine watch feature, and it is the one an H10 owner is most likely to want, because the H10 is exactly the sensor it requires. It needs an H10 or H9 and the Orthostatic test on at least three mornings a week, and returns a specific daily training advice from a fixed set that runs from Train more through Take it easy! [Polar Recovery Pro]. It runs only on the Grit X Pro, Grit X2, Grit X2 Pro, Street X, Vantage M3, Vantage V2 and Vantage V3. You own the sensor; you do not own the thing that reads it. The third is guidance during the session. A watch can walk you through a phased target on the wrist; Beat gives real-time voice guidance [Polar Beat] but not a structured interval sequence. For steady endurance and long tempo work the voice cues are enough. For a set of short reps with tight recoveries, a phone in a jersey pocket is a worse instrument than a watch, and that is the honest case for upgrading. None of the three is the plan itself, which is the pattern that repeats across the whole question of training with a Polar. Cardio Load tells you how hard the last six weeks were. Recovery Pro tells you whether today should be easier. Neither writes the session, and no watch in the range does. That layer has to come from somewhere else, and reading your rides is all it needs from your hardware — which an H10 and a free app already supply. ## Common questions **Can I train with a Polar H10 and no Polar watch?** Yes, and it is a complete setup. Polar Beat is free, records live heart rate with voice guidance, tracks route and distance from the phone's location services, and syncs to Polar Flow [Polar Beat]. The resulting session produces full Cardio Load, because that is calculated from heart rate and session duration [Polar Training Load Pro]. **Can the H10 record a ride without my phone?** One ride at a time, and the phone has to bookend it. The strap stores heart rate data for one training session, started from Polar Beat by choosing Save HR with sensor, and downloaded automatically when you return to Bluetooth range [Polar H10 Memory]. It records heart rate only — no GPS, route or distance. **How long can the H10 record for on its own?** Beyond 30 hours. It samples heart rate once per second and a stored session can hold data up to more than 30 hours [Polar H10 Duration]. The binding limit is not duration but count: the memory holds a single session, so download it before the next ride. **Will my H10 rides appear on Strava?** Yes, once Flow and Strava are linked. The automatic sync needs a session to carry sample data from GPS, heart rate, altitude, cadence or power [Polar Strava Sync], and a Beat-recorded ride has heart rate at minimum. Note that only rides recorded after you link the accounts sync automatically. **Would buying a Polar watch give me more than the strap does?** Three things, and one of them you may not get. Recovery Pro needs a watch from a specific list even though it requires your H10 as the sensor [Polar Recovery Pro]. On-wrist structured workout guidance is a watch feature. Muscle Load is not — it is average power multiplied by duration [Polar Training Load Pro], so a watch without a power meter produces none either. ## References 1. **Polar Beat.** [Polar Beat](https://support.polar.com/en/beat). Polar Support. 2. **Polar H10 Memory.** [H10 user manual - Sensor memory](https://support.polar.com/e_manuals/h10-heart-rate-sensor/polar-h10-user-manual-english/sensor%20memory.htm). Polar Support. 3. **Polar H10 Duration.** [How long a training session can I record with H10?](https://support.polar.com/us-en/how_long_a_training_session_can_i_record_with_h10). Polar Support. 4. **Polar Training Load Pro.** [Training Load Pro](https://support.polar.com/en/training-load-pro). Polar Support. 5. **Polar Recovery Pro.** [Recovery Pro](https://support.polar.com/en/recovery-pro). Polar Support. 6. **Polar Strava Sync.** [How can I automatically sync my data from Polar Flow to Strava?](https://support.polar.com/en/support/how_can_i_automatically_sync_my_data_from_polar_flow_to_strava). Polar Support. --- # Is Polar FitSpark any good for cyclists? URL: https://www.adaptcycling.com/guides/polar-fitspark-cycling Updated: 2026-08-27 Author: Jim Camut Yes for what it is, and it is not what most people searching this question want. FitSpark offers two to four ready-made workouts a day chosen against your recovery and training history, and Polar states plainly that its cardio sessions can be completed with any sport profile on your watch [Polar FitSpark]. So it genuinely works on a bike. What it cannot do is build toward anything, because a daily suggestion has no goal date and no memory of the eleven weeks in front of it. ## What FitSpark actually offers, and yes it runs on the bike Three workout types and two to four options each day [Polar FitSpark]. Cardio is heart-rate-zone training with warm-up, work and cool-down phases. Strength is circuit work with bodyweight or equipment. Supportive combines strength and mobility. Only the cardio type is relevant to cycling, and it is the one that is sport-agnostic. Polar's wording on the sport question is unambiguous: cardio sessions can be completed with any sport profile on your watch [Polar FitSpark]. Pick the cycling profile, start the suggested session, and the watch guides you through the heart-rate phases exactly as it would for a run. There is no cycling-specific version of a FitSpark session, but there does not need to be — a heart-rate zone is a heart-rate zone, and the structure carries across. That makes FitSpark more useful to a cyclist than its marketing implies. Polar files it under daily inspiration on its own training-plan menu, alongside Flow manual planning, the Running Program, TrainingPeaks and the Fitness Program [Polar Training Plans], and inspiration undersells a structured heart-rate session delivered to your wrist with no setup. The strength and supportive types are worth a sentence because they are the reason a lot of riders dismiss FitSpark after a week. Circuit sessions dominate the suggestions for some users, and a cyclist looking for bike work sees a list of things they will not do. The cardio option is there every day; you have to pick it. ## What the suggestion is actually based on More than a random workout generator and less than a coach. Polar bases the recommendations on your training history — average weekly heart-rate zones from the past 28 days — plus your VO2max from the fitness test and your training background, adjusted by your Nightly Recharge status from the previous night [Polar FitSpark]. The 28-day window is the interesting number. It means FitSpark is looking backward at roughly a training month and asking what is under-represented, then offering something that fills the gap. That is a real and defensible logic, and it is the same reasoning a coach applies when they notice you have not been above threshold in three weeks. It is also entirely retrospective. The recovery input is genuine rather than cosmetic. Nightly Recharge combines heart rate, heart-rate variability as RMSSD and breathing rate measured during roughly the first four hours of sleep, compared against your own 28-day baseline [Polar Nightly Recharge]. So a poor night produces a softer suggestion, from a physiological measurement rather than a self-report. Very few daily-suggestion systems are grounded that well. The gap in the inputs is what it does not read. There is no goal, no event date, and no forward structure — nothing in the list of inputs is about where you are trying to get to. A system that reads the last 28 days and last night can tell you what today should probably look like. It cannot tell you that this is week seven of eleven and the volume needs to come down next week so the peak lands correctly. ## Why a good daily suggestion is still not a plan The difference is not quality, it is direction. Periodization works by sequencing blocks so that each one builds on the adaptations of the last and the whole thing converges on a date [Issurin 2010]. A day-at-a-time system has no mechanism for sequencing, because each decision is made without reference to the decisions after it. Run the thought experiment. FitSpark offers a sensible session every day for eleven weeks, and you take the best one each time. What you get is eleven weeks of individually reasonable training with no deliberate build, no planned recovery week, and no taper. It will not be bad training. It will not be a plan for the event in week twelve, and if the event is the point, the difference shows up on the day. There is a second-order effect worth naming because it is the classic self-coached failure mode. Left to daily choice, most riders drift toward the moderate middle, and whether a training year works is largely decided by whether the easy rides stayed easy and the hard ones stayed hard [Seiler 2010]. A system that asks what is missing from the last 28 days tends to answer with intensity, which is the direction the drift already runs. Where FitSpark earns its place is in the gaps a plan leaves. An unplanned free day, an off-season week with no structure, a session where you genuinely do not know what to do — those are exactly the situations a good daily suggestion solves, and it solves them with a real recovery measurement behind it. Used that way it is a useful part of training with a Polar rather than a substitute for the part Polar does not build. That missing part is the plan itself, and it is a documented vacancy rather than an oversight — Polar's own menu of five training options contains nothing cycling-specific [Polar Training Plans]. AdaptCycling fills it from the other direction: we read the rides your Polar records, build a periodized plan with a goal and an arc, and rewrite it when the week collapses. FitSpark still answers today better than we do on the days you are off-plan, and the two coexist without conflict. ## Common questions **Can I use FitSpark for cycling workouts?** Yes. Polar states that FitSpark cardio sessions can be completed with any sport profile on your watch [Polar FitSpark], so selecting the cycling profile works. The sessions are heart-rate-zone based with warm-up, work and cool-down phases. There is no cycling-specific FitSpark session, but heart-rate structure carries across sports without modification. **How does FitSpark decide what to suggest?** From your average weekly heart-rate zones over the past 28 days, your VO2max from the fitness test, and your training background, adjusted by the previous night's Nightly Recharge status [Polar FitSpark]. Nightly Recharge itself combines heart rate, heart-rate variability and breathing rate measured during the first hours of sleep against your own 28-day baseline [Polar Nightly Recharge]. **Is FitSpark a training plan?** No, and Polar does not present it as one — its own menu files FitSpark under daily inspiration [Polar Training Plans]. It offers two to four options each day based on the recent past. It has no goal date, no forward structure and no way to sequence blocks so they converge on an event [Issurin 2010]. **Does FitSpark use power targets?** No. FitSpark cardio sessions are built on heart-rate zones [Polar FitSpark]. If you want power-based interval guidance on a Polar watch you need a phased training target, and those support power on only five models. Nothing in FitSpark is expressed in watts. ## References 1. **Polar FitSpark.** [FitSpark daily training guide](https://support.polar.com/us-en/fitspark-daily-training-guide). Polar Support. 2. **Polar Training Plans.** [Which Polar Training Plan Is Right for You?](https://www.polar.com/en/guide/polar-training-plans). Polar. 3. **Polar Nightly Recharge.** [Nightly Recharge recovery measurement](https://support.polar.com/us-en/nightly-recharge-recovery-measurement). Polar Support. 4. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. 5. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. --- # Is there a Runna for cycling? What the equivalent actually looks like URL: https://www.adaptcycling.com/guides/runna-for-cycling Updated: 2026-09-01 Author: Jim Camut Yes, the category exists — but not as one app with one name, and the closest literal answer now belongs to Strava. Runna is running-only. The app most often named as its cycling counterpart, The Breakaway, had its core assets acquired by Strava in May 2025, a month after Strava agreed to buy Runna itself. The deeper reason cycling never had a clean Runna analog is not neglect. A bike plan is a harder object than a run plan. ## The short answer, and who owns what Runna does not do cycling. Strava announced a definitive agreement to acquire it in April 2025, then acquired the core assets of the cycling coaching app The Breakaway a month later. Strava is assembling the category. Today there is still no single app that is Runna, for bikes, under one subscription. Runna is a running app in the strict sense. It writes plans for 5K, 10K, half marathon, marathon and ultra distances, plus strength and mobility work for runners, and cycling is not a supported primary sport. It costs $19.99 per month or $119.99 per year with the first week free. Dom Maskell and Ben Parker founded it in 2021, it launched in March 2022, it is available in more than 180 countries, and it was one of three finalists for Apple's App of the Year in 2024. It earned its position — none of what follows is a knock on it. On 17 April 2025 Strava announced it had entered a definitive agreement to acquire Runna [Strava Runna 2025]. Strava said it planned to keep the apps separate for the foreseeable future and to invest in growing the Runna team. Strava now sells a combined Strava and Runna bundle, and the training plans inside it are running plans. So the answer to "who owns Runna" is Strava, but the answer to "is Runna now inside Strava" is not yet. A month later, on 22 May 2025, Strava announced it had acquired the core assets of The Breakaway, a Y Combinator incubated cycling training app, with co-founders Jordan Kobert and Kyle Yugawa joining the company [Strava Breakaway 2025]. The Breakaway bills itself as an AI personal performance coach — weekly targets, daily guidance, team features — which is about as close to "Runna for cycling" as the market has produced. Strava said its personalized cycling training, ride analysis and achievement-tracking tools would be integrated into Strava's subscription offering. That is the plan Strava stated; it is worth watching rather than guessing at. ## Why a running plan does not port to a bike A running plan is anchored to pace and distance, which behave almost like a single axis. A cycling plan is anchored to power, and the same hour on a bike can mean wildly different work. The gap is not app maturity. It is the shape of the data. Start with the most concrete difference: a runner is producing power essentially the whole time they are moving, while a cyclist can stop pedalling and keep rolling. Elapsed time on a bike systematically overstates work done. An hour of rolling terrain with twenty minutes of descending is not an hour of training, and no distance-based rule can tell the two apart. That is why cycling load has to be intensity-weighted rather than distance-based: volume indices like kilometres or hours per week fail to account for how hard the work actually was [Jobson et al. 2009] — the mileage number that anchors a marathon plan simply is not the load number on a bike. The variance compounds from there. The same sixty minutes can be a wheel-sucking group ride, a solo tempo effort into a headwind, or an indoor session where the trainer holds you at a fixed target with no coasting at all. Foster's monitoring work established the case for tracking session intensity alongside duration precisely because duration alone misrepresents the stress an athlete absorbed [Foster 1998]. A running app can get a long way on pace and distance. A cycling app that tried the same thing would be wrong about most of the week. ## What a cycling plan has to do that a running app does not When a week breaks, a cycling plan cannot simply be re-dated. It has to be re-periodized — the next block adjusted to what the last block actually delivered, and the weekly balance of easy and hard work preserved rather than quietly inverted. Block periodization only works if each block responds to what the previous one actually produced, not to what it was scheduled to produce [Issurin 2010]. If a base block ends 30% under prescribed volume because of a work trip, starting the build block on the original date is not adaptation — it is the same plan, later. Shifting dates is the cheap version of the problem and it is what most calendar-based tools do. The harder constraint is intensity distribution. The most replicated finding in endurance training is that the ratio of easy to hard work across the week drives the adaptation, with roughly 80% of time at low intensity and 20% genuinely hard [Seiler 2010]. That makes the obvious repair the wrong one: when a rider misses an easy Tuesday and the plan compensates by adding threshold work on Thursday, the week's ratio moves in exactly the wrong direction. A rebuilt cycling week has to preserve the intensity balance, not just the total. This is the whole subject of our wider guide to adaptive cycling training plans, which sets out what it takes for a plan to survive real life rather than assume an athlete who never misses a session. The short version is that reading your rides is table stakes; restructuring the plan around them without breaking the periodization rules is the part almost nobody ships. ## So what should you actually use Three honest answers, depending on what you want. Wait for Strava if you want everything in one place. Pick an established cycling app if your training week is stable. Pick something life-adaptive if your week is the thing that keeps breaking. If you want the Runna experience inside the app you already use, waiting is reasonable. Strava has said The Breakaway's cycling training and ride-analysis tools are being integrated into its subscription [Strava Breakaway 2025], and Strava is the best in the world at the parts of this it already does. If a single subscription and a single social graph matter more to you than having the plan today, that is a legitimate call. If your training week is a metronome — the same slots, the same hours, the same recovery rhythm — the established cycling apps are mature and deep, and an intensity-adaptive tool inside a fixed weekly skeleton will serve you well. Our head-to-head on TrainerRoad, JOIN and AdaptCycling walks through where each one stops. AdaptCycling is the answer for the third case: the rider whose week is the unstable variable. It costs $15 per month or $150 per year, reads Wahoo, Hammerhead, Polar or intervals.icu, writes the plan from your real ride history, and rebuilds it when the week falls apart. Strava seats are capped, so Strava riders join a short waitlist. The tradeoff is real and worth stating plainly. It is cycling only — if you run as well and want one plan across both sports, this is the wrong tool and Runna plus a cycling app is the better stack. It does not have the workout-library depth of the big indoor platforms. And the free tier is read-only once the 14-day trial ends, so the adaptive plan is the paid product, not the free one. ## Common questions **Does Runna support cycling?** No. Runna writes running plans from 5K to ultra distance, plus strength and mobility work aimed at runners. Cycling is not a supported primary sport. Riders sometimes use it for a running block alongside a separate cycling plan, but there is no bike plan inside Runna to switch to. **Is The Breakaway the cycling version of Runna?** It is the closest analog the market produced — an AI coaching app giving cyclists weekly targets and daily guidance. Strava acquired its core assets in May 2025 and said those tools would be integrated into Strava's subscription [Strava Breakaway 2025]. Where that leaves the standalone product is Strava's to announce, not ours to predict. **Why can't I just follow a running plan structure on the bike?** Because the load numbers do not transfer. A runner produces power almost continuously; a cyclist coasts, so elapsed time overstates the work. Cycling load has to be intensity-weighted rather than distance-based [Jobson et al. 2009], which means the weekly structure a marathon plan is built on has no clean equivalent in watts. **Is there a free Runna equivalent for cycling?** Not for a plan that writes and rebuilds itself. Intervals.icu is free and excellent for analysis and manual plan-building, and Strava's own training features are included in its subscription. AdaptCycling's free tier is read-only after a 14-day trial with no card — the adaptive plan itself is the paid tier at $15 per month. ## References 1. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 2. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. 3. **Seiler 2010.** [What is best practice for training intensity and duration distribution in endurance athletes?](https://pubmed.ncbi.nlm.nih.gov/20861519/). International Journal of Sports Physiology and Performance. 4. **Jobson et al. 2009.** [The analysis and utilization of cycling training data](https://pubmed.ncbi.nlm.nih.gov/19757861/). Sports Medicine. 5. **Strava Runna 2025.** [Strava to Acquire Runna — A Leading Running Training App](https://press.strava.com/articles/strava-to-acquire-runna-a-leading-running-training-app). Strava Press. 6. **Strava Breakaway 2025.** [Strava Announces Acquisition of Cycling Training App The Breakaway](https://press.strava.com/articles/strava-announces-acquisition-of-cycling-training-app-the-breakaway). Strava Press. --- # TrainingPeaks vs Intervals.icu: what each one is actually for URL: https://www.adaptcycling.com/guides/trainingpeaks-vs-intervals-icu Updated: 2026-09-01 Author: Jim Camut Start here: Intervals.icu is the deeper analytics platform and its core is free, while TrainingPeaks is the better place to work with a coach. That settles it for most riders. The complication is that neither one writes your plan or rebuilds it when the week falls apart — TrainingPeaks expects a coach to do that job, and Intervals.icu expects you to. Deciding between them is really deciding which of those two jobs you are hiring for. ## They are not competing on the axis you think Both draw fitness and fatigue curves, which makes them look like rivals. They are not. TrainingPeaks is a coach-delivery platform with excellent analytics attached; Intervals.icu is an analytics workbench with planning tools attached. Almost every bad purchase here comes from comparing the charts instead of the jobs. TrainingPeaks Premium runs $19.95 a month or $134.99 a year [TrainingPeaks 2026], and the Performance Management Chart is the reason most people know the name. It is a direct consumer implementation of the Banister fitness-fatigue model: performance as a fitness term minus a fatigue term, each decaying on its own time constant [Morton et al. 1990]. Layered on top is Training Stress Score — duration times intensity squared, scaled to your FTP — feeding the chronic and acute load curves [Mujika 2017]. But the product around that chart is a workspace shared with a coach: the calendar they write into, the structured files that push targets to your head unit, the plan Marketplace. TrainingPeaks has been part of Garmin since the acquisition announced in July 2026. Intervals.icu comes at it from the opposite direction. The core platform is free, with an optional Supporter tier at $4 a month that adds weather analysis, the annual training plan builder, and early access to betas [Intervals.icu 2026]. On raw analytical depth it is not merely competitive, it is ahead: eFTP, several power-curve models, more than 140 custom metrics you can define yourself, an open API, and integrations with 250-odd services. It covers cycling, running, swimming, and triathlon with separate zones per sport. One practical decoder note for anyone reading both at once. The Performance Management Chart vocabulary — CTL, ATL, TSB, TSS — is TrainingPeaks terminology, so Intervals.icu names the same underlying quantities Fitness, Fatigue, and Form. The curves are computed from the same modeling lineage; it is the labels that differ, not the methods. Riders regularly assume the two platforms disagree about their fitness when they are looking at the same math under different names. ## What the $19.95 actually buys, given the free one is deeper The uncomfortable comparison for TrainingPeaks is that a free tool models fitness at least as well and customizes far better. What the subscription buys is not superior charts. It is the coach ecosystem, the device-integration breadth, the Marketplace, and the shared professional vocabulary of the sport. Take the analytics head-on, since it is the axis most people compare. For a rider who wants custom metrics, a power curve they can interrogate several ways, or their own data out through an API, Intervals.icu wins outright and costs nothing [Intervals.icu 2026]. Multi-sport is a genuine second win: if you run in winter or train for a triathlon, separate zones per sport in one place is worth real money, and here it is free. A third is openness — the API means your training history is never hostage to a subscription lapse. TrainingPeaks earns its price on a different set of things, and they are not trivial. It is the rail the paid-coaching industry runs on, so if a coach is delivering your training, they are almost certainly delivering it here and the subscription is the cost of the shared workspace. Its device integrations are the broadest in endurance sport. The Marketplace is a real inventory of purchasable structured plans. And the PMC vocabulary is the language coaches, podcasts, and training books actually speak — learning it has transfer value beyond the app. One concrete limitation worth knowing before you buy. The TrainingPeaks Workout Builder is web-only; the mobile app cannot create or edit structured workouts. If your habit is to reshape Thursday from your phone on Wednesday night, that is a laptop trip every time. ## The job neither one does Both platforms will show you that your load ramp is too steep. Neither will redesign next week to fix it. TrainingPeaks leaves prescription to a human coach by design. Intervals.icu hands you a calendar, a workout builder, and a phase structure, then leaves the dragging to you. This is a property of the metrics, not a gap in the software. Mujika's review of load quantification is explicit that external-load measures like TSS exist to estimate the biological stress an athlete already absorbed, not to prescribe the stress they should take on next [Mujika 2017]. Foster's monotony work is the same shape: it can tell you that low day-to-day load variance plus accumulated load is a risk pattern, and it will not restructure the week that produced it [Foster 1998]. Both charts are measurement instruments pointed backward. They are excellent dashboards and poor autopilots. Be precise about Intervals.icu here, because it is the claim people most often get wrong. It is not analysis-only. It has a drag-and-drop calendar workout builder and a deterministic Annual Training Plan Builder that lays out base, build, and peak phases — real planning tools, and more than TrainingPeaks gives a rider with no coach. What it does not do is generate adaptive day-to-day workouts, or rebuild the plan when you miss a block. When a work trip eats Tuesday through Thursday, nothing recomputes. You open the calendar and move the workouts yourself, and you decide what the block should now look like — which is exactly the periodization judgment that is hardest to make about your own training [Issurin 2010]. The clearest evidence that this is deliberate is the ecosystem that has grown on top of it. IntervalCoach, Intervals Pro and Athletica all exist in part to bolt prescription onto Intervals.icu's open API, and its own forum runs a dedicated AI Tools section. A platform does not accumulate that many third-party layers for a job it was trying to do itself. For the self-coached cyclist the consequence is specific and worth sitting with: both of these tools solve the easy half of training without a coach, which is seeing what happened, and leave the hard half — periodizing the year and enforcing recovery when you would rather not — entirely to you. ## Who should pick which Pick TrainingPeaks if a coach is in the loop, or you want the vocabulary the sport speaks. Pick Intervals.icu if you are analytically self-sufficient and want the deepest free modeling in endurance sport. Pick neither if what you actually need is the plan written and rebuilt for you. Choose TrainingPeaks, over Intervals.icu and over us, in three situations. You work with a coach, or expect to this season — it is their delivery surface. You intend to buy a structured plan from the Marketplace and follow it as written, accepting that even the Dynamic ones only shift dates onto your calendar rather than reshaping the week around what you rode. Or you are deep enough into the analytical craft that the PMC lineage and the coach-facing tooling are the point. Choose Intervals.icu, over TrainingPeaks and over us, when you are happy making your own calls. If you know what a base phase should look like, you enjoy building the week, and you want to own your data through an API rather than rent a dashboard, nothing comes close on value — it is free and it is better at analysis than the paid option. Add multi-sport and the case strengthens: AdaptCycling is cycling-only, so a triathlete has a straightforward reason to pick Intervals.icu and not look back. The third option only makes sense if the thing you are missing is the deciding. AdaptCycling is $15 a month or $150 a year, cycling only, and it writes the periodized plan from your real ride history then rebuilds it when the week breaks — with a chat coach that remembers your season across months. It does not try to out-analyze either platform above, and it will not replace a good human coach. Worth being straight about the pricing: there is a 14-day trial with no card, after which the free tier is read-only. It answers a different question: not what did my training do, but given that, what should the next eight weeks be. ## Common questions **Is Intervals.icu really free?** Yes, the core platform genuinely is — fitness tracking, activity and interval analysis, the training calendar, the workout builder, and the integrations all sit in the free tier. There is an optional Supporter tier at $4 a month that adds weather analysis, the annual training plan builder, and early beta access [Intervals.icu 2026]. It is not a time-limited trial and there is no card gate on the core. **Does either one write a training plan for you?** Neither writes and adapts one on its own. TrainingPeaks expects a coach to author the plan or expects you to buy a static one from its Marketplace. Intervals.icu gives you real planning tools — a workout builder and a base-build-peak annual plan builder — but you make the periodization decisions and you move the workouts by hand when life disrupts the week [Mujika 2017]. **Why do the two platforms use different names for the same charts?** Terminology, not method. CTL, ATL, TSB, and TSS are TrainingPeaks terminology, so Intervals.icu calls the equivalent quantities Fitness, Fatigue, and Form. Both descend from the same fitness-fatigue modeling lineage [Morton et al. 1990]. If your numbers differ slightly between platforms it is usually down to FTP settings or which activities each one ingested, not a disagreement about the underlying math. **Can I use both at the same time?** Plenty of riders do, and it is coherent if you assign each a job. TrainingPeaks becomes the surface you share with a coach; Intervals.icu becomes the place you actually interrogate the data, because the custom metrics and power-curve models go deeper. Both ingest from the same upstream sources, so the duplication costs you setup time rather than data quality. **Which is better for a rider with no coach and limited time?** Intervals.icu, on price and depth, if you are willing to make the training decisions yourself. The caveat is that limited time usually means an unpredictable week, and a manual calendar is precisely what stops getting maintained when the week goes sideways. If you find yourself repeatedly not rebuilding the plan after a disrupted week, the problem is not the analytics tool you chose [Foster 1998]. ## References 1. **Morton et al. 1990.** [Modeling human performance in running](https://pubmed.ncbi.nlm.nih.gov/2246166/). Journal of Applied Physiology. 2. **Mujika 2017.** [Quantification of training and competition loads in endurance sports: methods and applications](https://pubmed.ncbi.nlm.nih.gov/27918666/). International Journal of Sports Physiology and Performance. 3. **Foster 1998.** [Monitoring training in athletes with reference to overtraining syndrome](https://pubmed.ncbi.nlm.nih.gov/9662690/). Medicine & Science in Sports & Exercise. 4. **Issurin 2010.** [New horizons for the methodology and physiology of training periodization](https://pubmed.ncbi.nlm.nih.gov/20199119/). Sports Medicine. 5. **TrainingPeaks 2026.** [Premium Pricing for Athletes](https://www.trainingpeaks.com/athlete-pricing/). TrainingPeaks. 6. **Intervals.icu 2026.** [Intervals.icu — free training analysis and planning platform](https://intervals.icu/). Intervals.icu.