How long should your taper be? A cyclist's guide to what the evidence actually says

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.

Quick answers

Sources cited in this guide

1. 01
   Bosquet et al. 2007. Effects of tapering on performance: a meta-analysis. Medicine & Science in Sports & Exercise.
2. 02
   Mujika & Padilla 2003. Scientific bases for precompetition tapering strategies. Medicine & Science in Sports & Exercise.
3. 03
   Neary et al. 2003. Effects of different stepwise reduction taper protocols on cycling performance. Canadian Journal of Applied Physiology.
4. 04
   Thomas et al. 2008. A model study of optimal training reduction during pre-event taper in elite swimmers. Journal of Sports Sciences.
5. 05
   Thomas et al. 2009. Computer simulations assessing the potential performance benefit of a final increase in training during pre-event taper. Journal of Strength and Conditioning Research.
6. 06
   Hellard et al. 2006. Assessing the limitations of the Banister model in monitoring training. Journal of Sports Sciences.

More inside The self-coached cyclist

Other articles in this series

1. 01

   How to restart cycling training after two weeks off

   When you miss two weeks on the bike, the question isn't whether to resume — it's how to restart without overshooting the next two.

   →
2. 02

   How often to take a recovery week as a self-coached cyclist

   The 3-on/1-off recovery-week cadence is a guideline, not a law. When self-coached riders should hold it, push it, or deload sooner.

   →
3. 03

   Am I overtrained or just tired? A self-coached cyclist's guide

   Functional overreaching, non-functional overreaching, and OTS — what each looks like, and when fatigue crosses into something more serious.

   →
4. 04

   How to spot and fix junk intensity as a self-coached cyclist

   Most amateur rides drift into the medium-hard zone you can't recover from but don't adapt to. How to spot junk intensity — and three structural fixes.

   →
5. 05

   Why year two stalls: junk intensity and the amateur plateau

   Year-one gains came from any training. Year two stalls when intensity drift inverts the 80/20 rule. The mechanism — and the fix.

   →
6. 06

   Cycling training plan for an irregular schedule

   Shift work, a new baby, and weekly travel break the standard 7-day training week. How to plan around an unpredictable schedule.

   →
7. 07

   What weekly TSS should an amateur cyclist actually target?

   TrainingPeaks charts make weekly TSS look like the goal. For most amateurs it's the wrong primary target — here's what to use instead.

   →
8. 08

   Safe CTL ramp rate for amateur cyclists: how fast is too fast

   How fast a self-coached cyclist can grow CTL without crossing into non-functional overreaching — the numbers, the signals, and the override rules.

   →
9. 09

   The cheapest way to self-coach cycling in 2026

   Free apps cover most of self-coaching. The one tool worth paying for, where the false economies hide, and the cheapest credible stack in 2026.

   →
10. 10

    When self-coached stops working: signals it's time for a coach

    Self-coaching works for most amateur cyclists — until it doesn't. The four honest signals you've outgrown DIY, and the cheaper fixes to try first.

    →
11. 11

    FTP test vs power curve: should self-coached riders test?

    When the self-coached cyclist actually needs an FTP test versus when the power curve is enough — a judgment-call framework with research backing.

    →
12. 12

    Self-coached 12-week training plan for a goal event

    How a self-coached cyclist builds a 12-week base/build/peak/taper block for a goal event — the math, the workouts, and the taper that holds the peak.

    →
13. 13

    Polarized training for amateurs: making 80/20 actually work

    What 80/20 means in the three-zone model, why it's hard to run on 5-7 hours a week, the pyramidal counter-evidence, and how to execute it.

    →
14. 14

    Is TrainingPeaks worth it for a self-coached cyclist?

    TrainingPeaks is the gold-standard analysis platform — but for a self-coached rider with no coach, is it worth $20/mo? An honest verdict.

    →
15. 15

    TrainingPeaks free vs Premium: what the free tier gives you

    What the free TrainingPeaks Basic tier does for a self-coached cyclist, what it withholds, and whether Premium or a free alternative wins.

    →