Indoor vs outdoor FTP: why the same legs make less power on the trainer

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 Strava 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.

Quick answers

Sources cited in this guide

1. 01
   Allen et al. 2019. Training and Racing with a Power Meter (3rd ed.). VeloPress.
2. 02
   Borszcz et al. 2019. Is the Functional Threshold Power Interchangeable With the Maximal Lactate Steady State in Trained Cyclists?. International Journal of Sports Physiology and Performance.
3. 03
   Lipski et al. 2022. Differences in Performance Assessments Conducted Indoors and Outdoors in Professional Cyclists. International Journal of Sports Physiology and Performance.
4. 04
   Galloway and Maughan 1997. Effects of ambient temperature on the capacity to perform prolonged cycle exercise in man. Medicine and Science in Sports and Exercise.
5. 05
   Gonzalez-Alonso et al. 1999. Influence of body temperature on the development of fatigue during prolonged exercise in the heat. Journal of Applied Physiology.
6. 06
   Coyle and Gonzalez-Alonso 2001. Cardiovascular drift during prolonged exercise: new perspectives. Exercise and Sport Sciences Reviews.
7. 07
   Bouillod et al. 2022. Caveats and Recommendations to Assess the Validity and Reliability of Cycling Power Meters: A Systematic Scoping Review. Sensors.

See the direct comparisons

• AdaptCycling vs Zwift →

More inside FTP without a test

Other articles in this series

1. 01

   How to estimate FTP without a power meter

   Estimating FTP from heart rate, RPE, and Strava when you don't own a power meter — how close you can get and where the method breaks down.

   →
2. 02

   20-minute vs 8-minute FTP test: which to use

   How the 20-minute and 8-minute FTP tests differ, the multipliers each uses, and which one fits your riding — plus why both are only protocols.

   →
3. 03

   Why cycling apps show you different FTP numbers

   Strava, Xert, Intervals.icu, and TrainingPeaks can each report a different FTP. Why the estimates diverge and which one to trust.

   →
4. 04

   Is your ramp test FTP too high? Why it happens

   Ramp tests overestimate FTP for anaerobically-gifted riders and underestimate it for diesels. Why the 75% rule misfires and how to correct it.

   →
5. 05

   How often should you test your FTP?

   How often to re-test FTP as a self-coached cyclist — twice a year, not every six weeks — and why modeled estimates change the cadence.

   →