Your FTP went down. Six reasons, and only one of them is fitness
The Moveee team
Free coaching · a real route for every ride
You did the same test on the same trainer as last time. You suffered at least as much. The number came out 12 watts lower, and the rest of the evening was spent wondering where the winter went.
Before you rewrite your plan: a lower test result is not the same thing as lost fitness, and in most cases it is not lost fitness. Here are the six explanations in order of how often they turn out to be the answer.
What an FTP test actually measures
A threshold test measures your best sustained effort on that day, in those conditions, under that protocol, with that motivation. Fitness is one input among several. Treating the output as a pure fitness readout is the mistake that turns an ordinary bad day into a crisis of confidence.
The number also has real error bars. Repeat the same test on the same rider across a week and you will not get the same figure — which means a small change between tests may be indistinguishable from noise. FTP explained: how to test it and raise it covers the protocols; this article is about what to do when one disappoints you.
The six causes, most likely first
The test was worse, not you
Very commonPacing, motivation, a warm room, a poor warm-up, a bad night's sleep, a heavy meal, a different position. A 20-minute test is a performance, and performances vary by several percent for reasons that have nothing to do with fitness.
You tested fresh last time and tired this time
Very commonThe single most frequent cause. A test taken on the back of a hard block measures a fatigued rider. That is useful information, but it is not a fitness decline.
The protocol changed
CommonA ramp test, a 20-minute test and an 8-minute test do not produce the same number, and the same protocol on a different trainer or with a different power meter does not either. Compare like with like or do not compare at all.
You have been doing the wrong kind of work for it
CommonSix weeks of pure endurance, or of short sprint work, will not raise a threshold number and may let it drift. That is not a failure of training if endurance was the goal — but it will show up here.
You are under-fuelled or unwell
Worth checkingIllness, low iron and chronic under-eating all cut sustainable power, and all of them show here before they show anywhere else obvious.
You genuinely lost fitness
Least likely of the sixTime off, a long illness, a season of maintenance. Real, and far rarer than the five above — which is why it should be the last thing you conclude, not the first.
The test that answers the question
There is a simple way to find out whether the drop is real: look at your ride data instead of your test data.
If your best 20-minute power from actual rides over the last month is as high as it was three months ago, your threshold has not moved and the test was the problem. If those real-world efforts have fallen too, something has genuinely changed and it is worth investigating. This is one of the strongest arguments for a power curve built from every ride rather than from tests — the rides are a much larger sample, and nobody games them.
- Compare your best real-ride 20-minute power this month against three months ago.
- Check what your fatigue was doing in the three days before each test.
- Check the protocols were identical — same test, same trainer, same calibration.
- Only if all three look clean should you treat the drop as real.
If it is real, what then
- Check the boring things first. Sleep, food, illness and stress explain most genuine declines in amateur riders, and none of them are solved by training harder. Low energy availability and REDs and iron and ferritin are the two worth ruling out properly.
- Look at what you have actually been doing. If the last eight weeks were long easy rides, a flat threshold is the expected outcome, not a surprise. Add threshold work — over-unders are a good re-entry — and retest in six weeks.
- Check you are not permanently tired. A rider who never comes back to positive form will test badly forever. Recovery is training is the fix, and it is less popular than it should be.
- Accept the seasons. Nobody holds peak threshold for twelve months. A lower number in November is a plan working, not failing — see how to peak twice in one season.
A better relationship with the number
FTP is a useful anchor for setting training zones and a poor scoreboard for your worth as a cyclist. The riders who improve most steadily tend to be the ones who test occasionally, adjust their zones, and then forget about it — rather than the ones who test monthly and ride the emotional wave.
Moveee builds your power curve from every ride you upload rather than from tests alone, so you can see whether a disappointing test is contradicted by your actual riding. Have a look at your power profile — and if your best real efforts are fine, the only thing that went down was one Tuesday evening.
Sources 7
Where this article summarises a study, the study itself is linked — not a write-up of it.
- 1 Borszcz FK, Tramontin AF, Bossi AH, Carminatti LJ, Costa VP Functional threshold power in cyclists: validity of the concept and physiological responses · International Journal of Sports Medicine · 2018
- 2 Mackey J, Horner K What is known about the FTP20 test related to cycling? A scoping review · Journal of Sports Sciences · 2021
- 3 Currell K, Jeukendrup AE Validity, reliability and sensitivity of measures of sporting performance · Sports Medicine · 2008
- 4 Meeusen R, Duclos M, Foster C, et al. 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 · Medicine & Science in Sports & Exercise · 2013
- 5 Mujika I, Padilla S Detraining: loss of training-induced physiological and performance adaptations. Part I · Sports Medicine · 2000
- 6 Mountjoy M, Ackerman KE, Bailey DM, et al. 2023 International Olympic Committee's (IOC) consensus statement on Relative Energy Deficiency in Sport (REDs) · British Journal of Sports Medicine · 2023
- 7 Sim M, Garvican-Lewis LA, Cox GR, et al. Iron considerations for the athlete: a narrative review · European Journal of Applied Physiology · 2019
See this in your own numbers
Moveee builds your power curve, fitness and fatigue from every ride you upload, and writes a training plan around them. It is free, it takes one Strava connection, and it will tell you more about your riding in a week than any article can.
Start freeKeep reading
Why you stopped getting faster
For two years cycling gives you progress for free. Then one spring it stops. The six causes of a plateau in roughly the order they turn out to be the answer — and the uncomfortable sixth one, which is that you may be measuring the wrong thing.
Indoor training with heart rate only (no power meter, no problem)
No power meter, no smart trainer — just a heart-rate strap? You can still train properly. How to find your threshold heart rate with one honest 30-minute test, set five zones from it, structure a week around what HR measures well, and work around what it doesn't (lag, drift and hot rooms).
Cycling economy: why the same watts cost two riders different amounts
Two riders hold 250 watts; one is burning noticeably more oxygen to do it. Gross efficiency sits between roughly 18 and 24%, it is largely not trainable, and it decides more about a five-hour day than any interval session will.

Winter training for cyclists: how to come out of winter faster than you went in
Winter is when next season is won or lost. Here's a month-by-month indoor-first plan that keeps you fit through the dark months without burning out by February.
30/30s: the shortest intervals that still build your ceiling
Thirty seconds hard, thirty seconds easy, twenty times. Microintervals chase the same adaptation as five-minute VO2max efforts by a route most riders can actually finish — provided you get the first three reps right.
Testing your own lactate: what a meter tells you that power can't
A finger-prick and a €200 analyser can find the two thresholds your power meter only estimates. Designing a step test that gives comparable numbers, the errors that quietly ruin one, and whether you actually need this.
Sprint work for riders who are not sprinters
Most amateur plans have a hole at the top end, usually because the rider decided they are not a sprinter. Four sessions, why the rests are far longer than feels reasonable, and the one setting you must turn off to do these indoors.
Junk miles: what they are and when they aren't
Easy riding is not junk; the fittest riders do most of their hours there. Unplanned medium riding usually is. A three-question test for any ride, and the evidence behind it.
Over-unders: the interval that teaches you to ride through lactate
The session that turns "I blew on the third climb" into "I was there at the top" — and one of the most frequently botched workouts in cycling. Four versions in order of difficulty, and the four ways riders ruin them.
Durability: why your power at hour four is the number that matters
Two riders with identical FTPs can be 55 watts apart four hours into a sportive. That gap has a name — durability — and no fresh test will ever show it to you. How to measure your own, and the training that actually builds it.
Consistency beats everything: the evidence for boring training
The strongest predictor of an endurance result is how much of the planned training got done. Why gaps cost more than they look, why the spike after a gap is the enemy, and rules for boring training.
How to fix a flat in five minutes (and stop dreading it)
A roadside puncture repair, step by step, and the two things beginners skip that turn a five-minute job into twenty: finding the cause, and seating the bead before you pump it hard.