FTP is the number every power-based plan is built on: the effort you could hold for about an hour. Nobody tests it by riding all-out for an hour, so it gets estimated. These are the three ways that estimate happens, and they do not always agree.
The field test most cyclists have done. Warm up properly, then ride 20 minutes as hard as you can hold without fading, and take the average power. Your FTP is 95% of that number.
The 5% haircut exists because 20 minutes is shorter than an hour, so you can push above your true threshold for it. Ride it wrong — start too hard and fade — and the average still looks fine while the number is wrong. Even pacing is the whole test.
Power goes up in steps until you cannot hold the next one. Your FTP is 75% of the average power in the final full minute you completed.
It is shorter, it hurts for less time, and it is far harder to pace wrong — there is no pacing to get wrong. The trade-off: it leans on your anaerobic capacity, so a sprinter tends to come out flattered and a diesel undersold.
Straight to the zones. Use this if you have a recent test, a number from your head unit, or a figure your training app already keeps.
One caveat worth stating: an FTP older than about six weeks of consistent training is a guess about a rider you no longer are.
The two tests do not have to agree, and usually do not. A 20-minute test and a ramp test on the same rider in the same week can land 10 to 20 watts apart, because they load different systems. Neither is the true number — an hour-long time trial is the only thing that would give you that, and almost nobody does one. Pick the test you will repeat, and compare yourself only against your own previous result on the same test.
The same rider, both tests, same week. This is what "the protocols disagree" looks like in watts.
At 72 kg that is 3.5 W/kg — the median for a male cyclist in his twenties in our data, and a shade above the median for his thirties.
Three watts apart from Example A, which is closer than most riders get. A 15-watt gap between the two protocols is normal and does not mean either test failed. What it means is that you should not mix them: switch protocol and you have restarted your baseline.
Zone 2 is the widest band on the table — from 55% to 75% of FTP, a 50-watt spread for most riders — and that width is exactly why people get it wrong. The top of Zone 2 and the bottom of Zone 3 feel almost identical for the first hour. They stop feeling identical at hour three.
Seiler's review of how competitive endurance athletes actually train found they converge on roughly 80% of sessions at low intensity and about 20% genuinely hard. The middle is where the least useful work lives: hard enough to need recovery, not hard enough to force adaptation.
Power makes this easier to police than heart rate does. Heart rate lags the effort by a minute or two and drifts upward on long rides; watts do neither. If your plan says Zone 2 and your average is sitting at 74% of FTP, you are riding the top of the band on purpose or by accident — and over a four-hour ride, that difference is what decides whether tomorrow's session happens.
One thing power cannot do: it does not know how you feel. A day when 200 watts costs what 240 usually costs is real information, and the number on the screen will not show it. That is what heart rate and RPE are for, and why a good plan reads all three.
"Is my FTP good?" is the most common question that follows a test, and the honest answer is that it depends on what you weigh and how old you are. We can do better than a shrug, because we can see the numbers.
We took the most recent FTP of every cyclist training with us who has one, kept a single figure per rider, and split it by age and sex.
| Age | 25% | 50% | 75% | 90% |
|---|---|---|---|---|
| Men 18–29 | 3.0 | 3.5 | 4.3 | 5.1 |
| Men 30–39 | 2.7 | 3.3 | 3.8 | 4.4 |
| Men 40–49 | 2.7 | 3.3 | 3.7 | 4.2 |
| Men 50–59 | 2.4 | 2.9 | 3.4 | 3.9 |
| Men 60+ | 2.1 | 2.7 | 3.2 | 3.5 |
| Women 18–29 | 2.4 | 3.3 | 4.1 | 4.4 |
| Women 30–39 | 2.0 | 2.8 | 3.4 | 4.2 |
| Women 40–49 | 2.3 | 3.0 | 3.5 | 4.0 |
| Women 50–59 | 1.1 | 2.0 | 2.3 | 3.4 |
| Age | 25% | 50% | 75% | 90% |
|---|---|---|---|---|
| Men 18–29 | 210 W | 260 W | 300 W | 332 W |
| Men 30–39 | 214 W | 254 W | 285 W | 315 W |
| Men 40–49 | 220 W | 250 W | 284 W | 310 W |
| Men 50–59 | 206 W | 237 W | 266 W | 295 W |
| Men 60+ | 193 W | 210 W | 240 W | 261 W |
| Women 18–29 | 181 W | 215 W | 246 W | 276 W |
| Women 30–39 | 147 W | 183 W | 213 W | 273 W |
| Women 40–49 | 159 W | 197 W | 240 W | 286 W |
| Women 50–59 | 134 W | 150 W | 191 W | 212 W |
Raw watts barely move between 30 and 50. A man in his forties has a median FTP of 250 watts against 254 in his thirties: four watts across a decade. The decline people expect does not show up until the fifties, and even then it is 13 watts.
W/kg is flatter still, and that is the more useful reading. From 30 to 50, the median sits at 3.3 both times. What separates riders at any age is not the birthday — it is the spread within each decade, which runs from 2.7 to 4.2 in the same ten years.
The gap between a rider getting started and a top rider is bigger than the gap between any two age groups. Which is another way of saying that training explains more of your number than age does.
Methodology: our cyclists with a recorded FTP and a usable weight, measured on 4 September 2026. One figure per rider — the most recent — so that riders who test often do not weight the result. Values outside 50–600 W were discarded as data-entry errors, a known issue where riders type their W/kg into the watts field. Two biases worth stating: these are cyclists who train with a structured app, not the general population, and the women's groups are small enough that their percentiles should be read as indicative.
Worth saying plainly, because a number you trust too much is worse than no number.
It does not test you. It applies a convention to a figure you provide. Ride the test badly and the arithmetic is still perfect while the answer is wrong.
It does not know which protocol you used last time. The 20-minute and ramp tests can land 10 to 20 watts apart on the same rider. Comparing across protocols tells you nothing about your fitness.
It does not account for the day. Heat, altitude, poor sleep, a hard week: all of them move what you can produce, sometimes by 5%. A test on a bad day sets your zones too low for weeks.
It does not check your power meter. A meter that reads 3% high gives you zones that are 3% high, and every session lands slightly harder than the plan intended. Calibrate before you test.
It does not replace the hour. FTP is defined as roughly one hour of maximal effort, and both tests are estimates of a thing almost nobody measures directly. Treat the number as a working baseline, not a physiological truth.
The claims above that come from published work, with the work named.
The questions that come up most often after a test.
It depends on your weight and your age, which is why the table above splits both. As a rough anchor: the median male cyclist in our data sits at 3.3 W/kg and the median woman in her thirties at 2.8. But the spread inside any single decade is wider than the gap between decades, so your training explains more of your number than your birthday does.
The one you will actually repeat. The 20-minute test is more faithful to what FTP means but punishes bad pacing; the ramp test is shorter and harder to get wrong, though it flatters riders with a big anaerobic engine. Pick one and stay on it — switching protocols resets your baseline.
Because they load different systems and each applies its own correction factor. A 10 to 20 watt gap between protocols on the same rider in the same week is normal. It is not a sign that either test failed.
Every 6 to 8 weeks of consistent training, or whenever sessions start feeling systematically easy or impossible. Testing more often than that mostly measures your freshness on the day.
For this calculator, yes — all three methods start from watts. If you train by heart rate instead, our heart rate zones calculator builds the same five-zone structure from your maximum or your threshold.
Because 20 minutes is shorter than the hour FTP is defined by, so you can hold more power for it than you could for sixty minutes. The 5% correction is the convention from Allen and Coggan, and it is an average across many riders — your personal factor might be 93% or 97%.
Not necessarily. A lower number can come from a bad test day, a hard training block you have not recovered from, heat, or a power meter that has drifted. Look at the trend across several tests rather than at one result, and check whether the conditions were comparable.
