How to Do an FTP Test Right — The Most Common Mistakes
Your FTP is the most important number in your cycling training — it defines your zones, determines your TSS, calibrates your power curve. A wrong FTP leads to wrong training. And yet most amateur athletes have an FTP that's 10–15% too high.
Why? Because they did the wrong test, executed it wrong, or haven't updated the value in months. Let's look at how to do it better.
What FTP actually is
FTP (Functional Threshold Power) is the highest power you could theoretically sustain for 60 minutes. Andrew Coggan defined it in 2003 as a practical approximation of the lactate threshold — easy to measure, easy to understand.
The key word: theoretically. Nobody voluntarily does a 60-minute all-out test. So there are several approximation methods:
Method 1: 20-minute test (Coggan classic)
Probably the best-known test. After a warm-up: 20 minutes as hard as possible, evenly paced. Average power × 0.95 = FTP.
You don't need to do the math in your head — the FTP calculator takes your 20-min power and weight and gives you FTP, W/kg and your Coggan classification directly.
Pros:
- Well known, well documented
- Reproducible if you don't do it too often
- Also gives you "20-min power" directly as a training anchor
Problems:
- Pacing is everything. Out too fast — you blow up. Too conservative — value too low. Mentally demanding.
- The 0.95 factor is optimistic for the untrained. Realistic for well-trained athletes, but for beginners it often gives values that are too high — because they can't hold a true "all-out" for 20 minutes.
- Needs preparation: 15-min warm-up, 5-min all-out before the test (to "pre-fatigue"), then 10-min easy, then the 20-min test. Realistically 60 minutes total.
Skipping the 5-min all-out before the test. Coggan built it in for a reason: it provides some pre-fatigue so the 20-min value doesn't come out systematically too high. Without it you'll produce more like 95–97% of your true 20-min power in those 20 minutes — and then the × 0.95 factor becomes an overestimate.
Method 2: Ramp test (Zwift / TrainerRoad style)
Power increases in steps, usually by 20W every 60 seconds, until you can't keep up. FTP = last complete minute × 0.75.
Pros:
- Shorter: 25–30 minutes total
- No pacing needed — the device sets the pace
- Mentally easier — just hold on to the end
- Reproducible if you always use the same step length
Problems:
- The 0.75 factor varies a lot between athletes. Sprinters with a high anaerobic share often get FTP values that are too high; long-distance-strong athletes get values too low.
- Training state matters: for athletes in a "diesel" phase (lots of Z2 training), the ramp gives lower values than the 20-min test.
Method 3: Critical Power (CP) from multiple best efforts
More mathematical than the other two. You need 3–5 maximal efforts of different durations (e.g. 3, 5, 8, 12, 20 minutes) — they can be spread across several weeks. From these you compute, via linear regression:
CP is the asymptotic "critical power" — very similar to FTP, but not interchangeable: CP usually sits slightly above FTP and is better calibrated. W' (say "W prime") is your anaerobic work reserve in joules.
Pros:
- Most accurate method — uses multiple data points
- Predicts not just FTP but performance across different durations
- Works with real training data — no dedicated test needed if you ride hard regularly
Problems:
- Needs genuine all-out efforts — not every hard interval qualifies
- Plausibility check matters: if all efforts are too similar (e.g. all at 5–10 min), the regression becomes unstable
- Yama implements this automatically — many other apps don't
Which method for whom?
| Profile | Recommendation |
|---|---|
| Beginner, indoor setup | Ramp test (Zwift) |
| Disciplined pacer, experienced | 20-min test |
| Regular hard training, power meter | CP from training data |
| Race preparation | 20-min on the road |
| Climber / TT specialist | CP model — accounts for your distance profile |
The most common mistakes
1. FTP not updated for too long
FTP changes by ±10–20% over the year. If you entered 250W in spring and are still using it in late summer, you get completely distorted TSS values and zones. Recommendation: retest at least every 8–12 weeks.
2. Wrong FTP entered in your profile
Sounds trivial: a user recently tested 268W on his device. But the app still had the old FTP of 249W — because it was never updated. Result: all TSS values about 15% too high.
3. Testing without calibration
If your power meter (or smart trainer) isn't calibrated, all values are garbage. Before every test: do a zero offset. For pedals with torque calibration, that too.
4. Testing under unfamiliar conditions
You've trained indoors the last 4 weeks and suddenly do an FTP test outdoors? The values will be completely different. Always test in the environment where you mainly train — comparable values over time matter more than the absolute number.
5. Not mentally ready
An FTP test is 95% mental. You have to be ready to push yourself into a painful zone for 20 minutes. Pick a day when you're rested and motivated, not tired after a long workday.
Yama calculates your CP automatically
From your Strava training data, Yama extracts your maximal efforts over 3–20 minutes and estimates your critical power via linear regression — without you doing a dedicated test. Plus: a power curve with season comparison for trend analysis.
Try YamaConclusion
There's no perfect FTP test method. Each has weaknesses. More important than the method is consistency: use the same method regularly, compare values over time, and accept that FTP is an estimate — not a hard physiological constant.
And if you have a feeling your FTP is too high: it probably is. Do the test again, this time with honest full-gas effort — and accept the result, even if it comes out lower than you'd like.