FTP and power tests for cycling
Every power-based session you ride is written as a % of your FTP, the power you can hold for about an hour: a test is how you measure it, or more accurately estimate it. The 6 protocols on this page run from the ramp test, half an hour with nothing to pace, to the 20-minute FTP test, the reference field test. Each applies its own factor: 75% of your best minute on the ramp, 90% of your better 8-minute average, 95% of your 20-minute average.
Other tests measure something else: your critical power and W′ reserve, a fuller model than FTP, or your sprint power over 30 seconds. Each page gives the exact protocol, the conditions, the calculation step by step and what the sources say, with how strong the evidence is.
·Testing · every zone
How to choose
Your first test
the ramp test, 34 minutes at most, with the power set by your smart trainer: you just hang on.
The reference test
the 20-minute FTP test, longer (75 minutes) and harder to pace, but closer to riding at threshold.
Twenty minutes flat out sounds grim
the 8-minute FTP test, two shorter efforts that are easier to get your head round.
You want more than an FTP
critical power 12 + 3 min out on the road, or Vanhatalo's 3-minute all-out test, for your critical power and W′ reserve.
You're working on your sprint
the 30-second sprint test, based on the Wingate, for your peak power and how well you resist fatigue. It doesn't give you an FTP.
The sessions
30-second sprint test (Wingate-style)
Thirty seconds flat out against a heavy resistance: your peak power, your mean power and how much you lose along the way. A profile test, not an FTP test.
Critical power: 3-minute all-out test (Vanhatalo)
Three minutes flat out from the first pedal stroke, never pacing: power drops, levels off, and that plateau gives you your critical power.
Ramp test (FTP)
A harder step every minute until you can't hold it: your FTP is 75% of your best minute. The easiest FTP test to pace, in about half an hour.
8-minute FTP test (Carmichael, 2 × 8 min)
Two all-out 8-minute time trials, 10 easy minutes apart: your FTP is 90% of the better of the two averages. 51 minutes in all.
Critical power: 12 min + 3 min
Two all-out efforts, 12 then 3 minutes, with 30 easy minutes between: two averages are enough to plot your critical power and your W′ reserve.
20-minute FTP test (Allen & Coggan)
Warm-up, 5 minutes all-out, then 20 minutes as hard as you can: your FTP is 95% of the average. The reference field test, in 75 minutes.
Frequently asked questions
Which FTP test should I do?
For a first test, the ramp test: short, with the power set for you and nothing to pace. If you can pace a long effort, the 20-minute test is the reference, closer to riding at threshold. The 8-minute test sits between the two. All three give an estimate; what matters most is sticking with the same one each time.
Why do two FTP tests give different results?
Because each measures a different effort, then brings it back to an hour with an average factor: 75% of your best minute on the ramp, 90% over 8 minutes, 95% over 20. Depending on whether you're punchy or more of a diesel, an average factor fits you more or less well. Two FTPs from two different tests don't compare directly.
How often should I retest my FTP?
Every 4 to 6 weeks, or at the start of a training block: a common guideline, not a rule drawn from studies. Do it fresh, in the same conditions each time. Once your FTP is updated, every session set as a % of it recalculates.
FTP or critical power: what's the difference?
Both describe the power you can sustain for a long time, but not in the same way. FTP is an estimate of your one-hour power. Critical power comes from a two-parameter model, with your W′ reserve for efforts above it. They are close without being identical: don't swap one in for the other without thinking it through.
Other types of cycling sessions
Take it up a level with Wattson
An intuitive workout editor to build cycling and running sessions in no time, then sync them to your watch or bike computer: Wahoo, Garmin, Suunto or Coros.
Try a session