Vanhatalo's 3-minute all-out test: 3 minutes flat out from the first second, against a fixed resistance. Power drops then levels off; your last 30 seconds give your CP, and the work done above it, your W′. 31 minutes, for experienced riders.
| Target | Zone | Time | FTP 200 W | FTP 250 W | FTP 300 W |
|---|---|---|---|---|---|
| 125–160% | 6Anaerobic | 3 min | 250–320 W | 313–400 W | 375–480 W |
Put this session on your bike computer or watch
Do the test from your bike computer, guided block by block. Then enter your new FTP and every session follows.
This test doesn't make you fitter: it measures both parameters of the critical power model in a single effort. Your CP is the highest power you can hold in a steady state; your W′, the amount of work you can do above it.
The idea is clever. Going genuinely all-out empties your W′ in a little over two minutes. Once the reserve is gone, all that's left is what your aerobic engine can sustain: power stops falling and settles into a plateau, which corresponds to your CP. Everything you produced above that plateau during the 3 minutes is your W′. So the test only works if you pace nothing at all: the slightest saving early on leaves reserve for the end, and the plateau comes out too high.
Experienced riders, used to maximal efforts, on a smart trainer that can hold a fixed resistance or on a laboratory ergometer. It's the one to pick if a 70-minute test sounds like too much: it fits into 31 minutes. If you'd rather ride paced efforts, closer to racing, critical power 12 + 3 min gives you the same parameters with no special trainer setup.
At the start of a training block, or every 4 to 6 weeks: a common guideline. Do it fresh, after one or two easy days, and fuel as you would for a hard session. The first attempt often doubles as a familiarisation run: holding back at the start without meaning to is common first time round. After that, keep exactly the same resistance setting and gear from one test to the next, or your plateaus won't compare.
5′ at 45–55% FTP + 5′ at 56–65% FTP + 5′ at 66–75% FTP
15 minutes building from 45 to 75% of FTP, in 5-minute steps. Use the warm-up to check your setup: ERG mode off, fixed resistance (constant level or gradient), and a gear you won't touch again until the test is over.
3′ at 40–55% FTP
3 easy minutes at 40–55% of FTP: the baseline pedalling from the published protocol. Over the last few seconds, wind your cadence right up so you hit the start already rolling.
3′ at 125–160% FTP
3 minutes all-out, from the first pedal stroke to the last second, without looking at the clock. Power starts very high, falls away over the first two minutes, then levels off. Don't chase a target: the 125–160% FTP range, or 313 to 400 W for an FTP of 250 W, is only a guide for the 3-minute average.
10′ at 40–55% FTP
10 very easy minutes at 40–55% of FTP, without stopping dead. Then the maths, from your ride file: CP = average power over the last 30 seconds; W′ = (3-minute average − CP) × 180, in joules. 380 W over the 3 minutes and 280 W over the last 30 seconds give a CP of 280 W and a W′ of 18 kJ.
Maximal from the first minute, with no comfortable phase at all. Then an odd sensation: you want to give more, and you simply can't, however hard you grit your teeth. The last minute lasts a good deal longer than sixty seconds. Guide: 10 out of 10. If you managed to wind it up at the end, you paced it, and your CP is overestimated.
The mistake that ruins the test. Keep a little in reserve early on and you don't empty your W′, so it's still contributing over the last 30 seconds: your CP comes out too high and your W′ too low. Go off as if the test lasted 30 seconds.
In ERG the trainer sets the power: it can't start high or fall away, so there's no plateau to read. Same problem with a fixed cadence: Karsten and colleagues found the test invalid on an isokinetic ergometer.
Too light and you spin out at the start, with the end of the test limited by your cadence. Too heavy and you're grinding, with your cadence collapsing before the plateau. Aim for a setting where, by the end, you're turning the pedals at roughly your usual cadence.
Knowing there are 90 seconds left nudges you into pacing, even without meaning to. Hide the timer on your head unit, or have someone call out only the finish.
The test grew out of work by Burnley, Doust and Vanhatalo (Med Sci Sports Exerc 2006). Vanhatalo, Doust and Burnley (Med Sci Sports Exerc 2007) validated it in 10 active subjects: end-test power from a 3-minute all-out test against a fixed resistance (287 ± 55 W) did not differ from CP measured with five constant-power tests (287 ± 56 W), with an estimation error of about 6 W. Karsten and colleagues (2013) found it invalid on an ergometer in isokinetic mode, where cadence is fixed. The 3 minutes of baseline pedalling come from the published protocol; the % FTP range, the warm-up and the cool-down are Wattson choices. Bottom line: good laboratory evidence in linear mode, less certain on a smart trainer.
In your ride file, take the average power over the last 30 seconds: that's your CP, known as end-test power. For W′, subtract CP from your 3-minute average, then multiply by 180. A 380 W average and 280 W at the end give a CP of 280 W and a W′ of 18 kJ.
Turn ERG mode off and choose a fixed resistance: a constant level or a constant gradient. Pick a gear where, by the end of the test, you're turning the pedals at roughly your usual cadence, then leave it alone. Keep exactly the same setup from one test to the next: it's the only way to compare your results.
Because you're emptying your W′, your reserve above CP. Going all-out, it's used up in a little over two minutes. What's left is the power your aerobic engine can hold steadily: the curve stops falling and flattens into a plateau. That plateau, measured over the last 30 seconds, is your CP.
The 3-minute test is shorter, 31 minutes in all, and done in one effort, but it depends heavily on the trainer setup and on a genuinely all-out start. The 12 + 3 minute version takes 70 minutes, with paced efforts and no special setup. Choose one, and stick with it to compare your tests.
It's recommended. Going flat out with nothing in hand runs against every instinct of a rider used to pacing: the first go often comes out with a start that's too cautious. Treat your first test as a rehearsal, and keep the result of the second, done in the same conditions.
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 this session