FTP stands for functional threshold power: roughly the highest power you could hold for about an hour before the effort falls apart. You almost never ride a flat hour at threshold to find it. Instead you run a shorter test and apply a correction, because a shorter maximal effort is easier to pace and less soul-destroying than sixty minutes on the rivet.
Get the number right and every ERG target the app sets lands where it should. Set it too high and your “sweet spot” sessions turn into threshold grinders you cannot finish; too low and you never actually stress the system. A test two or three times a year keeps it honest.
Before you test
A test is only as good as the ride around it. Two things matter more than the protocol you pick.
- Be reasonably fresh. Do the test on legs that have had an easy day or a rest day, not the morning after a hard session. A fatigued test lowballs your FTP and drags every future workout down with it.
- Warm up properly. All three protocols below include a warm-up, but the short, sharp openers matter — a few 20-second bursts wake up the top end so your first minute of the test is not spent getting going.
Run the test in resistance mode or free-ride, not ERG. You are trying to find your ceiling, and ERG would cap you at whatever target it was holding. The test workouts in CrankPilot open the test blocks as free rides for exactly this reason: you push, the trainer measures, nothing holds you back.
The ramp test
The ramp is the shortest and least painful, which is why most people who test regularly settle on it. Power starts easy and steps up by a fixed amount every minute until you cannot hold the next step and stop.
CrankPilot’s ramp is 30 minutes: a 5-minute warm-up, then 25 one-minute steps climbing from 46% to 166% of your current FTP estimate. You ride until you blow, usually somewhere in the last third. Your FTP is estimated at 75% of the best one-minute power you managed before quitting.
- Cost: low. One truly hard minute at the end rather than a sustained ordeal.
- Best for: frequent testing, riders new to testing, and anyone who paces sustained efforts badly.
- Watch out for: the ramp leans on your anaerobic top end, so very punchy riders can read a little high and diesel-engine time-triallists a little low.
The 20-minute test
The classic. You ride a single 20-minute maximal effort, as steady and as hard as you can sustain, and take 95% of your average power as your FTP. The 5% haircut accounts for the fact that 20 minutes is shorter than an hour, so you could hold a bit more for the shorter time.
CrankPilot’s version runs 55 minutes end to end: a warm-up with openers, a 9-minute settle, then the 20-minute effort. Pacing is the whole skill. Go out at a number you can hold to the last minute; if you are dying at minute 12 you started too hard and the result is junk. A good 20-minute test has your last two minutes as your hardest, not a survival crawl.
- Cost: high. Twenty minutes of sustained discomfort with nowhere to hide.
- Best for: riders who pace well and want a number rooted in a real threshold-length effort.
- Watch out for: pacing errors ruin it, and the mental load means many people quietly underperform the second time they face it.
The 8-minute test
A middle path. You do two 8-minute maximal efforts with recovery between them, and take 90% of the better effort’s average power. The shorter blocks are easier to commit to than a single 20, and having two goes means a bad pace on the first does not sink the whole test.
CrankPilot’s 8-minute protocol is 51 minutes: warm-up, then two 8-minute all-out efforts separated by recovery. Ride the first hard but repeatable, then empty the tank on the second.
- Cost: medium. Two sharp efforts rather than one long one.
- Best for: riders who want a sustained-effort number without the full 20-minute commitment.
- Watch out for: eight minutes leans slightly more anaerobic than twenty, so like the ramp it can read a touch high for pure endurance types.
Turning the result into your FTP
None of these tests set your FTP automatically. That is deliberate — you should look at the number, sanity-check it against how the test actually felt, and enter it yourself. After the test, read your result:
- Ramp: your best one-minute power, times 0.75.
- 20-minute: your 20-minute average, times 0.95.
- 8-minute: the better effort’s average, times 0.90.
Round to the nearest 5 watts and set it in the app’s profile. From that moment every workout target, every zone boundary, and the intensity factor on every ride recalculates against the new number.
How often, and reading the trend
Test every 6 to 8 weeks if you are training with intent, or at the start and end of a training block. More often than that and you are testing noise; the number does not move week to week.
One caution: do not chase a personal best on every test. A flat or slightly lower FTP after a hard block can mean you are fatigued, not weaker — retest after a couple of easy days before you conclude anything. The trend across a season tells the real story, and it is the trend, not any single test, that proves the training is working.