The 20 minute FTP test is the most popular way to measure your threshold without a lab. It is also the one most amateurs get wrong, walking away with a number that does not reflect their real fitness — and then building weeks of training on top of that error.

What FTP is, and why a clean test matters

FTP is the power you could in theory hold for roughly an hour at a near-steady intensity. It is the reference point your zones, your threshold sessions and your race pacing are all built around. If you want to understand exactly how it is defined and used, start with this complete guide to cycling FTP.

The logic is simple: a plan rests on your zones, and your zones rest on your FTP. A number that is off by 10% is not a rounding error: it means systematically training too hard (and digging a hole) or too easy (and never adapting). That is why how you run the test matters as much as the test itself.

Before the test: the prerequisites that decide the outcome

A good 20 minute FTP test starts long before you clip in. Skip these and you are measuring your fatigue, not your fitness:

  • Show up rested. No hard sessions in the preceding 48 hours — ideally a day or two of easy riding or rest. Testing at the end of a heavy block hands you a number lower than the truth.
  • Show up fuelled and hydrated. Eat normally in the hours beforehand, including carbohydrate. A fasted test does not measure threshold, it measures hunger.
  • Calibrate the meter. Run a zero-offset on your power meter after a few minutes of riding, once it is at working temperature. An uncalibrated reading makes everything else meaningless.
  • Pick the right course. A steady climb or an indoor trainer is ideal: no junctions, no descents, nothing to think about except holding a smooth effort.

The warm-up: never skip it

A 20-minute, near-maximal effort needs a body that is already switched on. A sloppy warm-up means starting cold and falling apart halfway through. A proven protocol takes about 20–25 minutes:

  • 10 minutes very easy, building gently.
  • 3 short openers of around 1 minute each at a high intensity (above threshold), with 1 minute easy between them, to wake the system up.
  • 5 minutes easy spinning to clear out before you begin.

Many protocols also add a single 5-minute all-out effort before the real test, followed by a generous recovery, to take the edge off legs that would otherwise start far too hard. If you are new to testing, you can leave it out — what matters is arriving at the 20 minutes warm but not fatigued.

The 20 minutes: how to pace the effort

This is where the test is won or lost. The target is the highest average power you can hold steady for the full 20 minutes — steady being the key word.

Mistake number one is going out too hard. Fresh legs push you to overcook the first three minutes; then you pay a heavy bill over the last ten and your average collapses. The result is a brutal effort that underestimates your threshold.

Set off as if you had to hold the pace for half an hour, not for five minutes. If you can lift in the final minutes, you paced it well.

A strategy that works: for the first 1–2 minutes, deliberately hold back, slightly under what you feel you could do. Settle on a number you believe is sustainable and protect it. In the final third, if there is anything left, lift it gradually. Watch your average power or perceived effort, not the jumpy instant reading.

From the number to your FTP: the ~5% subtraction

The test lasts 20 minutes, but FTP is defined over roughly an hour. That is why you do not use the raw 20-minute average directly: you correct it downward, because over a real 60 minutes the sustainable power is lower.

The classic rule is simple: FTP ≈ 20-minute average power × 0.95. In other words, subtract about 5%. If you averaged 280 W, your estimated FTP is around 266 W. It is an estimate, not an absolute truth — but a solid, repeatable base from which to set your zones.

The alternative: the ramp test

If the idea of managing 20 minutes flat out is daunting, or if you are new to pacing your own effort, the ramp test is a strong alternative. It is shorter and self-regulating: you start at a low power and step it up at regular intervals (every minute, for example) until exhaustion — that is, until you can no longer hold the required pace.

FTP is then estimated as a percentage of the power in the last completed step (75% is a commonly used figure). The upside: far less dependent on your pacing and harder to get wrong. The downside: it tends to favour riders with a strong kick on short efforts. Good practice is to pick one protocol and repeat it identically every time: consistency over time matters more than which method you choose.

The mistakes that ruin an FTP test

  • Going out too hard. By far the most common. Underdo the opening minutes and finish strong.
  • Testing fatigued. You measure tiredness, not threshold. Rest first.
  • Testing too rarely — or too often. Once a year tells you nothing about progress; every two weeks is just needless stress. Every 6–8 weeks, under similar conditions, is a good rhythm.
  • Changing the variables every time. Different course, wind, heat, an uncalibrated meter — you are comparing numbers that are not comparable. Standardise the conditions.

Where Per Pro fits in

An FTP number on its own does not make you faster: what counts is what you build on top of it, and how well it fits your real life. Per Pro takes the training methodology of genuine World Tour coaches and rewrites it around your actual week — using your threshold to set sensible zones and sessions, and reshaping the plan when a session gets missed or your form shifts. The same approach that brings professionals to the start line, redrawn around you. You can see how it works at per-pro.com.

Key takeaways

  • Arrive rested, fuelled and with a calibrated meter: the prerequisites decide the result.
  • Warm up for 20–25 minutes with a few openers — a long effort needs a body already switched on.
  • Keep the 20 minutes as steady as possible; finish strong rather than starting flat out.
  • FTP ≈ your 20-minute average minus about 5% (× 0.95).
  • The ramp test is a good alternative if you cannot yet pace your own effort.
  • Test every 6–8 weeks, always under the same conditions, so you compare real numbers.