Blog · Training

What Is FTP — and How Do You Actually Find Yours?

The short version: FTP (functional threshold power) is roughly the power you can hold for about an hour — the line between sustainable and unsustainable. It sets your training zones and training load, so getting it wrong quietly throws off everything. There's no single "true" test without a lab; the honest options are a 20-minute test, a ramp test, or modelling your power curve. The best number is one you can see the working behind.

"What's your FTP?" is the one number cyclists ask each other about most. It's also one of the most misunderstood. Here's what it actually is, why it matters more than you'd think, and how to find yours without kidding yourself.

What FTP actually is

Functional threshold power is, in plain terms, the highest power you can hold for roughly an hour. It's the fence between two worlds: just below it, your effort is steady and repeatable — your body clears fatigue about as fast as it builds. Just above it, the clock starts: fatigue accumulates and you're on borrowed time.

That's why it's such a useful anchor. Nearly everything in structured training is defined relative to it — "endurance" is a fraction of FTP, "threshold" work sits right around it, "VO₂" work is above it.

Why the number matters

Your FTP sets two things that shape every ride:

Get it too high and your prescribed workouts are secretly too hard — you fail sessions, dig a hole, and wonder why you're always tired. Too low and your training is too easy to drive adaptation. The number isn't vanity; it's the ruler you measure everything else with.

The honest part: FTP is a model, not a measurement

The benchmark physiologists actually chase is the metabolic (lactate) threshold — measured in a lab, with blood samples, across several visits. Almost nobody has that. FTP is a practical field proxy for it: close enough to train by, but an estimate. So every method below is an approximation. Pick one, be consistent, and re-test the same way — a change of method can look like a change of fitness when it isn't.

The three ways riders find it

1. The 20-minute test (× 0.95)

Warm up, then ride as hard as you can hold for a full 20 minutes. Your FTP is about 95% of that 20-minute average power (the 5% accounts for the fact that 20 minutes is a bit shorter and a bit harder than an hour). Simple, and the most widely used. Downsides: it's genuinely brutal and hard to pace, and the 0.95 is a population average — punchy, sprinty riders often need a bigger discount, steady "diesel" engines a smaller one.

2. The ramp test

Ride an ever-increasing power until you can't hold the next step; FTP is estimated at roughly 75% of your best one-minute power. Short and much less painful to pace. The catch: a ramp leans heavily on anaerobic capacity, so it can read high for explosive riders — the classic reason a test says one number but the threshold workouts built on it feel impossible.

3. Model your power curve (Critical Power)

Instead of one dedicated test, fit a curve to several of your maximal efforts of different lengths — roughly 2 to 20 minutes. The Critical Power model describes your power-duration relationship as:

power = CP + W′ ÷ time

CP (Critical Power) is the asymptote — the sustainable ceiling your curve trends toward. W′ ("w-prime") is the finite battery of work you have above CP — your anaerobic reserve. The beauty of this method is that it uses efforts you already do on hard rides, so it doesn't need a dreaded test day — as long as your rides actually contain some hard efforts.

CP and FTP aren't the same number

A subtlety worth knowing: Critical Power sits a little above the true metabolic threshold for most riders, so FTP usually lands a touch below CP. A sensible, conservative conversion is FTP ≈ 0.95 × CP — erring low on purpose, because an FTP set too high is the one that costs you workouts. (That's Crux's baseline conversion — though there's one kind of evidence that beats any model, below.)

The best evidence: you already held it

All three methods above estimate what you could sustain. But if one of your rides already contains a genuinely steady 35–75-minute effort — a long climb, a solo TT, a hard tempo block — then the question is answered by demonstration, not modelling: you held that power, for that long. The only honest check needed is steadiness (a surgy group ride can post a big 40-minute average without any steady threshold demand — a variability check catches that).

That's how Crux decides your FTP when the evidence exists: a steady 35–75-minute effort is used directly, with the time you held it shown alongside; a steady 25–35-minute effort is projected conservatively to the 45-minute mark along your own power-duration curve; and only when neither exists does the Critical Power model (× 0.95) decide. Evidence beats modelling, every time it's available.

How to get a good estimate without a lab

Trust the working, not just the number

The real trap with FTP tools — especially the "it just knows" black-box kind — is a number appearing with no explanation. When it feels wrong, you can't tell why, so you either follow it off a cliff or ignore it entirely. A number you can act on is one that shows its working: the efforts it used, how it was calculated, and how confident it is. If you can't see those, treat the number with suspicion.

Frequently asked questions

What is FTP in cycling?

Functional threshold power is roughly the highest power you can hold for about an hour — the boundary between an intensity you can sustain and one that fatigues you quickly. It's used to set your training zones and to score training load.

Is FTP the same as lactate threshold?

Not exactly. The physiological benchmark is the metabolic or lactate threshold, measured in a lab. FTP is a practical field proxy for it — close enough to train by, but an estimate.

How often should I re-test?

Every four to six weeks, or after a training block, is plenty. Re-test the same way each time so a change reflects fitness, not method.

Can I find my FTP without doing a test?

Yes — if your rides already contain hard efforts of a few different lengths, the Critical Power method can model your FTP from them, no dedicated test day required.

Why did my FTP test feel wrong?

Usually it over-read for your physiology — a ramp test in particular can set FTP too high for explosive riders. If threshold sessions keep failing at the top of your zones, your FTP is probably set too high.

Sources & method

Definitions here follow standard training-with-power practice. The FTP ≈ 0.95 × CP conversion and the 2–20-minute fitting window are how Crux's on-device baseline is implemented; when a ride contains a genuinely steady 35–75-minute effort, Crux uses that demonstration directly instead of the model (steadiness-checked against the ride's own samples). The 20-minute (× 0.95) and ramp (≈ 75% of one-minute power) figures are the common field conventions. All are starting points — your own testing and how workouts actually feel always take precedence.

Let Crux build your FTP from evidence — and show its working

A genuinely steady 35–75-minute effort becomes your FTP directly, with how long you held it; otherwise Crux fits the Critical Power model to your own power curve, entirely on your phone, with the efforts, the model and its confidence all on show. It'll even answer "how long could I hold X watts?" from your own numbers, chart your FTP's journey over time, and — new — show your durability: how much of your fresh power survives once you're 1,500 kJ deep into a ride. When your recent rides don't have the efforts an estimate needs, it says so instead of guessing. Private by design — your rides never leave your phone.

See Crux