Critical power (CP) and functional threshold power (FTP) are two attempts to name the same boundary — the hardest effort you can hold without your body falling behind. They track each other closely across a group of riders, but for one individual they can differ by roughly twenty to thirty watts in either direction, which is why the researchers who compared them directly concluded the two should not be used interchangeably.
What FTP actually measures
FTP is the power you could hold in a quasi-steady state for about an hour. In practice almost nobody tests it that way, because riding flat out for an hour is miserable and hard to pace. The common shortcut is a twenty-minute maximal effort, with FTP taken as roughly 95% of that power.
That shortcut is the whole reason FTP became the industry default. One hard effort, one number, one set of zones. It is not that FTP is physiologically superior — it is that it is cheap to measure and easy to explain, and a number a rider will actually go and test beats a better number they never test.
What critical power measures instead
Critical power comes at the same question from the other side. Plot your best power against how long you could hold it, and the curve flattens toward an asymptote. That asymptote is CP: the highest output at which your physiology can still find a steady state.
Below CP, muscle metabolites and blood lactate settle at an elevated but stable level — you are working hard, but you are not progressively digging a hole. Above CP, nothing settles. Oxygen uptake drifts upward toward maximum, efficiency leaks away, and exhaustion arrives on a fairly predictable schedule.
That schedule has a name: W′ ("W prime"), the finite amount of work you can do above CP before you run out. It is measured in kilojoules rather than watts, and it behaves like a battery. In one study of trained and untrained male cyclists, W′ averaged about 12.7 kJ in the trained group and 9.6 kJ in the untrained one, with critical power around 301 W and 242 W respectively.
The battery recharges when you drop back below CP — but not on a flat rate. In that same study, riders reconstituted noticeably less W′ after the second recovery period than after the first. Anyone who has gone deep three times in a hilly group ride already knows this in their legs: the third time back is never like the first.
What happens when you measure both on the same rider
This is where the comparison stops being theoretical. Seventeen moderately trained cyclists and triathletes were tested for both, with CP taken from three maximal self-paced efforts of twelve, seven and three minutes, and FTP taken as 95% of twenty-minute power.
The averages were almost the same: CP about 256 W, FTP about 249 W, with the two correlating at r = 0.969. Read only that far and you would conclude the argument does not matter.
But the agreement between methods for an individual rider spanned roughly −19 to +33 W. In other words, the group mean agreed while single riders did not, and the authors' own conclusion was that the two values "generally should not be used interchangeably". If your training zones are built on one number and you swap in the other, some riders will not notice and some will find every zone shifted by more than a whole zone's width.
What each test actually costs you
This is the part usually skipped in the "CP is better" argument. The textbook CP protocol is three to five separate maximal tests on different days, each paced to bring you to exhaustion somewhere between roughly two and fifteen minutes. Practical field versions compress this — the study above used three efforts of twelve, seven and three minutes in one visit with thirty minutes of rest between them — but that is still three maximal efforts in an afternoon.
An FTP test is one hard twenty-minute effort. So the honest framing is not "CP is more accurate, therefore use CP". It is: CP asks for several times the testing effort, and buys you a model that also describes what happens above threshold. Whether that trade is worth it depends entirely on whether you race there.
Where both numbers run out
The critical power model is built from efforts lasting roughly two to fifteen minutes, which is exactly why it is sharp for short, hard racing and why you should not read too much into it for a five-hour day. CP is a boundary you can theoretically hold in a steady state — it is not a promise that you can hold it all afternoon.
FTP has the mirror-image problem: anchored near the one-hour mark, it tends to flatter you on very short efforts and mislead you on very long ones.
And neither number answers the question that decides most long events, which is how much of your threshold is still there in hour four. That is durability, and it is invisible to any test done on fresh legs.
So which one should guide your training?
For most riders, the honest answer is disappointing and useful: consistency of method matters more than choice of method. A number you test the same way, on the same course or trainer, every six to eight weeks, will guide your training better than a theoretically superior number measured once and never repeated.
Beyond that, a rough division: if your riding is long and steady — gran fondos, endurance events, hard group rides — FTP is a perfectly adequate anchor and the extra testing buys you little. If you race criteriums, do cyclocross, or spend your season making and closing repeated hard efforts, the CP and W′ pair genuinely describes something FTP cannot, because it puts a number on the work available above threshold and on how quickly it comes back.
What you should not do is quietly switch between them and compare the results as if they were the same measurement. They are two different estimates of the same boundary, and the research says the difference on any given rider is large enough to matter.
Your numbers, explained the moment you ask.
Wherever your threshold number comes from, the thing that makes it useful is knowing what it is and when it changed. Peakfy keeps one explainable source of truth for your zones and thresholds: the engine computes them from your actual rides, and when a number moves, the coach tells you which rides moved it and what that changes about tomorrow — in plain language, not as a silent update to your zones overnight.
That also means you can argue with it. Asking "why is my threshold lower than last month?" gets you the reasoning behind the number, not just the number.
Questions riders ask
Keep reading
- What is FTP in cycling, and how do you test it?
- Durability in cycling: why your FTP does not predict hour five