Every number in your training — every zone, every TSS score, every line on your Performance Management Chart — flows from a single value: your Functional Threshold Power. If that number is wrong, everything downstream is wrong. Your zones are off. Your training stress calculations drift. Your coach, human or AI, is setting targets based on a fiction. And the frustrating part is that most riders are walking around with an FTP that's wrong by 10-20 watts in either direction, and they don't know it. Understanding how to improve your cycling FTP starts with a question most riders skip: do you actually know what your FTP is?
What is Functional Threshold Power?
FTP is the highest average power you can sustain for approximately one hour. That "approximately" matters. FTP isn't literally your best 60-minute power — it's a physiological marker representing the boundary between sustainable and unsustainable effort. Below it, your body can clear lactate roughly as fast as it produces it. Above it, lactate accumulates faster than you can process it, and the clock starts ticking on how long you can hold on.
The concept was developed by Dr Andrew Coggan and Hunter Allen, and it has become the foundation of modern cycling training. Your FTP is the anchor point for your power zones — every zone is expressed as a percentage of it. It's also the basis for Training Stress Score (TSS), which feeds directly into CTL, ATL, and TSB — the three metrics that tell you whether you're building fitness, accumulating fatigue, or ready to perform.
When your FTP is accurate, all of these metrics work. When it's not, none of them do.
Why your estimated FTP is probably wrong
Here's the awkward truth: many riders have never formally tested their FTP. Instead, they rely on an estimate. Garmin, Strava, Veloi — most platforms that work with power data will estimate your FTP by looking at your best outputs across various durations and applying a mathematical model. It's clever. It's convenient. And it's a useful starting point. But it's almost certainly not accurate enough to build your entire training structure on.
The problem isn't the maths. The models are sound. The problem is the data feeding them. These estimates work best when you've recently done a sustained, near-maximal effort of the right duration — but most riders haven't. Your hardest 20-minute effort might have been months ago, or it might have been 12 minutes at 80% as part of an interval session, or a group ride where you were drafting half the time. The model is extrapolating from incomplete data, and the less representative your efforts are, the wider the margin of error.
Even when you do have a genuine hard effort on file, conditions shape the result. A tailwind, competitive adrenaline, perfect sleep, terrain that suited a sustained push — change any of those and the number shifts. Your estimated FTP is a product of your best day, your best conditions, and your best motivation — not a reliable reflection of what you can actually sustain in a controlled setting.
These estimates are genuinely useful — as a starting point and as a way to detect drift over time. If Garmin says your estimated FTP has jumped 15 watts in a month, that's a meaningful signal worth investigating. But treating any estimate as a precise number and building your entire training structure on it is like setting your car's speedometer from a rough guess and then driving based on what it says. Close enough for general awareness. Not close enough for precision.
How to test your FTP accurately
If you want a number you can trust, you need to test. The good news is that you don't need a sports science lab. The bad news is that every testing protocol involves suffering. The question is how much, and what kind.
The 60-minute test
This is the theoretical gold standard: ride as hard as you can sustain for one full hour, and your average power is your FTP. Simple. Also brutal. An hour of maximal sustained effort is as much a psychological test as a physical one. Pacing is treacherous — start too hard and you blow up at 35 minutes; start too conservatively and you leave watts on the table. Very few riders outside of professional teams do this test, and even fewer do it well. It exists as a reference point, not a practical recommendation.
The 20-minute test
Developed by Hunter Allen and Dr Andrew Coggan in Training and Racing with a Power Meter, this is the most widely used protocol, and for good reason. After a thorough warm-up that includes a short hard effort to open the legs, you ride as hard as you can sustain for 20 minutes. Multiply your average power by 0.95, and that's your estimated FTP. The correction factor accounts for the fact that most riders can hold slightly more power for 20 minutes than they can for a full hour.
Pros: Well-validated across the research, manageable duration, plenty of guidance available on execution. Twenty minutes is long enough to require genuine pacing but short enough that you can hold yourself together mentally.
Cons: The 0.95 correction is a population average. Allen and Coggan's own data showed individual variation ranging from around 0.90 to 0.97 — riders with a strong anaerobic engine can hold a much higher percentage of their FTP for 20 minutes, while diesel-engine steady-state riders sit closer to 0.96 or 0.97. You won't know where you fall until you've validated against sustained efforts in training.
Pacing remains a skill. If you go out too hard in the first five minutes, the back half collapses and the result underestimates your true FTP. If you're too conservative, same problem in the other direction. It typically takes two or three attempts before you learn how to pace a 20-minute effort properly.
The 8-minute test
Originally developed by Chris Carmichael and exercise physiologist Dr Edmund Burke as the CTS Field Test, this protocol uses two maximal 8-minute efforts with ten minutes of recovery between them. Take the average power of the higher effort and multiply by 0.90. The larger correction factor reflects the shorter duration and higher anaerobic contribution.
Pros: Shorter, less daunting for newer riders or those who find the mental game of 20 minutes overwhelming. Two efforts also give you a consistency check — if the two numbers are wildly different, your pacing was off.
Cons: The 0.90 correction factor has a wider margin of error. An 8-minute effort draws more heavily on your anaerobic system, which means riders with a disproportionately strong anaerobic capacity will test higher than their true sustainable threshold. The correction factor was designed to account for this, but it's a blunter instrument than the 20-minute protocol.
The ramp test
Start at a low wattage — typically 100-150 watts — and increase by a fixed amount every minute (usually 20 watts) until you physically cannot maintain the target. Take 75% of your best one-minute power during the test, and that's your estimated FTP. The 75% figure comes from research by cycling coach and sports scientist Ric Stern, who quantified the relationship between Maximal Aerobic Power (MAP) and threshold power in the late 1990s, finding that FTP typically falls between 72% and 77% of MAP.
Pros: No pacing decisions whatsoever. You simply ride until you can't. This makes it the most repeatable protocol — there's no psychological game of deciding how hard to start or when to push. If you have a smart indoor trainer with ERG mode and a platform like Zwift, the ramp test is essentially automated. The trainer sets the resistance, you pedal until you can't, and the software does the maths. For riders who have access to this setup, it removes almost every source of testing anxiety.
Cons: The ramp test tends to overestimate FTP for riders with a strong anaerobic engine — they can spike hard in the final minutes, inflating the peak one-minute power that the 75% calculation is based on. Conversely, if you're a diesel-engine rider whose strength is grinding out steady power for long periods, a ramp test may underestimate what you can actually sustain. Either way, the 75% correction factor is a population average, not personalised to you — individual variation of 3-5% in either direction is common.
Which test should you use?
| Protocol | Duration | Equipment needed | Pacing difficulty | Best for |
|---|---|---|---|---|
| 60-minute | 60 min | Any power meter | Very hard | Reference only |
| 20-minute | 20 min effort | Any power meter | Moderate | Experienced riders comfortable with pacing |
| 8-minute | 2 x 8 min efforts | Any power meter | Lower | Newer riders, those wanting a shorter test |
| Ramp test | 15-25 min total | Smart trainer + ERG mode | None | Indoor riders with a smart trainer |
If you have an indoor trainer with ERG mode, the ramp test is a strong default. It's repeatable, low-stress to execute, and eliminates pacing as a variable. For outdoor riders or those without a smart trainer, the 20-minute test is the most established proxy — long enough to reflect your threshold physiology, short enough that pacing mistakes don't completely derail the result. Just expect that your first attempt will be a learning experience.
Any of these protocols will give you a substantially more accurate FTP than a software estimate based on uncontrolled ride data. Beyond that, the test you'll actually do consistently is the best one — a ramp test repeated every six weeks is more valuable than a 20-minute test you never get around to.
Why accurate FTP matters more than you think
This isn't an abstract concern. When your FTP is wrong, the errors compound across every training metric you track.
Training Stress Score (TSS) is calculated using FTP as the denominator. If your real FTP is 260 watts but you're training off a stale estimate of 240, every ride registers more stress than it should. A recovery ride looks like an endurance ride. An endurance ride looks like tempo. Your CTL, ATL, and TSB all shift — and the training decisions you make based on them are being made with corrupted data.
Power zones are expressed as percentages of FTP. If your FTP is 20 watts too low, what you think is zone 4 (threshold) work is actually mid-zone 3 (tempo). You're spending sessions accumulating moderate fatigue without the stimulus that actually raises your threshold. Conversely, if your FTP is set too high, your "sweet spot" work is actually threshold, your threshold work is VO2max, and you're burying yourself without the recovery to absorb it.
Workout intensity targets — whether prescribed by a coach, a training plan, or an AI cycling coach — are only as good as the FTP they reference. Every target wattage is a percentage of FTP. Get the input wrong, and every output is wrong.
The compounding effect is what makes this matter. A single workout at the wrong intensity is trivial. Six weeks of training at systematically incorrect intensities — because your zones are all shifted — means suboptimal adaptation. You're putting in the work but not getting the return. And the most frustrating part is that you won't necessarily know it's happening. You'll just feel like the training isn't working as well as it should.
How to actually improve your cycling FTP
Once you have an accurate FTP, you have a foundation to build from. Improving it is straightforward in principle — though not easy in practice. The training approaches that raise FTP are well-understood, and they all centre on spending meaningful time at and around your threshold.
Sweet spot intervals (88-94% FTP) are the workhorse of FTP development. Efforts like 2x20 minutes or 3x15 minutes at sweet spot accumulate significant training stress at an intensity that's hard enough to drive adaptation but sustainable enough to recover from. For time-crunched riders who can't do long endurance rides, sweet spot work delivers a strong return on investment. It's the "high yield, moderate cost" zone.
Threshold intervals (95-105% FTP) target your FTP directly. Classic sessions like 2x20 minutes at FTP or 4x10 minutes at 100-105% FTP push your lactate clearance capacity and teach your body to sustain higher power outputs for longer. These are harder to recover from than sweet spot, so they need to be balanced carefully within a training week.
VO2max intervals (106-120% FTP) might seem unrelated to threshold, but they raise the aerobic ceiling above your FTP. Think of it as creating headroom. If your VO2max capacity improves, your threshold power has room to grow underneath it. Sessions like 5x4 minutes at 110-115% FTP with equal recovery are brutal but effective. Riders who plateau on FTP often find that a block of VO2max work unlocks the next step.
Consistency and base volume underpin all of the above. A deep aerobic foundation — built through regular zone 2 riding — supports a higher threshold. The riders who sustain long-term FTP improvements are the ones who combine targeted intensity with consistent overall volume. There are no shortcuts here. For a deeper look at how these approaches fit together, see our guide to cycling training methods.
Realistic expectations: An intermediate rider following a focused 8-week training block might see a 5-8% FTP improvement — roughly 13-20 watts on a 250-watt FTP. Individual results vary considerably depending on training history, genetics, and how much low-hanging fruit remains. A rider with an FTP of 180 has more room to grow than a rider at 320. Patience and consistency matter more than any single session.
Common mistakes: Testing too often (every 4-6 weeks during a focused block is sufficient — more frequently adds fatigue without information). Chasing FTP at the expense of overall fitness (an impressive FTP means nothing if you can't ride for more than 90 minutes). Neglecting recovery between hard sessions (adaptation happens during rest, not during the intervals themselves).
When to retest your FTP
FTP isn't static. It changes with training, detraining, fatigue, and season. Knowing when to retest is as important as the test itself.
Every 4-6 weeks during a focused training block. This is the sweet spot for tracking progress without the test itself becoming a source of fatigue. If you're following a structured plan with progressive overload, retesting at the end of a block tells you whether the work is producing the expected adaptations.
After a break or recovery period. If you've taken two or more weeks off the bike — whether from illness, travel, or a deliberate off-season — your FTP has almost certainly shifted downward. Retest before resuming structured training so your zones reflect your current fitness, not your pre-break fitness.
When efforts start feeling mismatched. This is the most important signal to watch for. If your zone 4 intervals feel surprisingly manageable — you're completing them without the expected difficulty — your FTP has probably risen and your zones are now too easy. Conversely, if prescribed endurance rides are leaving you unusually fatigued, your FTP might have dropped and your zones are too aggressive. Either way, it's time to retest.
Veloi tracks your power data continuously and flags when your FTP has likely shifted — identifying patterns like sustained efforts above threshold that feel routine, or normalised power creeping upward on easy rides. It's the kind of ongoing monitoring that catches drift before it silently corrupts your training for weeks.
The foundation everything else depends on
FTP is a single number. It takes less than half an hour to test properly. And yet it's the foundation that your entire training structure rests on — your zones, your stress scores, your fitness tracking, your prescribed intensities. Getting it right is one of the highest-return investments you can make in your cycling.
Whether you're setting your FTP for the first time or chasing a number you've been stuck on for months, the path starts the same way: an honest measurement. Know where you stand. Then build from there.
