FTP estimator & power zones
Turn a 20-minute, 8-minute, ramp or 1-hour effort into Functional Threshold Power, then into all seven Coggan power zones. Every multiplier here is traced to a named source, and the answer comes back as a range — because the conversion behind it is an approximation, and one watt of precision would be a lie.
Protocol. 20 min warm-up, a 5-min all-out blowout, 10 min easy, then 20 minutes as hard as you can hold evenly.
Source. The ×0.95 is from Allen & Coggan, Training and Racing with a Power Meter.
Coefficients from Sitko et al. (2023), who sorted 120 cyclists by measured VO2 max. The paper does not publish its cut-offs, so the one-line descriptions above are ours, not the study's. Pick the nearest and read the shift as a sensitivity check, not a diagnosis.
Your FTP
260 W × 0.95
Sitko et al. (2023) ran the 20-minute and a real 60-minute test on 120 cyclists and found the coefficient depends on how trained you are: 0.88 recreational, 0.92 trained, 0.95 well trained, 0.96 professional. The textbook 0.95 turned out to be right for well-trained riders specifically.
Power to weight
3.05–3.33 W/kg across the FTP range above
- World class5.93–6.60
- Exceptional5.36–5.93
- Excellent4.79–5.36
- Very good4.22–4.79
- Good3.65–4.22
- Moderate3.08–3.65
- Fair2.51–3.08
- Novice1.93–2.51
- Untrained1.27–1.93
Bands are the FT column of the Coggan power profile chart from Allen & Coggan's Training and Racing with a Power Meter, taken from the transcription at wattkg.com. Two cautions. Coggan built the chart by anchoring the top to world champions and the bottom to untrained people and interpolating between them — it is not a percentile of real riders, so “moderate” does not mean you are at the 50th percentile of anything. And secondhand transcriptions of this chart disagree with each other by up to half a W/kg across editions, so treat the edges as soft. For a real-world contrast, Cycling Analytics found the median 20-minute power among its uploaders was about 3.8 W/kg — a self-selected crowd who own power meters.
Power zones
Anchored to 247 WCoggan's seven levels. Bar length is the real width of each zone on a shared axis running to 160% of FTP — the dashed line is 100%, your threshold.
- Z1Active recoveryEasy spinning, very short sessions0–136 W0–55%
- Z2EnduranceLong rides, conversational138–185 W56–75%
- Z3TempoSteady, purposeful, 1–3 hr188–222 W76–90%
- Z4Threshold8–30 min intervals at or near FTP225–259 W91–105%
- Z5VO2 max3–5 min hard intervals262–296 W106–120%
- Z6Anaerobic30 sec to 2 min max efforts299–371 W121–150%
- Z7NeuromuscularSprints, under 15 sec373–∞ W151–∞%
Notice how narrow threshold is and how wide endurance is. That is not a drawing choice — it is what the percentages say, and it is why an easy ride is hard to get wrong and a threshold session is easy to get wrong. Every one of these boundaries moves with your FTP, so an FTP that is 15 watts optimistic makes every zone 15-ish watts optimistic too.
How this works
The 0.95 is an average, not a law
A maximal 20-minute effort is not an aerobic effort. Some of that power comes out of anaerobic reserves that you cannot keep drawing on for an hour. The 0.95 in Allen and Coggan's book is an estimate of how much to subtract — averaged over riders. It was never a claim about you.
FTP = P20 × 0.95In 2023 Sitko and colleagues put 120 cyclists through both a 20-minute test and a genuine 60-minute test and solved for the coefficient in each group. It was not 0.95 for everyone. It ran from 0.88 in recreationally trained riders to 0.96 in professionals, with 0.95 landing on the well-trained group. On a 260 W twenty-minute effort that is the difference between 229 W and 250 W, which is not a rounding error — it is a different training plan.
How to tell which rider you are
- Start with how trained you are. That was the strongest single determinant in the 2023 data, and it points the same way for most readers: if you ride for fitness rather than in structured blocks, 0.95 is flattering you and something nearer 0.88 fits better.
- Then check your shape. Riders with a big anaerobic engine put a lot of non-sustainable work into a 20-minute number, so the conversion over-estimates them. Diesel riders — strong 5-minute power barely above their 20-minute power, comfortable on long steady efforts, poor sprint — get under-estimated. The ramp test amplifies this in the same direction, which is why an anaerobically strong rider often gets their highest number of all from a ramp.
- Then settle it by riding the number. Go out fresh and hold your estimate for 45 to 60 minutes. Falling apart at 30 minutes means the estimate is too high. Finishing the hour able to push on for another 30 minutes means it is too low. FTP is defined by that hour, so this outranks any multiplier on this page — including ours.
- And check you ran the protocol. The Coggan 20-minute protocol includes a 5-minute all-out effort before the 20-minute effort, specifically to burn off anaerobic reserve. Skip it and you arrive at the 20 minutes fresher, post a higher number, and 0.95 over-shoots. MacInnis and colleagues measured 60-minute power at about 90% of a standalone 20-minute trial, not 95%.
None of this makes the 20-minute test bad. It has more published validation behind it than any other field protocol here, and a repeatable number you can compare to yourself in eight weeks is worth more than an accurate number you only ever collect once. What we will not do is print it to the watt and let you think that is precision.
Every multiplier, and where it came from
Four protocols, four conversions, four different levels of evidence behind them. The badges are not decoration — they are the difference between a number somebody measured and a number somebody published.
20-minute test
Range measuredFTP = P20 × 0.95 (0.88–0.96)20 min warm-up, a 5-min all-out blowout, 10 min easy, then 20 minutes as hard as you can hold evenly.
Source. The ×0.95 is from Allen & Coggan, Training and Racing with a Power Meter.
Range. Sitko et al. (2023) ran the 20-minute and a real 60-minute test on 120 cyclists and found the coefficient depends on how trained you are: 0.88 recreational, 0.92 trained, 0.95 well trained, 0.96 professional. The textbook 0.95 turned out to be right for well-trained riders specifically.
8-minute test
Range assumed by usFTP = P8 × 0.90 (0.86–0.94)Two 8-minute maximal efforts separated by 10 minutes of easy spinning. Take the higher of the two averages.
Source. The ×0.90 is Carmichael Training Systems’ own coefficient for its field test.
Range. CTS publishes 0.90 and no spread, and we found no independent validation of the 8-minute coefficient. This band simply mirrors the relative spread that was measured for the 20-minute test. Treat it as our assumption, not a finding.
Ramp test
Range widely reportedFTP = P1_best × 0.75 (0.72–0.77)Power steps up about 20–25 W every minute until you cannot hold the target. Use the best one-minute average, which is normally the last complete step.
Source. The ×0.75 is what TrainerRoad and Zwift both apply to your best one-minute ramp power.
Range. The 75% is the midpoint of a 72–77% relationship between maximal aerobic power and threshold, attributed to coach Ric Stern and repeated across coaching literature. We could not find a peer-reviewed paper establishing that range, so it is sourced but not evidence-graded.
1-hour effort
No conversionFTP = P60A genuine maximal 60-minute effort, ideally solo on a steady climb or a trainer.
Source. No conversion. FTP is defined as roughly the highest power you can hold for an hour, so this is the number itself.
Range. There is no conversion error to carry here. What is left is the variation in you — pacing, sleep, heat, motivation — which every method has on top of whatever its multiplier adds.
One thing we could not verify: nobody appears to have independently tested the CTS 8-minute coefficient against a real 60-minute effort. The 0.90 is CTS's own, and the band we draw around it is inferred from the 20-minute literature rather than measured. If you know of a study, we would genuinely like to read it.
Frequently asked questions
What is FTP?
Functional Threshold Power is roughly the highest average power you can sustain for about an hour. It is the anchor for power-based training — every cycling zone is defined as a percentage of it. The word "roughly" is doing real work: FTP is a training construct, not a measured physiological boundary, and studies comparing it to critical power and to maximal lactate steady state find it lands near them without being the same thing.
Why is the answer a range and not one number?
Because every test except the 60-minute one reaches FTP through a multiplier, and those multipliers are population averages. Sitko et al. (2023) tested 120 cyclists on both a 20-minute and a real 60-minute effort and the correct coefficient ran from 0.88 for recreational riders to 0.96 for professionals. Reporting a single watt figure to the nearest watt hides a spread that can be 20 watts wide. The range is the honest shape of the answer.
Why multiply the 20-min test by 0.95?
A well-paced 20-minute maximal effort pulls partly from anaerobic reserves. The 0.95 is Allen and Coggan’s estimate of how much to strip out. It is a population approximation, not a law — riders with strong anaerobic capacity over-estimate their FTP from it, and diesel riders under-estimate. One detail people miss: the protocol includes a 5-minute all-out blowout before the 20-minute effort. Skip that and 0.90 fits better than 0.95, which is close to what MacInnis et al. (2019) measured on standalone 20-minute trials.
How do I tell whether 0.95 flatters me or shortchanges me?
Two checks. First, how trained you are: in the 2023 data that was the strongest single determinant, and recreational riders came out at 0.88, not 0.95. Second, ride the number. Go out fresh and hold your estimated FTP for 45 to 60 minutes. If you fall apart before 45 minutes the estimate is too high; if you finish able to carry on for another half hour it is too low. That is the definition of the metric, so it outranks any multiplier.
Which FTP test should I use?
The 20-minute test has the most published validation behind its conversion. The ramp test is the gentlest on untrained riders and needs no pacing skill, but it is a maximal aerobic power test being converted into a threshold number, so it flatters explosive riders. The 8-minute test is quicker and its coefficient comes from CTS with no independent validation we could find. The 1-hour test needs no conversion at all and is the most demanding. Pick whichever you will actually repeat every 6–8 weeks — comparability over time beats theoretical accuracy once.
How often should I retest?
Every 6–8 weeks during a structured training block, using the same protocol each time. More frequently than that and you are mostly testing sleep, heat and motivation. Less frequently and your zones drift out of date.
How is FTP different from threshold heart rate?
FTP is an external metric (what you produce), threshold HR is an internal one (how hard your body is working to produce it). They track each other closely in fresh, trained athletes, but fatigue, heat, caffeine and dehydration decouple them. Train to power, monitor with heart rate.
What is a good FTP for a recreational cyclist?
Power-to-weight matters more than raw watts on anything that goes uphill. On Coggan’s power profile chart, 3.08–3.65 W/kg is "moderate" for men and 2.58–3.09 for women; "good" starts at 3.65 and 3.09 respectively. For a real-world contrast, Cycling Analytics found the median 20-minute power among its users was about 3.8 W/kg — but those are people who own power meters and upload files, which is not the general population.
Sources
Everything on this page traces to one of these. To be explicit about what is not here: we have no cycling power data of our own. Our published studies cover race prediction, training readiness, VO2 max drift, recovery-signal overlap and sleep — all running and recovery. Nothing on this page is our finding.
- 01Allen & Coggan, Training and Racing with a Power MeterVeloPress. Origin of the ×0.95 twenty-minute conversion, the seven power levels, and the power profile chart.
- 02Sitko, Cirer-Sastre & López-Laval (2023)An Update of the Allen & Coggan Equation to Predict 60-min Power Output in Cyclists of Different Performance Levels. Int J Sports Med 44(13):983–987. 120 cyclists; the coefficient ranged 0.88 to 0.96 by performance level.
- 03MacInnis, Thomas & Phillips (2019)The Reliability of 4-Minute and 20-Minute Time Trials and Their Relationships to Functional Threshold Power. Int J Sports Physiol Perform 14(1):38–45. 60-minute power came out at roughly 90% of 20-minute power, standard error of the estimate 3.0%.
- 04Carmichael Training Systems — the CTS Field TestTwo 8-minute maximal efforts split by 10 minutes easy. CTS coaches take the higher average and multiply by 0.90.
- 05TrainerRoad — Ramp Test FAQsFTP is taken as 75% of the best one-minute power reached in the ramp.
- 06Zwift Insider — Zwift FTP testsConfirms Zwift uses the same rule: 75% of the highest one-minute average power in the ramp.
- 07Ric Stern — the 72–77% maximal aerobic power relationshipThe coaching-literature origin of the 75%: threshold power sits at roughly 72–77% of maximal aerobic power, 75% being the midpoint used in software. Widely repeated by coaches; we did not find a peer-reviewed paper establishing the range.
- 08wattkg.com — transcription of the Coggan power profile chartThe W/kg band edges used on this page. Secondhand transcriptions of this chart disagree with each other; see the note under the scale.
- 09Cycling Analytics (2018) — how does your cycling power output compare?Percentiles from real uploaded power files. Median 20-minute power among its users was about 3.8 W/kg; the 90th percentile about 4.86 W/kg for men and 4.61 for women.
More on how we handle numbers we cannot verify: our methodology.
Read more
FTP anchors every power number you will ever look at — normalised power, intensity factor, TSS. Here is where those come from, and what the evidence says about training to them.