REF-747 · Sitko, Cirer-Sastre & López-Laval, 2022 · Journal of Science and Medicine in Sport, · Paper

"FTP" from a 20-min test is not an hour of power: time to exhaustion at 95% of 20-min power had a median of 35 min in recreationally trained cyclists, rising to 51 min in professionals, with wide individual spread at every level

Blimited evidenceMeasured on recreational to elite athletesWritten before October 2026, being re-checked

Protocol

87 road cyclists rode two tests 72 h apart on their own bikes mounted on a Tacx Neo 2T: (1) the Allen & Coggan 20-min FTP test — a ~50-min warm-up including 3 × 1-min fast-pedalling efforts and a 5-min maximal opener, then a 20-min maximal effort, with FTP defined as 95% of the 20-min mean power; (2) a time-to-exhaustion (TTE) ride at that fixed FTP power until it could no longer be held.

Riders were split into four groups by V̇O₂max: recreationally trained, trained, well-trained and professional. The newsletter summary (Knowledgeiswatt #190) reports group means of about 47, 59, 66 and 74 mL·kg⁻¹·min⁻¹ and 3, 4, 7 and 15 years of training experience; those figures are from the full paper, not the abstract.

Key Findings

Median (IQR) time to exhaustion at FTP20: 35 (31–38) min recreationally trained · 42 (38–51) min trained · 47 (41–56) min well-trained · 51 (44–59) min professional.
Time to exhaustion increased with performance level and experience (P < 0.001). Even the professional median falls short of the "approximately one hour" FTP is defined as.
Spread within each group was large — the interquartile ranges alone span 7–15 min, and the newsletter reports individual values from roughly 20 to 80 min across the cohort (full-paper figures, not in the abstract).
The authors' conclusion: FTP and time to exhaustion "should be assessed and reported independently for each subject", and level and experience must be factored in when TTE at FTP is studied.

What this does and does not say

It does not say the 20-min test is useless as a training anchor: a consistently measured 95%-of-20-min number still tracks fitness within one rider over time. What it removes is the DURATION claim attached to the number — that FTP is "what you can hold for an hour".

For a recreational rider the honest translation of FTP20 is "about 35 minutes, give or take a lot"; for a professional it is closer to 50. The gap is durability and experience, not a flaw in the arithmetic of any one rider.

The study did not measure critical power or MLSS in the same riders, so it cannot say which construct FTP20 lands nearest; that comparison is REF-038.

Practical Application

Never tell an athlete their FTP is the power they can hold for an hour. Say what the test measured: 95% of their best 20 minutes. If they want to know what they can hold for 60 min, that is a separate measurement or a modelled estimate from a power–duration curve (REF-024), and for a Tier 1–2 rider the true 60-min power will usually sit BELOW FTP20.
Pacing a 40–60-min effort off FTP20 will overcook a recreational rider. The median rider at that level fails before 40 min at FTP20. Prescribe hour-long steady efforts below FTP20, or from a duration model, and read the fade as data rather than as a failed session.
Threshold intervals (2 × 20, 3 × 15) at 100% FTP20 are near the ceiling for a less-trained rider. 2 × 20 at FTP20 with a short rest is a large fraction of that rider's total TTE; if the second rep collapses, the target is right and the session shape is wrong for the athlete's durability, not the other way round.
Time to exhaustion at threshold is a trainable, level-dependent quality in its own right — the 35 → 51 min progression across levels is what "durability at threshold" looks like in a number. Track it within the athlete (same protocol, same conditions) rather than against these group medians.
Prefer CP for the physiological threshold and zone boundaries (REF-024, REF-038); use FTP20 as a consistent anchor when that is what the athlete has, and never mix the two in one prescription.

Limitations

Time to exhaustion at a fixed power is one of the most variable measurements in exercise physiology (day-to-day CV commonly 10–20% or more in constant-load TTE literature); each rider rode ONE TTE, so individual values carry that noise and only the group medians and the level trend are robust.
FTP20 here follows the Allen & Coggan protocol with a 5-min opener; other 20-min protocols yield different FTPs (REF-038), and the TTE would shift with them.
Groups were formed by V̇O₂max, not by competitive category; "professional" is a physiological label here.
No CP, MLSS or lactate measurements in the same riders, so the study cannot say where FTP20 sits relative to the metabolic steady state.
Indoor trainer, fixed power, no drafting or terrain — the ecological question of how long a rider holds "threshold" in a race is different again.
Group means for V̇O₂max and experience, and the 20–80-min individual range, come from the full paper via the newsletter, not from the abstract that was verified here.
How this entry was checked and changed
Content correction 2026-09-27
Cross-reference descriptions checked against the cited entries' current content; wrong years, wrong entries and claims the cited entry does not make were corrected.