Estimated FTP from Cycling Power

eFTP is a constrained estimate from cycling power-duration evidence, separate from the configured FTP.

Evidence and model selection

The model considers qualified efforts over the preceding 180 days, grouped by cycling power source. Evidence older than 90 days is stale. Target durations include 300, 600, 1,200 and 1,800 seconds. A multi-point critical-power fit needs at least three points spanning a short duration of at most 600 seconds and a long duration of at least 1,200 seconds, from at least two activities. Otherwise a five-/twenty-minute pair can be used, or a longest available anchor of at least 600 seconds with lower-confidence prior assumptions.

Formula and cap

For a critical-power fit, power P at duration t is modeled as P = CP + W′/t, with CP in watts and W′ in joules. The anchor is converted to a twenty-minute equivalent:

P20 equivalent = anchor power × (anchor seconds / 1200)^0.07
safety cap = 0.95 × P20 equivalent
eFTP = min(CP, safety cap), or safety cap for the single-anchor model

The final estimate is rounded to whole watts. Values outside 50–2,000 W, invalid/nonpositive CP or W′, or insufficient evidence are rejected. The 0.07 exponent and cap are product assumptions, not proof that all athletes share a fixed power-duration relationship.

Verified example

A single twenty-minute anchor at 300 W has a P20 equivalent of 300 W and cap 285 W, so the single-anchor estimate is 285 W. It does not obtain multi-point confidence merely because this arithmetic is exact.

Decision and limitations

Source disagreement, recency, duration coverage and repeated evidence affect confidence and whether a suggestion is surfaced. The configured FTP is not an input to the estimate itself. Applying an estimate is a separate explicit decision in Power zones; viewing the number does not change training targets. This is cycling evidence, not a laboratory measurement or automatic coaching.

Sources and implementation scope

Clarke and Skiba (2013), performance models and critical power provides background, not a specification of every EffortSignal rule. The formulas, windows, quality gates and examples above describe EffortSignal 0.1.13, checked September 10, 2026. Product-specific thresholds and rounding should not be attributed to the background source.

Content checked September 10, 2026 · EffortSignal release 0.1.13

All explanations · All features