REF-751 · Black, Burnley, Skiba, Poole, Vanhatalo & Jones, 2026 · International Journal of Sports Medicine, epub ahead of print · Paper

Ten common criticisms of the critical power model, answered by the group that built it

Including the one that matters most in practice: "indefinitely sustainable" was always the mathematician's "undefined", never the athlete's "forever"

B+moderate evidenceMeasured on a mixed groupWritten before October 2026, being re-checked

What it is

The six authors identify ten pervasive critiques of the critical power model and evaluate each against its assumptions and the available evidence. The abstract names three: that CP is a purely mathematical construct without physiological basis; that CP represents an indefinitely sustainable power output; and that newer performance models have made CP obsolete.

Their verdict: the model "is congruent with the current understanding of exercise physiology", its limitations "highlight the complexity and manifestation of fatigue dynamics across different exercise intensity domains", and applied in context it "remains a useful and informative tool for characterising exercise tolerance".

Key Findings — what can be stated from the abstract and quoted text

"Indefinitely" is a term of art. In the mathematics of the hyperbolic model the asymptote is the power at which the predicted time to exhaustion is undefined; in lay reading it became "forever". The authors write, as quoted: in lay language "indefinitely" is often taken to mean "forever"; in mathematics "indefinite" means "undefined". CP has never been a claim about holding a power for an hour.
CP is a domain boundary with a physiological basis, separating the intensity range in which V̇O₂ and blood lactate can settle from the one in which they cannot — the same framing as REF-168 — and not an artefact of curve-fitting.
Above CP, the model predicts tolerance; at and below it, it predicts nothing about duration. The CP/W′ framework gives the time a rider can hold any power in the severe domain; it does not assign a sustainable duration to CP itself, to MLSS, or to any other second-threshold marker.
Newer models (power-law, alternative fatigue models) are addressed as complements or alternatives with their own assumptions, not as a replacement — the paper's stated position, not evaluated here.

Read the direction carefully

Two things this entry cannot do. It cannot list all ten criticisms — only the three the abstract names and the one passage a secondary source quoted have been verified. And it cannot be used to argue that CP is a bad threshold: the paper's conclusion is the opposite. What it settles, for the coach, is a single sentence that has been misquoted for thirty years: CP is where the physiology changes, not how long anyone lasts.

Practical Application

Never present CP, MLSS or "threshold" as a duration. The correct sentence is "the highest power your body can still settle at"; the measured TTE for trained riders sits around 20–40 min and is individual (REF-750).
Use W′ above CP as the duration tool, and say so. "About 6 minutes at 320 W" is a prediction the model makes; "about an hour at CP" is one it never made.
When an athlete cites a newer model against CP, the answer is "different question, not better answer": the power-law curve describes the whole range's shape, the CP model describes the boundary and the severe domain's tolerance (REF-442).
Keep the ±3–5% bandwidth in every prescription "at threshold" (REF-168, REF-750): a target set exactly on the estimate lands in either domain on a coin toss.

Limitations

Narrative review by the model's own authors — an authoritative defence, not an independent audit.
Entered from the abstract plus a quoted passage; the substance of seven of the ten rebuttals has not been read and is not represented here.
Epub ahead of print; volume and pages not yet assigned.