Cycling Repeated-Sprint Fatigue Test: Protocol, Power-Decay Scoring, and Norms

A rider can put out 850 watts on a single all-out sprint test and still get dropped on lap nine of a criterium — not on the first attack, but around the fourth or fifth, when the field keeps surging and nobody lets the pace settle. Coaches see this mismatch constantly. The athlete who tops the leaderboard on one maximal effort is not always the one still in the group when the race actually breaks apart. What separates them usually isn't peak wattage. It's how much of that peak survives six or seven maximal efforts stacked back to back, each one launched before the legs have finished recovering from the last.

Most cycling power testing answers one question well: how hard can this rider push, once, from fresh legs. A single max sprint or a standard Wingate test nails that number. Neither says anything about what happens to it when it has to repeat nine or ten times with 20-30 seconds of soft pedaling in between — closer to what a crit, a cyclocross race, or a track omnium actually demands. The protocol below exposes that gap: a repeated-sprint test scored on how much power decays across a set, not on the single best number buried inside it.

Why a Single Peak-Power Number Hides Race Fitness

Two different physical qualities get flattened into the phrase anaerobic power. One is how much force fast-twitch fibers can generate and how fast the nervous system recruits them in a single burst — what a 6-15 second maximal effort measures well. The other is how quickly phosphocreatine stores rebuild during a short recovery window and how well a rider tolerates rising acidity across repeated efforts. That second quality decides who is still attacking near the finish and who quietly slides off the back after the fourth hard dig.

A rider with excellent single-sprint power but sluggish phosphocreatine resynthesis can look identical to a well-rounded sprinter on test day, then fall apart in an actual race by the sixth effort, because a single-sprint test never asks the body to repeat under incomplete recovery. Testing peak power alone is a bit like judging a boxer by how hard he can hit a bag once — it tells you the ceiling, and nothing about round nine.

Equipment and Setup

The test needs an ergometer or smart trainer that logs power at least once per second, a resistance or gear setting fixed and written down for every retest, and a way to call out timing without the rider watching a clock mid-sprint.

A resistance that's too light lets a fatigued rider spin through it and mask a real power drop; too heavy, and sprint one looks mediocre. Set it once at a level the rider can push through a full 90-100rpm sprint, then never change it for that athlete's retests.

Step-by-Step Testing Protocol

1. Warm-up (10-12 minutes): Five minutes of easy spinning, brief dynamic mobility off the bike, then three progressive 6-second efforts at roughly 70%, 85%, and 95% effort with full recovery between each.
2. Familiarization: One submaximal sprint at about 85% effort against the resistance or gear you plan to use, mainly to confirm the rider can hit 90-100rpm within the first two pedal strokes. Take 3 minutes of easy spinning after it.
3. Lock the resistance: Record the exact resistance setting, or gear plus cadence target on a smart trainer, and never change it for that athlete's future retests.
4. Maximal set: 10 sprints of 6 seconds each, maximal effort, from a rolling start reaching target cadence in the final two pedal strokes before the sprint signal. Each sprint is followed immediately by 24 seconds of easy spinning at low resistance (roughly 60-70 watts).
5. Record every sprint: Log peak power (highest 1-second reading) and mean power (average across the full 6 seconds) for all 10 sprints, not just the first and last.
6. Valid trial criteria: Rerun the set if the rider stops pedaling during recovery, needs an unscheduled pause, or badly misses the cadence target on sprint one — that usually signals an incomplete warm-up rather than genuine fatigue.

Total time, warm-up included, runs about 18-20 minutes; the maximal set itself is only around 4.5 minutes, but a rushed warm-up will wreck sprint one and inflate the apparent fatigue across the whole set.

Scoring: Fatigue Index and Power-Decay Math

Two numbers come out of the ten sprints, and they answer slightly different questions. Use mean power for both — reliability research on repeated cycle sprints found fatigue scores built from peak power considerably noisier session to session than the same calculation run on mean power.

Fatigue Index (FI%) compares only the best and worst sprint: FI% = [(P best − P worst) / P best] × 100. Same logic as the fatigue index inside a single 30-second Wingate test, applied instead across ten discrete efforts with recovery between them — sensitive to recovery capacity rather than single-bout muscular endurance.

Percent Decrement Score (%Dec) uses every sprint, not just the extremes: %Dec = 100 × [1 − (sum of all sprint mean powers) / (10 × P best)]. Because it factors in the whole set, %Dec is less sensitive to one unusually good or bad sprint and is the better number for comparing against normative bands.

Worked example: a rider posts mean power of 820W on sprint one and it drifts down to 590W by sprint ten, with the ten values summing to 6,975W. FI% = (820 − 590) / 820 × 100 = 28%. %Dec = 100 × [1 − 6,975 / 8,200] ≈ 14.9%. FI% is driven entirely by two data points while %Dec reflects the shape of the whole decline — a rider who fades steadily looks different under %Dec than one who holds steady for eight sprints and then collapses on the last two, even with matching FI% numbers.

What the Research Shows

Bishop, Spencer, Duffield, and Lawrence (2001), publishing in the Journal of Science and Medicine in Sport, ran trained team-sport athletes through a cycle-based repeated-sprint protocol closely resembling the one above and looked for what actually predicts resistance to fatigue across the set. Their result cuts against a common assumption: aerobic fitness (VO2max) showed no meaningful relationship with how much power a rider lost across the sprints, while muscle buffering capacity — the ability to tolerate and clear rising acidity — showed a clear, moderate association with holding power across repeated efforts. A big aerobic base, in other words, does not guarantee a good score here. The catch the authors flagged themselves: buffering capacity was measured from muscle biopsy, which no field coach can replicate, so a field test like this one stands in as a proxy rather than a direct measurement of the underlying mechanism.

Glaister, Stone, Stewart, Hughes, and Moir (2008), in the Journal of Strength and Conditioning Research, tested how reliable fatigue scores are from one session to the next when calculated different ways from repeated cycle sprints. Scores built from each sprint's peak power carried noticeably more day-to-day noise than the same score run on mean power or total work — enough that the authors recommended against using peak-power-based decrement scores to track an athlete over time. Their own limitation is worth carrying forward: the reliability numbers came from one ergometer and one sprint duration, so noise levels can shift with a different rig, which is one more reason to keep resistance and sprint length identical across every retest.

Reading the Score: Bands for Percent Decrement

The bands below are a field reference built from the repeated-sprint cycling literature's general findings on trained versus untrained decrement rates, not a strict pass-fail line. Weigh them against the rider's own history on this exact protocol before anything else.

A rider sitting in the moderate-to-poor bands with a strong single-sprint peak power number has a specific, coachable target: repeated-sprint capacity, not raw power. Piling on more max-effort sprint work without addressing recovery between efforts tends to move the peak number without touching %Dec at all.

Mistakes That Skew the Score

Building a Plan Around a Poor Score

A high %Dec is a training target, not a verdict on the rider's engine. Riders in the moderate or poor bands typically respond well to a block built around genuinely repeated-sprint-specific work rather than more long intervals or more single max efforts: sets of 8-12 short maximal sprints (5-8 seconds) with recovery matched to the test's 24-second window, done once or twice a week alongside — not instead of — normal aerobic volume. Retest every 4-6 weeks; %Dec moves more slowly than a single peak-power number, and testing sooner mostly captures noise rather than adaptation.

If a rider tests well on a single max sprint but poorly here, resist filing it under a general aerobic-fitness gap. It's a specific, trainable quality tied to phosphocreatine recovery and acid tolerance rather than aerobic capacity, and one of the few field tests that maps directly onto what happens when a race keeps attacking and nobody backs off.