Differences in Performance Assessments Conducted Indoors and Outdoors in Professional Cyclists
Participants
n = 14: male UCI Continental and UCI WorldTour cyclists
Age: 20.9 ± 2.8 years; mass 68.1 ± 7.3 kg
4 test sessions: 2 indoor + 2 outdoor, within a 14-day period
Protocol
Tests: maximum mean power (MMP) for 60, 180, 300 and 840 s, indoors and outdoors
Critical power (CP) and work capacity above CP (W') derived from the MMP set in each setting
Self-selected cadence recorded
Full text; not checked against the abstract: outdoor tests on a slight uphill grade
Key Results
Across all MMP durations, trimmed mean power was higher outdoors than indoors (P < .05)
Critical power was 19 W higher outdoors (P = .005); no difference in W' (work capacity above CP)
Self-selected cadence was 6 rpm higher indoors at 60 s (P = .038) and 300 s (P = .002)
Substantial individual variation in the indoor-outdoor difference across all durations
The abstract gives no per-duration percentage differences
Mechanisms (speculative — not measured in this study)
Inference, not tested: poorer cooling indoors, the absence of bike-plus-rider inertia on an ergometer, less complete neuromuscular engagement without balance and visual flow, and lower motivation indoors are all plausible contributors; the study measured none of them. The higher indoor cadence is consistent with riders compensating for the different inertial load, but that too is interpretation.
What the data do show: the shift is systematic at the group level (all durations, CP +19 W) and unpredictable at the individual level.
Practical Application
In n=14 male UCI Continental/WorldTour cyclists tested twice indoors and twice outdoors within 14 days, maximum mean power at 60, 180, 300 and 840 s was higher outdoors at every duration, critical power was 19 W higher outdoors while W' did not differ, self-selected cadence ran 6 rpm higher indoors, and the size of the indoor-outdoor gap varied substantially between riders — indoor and outdoor maximal tests are not interchangeable, and no group correction factor transfers to one athlete.
Do NOT mix indoor and outdoor tests for progress assessment: a retest must repeat the modality of the baseline (indoor vs indoor, outdoor vs outdoor), otherwise the setting difference can hide real progress or fake it.
FTP/CP indoor ≠ FTP/CP outdoor: use indoor numbers for indoor zones and outdoor numbers for outdoor zones. Never apply a group offset (such as this study's +19 W CP) as a personal correction — the per-rider spread is the finding.
When moving from trainer to road in spring, retest outdoors before assigning outdoor zones; copying indoor zones outdoors will likely underestimate them.
Indoor still has advantages for reproducibility (no descents, stops or traffic) — it is the comparison across settings that fails, not indoor testing itself.
Expect self-selected cadence to be a few rpm higher indoors; a cadence drop outdoors at the same power is normal, not a sign of fatigue.
Limitations
n = 14 young male professionals (UCI Continental/WorldTour), Tier 4-5. The indoor-outdoor gap in recreational riders is untested.
Outdoor tests on a slightly uphill section of road of approximately 2-3% gradient (full text, checked 2026-09-28: accepted manuscript, Methods), a condition that favours steady maximal effort. Flat or rolling outdoor testing is not the same comparison.
Environment cannot be blinded and outdoor motivation is plausibly higher; the design cannot separate physiology from context.
Two indoor and two outdoor sessions within 14 days — enough for a group mean, thin for per-athlete test-retest reliability.
The individual spread is the finding, not the mean. Differences varied substantially between riders, so group values must never be applied as a correction factor to one athlete's numbers.
Power meters may have differed between the indoor ergometer and the outdoor bike; part of the systematic gap may be instrumentation rather than the rider.
How this entry was checked and changed
Why this level
(n=14, UCI-level cyclists)
Content correction 2026-09-27
Per-duration % table (−4.2 to −8.8%) and individual example (−152/−48/−37/−36 W) -> not in the abstract; deleted. Abstract reports: outdoor power higher at all durations (P < .05), CP +19 W outdoors (P = .005), W' unchanged, cadence 6 rpm higher indoors at 60 and 300 s.
"Fan reduces the gap ~2-3%" -> not in the paper; deleted. Mechanisms (thermoregulation, inertia, neuromuscular, psychological) -> relabelled as speculative, not measured.
Outdoor "2-3% grade" -> marked full text, not checked against the abstract.
2026-09-28: lines labelled "full text; not checked" read against the open-access full text (accepted manuscript, newman.repository.guildhe.ac.uk): 1 confirmed (gradient made specific: ~2-3% uphill), 0 corrected, 0 removed.