REF-253 · Jeffries et al., 2019 · International Journal of Sports Physiology and Performance, 2019; · Paper
An Analysis of Variability in Power Output During Indoor and Outdoor Cycling Time Trials
Participants
n=20 male competitive cyclists, VO2max 60.4 ± 7.1 mL/kg/min.
Protocol
Two randomized 20-min maximal time trials separated by 7 days:
Indoor: laboratory electromagnetically braked training ergometer
Outdoor: cycle-specific racing circuit
Same bike and portable power meter in both tests, data sampled at 1 Hz. Frequency (spectral) analysis of power fluctuations.
Key Results
| Metric | Indoor | Outdoor | Significance |
|---|---|---|---|
| Mean 20-min power | 280 ± 44 W | 284 ± 41 W | NS (P = 0.256); r = 0.94 between conditions |
| Within-person SD of power | 33 ± 10 W | 69 ± 21 W | P < 0.001 |
Outdoor within-ride variability was roughly twice the indoor value at the same mean power.
Greater variability outdoors across all frequency bands (P < 0.001) except the very slowest bin (< 0.0033 Hz, P = 0.930) — the slow pacing curve is the same; the moderate- and high-frequency fluctuations are what the outdoors adds.
Authors' conclusion: constraints imposed by the external environment produce these fluctuations, so indoor testing protocols should be designed to reflect the external demands of outdoor cycling.
Mechanism
From the paper: the external environment imposes moderate- and high-frequency power fluctuations that the ergometer removes. Inference, not tested: wind, gradient, corners and traffic force continuous power adjustments; how the neuromuscular system handles these (recruitment patterns) was not measured.
Practical Application
Crossover, 20 male competitive cyclists (VO2max 60): a 20-min maximal TT gave the same mean power indoors and outdoors (280 vs 284 W, r = 0.94), but the within-ride power SD outdoors was about twice the indoor value (69 vs 33 W, P < 0.001), with more variability at every frequency except the slowest. A 20-min indoor test predicts outdoor 20-min power well; what it does not reproduce is the constant power fluctuation of riding outside.
Indoor and outdoor 20-min TT power can be treated as interchangeable for threshold estimation — the means and the ranking of riders agreed closely.
Extrapolation, not tested here: if a target event demands variable power (surges, corners, gradients), sessions in fixed-resistance ERG mode do not rehearse that pattern; use ERG-off / slope mode or ride outdoors for race-specific work, and keep ERG for precision-controlled intervals.
The paper's own recommendation is about testing: an indoor protocol meant to represent outdoor demands should build in variability rather than hold a constant load.
Limitations
Time trial only — not interval training; no training intervention, so effects of ERG training on outdoor performance are not tested
20-min TT — may differ for shorter/longer durations
Flat racing circuit outdoors — hillier or rougher courses would likely show even more variability
No EMG data — neuromuscular differences inferred, not measured
How this entry was checked and changed
Why this level
regraded 2026-09-27 from A: within-subject comparison, n=20
Content correction 2026-09-27
Lead bullet now the measured finding (same mean power 280 vs 284 W, r = 0.94; within-person SD 69 vs 33 W) instead of the "ERG creates a variability deficit" inference.
Training recommendations (ERG off for race-specific work, SIM mode for long rides) -> labelled extrapolation, not tested in this paper; "constrains fluctuations by > 50%" -> outdoor SD ~2× indoor.
Tier 3 -> Tier 2-3 (competitive cyclists, VO2max 60.4).
Cross-reference descriptions checked against the cited entries' current content; wrong years, wrong entries and claims the cited entry does not make were corrected.