REF-1198 · Rønnestad & Vikmoen, 2019 · Scandinavian Journal of Medicine & Science in Sports · Paper · PubMed
Can I peak in under two weeks?
In 17 elite cyclists, 6 days of daily 30/15 interval sessions then a 5-day step taper raised peak aerobic power 5% vs 1% after an 11-day taper. Small, not randomised, lab tests only.
Rønnestad and Vikmoen compared two ways to peak inside 11 days in 17 elite male cyclists, matched into two groups on VO2peak.
The compressed group (EXP, n = 9) did one high-intensity session every day for 6 days — three 9.5-min sets of 30 s hard / 15 s easy — with less easy and moderate riding, then a 5-day step taper (test day, rest day, 20-40 min easy, a short opener session, test).
The traditional group (TRAD, n = 8) kept its usual high-intensity sessions and shortened its easy and moderate riding across all 11 days. From the first to the last test, EXP raised VO2peak by 4.0% and peak aerobic power (Wmax) by 5.0%; TRAD changed by 0.8% and 0.9%, and none of TRAD's measures improved.
The difference in percentage change was significant for both (t-tests), but in the repeated-measures ANOVA the group-by-time interaction was significant for Wmax (P = .005) and not for VO2peak (P = .131); gross efficiency (P = .009, no within-group change) and peak heart rate (P = .025) also had significant interactions.
Power at 4 mmol/L lactate and knee-extension torque tended to improve more in EXP. EXP's fatigue score rose at the end of the block and fell below baseline by day 11, when its legs felt better than TRAD's. Total training load in EXP's "overload" block was no higher than in the previous four weeks; it was the daily high-intensity time that rose (+192 ± 81%).
Our reading: a small, non-randomised trial from one group showing that concentrating intensity into a 6-day block and then backing off can lift laboratory markers within 11 days in elite cyclists; it is not evidence that the block beats a taper on race day, and TRAD's lack of any change is unexplained.
Protocol
Participants: n = 17 male cyclists; EXP n = 9 (age 21 ± 7 y, 180 ± 5 cm), TRAD n = 8 (22 ± 8 y, 178 ± 8 cm); assigned and matched into two homogeneous groups by VO2peak from the previous 6 months; written consent, local ethics approval.
Baseline training (4 weeks before): EXP 10.6 ± 2.3 and TRAD 11.8 ± 2.3 h per week at low intensity (60-82% HRpeak), 1.6 h at moderate (83-87%) and 1.1 h at high intensity (88-100%) in both. Intervention started at the end of May, in the competition season.
Standardisation: 1.5-2.5 h easy riding on each of the two days before the first test.
EXP, days 1-6: one session a day of 3 sets × 9.5 min of 30-s work / 15-s recovery (recovery at 50% of work power), 3 min between sets, aiming for the highest average power across the three sets — typically 85% and 43% of Wmax for work and recovery, 72% of Wmax across a set; less low- and moderate-intensity riding.
TRAD, days 1-6: usual high-intensity sessions kept; low- and moderate-intensity sessions shortened.
Days 7-11 (Table 1): day 7 test (post-1; TRAD added a 2×4, 2×3, 2×2, 2×1-min interval session after it); day 8 EXP complete rest, TRAD 60-120 min easy; day 9 20-40 min easy (both, per Table 1); day 10 20 min easy + 2 × 5 min moderate + 3 × 1 min building towards Wmax (both); day 11 test (post-2).
Training load (TRIMP = minutes in zones 1-3 × 1, 2, 3; Table 2): days 1-6, EXP +9 ± 32% vs baseline with high-intensity time +192 ± 81%; TRAD −26 ± 8%. Last 4 days, EXP −57 ± 13% and TRAD −50 ± 13% (not different). Over the 11 days, TRAD −35 ± 8% vs EXP −17 ± 21% (P < .05).
Tests (pre, day 7, day 11): POMS; isokinetic knee-extension peak torque at 60°/s; incremental lactate test (125 W start, +50 W per 5 min) for power at 4 mmol/L and gross efficiency at 225 W; incremental test to exhaustion (+25 W per min) for VO2peak and Wmax (mean power of the last minute); perceived well-being in the legs on a 9-point scale.
Analysis: mixed-design ANOVA (group × time) with Bonferroni-adjusted follow-up; independent t-tests on percentage change pre to post-2; Cohen's d (0.6-1.2 moderate, 1.2-2.0 large); P ≤ .05 significant, P < .10 "tendency".
Funding and interests: no funding; no conflict of interest reported.
Key Results
Mean ± SD at pre → post-1 (day 7) → post-2 (day 11), and the between-group comparison of the percentage change from pre to post-2 (Table 3 and Results):
| Measure | EXP (n = 9) | TRAD (n = 8) | Change pre → post-2, EXP vs TRAD | Group comparison |
|---|---|---|---|---|
| VO2peak (mL/kg/min) | 76.8 ± 6.9 → 78.3 ± 5.3 → 79.7 ± 4.9 | 73.5 ± 4.1 → 73.3 ± 5.6 → 74.1 ± 4.4 | +4.0 ± 3.7% vs +0.8 ± 1.8% | t-test P = .041, ES 1.1; interaction P = .131 |
| Wmax (W/kg) | 6.43 ± 0.46 → 6.48 ± 0.54 → 6.75 ± 0.46 | 6.21 ± 0.33 → 6.25 ± 0.33 → 6.26 ± 0.29 | +5.0 ± 3.6% vs +0.9 ± 1.5% | t-test P = .009, ES 1.5; interaction P = .005 |
| Power at 4 mmol/L (W/kg) | 4.47 ± 0.35 → 4.59 ± 0.33 → 4.62 ± 0.38 | 4.25 ± 0.33 → 4.23 ± 0.35 → 4.27 ± 0.39 | +3.1 ± 2.8% vs +0.4 ± 3.4% | P = .088, ES 0.88 |
| Gross efficiency at 225 W (%) | 19.6 ± 1.5 → 19.4 ± 1.1 → 20.0 ± 1.8 | 19.1 ± 0.5 → 19.4 ± 0.5 → 18.9 ± 0.6 | no within-group change | P = .125, ES 0.81 |
| Knee-extension peak torque (Nm) | 184 ± 36 → 180 ± 34 → 196 ± 37 | 179 ± 28 → 174 ± 48 → 173 ± 46 | — | P = .060, ES 0.47 |
| Blood lactate after the VO2peak test (mmol/L) | 11.5 ± 2.2 → 12.0 ± 3.0 → 13.1 ± 2.1 | 12.2 ± 1.9 → 11.7 ± 1.7 → 12.0 ± 1.7 | +14.7 ± 13.6% vs −1.3 ± 9.7% | P = .014, ES 1.4 |
Absolute VO2peak: EXP 5396 ± 617 → 5570 ± 513 mL/min; TRAD 5192 ± 776 → 5218 ± 809. Body mass unchanged (70.6 → 70.1 and 70.5 → 70.3 kg).
After the block (day 7): no group-level decline in any measure; EXP's power at 4 mmol/L was already +2.7 ± 2.5% (P = .033). Some cyclists' Wmax fell after the block; of these, one improved 8.7% and another 0.3% from pre by day 11.
Mood and legs: EXP's POMS fatigue went 49.0 ± 8.1 → 53.8 ± 10.4 (P = .052) → 45.6 ± 6.7 (P = .007 vs day 7); no other POMS change in either group. On day 11 EXP rated its legs better than TRAD (3.0 ± 1.3 vs 5.1 ± 0.8 on the 9-point scale, lower is better, P = .002).
TRAD: no measured variable improved over 11 days despite a 35% cut in total load.
Where the reporting does not add up:
The abstract and conclusion rest the VO2peak claim on a t-test of percentage change (P = .041); the group × time interaction in the ANOVA was not significant (P = .131). For Wmax both agree.
Table 3 marks EXP's day-11 knee-extension torque as different from day 7 (P < .05); the text calls the same 9.1 ± 5.4% rise from day 7 to day 11 non-significant (P = .13).
The Methods give TRAD's 11-day load cut as "~40%"; Table 2 and the Results give −35 ± 8% (EXP's "~55%" matches its −57 ± 13% within rounding). The Discussion calls the −50% and −57% figures "the last 5 days"; Table 2 labels them the last 4 days. Table 4's footnote says EXP recovered over "the next 5 d", Tables 1-2 say 4 days.
Table 1 has TRAD riding 20-40 min on day 9; the text says days 8 and 9 were easy rides at 50% of normal duration (about 1-2 h).
The Discussion and Conclusion report "tendencies" toward a larger gain in cycling economy, but the between-group test was P = .125, above the authors' own P < .10 threshold for a tendency.
The abstract gives EXP's baseline VO2peak as 77 ± 5 mL/kg/min; Table 3 gives 76.8 ± 6.9.
Mechanisms
Measured: VO2peak, Wmax and peak post-test lactate rose in EXP; gross efficiency did not change significantly in either group; the block raised high-intensity time by 192% without raising total load (TRIMP +9 ± 32%).
Authors' interpretation: the larger VO2peak rise is the likely reason power at 4 mmol/L tended to rise more; the 6-day block was short enough not to depress performance at group level, so fatigue cleared quickly and the physiological gains showed within 5 days; whether the short-interval format itself or the compressed overload-and-taper produced the effect cannot be separated, because the same 30/15 format had already beaten 5-min intervals in their earlier work.
Authors' interpretation: TRAD's lack of change is unlikely to be too little recovery (its total load fell more than EXP's); elite cyclists well into their season have little room to improve.
Inference: "overload" in this design means concentrating intensity, not adding hours; the block matched baseline total load. With no group that kept its usual training, the study cannot say whether TRAD's taper did anything, only that EXP's protocol moved the markers more.
Practical Application
Can I peak in under two weeks? In 17 elite cyclists, 6 days of daily 30/15 interval sessions then a 5-day step taper raised peak aerobic power 5% vs 1% after an 11-day taper. Small, not randomised, lab tests only.
The tested protocol, for a cyclist with about two weeks to an important race (B, one trial; the authors call it "a starting point"): days 1-6 one session a day of 3 × 9.5 min of 30 s hard / 15 s easy with 3 min between sets and less easy riding; day 7 a hard day (in the study, incremental tests to exhaustion); day 8 off; day 9 20-40 min easy; day 10 20 min easy, 2 × 5 min moderate, 3 × 1 min building towards maximum; day 11 the target.Expect heavier legs and higher fatigue at the end of the block and fresher legs by day 11 (the direction the cyclists reported: the rise in fatigue was a tendency, P = .052; by day 11 their legs were rated better than TRAD's, P = .002).
Fitness gained inside two weeks can show up: here VO2peak and peak power rose within 11 days, 5 days after the last interval session (day 7's tests to exhaustion came after it). It is one reason SYNTH-34 (revised 2026-10-26) dropped its old convention that fitness gained inside two weeks is not available for the race.
Not for runners as written: the authors warn that a daily high-intensity block in running brings more impact and injury risk (their caution, not tested).
Close races: no trial here tests two races; for a single race about two weeks out this protocol and REF-048's are the controlled compressed peaks in the base (SYNTH-76 lists multiple peaking as unstudied beyond REF-109, and this trial does not change that); REF-048 (B) found a similar 6-day block of five 30/15 sessions plus 6 easy days lifted power at VO2max 3.7% vs 0.7% with regular training.
Do not go into it overreached: the block worked at group level because performance did not fall during it; in REF-1056, athletes who became functionally overreached before a taper gained less (inference linking the two).
Limitations
Small and not randomised. 9 vs 8 cyclists, assigned by matching on VO2peak; baseline values were not significantly different, but TRAD's VO2peak and Wmax were numerically lower.
No performance trial. Wmax from an incremental test is the authors' "primary indicator of performance"; no time trial or race was measured (the authors' limitation).
No usual-training control. Both groups changed their training, so the study compares two strategies; it cannot show that either beat simply carrying on.
Two things changed at once. EXP differed from TRAD in the interval format (30/15), the number of high-intensity sessions, the timing of the load and the total load reduction (17% vs 35%); which of these mattered cannot be separated (the authors name the interval format).
Statistics. The VO2peak result depends on the test chosen (t-test on percentage change significant, ANOVA interaction not); several other outcomes are "tendencies" at P < .10; no correction across outcomes.
One group, one season. Young male cyclists from one Norwegian laboratory, tested in May-June in season; the first author's group designed and earlier tested the 30/15 format.
Reporting inconsistencies (listed under Key Results): the torque significance, the load-cut percentages, the 4 vs 5 recovery days, and TRAD's day 9.
Funding: none; no conflicts reported.
How this entry was checked and changed
Why this level
parallel trial with n = 17 (9 vs 8), not randomised, physiological tests rather than a performance trial, one laboratory; read from the full text (open access, CC BY-NC-ND 4.0; publisher's version via the Norwegian research archive)
Full text (checked 2026-10-26)
From pre to post-2, EXP achieved a larger relative improvement than TRAD in VO2peak (4.0 ± 3.7% vs 0.8 ± 1.8%, P = .041) and the 1-min peak power output (5.0 ± 3.6% vs 0.9 ± 1.5%, P = .009) (full text)
For VO2peak there was a significant effect of time (F(2,30) = 4.77, P = .016) but no significant time*group interaction (F(2,30) = 2.18, P = .131) (full text)
EXP did one daily HIT session of 3 sets of 9.5 minute of 30-second work intervals interspersed with 15-second recovery, with 3-minute recovery between sets (full text)
During the first 6 days EXP's HIT time was +192 ± 81% and total training load (TRIMP) 9 ± 32% relative to baseline (full text)
The overall decrease in total training load during the 11-day intervention was larger for TRAD than EXP (35 ± 8% vs 17 ± 21%, P < .05) (full text)
At day 11 EXP had a better perceived well-being in the legs than TRAD (3.0 ± 1.3 vs 5.1 ± 0.8, P = .002) (full text)
No funding was received for this work; no potential conflict of interest was reported (full text)