SYNTH-46 · Guide
Sprint and anaerobic work
The format decides the quality: short sprints on full rest buy peak power, 30-s all-out sprints buy sprint and aerobic gains at tiny volume, supra-VO2max speed endurance buys 20 s–60 min event performance without moving VO2max — and outside burst-decided racing it is a small standing dose, not a block
The question it answers
Default (what to prescribe first)
Long-event athlete (gran fondo, marathon, ultra), or any athlete in a base or transition phase: one sprint session a week inside an easy ride — a transfer: tested only in elite male cyclists during a 3-week transition, not in long-event amateurs, runners or a base phase.The tested dose (REF-165, B, quasi-experimental, n=11 elite male cyclists, 5 vs 6): 3 sets × 3 × 30 s maximal, ~4 min between sprints, ~20 min between sets, inside a 3-week low-intensity transition with a 64% load cut — six weeks into the next preparation 20-min power was +7.3% vs −1.3% for low-intensity only, with no change in VO2max, gross efficiency or power at 4 mmol/L.Not from a study; the base's own practice: a shorter 4-6 × 8-10 s version is an acceptable substitute when 30-s efforts do not fit the ride.
Burst-decided racer (XCO, cyclocross, criterium, track endurance, 1500 m-5 km): Not from a study; the base's own practice: a 4-8 week block of 1-2 sessions a week of the format matched to the race (table below), counted inside the week's quality budget (SYNTH-26) alongside, not instead of, VO2max work (SYNTH-33).
Decision — format → quality
| Format | What it builds | What it does to VO2max / TT | Source |
|---|---|---|---|
| Neuromuscular short sprints on the bike (REF-191's entry gives no sprint length or rest; ≤10 s on long rest is the base's own practice) | Sprint power. In 28 competitive cyclists, after a shared 4-week strength block, 6 weeks of on-bike short-sprint training beat heavy strength on 6-s sprint (+4.7% vs +1.1%), 30-s sprint (+3.7% vs +1.3%) and a 55-m sprint on their own bike (+4.3% vs +0.2%) | No group difference in 5-min power, VO2max, power at 4 mmol/L or efficiency; 1RM fell 3.9% vs +9.3% with lifting | REF-191 (B+, RCT) |
| Same, as 10-s all-out on 4 or 2 min rest | Wingate peak power +8.5% (10:4), +4.2% (10:2) | VO2max +9.2% and 5-km TT +3.5% with 4-min rest; with 2-min rest the TT improved (+3.0%) but VO2max did not | REF-050 (B+, 48 participants of unstated training status, 2 weeks) |
| Sprint intervals (SIT), 30 s all-out (4-min rest in REF-050; 90-s rest in REF-742's 8-month trial) | Wingate peak/mean power +9.5%/+12.1% (30:4) | VO2max +9.3%, 5-km TT +5.2% (REF-050); aerobic capacity ~+8%, not different from endurance training, in healthy young non-athletes (REF-696, A-); in XCO riders SIT added for 8 months gave VO2max +6%, MAP +5%, sprint power +4%, and over 6 weeks SIT ≈ HIIT (race time −3% for SIT vs −5% for HIIT) | REF-050, REF-696, REF-742 (A review; single trials from its Table 3) |
| Speed endurance, above the VO2max intensity, bouts of 5 s-2 min — "production" (10-40 s, long rest, e.g. 30 s / 3 min) or "maintenance" (5-120 s, short rest, e.g. 1 min / 1 min; sub-types from the entry's full text) | Performance in 20-60 s, 1-10 min and 10-60 min events in trained athletes, even with markedly reduced volume, via economy and ion/pH handling | No change in VO2max (REF-697); with a normal aerobic base it also helped 40 km cycling and 10 km running (REF-695) | REF-697 (B), REF-695 (B), narrative reviews |
| Repeated sprint, ≤10 s on ≤60 s rest | Repeated-sprint ability (ES −0.62), sprint speed, jump, intermittent running (ES −0.61), 2-8 trials per outcome | The endurance outcome is intermittent running, not VO2max | REF-693 (B+, meta-analysis incl. non-controlled trials; trained, mostly team-sport participants — a transfer to endurance athletes) |
| Short supra-MAP repeats, 20 s @ 115% MAP / 40 s @ 50% MAP, 60-90 reps, 2×/week, 4 weeks | Not measured as such ("repeatable supra-maximal power" is the entry's practical reading) | TT time −4%, MAP +8%, power at 4 mmol/L +13% in XCO riders | REF-742 (A review; this is one trial from its Table 3) |
W′ / anaerobic capacity specifically: in the one entry that trains and measures it, 12 interval sessions over 4 weeks raised CP 230 → 255 W while W′ did not change significantly (17.2 → 15.5 kJ, P = 0.08) (REF-871, B, 9 habitually active people, uncontrolled). REF-743 (A, 17 trials, 262 cyclists) found heavy strength raised anaerobic POWER (ES 0.56) but not anaerobic CAPACITY, and REF-697 reports no W′ outcome.
So promise better short-effort performance, not a bigger W′ (inference).
Decision — who needs it
| Athlete / event | What to do | Source |
|---|---|---|
| XCO / MTB | Two pillars: aerobic ceiling AND repeatable supra-maximal power. The race is ~80-100 min, ~a quarter of the time above MAP, elite surges of ~3-10 s at ~130-140% MAP, ~15-20 per lap (REF-742); in 16 national/international riders most above-MAP periods lasted 5-30 s (REF-741) | REF-742 (A, 53 studies, mostly elite men); REF-741 (B) |
| Cyclocross | 93.6% of race time above the heart rate of 4 mmol/L, post-race lactate 8.3 mmol/L, no low-intensity time (n=8, HR and lactate only, no power). Inference: use the XCO burst formats on a threshold floor | REF-740 (C+) |
| Track endurance (pursuit) | Efforts beyond ~80 s are majority-aerobic (50/50 point 78.6 s, no running/cycling difference) — inference: anaerobic work is a slice, not the diet; a world-record team-pursuit squad trained only 5.3-7.7% above threshold, pyramidal | REF-460 (A-); REF-186 (B, 5 elite men) |
| Road / criterium finish | In 6 h of simulated racing, sprint peak power fell ~7% in the first 2 h and then held, while 5-km TT power kept falling (12 national elite men) — inference: a sprint finish depends more on fresh sprint power plus arriving with it than on sprint endurance. Demand: pro breakaways came after 5-10 min of 5-15 s surges at ~9.5-14 W/kg, then ~450-500 W for 30 s-5 min; flat women's World Cup races held more time above 500 W, hilly ones more 180-300 s power | REF-474 (B); REF-840 (C+, sample not stated); REF-839 (B, 15 women) |
| 1500 m-10 km runner | Speed-endurance periods help 1-10 min and 10-60 min events in trained athletes. 6-12 × 30-s sprint runs 3-4×/week with 25% less training for 6-9 weeks: 10 km 37.3 → 36.3 min, VO2max unchanged (REF-869); 7 weeks of 10-20-30 replacing all training (54% less distance): 5 km −48 s, VO2max +4% (REF-870). Both swapped volume for a short block | REF-697, REF-695 narrative reviews; REF-869 B (12 trained runners vs 5 controls); REF-870 B (18 moderately trained, women and men) |
| Gran fondo, marathon, ultra | Maintenance dose only (Default). Faster marathoners differed by easy volume while Z2/Z3 time was similar across levels, and SET evidence beyond 60 min is sparse | REF-197 (B+, 119,452 runners, cross-sectional); REF-697 |
| Time-crunched / health | SIT is an efficient aerobic stimulus in non-athletes — see SYNTH-33 | REF-696 (Tier 0-1 transfer) |
Session design — rest, sets, freshness
Rest for true max efforts is long. After one 30-s all-out sprint, muscle PCr was 65% of rest at 1.5 min and 85.5% at 6 min, and sprint-2 power was still below control at 6 min (REF-480, B, 14 men). 30 s / 90 s is therefore a capacity or tolerance session, not repeated peak power (inference). Above CP, recovery intensity and length decide how much W′ returns, and it comes back slower and more variably between athletes than it is spent (REF-809, B+, narrative).
Rest rules and national-team formats (Olympiatoppen guideline, level C — expert practice, not a study): rest at least 100-200% of the rep (sometimes up to 10 min) for I-6 reps of ~30 s-2 min, 5-15 min for I-7 reps of 15-60 s and 3-15 min for I-8 reps of 3-15 s; Norwegian cycling models: I-6 6 × 2 min / 5 min, I-7 1-3 sets of 3 × 30 s / 4 min, I-8 6-8 × 10 s / 3 min, maximal or near-maximal, often inside a long easy ride (REF-1077, C).Not tested against the ratios above.
Build-up: REF-050 started at 4 bouts and added one every second session up to 6, 3×/week for 2 weeks. Not from a study; the base's own practice: 1-2 sessions a week is the sustainable rate beyond two weeks.
Warm-up short, one opener. In track cyclists a ≥50-min warm-up with 4 sprints cut Wingate peak power to 1303 W vs 1390 W after 15 min to 70% HRmax with one sprint (REF-276, B).
Sprint fresh. In 14 professional cyclists, 5 × 8 min at 105-110% CP lowered 1-s and 15-s power more than ~2000 kJ of low-intensity riding (REF-193, B). Inference: put sprints early in the session or on a day not following hard intervals, never after the VO2max reps.
Do not use sprints as the opener before a severe effort: sprint priming gave no gain (324 vs 330 W) and started the 7-min trial at 5.9 mmol/L (REF-270, B, n=12).
Recovery between sessions is individual. After 3 × 30-s Wingates, fast-typology men's knee-extension torque was still not recovered at 5 h while slow-typology men recovered in 20 min (REF-307, B, n=20 men). Not from a study; the base's own practice: SIT and speed endurance count as quality days; a ≤10-s set inside an easy ride does not.
How to progress / periodise
Season: maintenance set in transition (REF-165; the base phase is a transfer); race-format block in the pre-competition phase (house practice); in the taper keep intensity and frequency and cut volume — the time-trial effect appeared only where they were kept (REF-3492, A, 14 studies). Inference in REF-695's entry, not tested: speed-endurance sessions are a candidate for holding fitness when volume drops.
Threshold blocks may cost a little sprint (trend, not significant): after a one-week moderate-interval block 10-s power was −1.5% vs +1.5% after a HIT block (p = 0.08) (REF-421, B+, 22 well-trained cyclists). Inference: sprint-dependent racers keep a few short sprints through threshold blocks.
On-bike sprinting and the gym are not substitutes (REF-191); heavy strength adds anaerobic power (REF-743), and after 25 weeks of lifting, stopping for 8 in-season weeks cut squat jump, Wmax and Wingate mean power relative to controls (REF-744, B, 7 per group, elite men). The lifting itself is SYNTH-43.
Masters
Keep the explosive element; SYNTH-35 already prescribes the REF-165 set. Inference in REF-599's entry (B–C, mechanistic review): ageing tendon adapts slower, so progress hard hills and sprints more gradually than in the young; applying this first to running sprints, which load tendon hardest, is our inference.
Fast intent matters for power: in adults ≥50, high-velocity resistance training raised leg power about three times as much as slow lifting (Δ 0.62 vs 0.20) (REF-597, A+, mostly untrained older adults — a transfer).
Explosive force falls with age even in sprint-trained men: across 91 sprinters aged 18-84, type II fibre area, maximal force and rate of force development declined (REF-808, B, cross-sectional); strength as the defence is that entry's inference.
No sprint-training trial in masters endurance athletes is in this base; widen spacing on what recovery shows (REF-307), not on age alone.
When NOT to use it / red flags
No consistent aerobic base: the lever is regularity (SYNTH-42 gate); sprint work is not the first lever except as a time-crunched aerobic stimulus.
Short-rest 30-s reps sold as "peak power" — PCr is far from restored at 1.5 min (REF-480, B, 14 men).
Pain in a tendon or a return from a running injury: not from a study; the base's own practice — no maximal running sprints until easy running is symptom-free.
What the evidence does and does not show
Does not show: which format is best for trained ROAD cyclists (the head-to-heads are XCO); running beyond two small controlled trials (REF-869, REF-870); women (REF-742 notes female data are scarce); masters; the best weekly frequency or block length; any W′ gain. REF-050's participants' training status is not stated and it lasted 2 weeks; REF-165 is 11 non-randomised elite men.
Evidence gaps
A W′ training study in trained cyclists; REF-871 is 9 habitually active people.
Sprint or SIT trials in masters endurance athletes, and in trained women.
SIT vs HIIT vs speed endurance head-to-head in trained road cyclists; weekly placement of sprint sessions.
How this entry was checked and changed
Why this level
for the format→quality table, C+ for the Default (one n=11 quasi-experiment plus house practice) (one A systematic review of XCO racing and training, an A- meta-analysis in healthy young non-athletes, small B+/B cycling trials and two B narrative reviews; doses and placement are mostly house practice)
Content correction 2026-09-27
Revision 2026-09-29: folded in REF-869 and REF-870 (B) into the 1500 m-10 km runner row (both swapped volume for a short block; the 'add, do not swap' default for long events is unchanged); REF-871 and REF-809 into the W′ paragraph and rest rule; REF-839/840/841 road-race demand; REF-808 into Masters; REF-887 and REF-804 as a red flag. Gaps, Tier line and 'does not show' updated. Level unchanged (B).
Verification 2026-10-01: all 36 cited REF entries opened in the snapshot; every number in the format table, W′ paragraph, session-design and red-flag lines matches its entry, and all levels match except REF-599 (B–C).
Verification 2026-10-01: no cited entry is unverified or without DOI/PMID. REF-597 is a three-paper composite (A+); the SYNTH uses only the Straight 2016 figures, which the entry sources to its DOI.
Verification 2026-10-01: the 'add, do not swap' rule is the card's organising claim and is practice: the only volume evidence cited (REF-697, REF-869, REF-870) improved performance with volume cut.
2026-10-02: aligned with the rewritten SYNTH-42 (missing stimulus, one lever per block, level C+): "the limiter is regularity" -> "the lever is regularity" (the consistency gate); cross-reference label "limiter" -> "lever".
Revision 2026-10-04: added Olympiatoppen's Intensity Scale v2 (2024) REF-1077 as practice lines labelled "(Olympiatoppen guideline, level C — expert practice, not a study)"; no default, decision row or level changed.