Road races are decided late, by the power a rider still has in the duration the course asks for
Build the engine, train that duration on top of fatigue, let the races be the intensity in season; criterium rows are transfer, not data
The question it answers
What the race asks for (as the base states it)
| Measure | Road race | Criterium |
|---|---|---|
| Demand by course | In professional men, more climbing means higher intensity and load and higher power over efforts >10 min; flat and semi-mountainous stages bring higher maximal power over efforts <2 min; single-day races carry a higher daily intensity and load than stages (REF-788, B+, narrative review, no numbers in the abstract) | No entry. Inference: the flat, short-effort end of that axis |
| Average vs peaks | Pro mass-start stages averaged ~210 W (flat) to ~270 W (mountain) but held 30-100 min at or above LT and 5-20 min at or above OBLA (REF-174, B+, 2001 narrative review, HR-based) | No entry |
| Finale | In 17 professional races six male sprinters finished top-5, power rose from 316 W over the last 10 min to 487 W in the last minute before the sprint; the final 5 min held more short hard efforts than the 5 min before; peak sprint 17.4 W/kg (REF-791, B, Tier 4-5) | Inference: the same pattern; no crit file |
| Women | In 27 top-20 World Cup finishes by 15 nationally ranked women, flat races averaged 192 W (3.3 W/kg) with more time above 500 W; hilly races 169 W (3.0 W/kg) with more time between lactate and anaerobic threshold and higher 180-300 s power (REF-839, B). Pro women's races were shorter than men's but spent largely more time in HR zones 4-5 (d = 1.38-1.55) and carried more load per km (TRIMP·km⁻¹ d = 2.31); total TSS only slightly lower (d = 0.53) (REF-841, B, 10 women vs 20 men, effect sizes only) | No entry |
| Breakaway | In pro race data, 5-10 min of repeated 5-15 s surges at ~9.5-14 W/kg preceded a successful break, then ~450-500 W for 30 s-5 min until the chase eased (REF-840, C+, sample not stated) | No entry |
| Terrain scaling | Level-ground ability tracks power/mass^0.32, uphill power/mass^1 (REF-174) | Inference: flat, so absolute power weighs more |
| Surges | Variable power (±15% every 2 min) cost more VO2 and lactate than constant power at the same mean (HR: no main effect in the Results (corrected 2026-10-12)), with similar whole-bout RPE at 95% LT (15 elite cyclists, REF-252, B+) | Nearest measured pattern is off-road: ~3-10 s at ~130-140% MAP, ~15-20 per lap in XCO (REF-742, A, systematic review) |
Durability is what separates riders of similar fresh power
Default (what to prescribe first)
Not from a study; the base's own practice, for a Tier 1-3 amateur 8-16 weeks from a target road race: two quality sessions a week (one engine session (SYNTH-33), one race-specific session from the table), one long ride with hard efforts placed late, heavy lifting per SYNTH-43, a short sprint set inside an easy ride, and a fatigued test at the start and end of the block — fresh 5- and 20-min, then 1,000 kJ at 70-80% of fresh 20-min power and repeat (REF-427, REF-790 protocols).
Crit-only racer: see the last row below.
Decision — rider profile → quality to build
Read the profile from target courses and own fresh-vs-fatigued numbers — a hypothesis, never a label (SYNTH-42). The profile → quality mapping in the table is coaching reasoning, not from a study; the refIds support the sessions, not the typing.
| Profile (course) | Quality to build | Sessions and source |
|---|---|---|
| Any profile, Tier 1-2 or an inconsistent base | The engine and regularity first | SYNTH-42 gate (house default); engine work SYNTH-33 |
| Sprinter (flat races, bunch finish, crits) | 10-s to 1-min power that survives a hard finale | On-bike short sprints beat heavy strength for sprint power in 28 competitive cyclists, including 6-s sprint after prolonged cycling (+6.1% vs +1.8%) (REF-191, B+); heavy lifting raised anaerobic power (ES 0.56) in a meta-analysis of 17 trials, mostly trained men (REF-743, A). Sprinting at the end of a hard session is practice built on REF-791 and REF-193 |
| Puncheur (short climbs, attacks, 1-5 min) | Repeatable efforts above threshold | 30/15 short intervals raised 20-min power +4.7% vs −1.4% for effort-matched 4 × 5 min in 18 elite cyclists (REF-045, B); W′ recovers bi-exponentially — only about half returns fast (τ 21.5 s) and a second exhaustive bout slows the rest (10 trained cyclists, REF-131, B). Transfer from XCO: 20 s @ 115% / 40 s @ 50% MAP, 2×/week for 4 weeks, gave TT time −4% and MAP +8% (one training study tabulated in REF-742, an A systematic review; the single trial is not A) |
| Climber (hilly road races, climbs >10 min) | Power/mass and 5-20-min power late in the race | Uphill scales with power/mass (REF-174); pro climbers' successful seasons kept 20- and 5-min power after work (REF-789); the 20-min drop was smaller in more successful amateurs (REF-790). VO2max shape SYNTH-33, threshold SYNTH-27, climbs late in the long ride (practice) |
| Rouleur (flat or rolling, breakaways, long pulls) | Absolute threshold power plus durability | Flat ability tracks power/mass^0.32 (REF-174); a one-week moderate-intensity block raised power at 4 mmol/L more than a HIT block (+4.5% vs +2.1%) in 22 well-trained cyclists (REF-421, B+); durability per SYNTH-32 |
| Criterium-only racer (short circuits) | Repeated accelerations and the sprint; durability less | Practice: SYNTH-32's matrix gives crit racers about one long ride a month; repeatability formats per SYNTH-46; all transfer, no crit trial |
Key sessions
How to progress / periodise, and racing weeks
Pacing and tactics — practice, labelled
No entry studies positioning or attack timing; one brief report describes successful pro breakaways (REF-840, above). Shielding in a group cut cyclists' external power by over 30% at racing speeds in an early measurement (REF-842, C+, 1979).
Continuous drafting versus swapping the lead every 500 m lowered VO2 (49.9 vs 59.8 mL/kg/min), HR (155 vs 173) and lactate (3.5 vs 6.3 mmol/L) in 10 national-level male triathletes (REF-657, B; transfer from draft-legal triathlon); late TT-power loss tracked accumulated time above LT (REF-474).
Not from a study; the base's own practice: stay in the draft, spend efforts only where the course decides, move up before known selection points rather than chase after them; a sprinter rides the first two-thirds as cheaply as possible.
Fuelling
SYNTH-02 sets grams per hour by duration (house practice): under 90 min — most crits — 30-60 g/h from about 1 h of racing, small amounts or a mouth rinse for shorter races (that row is consensus, REF-748); 90 min-3 h 60-90 g/h for amateurs; 3-6 h 80-100 g/h for amateurs.
The UCI statement indexes intake to the day the same way — a 90-min session a 30-60 g/h day, a 4-h ride a 90 g/h day (REF-748, A, elite-oriented consensus).
Carbohydrate versus none after 150 min halved the fall in VT1 power (−3% vs −6%) but did not significantly protect a 5-min finish effort (REF-181, B, 12 trained cyclists/triathletes) — fuel first (SYNTH-32), but it is not a durability plan.
When NOT to use it / red flags
What the evidence does and does not show
It shows that the course sets the power durations that matter (B+, pro men), that the finale of a sprint race is a rising, surging aerobic-plus-anaerobic effort (B, six pro men), that women's races are shorter but harder per km and terrain-shaped (B), and that holding the specialty duration after work separates better pro seasons, and that 20-min (not 5-min) power after work separated better amateur racers (B, observational).
It does not describe race-winning efforts (REF-788 says so), does not show that training durability improves race results (the REF-790 entry flags this as untested), nor that any session improves a mass-start result — every training outcome is a lab or field test. The profile table is coaching reasoning.