SYNTH-48 · Guide
The short time trial (prologue to 25 km, and the hill climb)
Pace from CP and W′ measured in the race position, spend W′ evenly, prime only short efforts, and treat aero and climbing-surge numbers as missing
The question it answers
What power, pacing, warm-up and position for an individual TT of roughly 5-40 min (prologue, 10-mile, 20-25 km) or an uphill TT, and which sessions and taper lead to it. Inference: for an amateur ≤8 km is roughly 8-15 min, 16 km 20-30 min, 20-25 km 30-40 min; plan from the athlete's expected time, not the distance.
Default (what to prescribe first)
1
Anchor on CP and W′, not FTP. Above CP, t = W′/(P − CP) (REF-023, B+; REF-024, B+), so the target for a planned duration t is P = CP + W′/t (REF-026, B+, which also advises against single-effort tests such as FTP). In 12 competitive men, CP and FTP matched on average but disagreed individually by +30 to −36 W (REF-040, B).
2
Measure CP from self-paced efforts in the race position. The self-paced half is inference from REF-034 and from REF-040's self-paced prediction trials, not a tested comparison; the position half is REF-201's finding. TT-position CP was significantly lower than on the hoods, W′ unchanged (REF-201, B, 7 trained cyclists, no watts in the abstract); self-paced CP was ~6% above constant-load CP (265 vs 250 W) in 10 young men (REF-034, B).
3
4
5
One or two fields on the head unit (one beat six in REF-396; "or two" is the base's own practice).
Decision
Target power by planned duration (Inference: arithmetic on P = CP + W′/t with REF-034's self-paced means, CP 265 W, W′ 18.1 kJ; a larger W′ relative to CP gives higher percentages)
| Planned duration | Target (example) | % of CP | Source |
|---|---|---|---|
| 5 min | CP + W′/300 s = 325 W | ~123% | model (REF-023, REF-026, B+) |
| 10 min | 295 W | ~111% | model |
| 15 min | 285 W | ~108% | model |
| 20 min | 280 W | ~106% | model; CP+W′ from self-paced lab trials (stationary bike, mobile power meters) predicted a 16.1-km TT at 27.5 vs actual 26.7 min, ns, r = 0.89, as a group mean (REF-040, B) |
| 30-40 min | at or a few % below CP; the model's premium is not usable | ≤100% | at CP, 7 trained men lasted 20:01-40:37 min with HR, V̇O₂ and lactate still rising (REF-750, B); REF-168 (B+) cites "approximately 20-30 min" only as the premise of a view it disputes; its body reports time to the limit at CP as highly variable (~15-40 min, occasionally ~60 min) and calls "how long" CP lasts an ill-conceived question (corrected 2026-10-12); the model predicts nothing about duration at or below CP (REF-751, B+) |
Only FTP20 known: median time to exhaustion at FTP20 was 35 min (recreational), 42 (trained), 47 (well-trained), 51 (professional), n=87, wide spread (REF-747, B). Inference: a recreational rider should start a 30-40 min TT below FTP20.
Keep ±3-5% around any "at CP" target (REF-751's practical rule): a target set exactly on the estimate can land either side of the boundary.
Ceiling, not target: male pro TTs ~400-420 W short (near OBLA), ~370-390 W long (near LT) (REF-174, B+, 2001 narrative, from race HR). In 18 world-class pros, race HR was 100 ± 3% of HR at OBLA in prologues (<10 km), 95 ± 7% in short TTs (<40 km), 89 ± 5% in long ones; the authors put events up to ~30 min at OBLA, longer ones at LT (REF-792, B, Tier 5, HR only). Inference: an amateur's 20-25 km falls in the 30-min-plus band (SYNTH-62).
Pacing
| Course / event | Prescription | Source |
|---|---|---|
| Flat, steady wind, > 2 min | Even power at target | In events > 2 min times may improve with more even pacing; all-out starts pay only ≤ 30 s (REF-010, B+); even pacing optimal > 4 km in steady conditions (REF-490, B+) — both narrative |
| ~4-7 min | Prime, ~12 s all-out to speed, then target | 9 men, ~6.7-min 4-km TT: 402 vs 411 s only with severe-intensity priming AND a 12-s all-out start (not tested after the below-CP priming of Default 4); either alone did nothing (REF-279, B). ~98-s trials: priming helped, the fast start added nothing (REF-273, B+). Pursuit model: all-out start, constant after ~12 s (REF-127, B, simulation) |
| 20-40 min | No fast start | After a sprint warm-up, riders started a 20-min TT harder (min 0-3) with no gain (REF-278, B+, 15 trained men); self-paced riders opened 14% above their mean over the first 1.61 km of a 16.1-km TT, then faded (REF-849, B, 7 men) |
| Rolling, windy | More power on rises and into headwinds, less on dips and tailwinds, same mean | Rule in REF-490 (B+). Riders: gradient-matched power was 12 ± 8 s (2.9%) faster than attempted constant power on an undulating 4-km field course (REF-848, B+, n=20); +5% into a simulated headwind half saved ~12 s over 16.1 km vs self-pacing, even power ~10 s (REF-849, B, 7 men, lab). Models (C): ±10% saved 126 s on a hilly 10 km (REF-795); ±15% in step with 6% gradients saved 188.89 s over 40 km at 200 W, out of step cost up to 43.90 s (REF-847) |
| Technical | Carry speed through corners | Time lost in slow sections was not won back on straights (40-km simulation, REF-125, C) |
| Head unit | One field beat six; "one or two" is the base's own practice | 30-min TT: 287.9 W with one field vs 227.99 W with six, 10 trained cyclists-triathletes (REF-396, B); no feedback did not change 20-km TT time significantly in 20 experienced cyclists (REF-395, B+) |
Surges cost more than their mean: in 15 elite cyclists, a 15% fluctuation every 2 min at 70% and 95% of LT raised VO₂ and lactate vs constant power (HR higher in the abstract, no main effect in the Results (corrected 2026-10-12)), with similar whole-bout RPE at 95% LT (REF-252, B+; submaximal, not a TT). Inference: keep the uphill surplus moderate.
Indoor targets transfer as means: 20-min TT 280 W indoors vs 284 W outdoors, outdoor variability twice indoor (REF-253, B+, 20 competitive men).
Judge a change against noise: repeat 20-km ergometer TTs varied 1.0-4.0% (typical error), the first and last km 6.6-6.8%, with starts blunting on repeats (REF-851, B, 17 well-trained men).
Warm-up and priming by planned duration
| Duration | What the trials support |
|---|---|
| ~3-7 min | Moderate or heavy priming ending 10 min before a 7-min trial: mean power ~2-3% higher (330 → 338/339 W); repeated all-out sprints did not help (324 W) (REF-270, B, 12 well-trained cyclists). 15-min progressive warm-up (5 min each at 70/80/90% VT): 3-km TT 2.8% faster than none; 3 min at RCP added nothing (REF-274, B, n=8). 5 × 10 s all-out priming: −5.8% vs self-selected; 10-s bouts at 100-150% of peak aerobic power: no difference (REF-272, B, 10 trained men) |
| ~8-15 min | Untested. Inference: between the 7-min trials that gained and the 20-min trial that did not; use the short-effort protocol |
| ~20 min+ | 10-min standard or 5-min + 3 × 10-s sprint warm-up: 20-min TT no better than no warm-up (REF-278, B+, 15 trained men, VO₂peak 66) |
| Wait at the start | Heated pants over a 30-min wait: 30-s sprint peak power +9.6%; insulation alone no effect (REF-745, B+, 11 trained men; no TT outcome) |
| Heat | Limit whole-body temperature rise and cool during the warm-up (REF-269, B+, expert review, no minutes) |
Not from a study; the base's own practice: warm up anyway before 20-40 min TTs, for readiness.
Aero position vs power. Test and set zones (HR included) on the TT bars and, as the base's own practice, train race-pace work there: TT-position CP is lower, the power gap grew with trial length (3 → 12 min), and in the 12-min trial HR was higher at lower power (REF-201, B, n=7).
A lower torso costs power: every maximal value fell as 19 well-trained men lowered it, flat-back 0° most (REF-793, B+); 20-min TT power was higher at a 24° torso than at 12° or 0° (REF-845, B, n=12, no drag measured). Position work cut drag 14% in five pros, pedalling drag area 31% above static (REF-794, B, relative).
A model on 19 TT cyclists' data: aero losses outweigh the power cost only above 46 km/h, upright is better below 30 km/h, a flat torso is never optimal (REF-844, C). No absolute CdA, watts saved or ridden net time in the base; the road power model itself is validated (R² = .97, SE 2.7 W; REF-846, B). SYNTH-17's "aero first on flat TTs" is house practice; Inference from REF-844: the slower the rider, the less a lower torso pays (SYNTH-62).
Race-day aids
Caffeine 3-6 mg/kg: TT +3.2 ± 4.3%, range −0.3 to +17.3% across 21 studies (REF-083, A). Rehearse it.
Sodium bicarbonate 0.3 g/kg, 60-180 min before, for 30 s-12 min efforts (REF-354, A): prologues and short climbs only; GI risk, rehearse.
No in-event fuel below ~40 min (Not from a study; the base's own practice). CHO mouth rinse: mean power SMD 0.25, completion time not improved (REF-483, A-). In 12 male amateur cyclists a maltodextrin or fructose rinse gave no more power in a 15-min time trial than placebo or no rinse (p = 0.989; REF-957, B).
How to progress / periodise
VO₂max shape → [SYNTH-33]. 30/15s raised 20-min power +4.7% vs −1.4% for effort-matched long intervals in 18 elite cyclists over 3 weeks (REF-045, B).
Threshold on CP → [SYNTH-27] (thin evidence, by its own account). One-week moderate- and high-intensity blocks both raised 15-min power (+4.9% vs +2.8%, ns) in 22 well-trained cyclists (REF-421, B+).
Speed endurance → [SYNTH-46]; with an aerobic base it also helped 40-km cycling (REF-695, B, narrative).
Low cadence: intervals ~20% below preferred cadence improved a preloaded 15-min TT more than high-cadence intervals (16 trained men, REF-259, B+); 50-70 rpm SIT/HIIT raised maximal aerobic and VT2 power (24 well-trained women, REF-084, B+); 60-rpm uphill intervals at RCP raised uphill 20-min TT power 4.4%, but no-interval controls gained 4.0% (REF-852, B, 18 men).
Heavy strength → [SYNTH-43]: TT/time-to-exhaustion ES 0.46, VO₂max unchanged (REF-743, A, low certainty).
Specificity: Not from a study; the base's own practice: in the last 6-8 weeks, race-duration efforts at target on the TT bars and one full rehearsal of warm-up and supplements.
Taper → [SYNTH-34]: cut volume, keep intensity and frequency.
Hill-climb TT
Power-to-weight: in male pros uphill ability tracks power/mass¹, level ability power/mass^0.32 (REF-174, B+, narrative); in 18 elite amateurs a 6.7-km, 6% climb tracked Wmax per kg (r = −0.66), a 14-km flat TT absolute Wmax (r = −0.90) (REF-850, B, cross-sectional). Targets use the same CP + W′/t rule from a CP tested upright (REF-201's entry: test in the position the event is ridden).
Pro uphill TTs sat near threshold heart rate: 101 ± 5% of HR at the lactate threshold, at the lowest share of HRmax of any TT type (78 ± 3%) (REF-792, B, Tier 5). A 9.2-km, 7.1% pass in 14 elite riders: 302 W (4.2 W/kg), ~90% of HRmax and VO₂max, lactate 10.87 mmol/L at the top (REF-853, B). The two samples disagree on %HRmax: compare terrain by lactate markers, not %HRmax.
Mass loss is not free: 4 weeks of time-restricted eating at estimated needs cut weight 2% in the TRE group of 16 elite U23 cyclists (two groups); performance did not differ from controls, PPO/BW rose, free testosterone and IGF-1 fell (REF-520, B). Keep energy availability near 40-45 kcal·kg FFM⁻¹·day⁻¹; body mass "as a metric rather than a goal" (REF-748, A). No entry shows a weight cut improving a climbing TT.
Gradient pacing: constant grade → even power (Inference from REF-010, REF-490). Variable grade → more on steep pitches, less on flatter ones; modelled gain grows with climb length and rider mass (REF-126, C, no numbers); in riders, shown only on a 4-km undulating course, not on a climb (REF-848, B+). Aero matters only on fast false flats (house practice, SYNTH-17).
Cadence: uphill and upright, freely chosen cadence was ~7 rpm lower (82 vs 89) and matched the lactate-optimal cadence; maximal power 2.0% higher (REF-262, B, 7 cyclists, treadmill, one gradient).
When NOT to use it / red flags
What the evidence does and does not show
Shows: CP/W′ model and CP as a boundary (B+); lab-derived CP+W′ predicting a ~27-min road TT as a group mean (B); priming helping 1.5-7-min efforts, not a 20-min TT (B/B+); TT position lowering CP (B); caffeine and bicarbonate (A); gradient-matched power beating constant power on a short field course (B+).
Does not show: individual prediction error; % of CP held in amateur or women's racing; surge sizes; net aero magnitudes (relative drag and a C model only); % of CP from racing power (pro race data are heart rate, REF-792).
Population: small samples, men only or sex not stated, except one women-only trial (REF-084); transfer to women is assumed.
Evidence gaps
Hill-climb TT pacing: REF-853 describes intensity in one elite group; no pacing trial on a climb.
Priming for 8-15 min and in women; weight loss → climbing performance; heat/pre-cooling for short TTs: not in the base or not assembled here.
How this entry was checked and changed
Why this level
(CP/W′ model from B+ narrative reviews; pacing, priming and position from small B/B+ crossovers of 7-20 riders; terrain pacing from one short B+ field trial and C models)
Content correction 2026-09-29
Revision 2026-09-29: folded in REF-792 (pro TT heart rate by length and terrain), REF-793 (lower torso costs power), REF-794 (relative drag in pros) and REF-795 (variable-power model, C); routing to SYNTH-62 for 40-80 km added. Level unchanged.
Content correction 2026-09-27
Revision 2026-09-29: rider data for terrain/wind pacing (REF-848 B+, REF-849 B; REF-847 C model) and TT noise (REF-851); aero paragraph with REF-845, REF-844 (C: aero pays only above 46 km/h), REF-846; hill climb with REF-850, REF-853, REF-852. Gaps narrowed. For length: ceiling, CP-test, population and cross-reference lines tightened and the earlier same-day note condensed. Default and level (B) unchanged.
Revision 2026-10-01: folded REF-957 (Mendonça 2026, B, crossover, 12 male amateur cyclists) next to the mouth-rinse line: a null result in a 15-min time trial, alongside the small pooled gain in mean power of REF-483. No default changed; level unchanged.
Verification 2026-10-01: all 58 cited REF entries opened in the snapshot; every level/design/n beside a refId matches the entry head block; the target-table arithmetic (CP 265 W, W′ 18.1 kJ → 325/295/285/280 W) is correct and labelled inference.
Verification 2026-10-01: numbers verified against entries for pacing (REF-279, 273, 848, 849, 795, 847), warm-up (REF-270, 274, 272, 278, 745), position (REF-793, 845, 794, 844, 846), aids (REF-083, 354, 483, 957) and hill climb (REF-850, 792, 853, 262) — no NUMBER finding.
Verification 2026-10-01: source hygiene, not flagged as findings: REF-125 is a re-attributed composite (original citation was a phantom; findings now rest on Zignoli 2020, DOI present, level C); REF-354's heading still names Hadzic 2019 / Grgic 2020 while the text is the Grgic 2021 ISSN stand (DOI present); REF-274 has a PMID only.
Verification 2026-10-01: no contradiction with SYNTH-62 (which starts where this card stops and sets ~1 h at 90-95% CP as practice). Level B is honest — no LEVEL-OVERALL finding.