The warm-up
A progressive block that ends warm but not tired, one below-threshold priming bout before short hard work, a few sub-maximal openers, never a set of all-out sprints, and a short gap to the start; the minutes, zones and heart-rate checks are practice, not evidence
The question it answers
What warm-up to put before a given session or race — minutes, structure, %FTP / pace / RPE, the heart rate it should reach by its end, the gap to the start — and how the trainer autopilot should correct a warm-up that turns out too easy or too hard.
Default (what to prescribe first)
Warm up before hard work. A meta-analysis of 32 studies found warm-up improved 79% of performance criteria with little evidence of harm (REF-263, A; mixed activities, so it does not say which warm-up).
Progressive, below threshold, then specific. A 15-min warm-up of 5 min each at 70/80/90% of the ventilatory threshold made a ~4.5-min 3-km TT 2.8% faster than none; 3 min more at the respiratory compensation point added nothing (REF-274, B, 8 well-trained cyclists).
Before short hard efforts, one priming bout below critical power, finished about 10 min before. Moderate or heavy priming raised 7-min trial power ~2-3% (330 → 338/339 W); repeated all-out sprints did not (324 W) (REF-270, B, 12 well-trained cyclists).
Openers sub-maximal, never an all-out set. 5 × 10 s at 100-150% of peak aerobic power matched a self-selected warm-up; 5 × 10 s all-out cost 5.8% of mean power on a 3-km TT (REF-272, B, 10 trained men).
Not from a study; the base's own practice (the cycling default the rest of this entry adjusts): 8-10 min ramp 50 → 70% FTP, 4 min steady at 75-80% FTP, the session-specific part, then 3-5 min at 50-55% FTP; 15-20 min in all.
Decision — by session and race
HR targets are house defaults for the END of the steady part, as % of a measured LTHR (%HRmax equivalents assume LTHR ≈ 90% of a measured HRmax; use the athlete's own anchors, SYNTH-38).
| Session / race | Warm-up (all minutes and %FTP are practice unless a refId is attached) | HR reached (house default) | Source and how sure |
|---|---|---|---|
| Easy / endurance | 5-10 min rising from ~50% FTP into the ride; no openers | none; the ride's own band | Practice only |
| Tempo / threshold | Default ramp + 4 min at 80% FTP; optionally 2-3 min at ~90% FTP or 2 × 20 s brisk; 3-5 min easy | 80-90% LTHR (≈72-81% HRmax) | A 20-min TT was no better after a 10-min or a 5-min + 3 × 10-s warm-up than after none (REF-278, B+, 15 trained men): comfort and pacing, not measured power |
| VO2max intervals | Default ramp + steady + one priming bout of 3-5 min at 88-95% FTP ending 6-10 min before rep 1; 1-2 × 15 s at rep power; easy spin | 85-90% LTHR before the bout; ~95-100% LTHR at its end | Priming ended 10 min before (REF-270, B) or a 6-min heavy bout 6 min before (REF-271, B, 9 men: raised primary VO2 amplitude, cut the slow component; 40 min passive warming did nothing). Single bouts, not sessions — inference |
| Sprints / SIT / anaerobic | ~15 min to ~70% HRmax, one sprint opener, then the set | ~70% HRmax | A ≥50-min warm-up (20 min rising 60 → 95% HRmax + 4 sprints) cut Wingate peak power to 1303 vs 1390 W after 15 min rising 60 → 70% HRmax + one sprint (REF-276, B, track cyclists). Short specific matched long specific and beat long general for 40-m sprints (REF-287, B, soccer — transfer) |
| Strength / gym | 5-10 min easy cardio, dynamic movement, ramp-up sets of the first lift; static holds <60 s per muscle, then dynamic work | n/a | Static stretching −3.7% immediately after, dynamic +1.3%; with dynamic activity afterwards no clear effect (REF-268, A). Ramp-up sets: practice. Before explosive work a conditioning activity gave a small (jumps) to moderate (sprints, ES 0.51) boost (REF-688, A, strength/power populations) |
| Tests (FTP20, CP efforts, ramp) | The same script every time: 15-20 min progressive, 5 min near but below threshold, 5-10 min easy (SYNTH-53, practice); ramp test: 10 min easy (the ramp warms up itself) | 85-90% LTHR before the hard part | SYNTH-53; the Allen & Coggan FTP20 protocol: ~50-min warm-up with 3 × 1-min fast pedalling and a 5-min maximal opener (REF-747, B — a protocol, not a tested warm-up) |
| SFT in the warm-up | 10-15 min easy, identical every time; the test doubles as the priming bout (inference) | not steered | SYNTH-41; an identical warm-up before the test (REF-752, B) |
| Short TT / prologue (≤~15 min) | Progressive 15-20 min, heavy priming ending ~10 min before, optional 3 × 10 s openers ending ~5 min before; after priming a ~12-s hard start | 85-90% LTHR; bout to ~95-100% | SYNTH-48; 3 × 10 s at 70% peak power: 1.7 ± 3.5 s, not significant (REF-277, B); severe-intensity priming plus a 12-s all-out start made a ~6.7-min TT 9 s faster, either alone did nothing (REF-279, B, 9 men) |
| 40 km TT (50-120 min) | 15-20 min with 2-3 min at target, then settle over the first minutes | 80-90% LTHR | SYNTH-62; REF-278 (warm-up optional for performance at 20 min); practice |
| Crit / XCO / cyclocross | 20-30 min: default ramp, one sustained priming bout (not sprints), 2-3 short openers; keep the legs warm on the grid | 85-90% LTHR; bout to ~95-100% | SYNTH-49, SYNTH-58; heated pants in a 30-min wait: +9.6% 30-s sprint peak power, insulation alone did not (REF-745, B+, 11 trained men); cyclocross: 93.6% of race time above the 4-mmol/L heart rate (REF-740, C+); no race-warm-up trial |
| Road race / gran fondo | 10 min easy + 1-2 openers if the start is hard; else the first 10-15 min progressive | 70-80% LTHR, or none | Practice (REF-278's entry); no trial |
| 5-10 km run | 10-15 min easy, drills, a few strides, one 1-2-min effort at race pace ending ~10 min before | 85-90% LTHR | SYNTH-47 (practice built on transfer); one 200 m at race pace instead of four strides made an 800 m 1.2 s faster (REF-285, B, 11 highly trained runners) |
| Half / marathon | 5-10 min jog and a few strides, or none; the first km at or below goal pace | none required | Fast starters finished slower (1.7 million recreational marathoners, REF-820, B+, observational); no tested warm-up changed time to exhaustion at 90% vVO2max (REF-286, B, 7 runners) |
| Trail / ultra | 5-10 min easy + mobility; start slow | none | Neglecting the warm-up was associated with trail injury (REF-783, A-, observational risk factors, no effect sizes) |
| Triathlon swim / T1 | 5-10 min in water with a few race-pace pickups if allowed, else dryland arm work; T1 needs none | none | Pool swimmers: 1,000-1,500 m with a brief race-pace approach, 8-20 min before (REF-290, B+, narrative); 10 min and 20 min both beat 45 min before a 200 m (REF-292, REF-291, B). Open-water transfer is inference; a slightly eased swim helped the bike and run (REF-834, C+) |
Environment and athlete
Autopilot rules (for the trainer)
Each rule is house practice unless a refId is named.
Drive: the warm-up is driven by POWER in ERG and never steered second by second by heart rate (current code, isWarmCool). Heart rate is read at one checkpoint and changes the rest of the warm-up by rule. Ridden at critical power, heart rate was still climbing at failure (REF-750, B, 7 trained cyclists), so openers and priming bouts are never judged by HR.
Default ramp: 0-3 min 50%, 3-6 min 60%, 6-9 min 70% FTP, then a 4-min steady step at 78-80% FTP (the check step), then the session part, then 3-5 min at 50-55%.
The check: mean HR over the last 30 s of the check step. HR lags power by ~1-2 min (measured on the owner's strap; code default 60 s up / 90 s down), so the check ends a steady part of ≥3 min, never a ramp. Only with a measured LTHR or HRmax (SYNTH-38, B: formula error RMSE 9.1-10.5 bpm).
Length: hard sessions 10-25 min; sprint sessions ≤ 20 min (practice; REF-276 compared 15 min with ≥50 min); races up to 30 min; endurance rides 5-10 min.
No HR strap: ask RPE at the end of the check step — aim 3-4 on CR-10, full sentences possible (the talk test is SYNTH-40's boundary marker below threshold, B; the RPE numbers are practice); RPE ≤ 2 → the "too easy" branch, ≥ 5 → "too hard". No power: prescribe by time and RPE, HR as a cross-check.
Cadence: free, ~85-95 rpm in the ramp; openers at the session's cadence; never add load when cadence is < 55 (existing code).
The owner's case
HRmax ~180: 100-105 bpm is 56-58% of HRmax — under the app's first bike zone boundary (70% HRmax) and below the 60% HRmax at which REF-276's shorter warm-up STARTED (it rose to 70%). With LTHR near 162 (assumed 90% of HRmax) it is 62-65% LTHR.
Why it under-prepares: the warm-up's effects come mostly from muscle temperature plus a raised baseline VO2 (REF-264, REF-265, B+ narrative reviews), and changing VO2 kinetics took prior heavy or sprint work — passive warming did nothing (REF-271, B).
Honest limit: before a 20-min effort even no warm-up did as well (REF-278), so for threshold work the cost is mainly a stiff, HR-lagging first rep (inference). First check the watts: if the step sat at ~50% FTP, this HR is expected and the plan, not the heart, was too easy.
How to progress / periodise
Not from a study; the base's own practice: rehearse the race warm-up in 2-3 key sessions, one script per race type; taper weeks keep it. Warm-ups are individual — let athletes explore (REF-267, C+ commentary).
When NOT to use it / red flags
What the evidence does and does not show
Evidence gaps
How this entry was checked and changed
(one A meta-analysis that warming up helps in general; priming and duration from small B/B+ crossovers of 7-15 mostly male riders or runners; long events, HR targets, session warm-ups and all autopilot rules are house practice)