Strength training for cyclists
Heavy, few reps, short of failure, twice a week to build and never zero in season: it buys efficiency, sprint and late-race power, not a bigger engine
The question it answers
Which lifts, at what load, how many sets and reps, how often and for how long a cyclist should strength-train; what changes in season; what to promise the athlete and what not to. WHERE the sessions sit relative to key rides is SYNTH-31; how much interference to expect is SYNTH-21. Not repeated here.
Default (what to prescribe first)
Trained cyclist, gym access, no injury:
Decision (who → what)
| Rider | Prescribe | Source and strength |
|---|---|---|
| Road, TT, gravel, long climbs | The default, built on bilateral lifts | Bilateral lifting improved economy ~8% where alternating unilateral did not (REF-148, B+, pilot, n = 24 trained triathletes/cyclists); after 12 weeks the 5-min all-out after a 185-min ride rose 371 → 400 W while endurance-only did not change (REF-146, B+) |
| Criterium, cyclocross, XCO, sprinters | The default plus a unilateral lift; fast concentric intent is practice | Anaerobic power had the largest pooled effect (REF-743, A); alternating unilateral lifting favoured sprint acceleration +20%, d = 0.5 (REF-148, B+, preliminary); XCO is raced as repeated short bursts above MAP (REF-741, B; REF-742, A), crit and cross by analogy (inference) |
| No gym, no time, or knees that dislike squats (the last is inference) | On-bike resistance: all-out efforts against very high resistance at very low cadence, 2×/week | In well-trained cyclists it did not differ from full squats on any parameter over 10 weeks; the squat group tended toward more injury-related symptoms (REF-158, B+, n = 37). Moderate-intensity 40-rpm intervals are NOT a substitute: they changed nothing, leg strength included, in veteran cyclists (REF-118, B+) |
| Time-crunched (one short session) | One hard set per main lift, 2-3×/week, or one denser session | A single set of 6-12 reps at ~70-85% 1RM 2-3×/week raised 1RM (REF-586, A-, trained men only, sets to failure — the one exception to the short-of-failure rule); with volume equated, frequency did not change strength gains (REF-583, A, mostly untrained); in people not lifting at all, one full session a week or one set per exercise 2-3×/week raised strength over doing nothing, though less than full programmes (REF-1141, C+, non-systematic overview, non-athletes) |
| Never lifted | Technique phase first, then the default | See progression below; practice, not a trial |
| Masters (50+) | The default, 2 sessions kept year-round, some fast-intent sets, slower ramp of tendon load | See masters paragraph |
| Women | The same programme, not a lightened one | See women paragraph |
Masters. In the one masters cycling trial, 3 weeks of knee extensions (3×/week, 10 × 10 at 70% of maximal repetition) raised strength more in 9 masters (51.5 y) than in 8 young endurance athletes (+17.8% vs +5.9%) and improved cycling delta efficiency only in the masters (REF-860, B, no non-lifting control; efficiency, not performance).
In adults aged 50 and over, resistance training raised leg power (d = 0.34, about 55 W) and high-velocity training worked better than slow traditional lifting for that outcome (REF-597, A+, mostly untrained, transfer to trained masters cyclists is indirect).
Heavy loading is not off-limits with age: in adults aged 50 and over, ≥70% 1RM built more leg strength than <70% (leg press SMD 0.95), with no significant difference in falls or adverse events (wide intervals) and no difference in lumbar-spine or femoral-neck BMD between the two intensities (REF-1086, A, general older adults, not athletes).
Older populations may need up to 2 sessions a week and 2-3 sets per exercise to hold muscle size, at maintained load (REF-587, B+, general populations) — so a masters rider's in-season floor sits higher than one session every 7-10 days (inference).
Ageing tendon loses modulus and strength, while resistance training can raise its stiffness (REF-599, B–C, mechanistic review) — Inference (the entry's own, not tested): tendon adapts slower than muscle, so progress load more gradually than the muscles ask for.
One supervised year of heavy lifting at retirement age left isometric leg strength at baseline three years after it ended while controls lost 7%, and light band circuits, mostly unsupervised at home, left no measurable advantage — but leg power fell in every group (REF-1139, A-, resistance-naive older adults, a secondary outcome): one more reason to keep fast-intent sets in a masters plan and to keep lifting year-round rather than bank a block (inference).
In general adults over 60, moderate-to-high weekly volume gave the largest strength gains, while a low volume was enough for function and muscle size (REF-1140, A-, from the abstract). The wider masters logic is SYNTH-35.
Women. The one female-cyclist trial in the base: in 19 well-trained women, 11 weeks of heavy strength added to endurance training improved 40-min all-out power, fractional utilisation of VO2max and cycling economy, while endurance-only changed nothing (REF-073, B, 11 vs 8).
In adults over 50 doing identical programmes, women gained more relative lower-body strength than men (g = −0.21) and the same relative muscle size (REF-532, A; transfer to young female cyclists is indirect). Do not schedule lifting by cycle phase: the evidence is inconsistent and does not support it (REF-571, B).
The pill did not change hypertrophy or strength gains (REF-729, A-, 8 studies). After menopause, strength rose normally at 6-8 sets per muscle a week but muscle mass did not grow, and the authors infer that more volume is needed (REF-523, B+, 41 healthy middle-aged women, not athletes).
Cyclists often have lower bone density than ball- and power-sport athletes, sometimes lower than inactive peers (REF-522, C+, narrative review), and a systematic review of 31 mostly observational studies found low bone density at key sites such as the lumbar spine in adult road cyclists in regular training (REF-896, A-, no countermeasure tested) — Not from a study; the base's own practice: axial-loaded lifting gives bone a load the bike does not.
How to progress / periodise
Never lifted: 2-4 weeks of technique (practice). Moderate loads, 2-3 sets of 8-12, stopping well short of failure, learning the squat and hinge. An identical bout disrupted 82% of fibres in the eccentric arm in untrained vs 45% in strength-trained men (REF-702, B, two separate studies, arm curls, not cyclists) — Inference: the first two weeks will hurt more than later ones, so keep them away from key rides.
Build (off-season / base): 8-12 weeks, moving to 4-6 reps at ≥80% 1RM, 2 sessions a week (practice on REF-743 and the trials above).
Do not take a full week off mid-block as a "deload". In resistance-trained young adults, a complete week off at week 5 cost lower-body strength and bought nothing (REF-217, B, not endurance athletes). Not from a study; the base's own practice: in the cycling recovery week, halve the sets and keep the load.
Year after year. In elite XCO riders, leg-press force and sprint power rose 16% after the first preparation period, held through the season, and the one rider followed for 3.5 years reached +44% force and +25% sprint power (sprint power plateaued in the last period), while two riders without lifting did not change (REF-047, B, case report, 4 riders).
Race week (practice): no new exercises in the 2-3 weeks before an A-race; the last heavy session 48 h or more before the key race (REF-744 entry, labelled practice there).
Practical Application
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 in cyclists with low GRADE certainty, REF-743, plus small B/B+ cyclist trials; exercise choice, progression and in-season spacing beyond once a week are practice)