SYNTH-82 · Guide
Low-cadence ("torque", "strength-endurance") riding
It does not build strength: lift for that, or with no gym ride all-out efforts against very high resistance; as an optional accent inside threshold or VO2max reps about 20% under preferred cadence it has small positive trials in well-trained riders, moderate grinding has a null trial, no low-cadence trial in the base shows an efficiency gain, and knee pain ends it (practice)
The question it answers
Default (what to prescribe first)
Not from a study; the base's own practice (dose, seated cue and placement): **strength comes from the gym (SYNTH-43).
If the athlete wants low-cadence riding, it goes inside a threshold or VO2max session already in the week: ride part of the reps seated at about 20% under the preferred cadence (roughly 65-75 rpm for a preferred 80-95 rpm; REF-084 rode 50-70 rpm), in one session a week for 2-3 weeks, then up to two sessions a week (the trials rode two or more) for the rest of a 6-8-week block, then rotate it out.
It is never sold as strength or efficiency work.** What that rests on:
It is not shown to be a strength stimulus (no entry measured strength after hard low-cadence intervals). In 11 cyclists, pedal force stayed below 54% of maximal dynamic force in every condition, including maximal aerobic power at 40 rpm; force rose with intensity and with lower cadence (REF-119, B, measurement only, no training).Twelve weeks of 40-rpm intervals at moderate intensity left leg strength unchanged in a trial of 22 well-trained veteran cyclists (REF-118, B+).
Inside interval sessions it has small positive trials. In 24 well-trained women, an identical 8-week polarized programme improved maximal aerobic power and VT2 power only in the group that rode its maximal 30-s sprints and 4-min reps at 90-100% MAP at 50-70 rpm, not in the group at a freely chosen cadence above 80 rpm (REF-084, B+; randomisation not stated in the abstract; women only).In 16 endurance-trained men, 18 interval sessions over 6 weeks at 20% below preferred cadence improved a preloaded 15-min time trial more than the same sessions 20% above it, though both groups improved (REF-259, B+; interval intensity and the size of the advantage are not in the entry).
The review is lukewarm. Across 7 heterogeneous studies there was no strong evidence of benefit and some favoured the freely chosen cadence; the authors only "tentatively" suggest low-cadence bouts at moderate to maximal intensity (REF-117, A-, trained cyclists).
Decision (who → what)
| Situation | Do this | Source and how sure |
|---|---|---|
| "I want to get stronger on the bike" | Lift heavy (SYNTH-43). Low-cadence intervals are not the strength lever | Heavy strength at ≥80% 1RM improved cycling performance (ES 0.46) and efficiency (ES 0.35) in 17 controlled trials, 262 cyclists (REF-743, A, GRADE low). Pedal force <54% of maximal dynamic force at any cadence tested (REF-119, B); 40-rpm moderate work did not change leg strength (REF-118, B+). No trial compares low-cadence intervals (as opposed to all-out on-bike resistance, next row) with the gym |
| No gym, no time, or will not lift | On-bike resistance instead of the gym: all-out efforts against very high resistance at very low cadence, 2×/week for about 10 weeks; it replaces a gym session, not adds to one (replacing is practice) | In 37 well-trained cyclists (male per the companion paper), it did not differ significantly from full squats on any parameter over 10 weeks (REF-158, B+; not proven equivalence; dose from the companion paper). In 46 well-trained men (VO2max 62.7), 2 weekly sessions of all-out pedal cycles against 70% of maximal dynamic force improved Wingate power, MAP, ventilatory thresholds and quadriceps CSA within the group; 4-min intervals against 35% beat control only on Wingate mean power, and high vs low load did not differ significantly (REF-428, A-; cadence not in the entry). Both trials are in men; for women it is transfer |
| Climber, hilly races, hill-climb TT | Optional: threshold reps ridden on a real climb at about 60 rpm, inside the threshold session already planned (placement is practice) | 18 male cyclists (VO2max 58.6), 4 weeks, 2×/week, 6 × 5 min at the power of the respiratory compensation point: uphill at 60 rev/min raised 20-min uphill TT power 4.4 ± 5.3% and flat 1.5 ± 4.5%; flat at 100 rev/min −1.3 ± 3.6% uphill, +2.6 ± 6.0% flat; but controls with no intervals gained 4.0 ± 4.6% uphill (REF-852, B; cadence and terrain changed together). Cadence falls on climbs anyway (REF-262, B; REF-884, B, professionals ~71 rpm on long passes) |
| Well-trained, polarized week, wants more from the hard days | Optional: ride the VO2max reps or 30-s sprints at 50-70 rpm for one 8-week block | REF-084 B+, 24 well-trained women — the only trial in women. Men: REF-259 B+ (intensity not stated). REF-1073 B, 18 road cyclists (sex not stated): sprints at 60-70 rpm beat 110-120 rpm, not the freely chosen cadence. Transfer between sexes is inference |
| Mountain bike, cyclocross, gravel | Same accent as the climber row; keep the heavy lifting (SYNTH-49, SYNTH-50) | Inference: no entry tests low-cadence training in off-road riders; every trial is road or laboratory riding |
| Loses power or stalls on steep grades | Check gearing first (practice): lower gears before torque training. Then the climber row if the athlete still wants it | No entry studies this rider. Less-trained non-cyclist men chose lower cadences that fell as power rose (~80 rpm at 75 W to ~65 rpm at 175 W) while cyclists held 90-100 rpm (REF-881, B, short lab bouts) — a lower cadence on hard efforts is what they did, not a proven fault |
| Wants better efficiency | Do not prescribe it for that (SYNTH-17) | Gross efficiency was unchanged after 40-rpm work (REF-118, B+); after REF-259 it rose in the high-cadence group at 90 and 110 rpm and is not reported as changed in the low-cadence group (B+); REF-084 and REF-428 do not report it |
| Wants tempo / sweet-spot "grinding" (40-60 rpm) | Not as a main protocol, and not as strength work (for 50-60 rpm this is practice); do not tell the athlete it does nothing | 12 weeks of 5 × 6 min at 40 rpm and 73-82% HRmax changed nothing in veterans, while the same 90 min a week at free cadence improved VO2max and 30-min power (REF-118, B+). Lower-force 4-min torque intervals (35% of maximal dynamic force, cadence not stated) beat control on Wingate mean power (REF-428, A-). 50-60 rpm at moderate intensity is untested in the base, and 73-82% HRmax sits below sweet-spot, so low-cadence sweet-spot is untested too (applying the 40-rpm null to it is transfer) |
| Novice (Tier 0-1), or masters | Novice: skip it; coach gearing (SYNTH-55). Masters: same default; no masters trial at high intensity | Practice. The only masters trial is the 40-rpm null (REF-118) |
| Crit, track, cyclocross starts: needs fast legs | High-cadence intervals rather than low (SYNTH-55) | High-cadence intervals raised preferred cadence 92 → 101 rpm and gross efficiency at 90 and 110 rpm (REF-259, B+), but the low-cadence group gained more on the 15-min TT; no trial tests crit, track or start performance, so the event use is inference |
How to progress and where it sits in the week
Not from a study; the base's own practice: start with part of one set, add reps before lowering cadence, cap at the trials' doses, run 6-8-week blocks.
Not from a study; the base's own practice: keep all-out torque efforts out of race week and away from the day before a key session.
Indoor (ERG) vs outdoor climbs
Same watts, different task. Relative pedal force rose as cadence fell (REF-119, B). At a constant 2.5 W/kg riders chose 81 rpm though metabolic cost was lowest at 60 rpm, and the chosen cadence kept vastus lateralis power capacity highest (REF-238, B); that a rep at 70 vs 95 rpm is a different muscular task is that entry's inference, not a tested training effect.
ERG holds the watts while cadence drops; all-out torque efforts are self-paced, not ERG (SYNTH-56; practice). Not from a study; the base's own practice: a gear that puts cadence in range, a cadence alert against drift back to the preferred cadence, and if cadence collapses late in a rep, shift easier and drop the target ~5% (the unmeasured "ERG spiral", SYNTH-56).
Outdoors (inference): a steady climb in a fixed gear supplies the torque on its own; REF-852's low-cadence reps were ridden uphill.
Testing: do not compare a ramp or threshold result done at an unusual cadence with one at the usual cadence — in 10 collegiate athletes the respiratory compensation point fell from 222 W at 60 rpm to 197 W at 100 rpm (REF-202, B).
When NOT to use it / red flags
Pain at the front of the knee, a diagnosed patellofemoral problem, or early return from any knee or back injury: skip low-cadence and torque work, and bring it back last. Not from a study; the base's own practice — no entry studies low cadence and knee pain.Inference: pedal force rises as cadence falls (REF-119); yet in 6 healthy men on an ergometer, modelled patellofemoral force rose with workload and a lower saddle but did not change significantly with pedalling rate (peak 905 N, 1.3 times body weight, at 120 W and 60 rpm; REF-1076, C+, modest loads, no pain or injury data).A lower saddle seems to raise patellofemoral compressive force (REF-882, B+, review of mechanics, not injury counts); among 109 professional road cyclists the knee caused 57% of time-loss overuse injuries (REF-921, B+, 12-month recall; Tier 4-5, transfer). Triage and return: SYNTH-73.
All-out torque sessions: not on top of a concentrated VO2max block or a full gym programme (practice; the low-cadence accent itself sits inside threshold or VO2max reps, as in the Default).
What the evidence does and does not show
Shows (small trials, Tier 2-3): positive inside interval sessions (REF-084, hard reps, women; REF-259, intensity not stated, men); positive uphill but matched by controls (REF-852); null at 40 rpm and moderate intensity (REF-118); all-out torque work not significantly different from squats (REF-158, men; equivalence not proven) and not proven better than lighter torque work (REF-428); sprints at 60-70 rpm beat 110-120 rpm on peak power by 3.6% (+/-3.7%) and 4-mM lactate power by 7.0% (+/-5.9%), jumps in both arms (REF-1073, B).
Shows (acute mechanism, not training): at the same metabolic cost, 50 rev/min emptied type II fibre glycogen more than 100 rev/min (49% vs 33%) in 8 men (REF-1074, C+); gross efficiency at 75% of peak power did not differ between level seated, uphill seated and uphill standing in 8 well-trained cyclists (REF-1075, B; cadence not reported).
Does not show: that low-cadence intervals build strength, raise efficiency, or add anything the gym does not (no head-to-head); crank torque profiles barely changed in either group of REF-259. A harder gear at a fixed cadence raised gross efficiency in 9 men in graded tests, but the study cannot say that the same power at a lower cadence is more efficient (REF-260, B, acute) — not a training effect.
Evidence gaps
No trial in novices, juniors, masters at high intensity, or off-road (MTB, cyclocross, gravel) riders.
No trial of seated versus standing low-cadence work (only acute efficiency by position, cadence not reported: REF-1075), of sweet-spot or tempo at 50-60 rpm, or longer than 12 weeks.
No head-to-head of low-cadence intervals against heavy lifting.
No measurement of cadence drift or the "ERG spiral" in low-cadence ERG reps.
How this entry was checked and changed
Why this level
(the accent rests on two small B+ parallel trials, REF-084 in 24 women and REF-259 in 16 men, and one B trial whose controls gained as much uphill REF-852, set against an A- review with no strong evidence of benefit REF-117 and a B+ null at 40 rpm REF-118; the strength verdict on the B+ leg-strength null REF-118, a B measurement study REF-119 and an A meta-analysis of heavy lifting (GRADE low) REF-743; the no-gym option on a B+ RCT REF-158 and an A- RCT that measured performance and quadriceps size, not strength REF-428, both in men; every dose, cue and placement rule is house practice)
Content correction 2026-09-27
2026-10-04: independent audit before posting — cadence accent corrected to ~65-75 rpm (20% under a typical preferred cadence), men-only samples labelled, "no different from squats" narrowed to "not significantly different, equivalence not shown", seated cue labelled practice, red flag reconciled with the Default.
Revision 2026-10-04: folded four new entries from the owner-approved candidate list. Paton 2009 REF-1073 (B, 18 road cyclists, sprints at 60-70 vs 110-120 rpm with jumps in both arms) added to the polarized-week row and the evidence bullets, labelled as a comparison against a very high cadence, not the freely chosen one.
Revision 2026-10-04: Ericson & Nisell 1987 REF-1076 (C+, 6 healthy men, modelled patellofemoral force rose with workload and lower saddle, not significantly with pedalling rate) added to the knee red flag next to the pedal-force inference; the red flag stays practice (no pain or injury data).
Revision 2026-10-04: gaps narrowed (seated vs standing; cadence and knee) and the candidate line removed; lead updated to five trials. Default and level unchanged.