Cut your cadence in half at the same power, and the torque through each pedal stroke roughly doubles. That’s just physics, no argument there. Whether it reliably shows up as more knee pain is a messier question. The single most-cited paper on cadence and kneecap force actually found no significant cadence effect at all.
That matters if a coach ever handed you a low-cadence “strength” interval, big gear, 50-60 rpm, and your knee complained by interval three. The torque math says worry. The research says: it depends.
The Physics: Why Halving Your Cadence Roughly Doubles the Force Per Pedal Stroke
Power on a bike is torque multiplied by how fast the cranks spin. Slow the spin at fixed power, and torque has to rise.
Torque (Nm) = Power (W) / (2 x pi x rpm / 60)
Force (N) = Torque / crank length (m)
Here’s what that looks like with real numbers, using a standard 172.5mm crank. At 250 watts and 90 rpm, a rider pushes roughly 154 newtons of pedal force per stroke. Drop to 45 rpm at the same power, and that climbs to roughly 308 newtons, almost exactly double.
At the same wattage, a low cadence means each stroke is a bigger shove. These are calculated numbers, not force measured off a real knee. Real peak forces run higher, since force spikes near the top of the stroke rather than staying flat.
Moving from 250 to 350 watts raises force at every cadence point, often more than cadence itself. Workload matters as much as rpm.
Torque is not the same thing as pain.
Cadence and Patellofemoral Load: What the Force Studies Actually Measured
The physics above predicts higher joint stress at low cadence. Whether that shows up in real riders is a separate question, and the answer is messier than most cycling articles claim.
Start with the paper almost every “low cadence hurts your knees” article points to: Ericson and Nisell, 1987. Six cyclists rode a lab ergometer while researchers modeled force through the patellofemoral joint (kneecap to thighbone). Peak load hit 905 newtons, about 1.3 times body weight, at 120 watts and 60 rpm.
In plain terms: at a moderate effort and a fairly normal cadence, that kneecap joint carried more force than the rider’s full body weight.
Here’s the part few people quote: pedaling rate did not significantly change the force in that study. Workload and saddle height did. Six riders is a small, 1987 modeling sample, not a live sensor in a real knee. Still, the paper everyone cites to prove cadence matters found, in its own data, that it didn’t.
The most-cited cadence study is also the weakest evidence for cadence.
A tighter, more modern study gets closer to a real answer. Bini and colleagues (2013) put 12 competitive cyclists through matched effort at two cadences, using instrumented pedals. At the same submaximal intensity, 90 rpm produced 29% less peak kneecap force than 70 rpm, a real, measured difference in the direction the physics predicts.
Same study, though: maximal effort produced 18% more kneecap force than moderate effort at 90 rpm. Workload moved the number more than cadence did. And the tibiofemoral joint, where shin meets thigh bone, barely changed with either variable. This story is specific to the kneecap, not the whole knee.
| Study | n | What changed | Cadence’s effect on knee force |
|---|---|---|---|
| Ericson & Nisell, 1987 | 6 | 60 rpm vs. higher, fixed workload | Not significant (workload & saddle height did matter) |
| Redfield & Hull, 1986 | model | 40-120 rpm, constant power | Torque falls as cadence rises (modeled, not measured) |
| Bini et al., 2013 | 12 | 70 rpm vs. 90 rpm, fixed submax effort | 90 rpm = 29% lower peak kneecap force |
| Kristoffersen et al., 2014 | 22 | 40 rpm intervals, 12 weeks | No performance gain vs. normal-cadence training |
Four studies, four different questions. Only one measured force at two matched cadences and found a clear difference.
How Common Is Knee Pain in Cyclists? The Epidemiology
Whatever the mechanism, cyclist’s knee is common. Clarsen, Krosshaug, and Bahr (2010) surveyed 109 riders on 7 pro road teams, a 94% response rate: 36% reported anterior knee pain in the prior year, and 19% saw a doctor for it.
More than one in three pro riders dealt with kneecap pain in a single year.
Recreational and amateur surveys report a wider range, roughly 28% to 42%, depending on the population and pain definition. One long-distance touring event survey found rates as high as 65%. These come from different studies with different definitions: common, sometimes very common, not one precise figure.
Knee pain is a cycling-wide problem, not a low-cadence-only one.
Does Low-Cadence Training Even Help? The Evidence Is Mixed
Does low-cadence “strength” work even improve performance? If not, the knee-load trade-off is easy to skip.
Kristoffersen and colleagues (2014) ran a 12-week test in 22 well-trained masters cyclists, average age 47. One group did 40 rpm intervals twice weekly; a control group trained at normal, freely-chosen cadence. The low-cadence group’s fitness barely moved: VO2max up just 1.4, 30-minute power down 3 watts. The normal-cadence group improved on both, VO2max up 3.3, power up 13 watts.
In this group of masters riders, grinding low cadence didn’t just risk the knee. It didn’t even work.
A different population tells a different story. Hebisz and Hebisz (2024) studied 24 young female cyclists doing low-cadence work inside short, hard intervals, not long moderate ones. That group improved max aerobic power 8.1% and VO2max 8.7%, both clearly beating the higher-cadence comparison group.
Put those together and the read is simple: low-cadence training isn’t a universal hack. It depends on intensity, duration, and who’s doing it. A masters rider grinding long moderate intervals got nothing extra for the added knee stress. A younger athlete doing short, hard intervals got more out of it. If you’re also stacking gym strength work on bike intervals, the same logic applies: concurrent strength and endurance training has its own timing rules worth knowing first.
Same idea, different rider, different answer.
Saddle Height and the Confounders That Matter More Than RPM
Cadence gets blamed first, but it’s rarely the only variable moving. Saddle height changes knee flexion angle at the bottom of the stroke; the old rule of thumb, keep it near 25-30 degrees, still holds up. Push it low enough to force more than 40 degrees, and discomfort tends to rise.
Even so, saddle height’s effect on measured force is smaller than folklore suggests. One comparison found it shifted kneecap force by only 5-13%, and shin-to-thigh joint force by just 1-7%. Riders with existing knee pain didn’t sit measurably lower than pain-free riders in that data.
Saddle height matters for comfort. For measured force, it’s a smaller lever than most riders assume.
Workload matters more, as the Bini study showed. So does terrain: a low cadence up a steep grade while seated stacks two load factors at once. Standing shifts load off the kneecap and spreads it through the leg differently. Pedal-stroke technique also changes shape and timing as cadence changes, even at constant power.
There’s also an efficiency gap. The most efficient cadence for most riders sits lower, often 50-80 rpm, than what trained cyclists choose, often 90 rpm or higher. Riders already trade efficiency for a faster spin.
RPM is one variable in a pile of them. It’s rarely the only one worth fixing.
A Practical Cadence Floor for Knee-Sensitive Riders
None of this puts low-cadence work off-limits. It means riders with a knee pain history need a floor and a plan, not a ban.
Think of it like an engine. Pushing 250 watts at 45 rpm is like pulling away from a stoplight in third gear: more torque per turn, more strain on whatever’s turning it. You can do it. You just can’t do it for long without complaining.
Start with a cadence floor around 60 rpm for structured low-cadence intervals, higher if you’ve had anterior knee pain in the past year. Keep early efforts short, 3 to 5 minutes, seated, moderate gear, well inside your aerobic base pace rather than maximal effort. Only lower the floor, or extend the duration, if the knee stays quiet 48 hours after the session, not just during it. Stop the block and get checked if you feel sharp or pinching pain at the front of the knee, not ordinary muscle fatigue.
Take a rider I’ll call Dave, 52, racing masters crits for six years, with a history of anterior knee soreness on hilly club rides. His coach prescribed 50 rpm big-gear strength intervals to build climbing power. Two sessions in, his knee ached on stairs the next morning. He raised the floor to 65 rpm, cut intervals to 3 minutes with full recovery, and capped effort at moderate aerobic power instead of near-max. Six weeks later he ran 55 rpm intervals pain-free, with 20-minute climbing power up 6%, no flare-ups.
That’s the trade-off AthleteOS tags for. Low-cadence strength intervals show up as their own session type in the training plan, separate from normal endurance or VO2max work. If a rider has anterior knee pain in their injury history, AthleteOS flags that a prescribed low-cadence block raises per-stroke torque at the same power before the session gets scheduled, not after the knee talks back. Pairing that with your readiness trend helps catch an early flare before it becomes missed weeks.
Cadence advice for cycling has the same one-size-fits-all problem cadence advice for running does. The physics is real. It just doesn’t point to one magic rpm for every rider.