Big-gear, low-cadence work: strength on the bike, or just slow?

Every winter the same session reappears in training plans: get into the biggest gear you own, drop your cadence to fifty revolutions a minute, and grind. It is usually sold as "strength on the bike" or "strength endurance", on the reasoning that pushing harder on each pedal stroke must build something that pushing lightly does not.

The physics behind the first half of that claim is unarguable. At the same power, halving your cadence roughly doubles the force on each pedal. The second half — that this force reliably transfers into more sustained power out on the road — is where the evidence gets awkward.

Here is what torque intervals do, what the controlled trials found when they tested them, and how to programme them if you still want to, which you reasonably might.

The arithmetic first

Power on a bike is torque multiplied by angular velocity. Hold the power constant and the two trade off exactly: drop the cadence and the torque has to rise to compensate. That relationship is not a training theory, it is a definition, and it is the only part of this subject that nobody argues about.

So the mechanical claim survives. A 50 rpm effort at 250 W really does load each stroke about 1.8 times as hard as the same 250 W at 90 rpm. The interesting question is whether your body cares about force per stroke, or only about the total metabolic demand — and on that, the answer is less flattering to the session.

What torque intervals actually train

Neuromuscular: probably something

Producing higher forces per contraction recruits more motor units, including higher-threshold ones that a light, fast pedal stroke may leave alone. That is a real, well-described feature of how muscle recruitment works. It is also the honest basis for calling these intervals "neuromuscular".

What it is not is strength training. The peak forces in a big-gear interval, even at 40 rpm, sit far below what your legs can produce in a squat. You are doing many hundreds of moderate-force contractions rather than a few maximal ones, and that is a different stimulus with different outcomes. If your goal is genuinely more force-producing capacity, the gym is the tool built for it — why every cyclist should lift weights covers the case, which is much better supported than the case for grinding.

Muscular and metabolic: mostly the same as any other interval

The metabolic demand of a 250 W effort is set by the 250 W, not by the gear. Oxygen cost at low cadence is typically slightly lower for the same power — gross efficiency tends to improve as cadence falls, because you are spending less energy moving your own legs around. Your heart rate will usually sit a few beats lower too. This catches people out: the session feels brutally hard in the legs while the cardiovascular system is barely working.

"Strength endurance" is a label, not a mechanism

The term is used constantly and defined rarely. It generally means something like "the ability to keep producing force when tired", which is real enough as a description of a race but does not map onto a specific trainable tissue property. When you see it in a plan, read it as "sub-threshold intervals in a big gear" and judge the session on that.

What the trials found

This is where honesty is required. Low-cadence training has been tested in controlled trials, and the results do not line up behind the marketing.

Read together, the picture is that low-cadence intervals work about as well as the equivalent intervals at a normal cadence, with one plausible exception: specificity to sustained climbing. If your event is a 40-minute mountain pass where the gradient forces you down to 60 rpm whether you like it or not, practising at 60 rpm is defensible on specificity grounds alone. If your event is a flat road race, the argument is much thinner.

One more finding deserves attention because it is rarely mentioned: riders who train at low cadence tend to adopt a lower freely chosen cadence. That is an adaptation, but it is not obviously a benefit, and it is worth knowing you are changing a habit as well as a tissue. Our piece on whether cadence matters at all goes into why self-selected cadence is more individual than most advice admits.

The knee, honestly

The caution here is not invented. Measurements of joint loading during cycling show that patellofemoral compressive force rises with workload and falls as pedalling rate increases at the same workload. Low cadence at high power is, mechanically, the worst corner of that map. Reviews of saddle height and knee injury risk reach a related conclusion: anterior knee pain in cyclists is associated with high forces at a flexed knee, and cadence is one of the levers that changes those forces.

None of this means torque intervals damage knees. It means the load is real and the dose matters. Practical implications:

• Do not start at 40 rpm. Begin at 60, spend a couple of weeks there, and only then go lower if you are going lower at all.
• Stay seated. Standing changes the loading pattern and defeats the point of the session.
• Stop the interval, not the season. Any sharp pain at the front of the knee means the set is over for the day. If it recurs, knee pain from cycling lists the usual causes, and bike fit is near the top.
• Check your saddle height before a block of these, not after. A saddle that is slightly low is tolerable at 90 rpm and much less tolerable at 50.

How to programme them

If you want these in your plan, treat them as a sub-threshold aerobic session with a cadence constraint, placed in base or early build, run for four to six weeks, and then stopped. They are not a year-round staple and there is no evidence that they should be.

Execution details that decide whether it works

• Hold the power target, not the gear. The cadence is a constraint on how you produce the watts. If the watts sag, the session has quietly become a very slow easy ride.
• Start from rolling speed, not from a standstill. Cranking a big gear up from nothing is a different exercise with a much higher peak load.
• Watch your upper body. Rocking hips, hauling on the bars and a tilting pelvis all mean the force has exceeded what your position can absorb.
• Do them fresh. The recruitment argument only holds if you can actually produce the force. On tired legs you get a slow tempo ride with extra joint load.

Indoors is the easier place to run these, because the resistance stays where you put it and the cadence is a choice rather than a consequence of the gradient. Moveee Indoor will hold a power target in ERG while you deliberately sit at 55 rpm, and it records cadence alongside power so you can check afterwards whether the fifth interval actually stayed in the gear or crept back up to 70 when it started hurting. It runs in a browser tab and is free while in alpha.

So: strength, or just slow?

Mostly the latter, with a real but narrow exception. The forces are genuinely higher — the arithmetic guarantees it — but they are not high enough, or few enough, to be strength training in the sense that word usually carries. When trials have compared low-cadence intervals against the same intervals at normal cadence, the low-cadence version has not reliably come out ahead, and in well-trained riders it has sometimes changed nothing at all.

What is left is a session that is specific to sustained climbing, tolerable in small doses, and slightly harder on the knee than the alternative. If you race up mountains, keep a block of it in late base and pair it with proper strength work in the gym, which has the stronger evidence for building force. If you do not, you are free to spend that session doing something with a clearer mechanism and spin up the climbs instead.