Overpronation in Running: What the Evidence Says

Walk into most running shops and you will be on a treadmill within ten minutes, being filmed from behind while someone watches your ankles. The verdict comes back as a single word: you overpronate. What follows is usually a stability shoe, sometimes an insert, and a quiet implication that your feet are built slightly wrong. Overpronation in running has been the sport’s default explanation for injury for something like forty years. It is also one of the least well-supported ideas in the whole field, and the research that has tried hardest to confirm it has mostly done the opposite.

This is not an argument that your feet do not roll inward. They do. It is an argument about what that rolling means, how reliably anyone can measure it, and what you should actually be looking at instead.

What pronation actually is

Pronation is the natural inward roll of the foot after it hits the ground. The arch lowers, the ankle rolls slightly toward the midline, and the foot flattens a little. This is not a malfunction. It is the foot’s shock absorption, the mechanism that lets a rigid-ish structure accept your bodyweight without transmitting the whole impact straight up your leg. A foot that did not pronate at all would be a worse foot.

“Overpronation” is the claim that some people do this too much, or too fast, and that the excess causes injury up the chain. The concept is intuitive. If the foot collapses inward, the shin rotates in, the knee follows, and you get the kind of inward collapse we cover in the guide to dynamic knee valgus. The mechanical story hangs together nicely, which is a large part of why it survived so long.

The trouble is that the story predicts something testable: people who pronate more should get hurt more. And when researchers went looking for that, they largely did not find it.

Is overpronation bad for running?

On current evidence, no. Pronation is a normal and necessary part of how the foot absorbs load, and larger-scale prospective studies have not found that runners who pronate more get injured more often. There is no good reason to treat pronation as a defect that needs correcting in a runner who is not in pain.

The most-cited work here is a Danish study by Nielsen and colleagues that followed 927 novice runners, 1,854 feet, for a full year. Foot posture was graded at baseline from highly supinated through neutral to highly pronated, everyone was put in the same neutral shoe, and GPS watches tracked exactly how far each person ran. Then they waited to see who got hurt and after how many kilometres.

If the overpronation model were right, the pronated groups should have broken down first. They did not. Pronated feet were not associated with a higher injury rate, and the highly pronated group actually recorded fewer injuries per 1,000 km than the neutral group.

Two honest caveats on that last detail, because it gets over-quoted in the other direction. The highly pronated subgroup was small, so the apparent protective effect is not something to lean on. And these were novice runners in neutral shoes, which is a specific population and a specific condition. The safe reading is not “pronation is good.” It is the more boring, more durable conclusion: pronation does not predict injury nearly well enough to justify the weight the industry puts on it.

Forty years of shoe fitting rests on a link that, when actually measured, is not there. Your foot rolling inward is a description of how feet work, not a diagnosis.

Why the label stuck anyway

Partly because it is easy to see. Film someone from behind and ankle roll is genuinely visible, unlike hip strength or ground contact time. Partly because it arrived with a solution attached, and a category of shoe was built to sell against it.

And partly because of the wet-foot test, that trick where you step on cardboard and read your arch height from the print. It has an enormous cultural footprint and close to zero predictive value. Static arch height in standing tells you very little about what your foot does at 3 metres per second under full bodyweight. Large military studies that assigned footwear by foot type on exactly this logic have generally failed to show a reduction in injuries compared to just giving everyone the same shoe.

None of which means stability shoes are bad. Plenty of runners find them comfortable, and comfort is a genuinely reasonable basis for choosing a shoe. It just means the diagnosis that sold them was doing less work than it appeared to.

Why a phone camera cannot settle this for you

Suppose you wanted to check your own pronation from video. You are up against a hard measurement problem.

A single 2D camera carries roughly 5 to 15 degrees of error depending on angle, lighting, and how square you are to the lens. Pronation angles of interest live inside that error band. Worse, the foot is the smallest and fastest-moving thing in the frame, so it is the part most degraded by motion blur, and it is partly hidden inside a shoe. Everything that makes joint tracking hard is concentrated exactly where you would need it to be easy.

This is why RunGait treats foot and ankle rotation as a low-confidence signal and says so, rather than reporting a decisive-looking pronation number it cannot actually stand behind. The general rule for reading any home video applies with double force here: trust the asymmetry between your two sides, because both legs were filmed from the same angle in the same clip and that comparison cancels most of the measurement noise, and distrust any absolute number in degrees.

What to look at instead

If pronation is a weak signal on a hard-to-measure joint, spend your attention further up the leg, where the signals are stronger and the fixes are real.

• Dynamic knee valgus. The knee diving toward the midline under load is the most reliable frontal-plane sign a single camera can catch, and it points at a trainable hip. See dynamic knee valgus in runners.
• Contralateral pelvic drop. The opposite hip dipping during stance, a direct read on glute medius control. See the glute medius sign.
• Left/right asymmetry. Differences between your sides in stance time and joint angles, which tell you where to send your effort. See muscle imbalances in runners.
• Foot placement relative to your body. Where your foot lands matters far more than which part of it touches down first. See overstriding.

Notice what these have in common. Each one is visible from a phone, each is a comparison rather than an absolute measurement, and each resolves to something you can train.

The part that actually helps: train the foot, do not correct it

If your arch does collapse hard on impact and you would rather it did not, the answer is not to brace it into a different shape. It is to build a foot and lower leg that can control the motion themselves.

That means calf and foot strengthening, short-foot drills, barefoot balance work on one leg, and general single-leg strength through the hip so the whole chain has something to push against. It is the same principle as everywhere else in running form: strengthen the structure that controls the motion, then let the pattern change on its own instead of trying to hold it in place with hardware.

Keep it honest

Pronation is a movement pattern, not a diagnosis, and seeing your ankles roll inward on a video does not mean anything is wrong with you. It is also, of all the gait signals, the one a phone camera reads least reliably. Treat it as the lowest-priority item on the list, well behind what your knee and pelvis are doing.

And if you have actual pain, in the arch, the shin, the knee, anywhere, that is a conversation for a physio or sports doctor rather than a video or a shoe wall. Screening is not diagnosis.

RunGait is built to be straight about this distinction. It records a short clip, tracks your joints on-device with Apple Vision, reports each signal with the confidence it actually deserves rather than a false precision, and turns the strong signals into strength priorities. For pronation, the honest answer is usually that there are better things to look at, and telling you that is more useful than selling you a shoe.