Running biomechanics analysis: do you really need accuracy to the degree?
In biomechanics, accuracy is the goal: the angle to the degree, repeatable studies, plenty of runners. I understand why, but it isn't how I look at a stride. I look for trends, and here is why.
What biomechanists look for
When a laboratory studies running, it wants accuracy. Angles are measured to the degree with infrared cameras and markers on the body. The study has to be repeatable, cover enough runners, and hold up statistically. That makes sense: it's how you know whether an effect is real or down to chance.
Look at the studies I cite in these guides. The one on cadence had 45 runners run in a laboratory at several cadences.[1] Another followed 30 on an instrumented treadmill.[2] This work is valuable, and it's what lets me write these pages.
Why that isn't what I'm after
I don't do research. What I want is for a runner filming themselves with a phone to see something useful in a minute. And very often we cling to accuracy and forget the essential. Knowing that your knee bends 42 or 45 degrees changes little about your running. Seeing that your foot lands far ahead of your pelvis, that your trunk is leaning back or that you bounce a lot, that does.
Those differences are visible to the eye, especially when overlaid on a reference stride. You don't need the degree to notice them. And from the sum of those differences comes a trend: which way to go.
What 2D measurements are worth
I owe you the truth about the limits. In a study where 20 runners were filmed in 2D (one camera) and 3D (motion capture) at once on a treadmill, agreement was good for foot angle at contact and knee flexion range, weaker for peak knee flexion, with differences that varied from runner to runner.[3] A phone video won't replace a laboratory, and I don't try to make it say more than it can.
On the treadmill side, a meta-analysis of 33 studies found most running measures comparable with overground, with differences at foot impact.[4] That's good news for a treadmill analysis, as long as you keep that nuance in mind.
No single number explains everything
In a study of 29 elite and 68 recreational runners, three variables together explained 39% of the differences in running economy, including the vertical bounce of the pelvis.[5] In other words, no isolated number makes a good stride. Fixating on one precise measure risks missing the whole, and the whole is easier to see by eye on a side-on video, as I explain in the posture analysis guide.
When accuracy becomes necessary
If you have pain that keeps coming back, if you're recovering from an injury or preparing a very ambitious goal, a clinic assessment with a deeper analysis makes sense. Gait analysis in a clinic, a store or by video doesn't have the same purpose, and one doesn't replace the other.
For the rest, to understand your running and know which way to go, you can start with the simplest thing: film yourself side-on and look.