Cricket biomechanics from a phone video

What a biomechanics lab actually measures

A traditional lab tracks reflective markers taped to a player's joints using several synchronised cameras. Because the marker positions are known in three dimensions, the system reconstructs the skeleton frame by frame and derives angles, velocities and the sequence in which body segments accelerate.

The output is genuinely valuable — it is how professional coaching understands whether a player's power comes from a well-sequenced kinetic chain or from compensating with one joint. The problem was never the data. It was that the setup costs more than most academies earn in a year, and a player has to travel to it.

How the same measurements come out of one camera

Markerless systems replace the markers with a pose model that locates joints directly in the image. Once joints are located in every frame, the same derivations apply: an angle is still an angle, and a sequence is still a sequence.

The hard part is scale and depth. A single camera sees a flat projection, so something of known size has to appear in the frame to anchor real-world measurements. In cricket that anchor is already there — stumps are a fixed, regulated size, which is why they matter as much as the player does.

This is also the honest limit of the approach. Movement across the camera plane measures well; movement directly toward or away from the lens measures less well, because the camera has least information along that axis. Filming from a semi-side-on, squarish view rather than head-on materially changes what you get back.

What the three cricket surfaces look at

Batting analysis looks at the body through a shot — how the base is set, how the trunk rotates, where the head stays, and whether the segments fire in a sensible order.

Ball and delivery analysis looks at the ball rather than the bowler: speed, the line and length it pitched on, and any seam or swing movement in flight.

Run-up analysis looks at the approach and delivery stride, which is where a lot of bowling problems actually originate even when they show up at release. What to measure in each phase, from the run-up to release, is set out in bowling action analysis.

All three come from ordinary phone footage in Stance's cricket video analysis app, which runs in a browser with nothing to install.

What the kinetic chain is, and why sequence beats strength

Almost every powerful cricket movement is a sequence rather than a single effort. Force starts at the ground, passes through the legs into the pelvis, through the trunk into the shoulder, and finally into the arm and the bat or ball. When each segment accelerates in turn and hands off at the right moment, the end of the chain moves far faster than any one segment could drive it. It is why a well-sequenced bowler of ordinary build can be quicker than a stronger one who is not.

When the sequence breaks, the body compensates. A pelvis that rotates too early leaves the trunk nothing to pull against, so the shoulder and arm supply the missing speed directly. The delivery may still come out fast, but the load has shifted onto tissue not built to produce it — the mechanism behind a great deal of shoulder and lower-back trouble in bowlers.

Sequence is visible in video in a way strength is not. You cannot see how strong someone is, but you can see the ORDER in which segments reach peak rotation, and whether those peaks are separated or bunched together. That ordering is what a kinetic-chain analysis reports.

The limits, stated plainly

A single camera sees a projection of a three-dimensional movement, and that has consequences worth knowing before trusting any number.

Depth is the weak axis. Movement across the camera plane is measured well; movement toward or away from the lens is measured least well, because that is the direction in which the image changes least. Filming from a semi-side-on, squarish view is not a preference, it is what makes the measurement possible. Our filming guide shows the set-up for each clip.

Occlusion breaks tracking. When the front arm hides the trunk, or the back leg passes behind the front, the joints underneath are inferred rather than observed, and inferred positions are less reliable than seen ones. Camera height and angle change how often this happens.

Absolute angles are harder than relative change. Recovering a joint angle to within a degree from one camera is difficult; detecting that the same joint moved 8 degrees further through its range than it did last month is considerably easier. Trends across repeated sessions are therefore more trustworthy than any single reading, which is the practical argument for measuring the same thing repeatedly rather than once.