snatchlike

Film it. Find out.

One snatch, filmed from the side. snatchlike reads where the bar went, whether the hips finished, how fast you got under it and how deep you caught it — then names the fault worth fixing and the drill that fixes it.

Your video never leaves this device.

The pose model runs in this tab. What comes out of it is twelve joints per frame with no face among them, and the reading below is worked out here, not on a server. There is nothing to sign into to use this page.

Signed out, the read and its landmarks stay in this browser and nowhere else. Sign in and the reading — and the twelve joints a frame it was taken from, never the video — are kept in your account instead, so a lift filmed on a phone is in the history you look at on a laptop.

How to film it so the reading is worth having

What it is measuring

Nine things, all of them in torso lengths and degrees rather than pixels, so a clip shot from three metres reads the same as one shot from one and a tall lifter's numbers mean the same as a short one's: how far the bar left the mid-foot, whether the arms stayed long, how far the hips opened at the finish, what the back angle did off the floor, how far your feet travelled, whether the hips got below the knees in the catch, how long the turnover took, whether the elbows were locked in the bottom, and how still the bar was overhead.

The tolerances are conventional coaching margins — the amount a coach would let go without saying anything. They are reasoned from practice, not fitted to anybody's data, and a measurement the clip cannot support is reported as unchecked rather than filled in. The rest of the method →