Running loads from two different vendors are not the same number

Sampling rates and hidden smoothing mean the same sprint produces different high-speed distance on different systems.

Run the same player, in the same session, wearing two different tracking units, and the high-speed distance will not match. Not by a rounding error. By enough to change a decision.

Most of the gap comes from choices made inside the software rather than anything on the pitch. A satellite unit sampling ten times a second and one sampling eighteen times see a change of direction differently. Every system smooths its raw signal before computing velocity, because raw position data from a moving athlete is jumpy, and the amount of smoothing is a design decision that vendors treat as their own business. Heavier smoothing clips the peaks, so peak speed drops and the time spent above a speed band shrinks. Lighter smoothing keeps noise that gets counted as sprinting.

Then there are the bands themselves. High-speed running above a fixed threshold means one thing for a winger and something else for a centre back, and some systems set the band relative to each player's maximum while others use one number for everybody. Two clubs can both report high-speed metres, define them differently, and publish the comparison anyway.

What this rules out is most of the cross-club conversation that happens in the sport. A player arriving from another team brings a training history expressed in units that are not yours. Treating his previous weekly distance as directly comparable to your squad's is a mistake that can point either way, and it is being made most often in exactly the situation where the stakes are highest, which is the first month at a new club with a new coach who wants to know what he can handle.

The fix inside a club is to never change vendor mid-season, and if you must, run both for a month and build your own conversion. It will not be a clean multiplier, because the difference varies with the type of movement, but even a rough correction beats pretending the series is continuous.

The fix between clubs does not exist without shared definitions nobody is incentivised to publish.

So when a number from another organisation lands on your desk, the right response is not to plot it next to yours. It is to ask what sampling rate produced it, and to notice how rarely anybody can tell you.