Two tracking vendors will give you two different sprint counts for the same match

Sprint totals depend on a threshold and a sampling rate that each provider chose independently, and neither prints them on the graphic.

Nineteen sprints, says one system. Twenty-six, says the other. Same player, same ninety minutes.

Both are correct, because a sprint is not a natural category. It is a definition. Somebody picked a speed above which movement counts, somebody picked a minimum duration, and somebody picked how to smooth the raw position data before the speeds were computed. Change any of those three and the count moves, sometimes by a third.

Threshold choice is the big one. A fixed cut off is common, and it will classify a quick centre back and a fast winger completely differently even when both are running at their own maximum. Some providers use a percentage of an individual's peak speed instead, which is more physiologically sensible and makes cross player comparison harder rather than easier. Neither approach is wrong. They are simply not the same variable wearing the same name.

Sampling rate compounds it. Position data collected at ten hertz and at twenty-five hertz produce different peak velocities, because a brief maximum can fall between samples. Filtering matters too, since the smoothing applied to remove positional jitter also removes some genuine acceleration, and every vendor smooths differently.

The result is that sprint counts are not portable. A player's figure from one league, collected by one supplier, cannot be compared with a figure from another league without knowing both definitions, and the definitions are commercial information that does not travel with the number.

Where this actually bites is recruitment and load monitoring. A club signs a player with a reported high intensity profile, brings him into a system using different capture, and finds the numbers drop. Somebody concludes he is not adapting. Nothing about the player changed except the instrument.

The fix is not technically hard and it is culturally impossible, which is to publish thresholds, sampling rate and filter alongside every figure, the way any laboratory would. Instrument specifications belong with the reading.

What clubs can do, and the good ones do, is recalibrate internally: keep a small set of players measured on both systems during a transition and derive a conversion. Unglamorous. Effective.

Anyone quoting a sprint count without naming the system that produced it is quoting a number with no units. We would not accept that from a speed gun. We accept it constantly from a vest.