Squad load data fails at the meeting, not the sensor

Clubs collect far more physical information than they can act on, then hand the decision to people whose incentives point in opposite directions.

The sensors are fine. The vests work, the numbers arrive, the reports are generated overnight, and by ten in the morning a sports scientist has a screen showing exactly which players are carrying more than they should. Then that screen has to become a sentence spoken out loud to a manager who is trying to win on Saturday, and this is where the whole system quietly falls apart.

Consider who is in the room. The medical staff carry the blame for injuries and none of the credit for results. The manager carries the blame for results and can usually survive an injury by calling it bad luck. Handed the same information, those two people rationally reach different conclusions, and the one with more authority tends to win. That is not a data problem. It is an organisational one, and no better wearable fixes it.

There is a second failure, which is that load reports are almost always presented as risk without a counterfactual. A player is flagged as at risk. Compared with what, and what happens if he plays anyway? Nobody says, because saying it would require committing to a probability, and if you commit to a probability you can be wrong in public. So the report hedges, the manager reads hedging as permission, and the player starts.

I think clubs would get more out of halving the number of metrics and forcing a single recommendation with a named owner. Play, manage, or rest. One word. Somebody signs it. That sounds crude next to a dashboard with forty variables and it would improve outcomes immediately, because the failure mode in these departments is not ignorance, it is diffusion of responsibility.

The players make it harder still. Self-reported wellness scores are the input that most influences the model and the one most easily gamed, because a professional fighting for a shirt learns within a month which answers get him rested. Everyone knows this. Almost nobody adjusts for it, and the model keeps treating a strategic answer as an honest one.

Buy fewer sensors. Fix the meeting. The most expensive thing in the building is the conversation nobody wants to have on a Friday.