Competition is the worst possible environment for analysis and the most valuable moment for intelligence. Everything is moving, everyone is occupied, and the decisions carry the most weight. Any system built for match day has to start from that contradiction.
The 90-minute problem
Across a match, an elite programme’s tracking and physiological systems generate orders of magnitude more data than any human can attend to, while the staff’s attention is, correctly, on the game. The paradox: the richer the data environment gets, the less of it anyone can actually consume in the moment.
Adding another screen to the bench doesn’t solve this. It makes it worse. Attention is the scarcest resource in the stadium.
Prediction as triage
The way through isn’t more information. It’s ruthless triage. Of everything happening right now, what actually changes a decision that’s still open? That’s a prediction problem:
- Which athlete is drifting away from their own baseline, not the squad average but their normal?
- Which pattern, if it continues, becomes a problem in the next twenty minutes?
- Which of the hundred things that just happened deserves one of the three moments of attention the staff can spare?
A live system’s job isn’t to show everything. It’s to interrupt exactly once, with exactly the right thing, at exactly the moment there’s still time to act.
The bar for live intelligence
For intelligence to survive contact with a touchline it has to clear three bars at once. It has to be trustworthy: staff will act on a signal twice; if it’s wrong once, it’s furniture. It has to be glanceable: absorbed in the seconds between phases of play, not studied. And it has to be timed: delivered inside the decision window, not after it closes.
Get those right and something shifts: data stops being the thing you review after the result and starts being part of how the result happens.