“Real-time” is the most overused phrase in sports technology. In practice it usually means “faster than last season”. Here’s the definition we hold ourselves to, and why it changes how you build.
Three latencies, not one
When people talk about real-time data they usually mean capture: the sensor saw the sprint, the camera saw the press. But between the event and the decision sit three separate delays:
- Capture latency. How long before the event exists as data at all.
- Processing latency. How long before that data becomes something interpretable: a load number, a risk flag, a pattern.
- Delivery latency. How long before it reaches the person who can act on it, in a form they can absorb mid-session.
Most systems optimise the first, tolerate the second, and ignore the third. A dashboard nobody opens until the debrief has infinite delivery latency, whatever the sensor spec sheet says.
Real-time means inside the decision window
Data is real-time if it arrives while the decision it informs is still open. Otherwise it’s history.
The useful definition is relative, not absolute. A substitution window might be three minutes. A drill adjustment between reps might be thirty seconds. A medical call on a returning athlete might be a full afternoon. The same pipeline can be genuinely real-time for one decision and useless for another.
That’s why we start from the decision and work backwards: who makes this call, when does it close, and what’s the minimum signal they need before it does?
What it forces you to build
Taking the decision window seriously has consequences:
- Streams, not batches. If intelligence has to land mid-session, overnight processing is already too late.
- Glanceable outputs. Staff on a touchline get seconds, not minutes. The interface has to respect that.
- Graceful degradation. Venues have dead zones, sensors drop out, feeds stutter. A live system has to say “here’s what I still know” rather than going quiet.
None of this is glamorous. All of it is the difference between a system that demos well and a system a performance team actually leans on when the score is close.