A single injury-risk score — whether displayed as a red gauge or a 76% probability — compresses completely different clinical situations into the same number. Worse, it hides uncertainty and missing measurements. That is why we do not build Omen Performance around a single injury probability.
Clinical judgement requires a transparent conversation. A head physiotherapist or club doctor must be able to return to the raw source in one click, inspect the trajectory over seven or twenty-eight days, and explain to the coaching staff why a player needs a modified session today.
Why black-box predictive scores lose the trust of club medical staffs
When an algorithm flags a player as "high risk" without stating whether the driver is an acute sprint spike, travel fatigue, or an old calf scar, clinicians face an impossible choice: trust the number blindly or ignore the system after its second false alarm.
- Loss of clinical nuance: travel fatigue and early tendon overload require completely different interventions.
- Hidden missing data: when two key measurements are missing, a numeric score projects false mathematical precision.
- No operational prescription: a percentage never tells the staff whether to cap high-speed meters, modify gym work, or examine the player on the table.
How qualitative, source-linked AI prepares the morning clinical review
In Omen Performance, every indicator keeps its source, time window and recent trajectory: training load, RPE, freshness, test results, reported pain and relevant history. Staff can assess player condition with the available observations without turning RPE into an objective form score. The AI formulates a concise qualitative summary stating what supports the observation and what measurements are still missing, leaving interpretation and action to the staff.