A new way to predict injury risk by grouping minute-by-minute football data into fixed check points.

Sports devices can track an athlete's body every minute during a match or practice. However, team records usually only note if an injury happened at some point during the entire session. Labeling every single minute as injured is wrong because nobody knows the exact minute the injury occurred. To fix this mismatch, researchers created a new method using fixed checkpoints called landmarkslandmarksFixed time points within a session, such as 10, 20, or 30 minutes, used as checkpoints., such as 10, 20, or 30 minutes into a session. At each checkpoint, they gather all the data recorded up to that specific time into a single summary for that athlete's session. They then test whether they can tell apart sessions with injuries from sessions without them, watching how clues change as more time passes. They tested their model on elite women's football records. Overall, detecting injury-linked sessions at these checkpoints proved very difficult, producing low accuracy scores with wide uncertainty.

Why it matters

For elite football athletes, accurate injury prediction could help coaches and medical teams manage training and prevent injuries that affect careers. This method tries to make better use of continuous data without misleading labels.

Questions answered