Follow by Email

Search This Blog

Nov 16, 2018

AI in football match analyis

Any football match (some call it soccer, but to me, it's football, and Amercian football is "handegg") in a major league produces millions of data points that are tracked, summarized and visualized. Heatmaps alone are based on tracking each player's movements on the pitch at any given time, the position of the ball is tracked, passes made, shots etc. All this data is mostly used to entertain us, the fans, and help coaches in their post-match analysis.

Chelsea is now working with scientists on a super-ambitious tpoic: evaluate decision making on the pitch. AI scientists try to determine what would have happened if the player would have made a different decision in any given situation. This is so complex (because it effects the decision subsequently made by others, too) that I am really sceptical about the (at least short term) outcome, but it shows the ambitions and possibilities of AI in sports. At some point in the future, we might have this "scenario analysis" in real time while we are watching football.