Sports

Artificial intelligence could determine what plays will win the game

Artificial intelligence could determine what plays will win the game
UC3M's technology analyzes the actions of players to determine which plays will work best
UC3M's technology analyzes the actions of players to determine which plays will work best
View 1 Image
UC3M's technology analyzes the actions of players to determine which plays will work best
1/1
UC3M's technology analyzes the actions of players to determine which plays will work best

If there’s one thing that sports fans love to debate, it’s coaching strategies. “Why didn’t he keep more players back to play defense?” “How come he had him pass instead of run with it?” “He should never have let that guy bat when the bases were loaded!” Such discussions could seemingly go on forever, as it’s impossible to definitively say what the right course of action would have been... or maybe not. Artificial intelligence researchers at Spain’s Universidad Carlos III de Madrid (UC3M) are developing technology that could analyze a team’s performance, then objectively determine the best plays for specific situations.

“In the near future, performance analysis of executions and decisions in real time could be made, providing precise feedback to improve performance during competition” stated head researcher Miguel Ángel Patricio.

The UC3M Artificial Intelligence Group started with the sport of basketball, using a series of cameras to record all the actions of all the players on the court during a game. They then analyzed that footage via “complex reasoning algorithms,” in order to determine what sort of tactics were taking place within it. Now, they are working on turning those observations into “If this team is in this situation, here’s what they need to do to win”-type recommendations. While the data would be invaluable for guiding teams within games, it could also be used to figure out why some teams generally win more often than others.

The researchers point out that their system is more objective than human analysts, and is capable of processing much more information than a human brain ever could.

A complimentary process for evaluating the performance of individual players is also in the works, utilizing time-of-flight cameras to obtain three-dimensional biomechanical models of the athletes. The UC3M group suggest that the technology could proceed to be used in the rehabilitation of people who have suffered injuries that affect their movements.

Both projects are being conducted in conjunction with the research group Deporte y Rendimiento (Sport and Performance) from the Universidad Politécnica de Madrid.

2 comments
2 comments
TogetherinParis
Even on the run direct coaching could prove effective, since in football at least, plays rarely run as designed. Players often wind up in formations from which they know no plays. Designing plays from formations formed after the ball is snapped would improve cohesion for teamwork.
HenryFarkas
The trouble with using a computer to decide what the best play is that the other team\'s computer can use the same information to figure out what play the other computer is going to suggest.