How AI is opening the playbook on sports analytics

Waterloo researchers Dr. David Radke and PhD student Kyle Tilbury developed a system that simulates and records unlimited soccer matches using Google Research Football’s reinforcement learning environment. By generating data from 3,000 games, they created a detailed dataset of passes, goals, and player movements. These types of resources are normally locked behind professional sports’ proprietary systems. The work opens new opportunities for researchers, smaller teams, and fans who lack access to expensive tracking data. Beyond sports, the datasets also advance AI research in multiagent systems, helping to model complex human behavior in dynamic environments. The study, "Simulating tracking data to advance sports analytics research," appeared in the 24th International Conference on Autonomous Agents and Multiagent Systems.