Google’s Deepmind AI claims to have mastered football tactics

Google DeepMind’s TacticAI system has figured out football tactics (Deepmind)
Google DeepMind’s TacticAI system has figured out football tactics (Deepmind)

Google’s leading AI division DeepMind has built a tool that it claims can match expert football tactics and predict the outcome of corner kicks.

DeepMind has previously used its artificial intelligence to defeat human champions in the notoriously complex board game Go, as well as beat professional gamers at the video game StarCraft II.

The London-based AI division unveiled TacticAI on Tuesday following three years of development alongside staff from Liverpool Football Club.

It uses a combination of generative and predictive AI to achieve “state-of-the-art results”, according to a blog post detailing the technology.

“TacticAI demonstrates the potential of assistive AI techniques to revolutionise sports for players, coaches and fans,” the blog post reads.

“Our system allows coaches to sample alternative player setups for each routine of interest, and then directly evaluate the possible outcomes of such alternatives.”

Tactics generated by the AI system were indistinguishable from human-designed tactics in a blind test, according to a study detailing the technology.

Setups proposed by TacticAI were also favoured 90 per cent of the time compared to the human-designed tactics.

Despite the success, DeepMind said the tool is not designed to replace human coaches, but rather assist them.

“TacticAI’s generative model also allows human coaches to redesign corner kick tactics to optimise probabilities of certain outcomes, such as reducing the probability of a shot attempt for a defensive setup,” DeepMind said.

“From these proposed adjustments, coaches can identify important patterns, as well as key players for a tactic’s success or failure, more quickly.”

The system, which was trained on data from 7,176 corner kicks in the 2020/21 Premier League season, offers coaches a way to simulate and analyse different player setups for routines.

Beyond that, DeepMind says it could be developed further to offer tactical advice in other sports, as well as general insights in human psychology that could translate to areas like robotics and traffic coordination.

The research was published in the journal Nature Communications in a paper titled ‘TacticAI: an AI assistant for football tactics’.