In a new collaboration with Liverpool Football Club, DeepMind has developed an artificial intelligence model aimed to improve football tactics.
The tool, aptly named TacticAI, leverages data from Premier League matches to predict the outcome of corner kicks and assist coaches in devising strategic plays.
TacticAI was spearheaded by DeepMind researcher Petar Veličković during a three-year research partnership with Liverpool FC. The model analyses various factors, such as player positions, height, weight, and historical data from corner kicks in past seasons.
According to FootyStats, the average Premier League match will see 10.7 corner kicks in one game. Despite this, Opta Analyst shows that just 13.6% of goals in the last five seasons have come from corner situations.
With an abundance of chances and low success rate, corners present the perfect opportunity for artificial intelligence to have a potential impact. On top of that, corner kicks, or any set plays for that matter, are more predictable than regular play, making it easier for an AI model to use data from past games.
Recommended reading
- Former FanDuel Execs Gain License for Blockchain Betting Platform
- The Premier League is Considering Jumping on the NFT Bandwagon
- Smart Data Could Help Scots Football ‘Build Back Better’ After Pandemic
TacticAI identifies patterns and trends from real-life corner kick scenarios, providing recommendations to coaches on optimising player positioning and movements. Impressively, in tests, the AI accurately predicted the recipient of the ball after a corner kick in 78% of cases.
When put through blind tests with football experts from Liverpool FC, they were unable to distinguish between AI-generated tactics and those devised by humans and were even favoured by the experts 90% of the time.
Petar Veličković emphasised that TacticAI is not intended to replace human coaches but rather to enhance their capabilities. “We are strongly in support of AI systems that amplify human capabilities and leave them more time for the creative part of their work, rather than a system that would replace them,” he said.
Furthermore, Veličković highlights the broader implications of the research beyond the realm of sports. He suggests that understanding football dynamics can shed light on various aspects of human psychology, positioning TacticAI as a stepping stone for future AI systems tackling complex real-world scenarios.





