Palmeiras tests DeepMind’s AI that predicts football plays 8 seconds early
▼ Summary
– Google DeepMind’s TacticAI uses geometric deep learning to predict football plays up to eight seconds ahead from broadcast video, and Brazilian club Palmeiras is the first to use it for live open-play analysis.
– In a study with Liverpool FC, football experts preferred TacticAI’s recommended tactical setups over real match configurations 90% of the time.
– Palmeiras uses a drag-and-drop interface to virtually reposition players and simulate how changes affect both teams’ collective behavior, turning gut feelings into quantified options.
– Google also partnered with Brazil’s football confederation CBF to use AI in World Cup preparation, and the technology has applications beyond sport in areas like autonomous robots and traffic systems.
– Football has been slower than other sports to adopt AI tactics due to complex 22-player dynamics, but TacticAI’s 90% expert preference rate signals a shift from background analytics to sideline tactical recommendations.
Google DeepMind has developed an artificial intelligence system capable of predicting football plays before they unfold on the pitch. Called TacticAI, it leverages geometric deep learning to model player movements, forecast game dynamics up to eight seconds in advance, and suggest tactical adjustments , all using standard broadcast-quality video footage. Brazilian club Palmeiras has become the first team to deploy this technology for live open-play analysis.
Originally built in collaboration with Liverpool FC, TacticAI was validated through a qualitative study involving the club’s football experts. These specialists compared the AI’s recommended tactical formations against actual match configurations. Impressively, they preferred the AI’s suggestions 90 percent of the time. The findings, published in Nature Communications, also showed that TacticAI outperformed existing models in predicting corner kick receivers and determining whether a shot would follow.
The partnership with Palmeiras, announced at Google’s Brasil event on June 10, represents a major milestone. Previously, TacticAI was limited to set-piece analysis, specifically corner kicks. Now, Palmeiras is the first club to apply it to open-play dynamics. The club’s data science team uses a simple drag-and-drop interface to virtually reposition players and observe how those changes affect the collective behavior of both their own squad and the opposition.
This means a coach can ask a hypothetical: what happens if we push the left back five meters higher? TacticAI then simulates the downstream impact on the entire defensive structure. It effectively quantifies tactical options that were once reliant on gut instinct. Google has also partnered with Brazil’s football confederation, CBF, to integrate AI into World Cup preparation.
The technology’s potential extends far beyond the pitch. The challenge of predicting coordinated movement from visual data is the same problem faced by autonomous robots, traffic systems, and logistics planners. TacticAI’s geometric deep learning approach, which treats players as nodes in a dynamic graph and models their spatial relationships, is architecturally closer to physical AI systems than to a standard language model.
Football has been slower than other sports to adopt AI-driven tactics. Baseball has Statcast, and basketball has Second Spectrum. Football’s continuous, 22-player dynamics make it inherently harder to model. However, TacticAI’s 90 percent expert preference rate suggests that gap is closing. The growing role of AI in professional sport is now moving from background analytics to tactical recommendations delivered on the sideline.
(Source: The Next Web)