Search papers, labs, and topics across Lattice.
TacticGen, a generative model for football tactics, is introduced, framing tactics as sequences of multi-agent movements conditioned on game context. A multi-agent diffusion transformer with specialized attention mechanisms captures player interactions, trained on a massive dataset of real game data. Classifier guidance allows for adaptable tactic generation based on rules, language, or neural models, validated by expert evaluation of realistic and strategically valuable tactics.
Forget trajectory forecasting – TacticGen generates *adaptable* football tactics, bridging the gap between predicting what *will* happen and prescribing what *should* happen to win.
Success in association football relies on both individual skill and coordinated tactics. While recent advancements in spatio-temporal data and deep learning have enabled predictive analyses like trajectory forecasting, the development of tactical design remains limited. Bridging this gap is essential, as prediction reveals what is likely to occur, whereas tactic generation determines what should occur to achieve strategic objectives. In this work, we present TacticGen, a generative model for adaptable and scalable tactic generation. TacticGen formulates tactics as sequences of multi-agent movements and interactions conditioned on the game context. It employs a multi-agent diffusion transformer with agent-wise self-attention and context-aware cross-attention to capture cooperative and competitive dynamics among players and the ball. Trained with over 3.3 million events and 100 million tracking frames from top-tier leagues, TacticGen achieves state-of-the-art precision in predicting player trajectories. Building on it, TacticGen enables adaptable tactic generation tailored to diverse inference-time objectives through classifier guidance mechanism, specified via rules, natural language, or neural models. Its modeling performance is also inherently scalable. A case study with football experts confirms that TacticGen generates realistic, strategically valuable tactics, demonstrating its practical utility for tactical planning in professional football. The project page is available at: https://shengxu.net/TacticGen/.
