We introduce a novel method for assessing agent teamwork based on their spatial coordination. Our approach models the influence of spatial proximity on team formation and sustained spatial dominance over adversaries using a Multi-agent Markov Decision Process. We develop an algorithm to derive efficient teamwork strategies by combining Monte Carlo Tree Search and linear programming. When applied to team defence in football using real-world data, our approach reduces opponent threat by 21%, outperforming optimised individual behaviour by 6%.

We introduce GAPP, a Graph Attention Network model that predicts football pass reception probabilities and provides interpretable insights into off-ball defending. Using attention mechanisms, GAPP captures player interactions and introduces two new metrics to quantify defender contributions. Tested on 306 English Premier League matches, GAPP outperforms baselines for pass reception prediction while offering unique insights for off-ball defender evaluation.

Player injuries in football significantly impact team performance, club financial stability and player welfare, with the ‘Big Five’ European leagues experiencing £513 million in injury-related costs during the 2021/22 season. We present a novel forward-looking team selection model, framed as a Markov decision process and optimised with Monte Carlo tree search, that balances team performance with the risk of long-term player unavailability due to injury.

This work tackles agent location imputation in environments with non-uniform timesteps and ~95% missing values. Using Long Short-Term Memory and Graph Neural Networks, we predict agent locations, applying this to football by imputing player locations from sparse event data. Our model increases accessibility to analysis such as physical metrics and pitch control without the need for costly optical tracking data.

We present a spatial Expected Threat (xT) model that computes goal likelihood using spatial event data. We propose a novel way of valuing attacking and defensive units in football by computing a ‘Threat Above Expected’ (TAx) metric which measures the change in goal probability resulting from spatial context. We demonstrate a concept for a positional defensive optimiser that can offer coachable insights by identifying improvements in defensive setups during post-match analysis.