Optimising Game Tactics for Football

Ryan Beal,Georgios Chalkiadakis,Timothy J. Norman,Sarvapali D. Ramchurn

DOI: https://doi.org/10.48550/arXiv.2003.10294

2020-03-23

Abstract:In this paper we present a novel approach to optimise tactical and strategic decision making in football (soccer). We model the game of football as a multi-stage game which is made up from a Bayesian game to model the pre-match decisions and a stochastic game to model the in-match state transitions and decisions. Using this formulation, we propose a method to predict the probability of game outcomes and the payoffs of team actions. Building upon this, we develop algorithms to optimise team formation and in-game tactics with different objectives. Empirical evaluation of our approach on real-world datasets from 760 matches shows that by using optimised tactics from our Bayesian and stochastic games, we can increase a team chances of winning by up to 16.1\% and 3.4\% respectively.

Artificial Intelligence,Computer Science and Game Theory,Multiagent Systems

What problem does this paper attempt to address?

This paper attempts to optimize tactical and strategic decisions in football matches by establishing a novel mathematical model. Specifically, the author models a football match as a multi - stage game, which includes a Bayesian game for simulating pre - match decisions and a stochastic game for simulating the transformation of states and decisions during the match. Through this modeling method, the paper proposes a method to predict the probability of match results and the benefits of team actions, and based on this, develops algorithms to optimize team line - ups and in - match tactics to achieve different goals.

The main contributions of the paper are as follows:
1. Propose a new mathematical model of football matches and the tactical decision - making process.
2. Use real - data from 760 matches in the English Premier League (EPL) over the past two seasons to learn the benefits of different team actions and state transitions. In particular, the paper shows that it can predict match state transitions with an accuracy of up to 90% and can accurately predict opponents' tactical decisions.
3. By learning the benefits of actions, it is possible to optimize pre - match and in - match tactical decisions and increase the probability of winning the match.

These achievements not only provide a benchmark for computational models in football matches, but also open up new research areas for data - driven tactical decision - making in team sports.