Modeling Matches as Language: A Generative Transformer Approach for Counterfactual Player Valuation in Football

TL;DR

This paper introduces ScoutGPT, a generative Transformer model that simulates football match events as language tokens, enabling counterfactual analysis of player transfers and their impact on match outcomes.

Contribution

The paper presents a novel language modeling approach for football match events, allowing counterfactual simulations to evaluate player transfer impacts beyond static statistics.

Findings

ScoutGPT outperforms baseline models in predicting match event sequences.

Counterfactual simulations reveal significant changes in offensive metrics.

Model captures player-specific impacts on match dynamics.

Abstract

Evaluating football player transfers is challenging because player actions depend strongly on tactical systems, teammates, and match context. Despite this complexity, recruitment decisions often rely on static statistics and subjective expert judgment, which do not fully account for these contextual factors. This limitation stems largely from the absence of counterfactual simulation mechanisms capable of predicting outcomes in hypothetical scenarios. To address these challenges, we propose ScoutGPT, a generative model that treats football match events as sequential tokens within a language modeling framework. Utilizing a NanoGPT-based Transformer architecture trained on next-token prediction, ScoutGPT learns the dynamics of match event sequences to simulate event sequences under hypothetical lineups, demonstrating superior predictive performance compared to existing baseline models. Leveraging this capability, the model employs Monte Carlo sampling to enable counterfactual simulation, allowing for the assessment of unobserved scenarios. Experiments on K League data show that simulated player transfers lead to measurable changes in offensive progression and goal probabilities, indicating that ScoutGPT captures player-specific impact beyond traditional static metrics.