Better Prevent than Tackle: Valuing Defense in Soccer Based on Graph Neural Networks

2025-12-11

AI Summary

Traditional football defensive evaluation relies heavily on explicit actions (e.g., interceptions, tackles), overlooking preventive defense—i.e., threat suppression through positioning and pressing *before* offensive actions unfold. This work introduces DEFCON, the first framework to systematically quantify defenders’ fine-grained contributions during the nascent phase of attacks. Leveraging a graph attention network (GAT), DEFCON models multi-agent spatial relationships by jointly encoding event and player-tracking data. It proposes a responsibility-aware, dynamic Expected Possession Value (EPV) attribution model that yields interpretable, moment-by-moment, region-specific, and matchup-resolved defensive impact scores. Validated on Eredivisie data, DEFCON-derived cumulative defensive credit exhibits a statistically significant positive correlation with market valuations (p < 0.01). The framework further enables real-time defensive timeline visualization, zonal heatmaps, and matchup-specific diagnostic analytics.

Technology Category

• Multiagent Systems: Adversarial Agents
• Game Theory and Economic Paradigms: Adversarial Learning
• Humans and AI: Planning and Decision Support for Human-Machine Teams

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Evaluating defensive performance in soccer remains challenging, as effective defending is often expressed not through visible on-ball actions such as interceptions and tackles, but through preventing dangerous opportunities before they arise. Existing approaches have largely focused on valuing on-ball actions, leaving much of defenders' true impact unmeasured. To address this gap, we propose DEFCON (DEFensive CONtribution evaluator), a comprehensive framework that quantifies player-level defensive contributions for every attacking situation in soccer. Leveraging Graph Attention Networks, DEFCON estimates the success probability and expected value of each attacking option, along with each defender's responsibility for stopping it. These components yield an Expected Possession Value (EPV) for the attacking team before and after each action, and DEFCON assigns positive or negative credits to defenders according to whether they reduced or increased the opponent's EPV. Trained on 2023-24 and evaluated on 2024-25 Eredivisie event and tracking data, DEFCON's aggregated player credits exhibit strong positive correlations with market valuations. Finally, we showcase several practical applications, including in-game timelines of defensive contributions, spatial analyses across pitch zones, and pairwise summaries of attacker-defender interactions.

Problem

• Quantifies defensive contributions beyond visible on-ball actions.
• Uses Graph Neural Networks to estimate attack success probabilities.
• Assigns credit to defenders based on opponent's Expected Possession Value changes.

Innovation

• Graph Attention Networks model defensive contributions
• Quantifies EPV changes to assign defender credits
• Correlates defensive metrics with player market valuations