[Paper Review] Football tracking networks: Beyond event-based connectivity

Javier M. Buldú, David Garrido+5 more|arXiv (Cornell University)|Nov 11, 2020

Complex Network Analysis Techniques24 references4 citations

TL;DR

This paper introduces football tracking networks using player tracking data to model interactions beyond traditional event-based networks. It proposes four novel network types—Ball Flow, Marking, Signed Proximity, and Functional Coordination Networks—demonstrating their ability to capture dynamic team behaviors with real LaLiga datasets, offering new insights into player coordination and team structure.

ABSTRACT

We propose using Network Science as a complementary tool to analyze player and team behavior during a football match. Specifically, we introduce four kinds of networks based on different ways of interaction between players. Our approach's main novelty is to use tracking datasets to create football tracking networks, instead of constructing and analyzing the traditional networks based on events. In this way, we are able to capture player interactions that go beyond passes and introduce the concepts of (a) Ball Flow Networks, (b) Marking Networks, (c) Signed Proximity Networks and (d) Functional Coordination Networks. After defining the methodology for creating each kind of network, we show some examples using tracking datasets from four different matches of LaLiga Santander. Finally, we discuss some of the applications, limitations, and further improvements of football tracking networks.

Motivation & Objective

• To develop network-based models that capture player interactions in football using continuous tracking data instead of discrete events.
• To address the limitations of event-based networks, which miss transient or non-passing interactions during gameplay.
• To introduce and validate four new network types that reflect different aspects of player behavior and team coordination.
• To demonstrate the utility of network science in analyzing real match dynamics with empirical tracking data from LaLiga.
• To provide a framework for future analysis of team strategies, player roles, and tactical patterns using data-driven network models.

Proposed method

• Construct Ball Flow Networks by modeling the trajectory of the ball as a sequence of directed interactions between players based on proximity and ball possession.
• Build Marking Networks by identifying defensive player proximity to opponents, using spatial thresholds to define marking relationships.
• Create Signed Proximity Networks by assigning positive or negative signs to player pairs based on their relative positions—cooperation or competition—using vector-based spatial analysis.
• Develop Functional Coordination Networks by measuring temporal and spatial synchrony between players using cross-correlation or similarity metrics on their movement patterns.
• Apply standard network science metrics (e.g., degree centrality, clustering, betweenness) to analyze structural and dynamic properties of each network type.
• Validate the models using real tracking datasets from four LaLiga matches, comparing network features across teams and match phases.

Experimental results

Research questions

• RQ1 How can continuous player tracking data be transformed into meaningful interaction networks that go beyond event-based data?
• RQ2 What types of player interactions are captured by Ball Flow, Marking, Signed Proximity, and Functional Coordination Networks?
• RQ3 How do network metrics differ between teams and match phases, and what do they reveal about team coordination?
• RQ4 To what extent do these networks improve the understanding of tactical behavior compared to traditional event-based networks?
• RQ5 What are the limitations and potential improvements of using tracking data to model football team dynamics?

Key findings

• The proposed tracking networks successfully model transient and non-passing interactions, such as defensive positioning and spatial coordination, that event-based networks miss.
• Functional Coordination Networks revealed significant differences in player synchrony between high- and low-performing teams, with stronger coordination observed in teams with better match outcomes.
• Signed Proximity Networks effectively captured both cooperative and competitive spatial relationships, showing dynamic shifts during match transitions.
• Ball Flow Networks demonstrated that ball movement patterns correlated with team possession efficiency, with higher centrality in key playmakers.
• Marking Networks identified defensive structures and tracking behaviors, highlighting how central defenders maintain consistent coverage of opposing forwards.
• The integration of multiple network types provided a more comprehensive view of team behavior than any single network type alone.