Analysis of collective team behaviour: sub-elite female rugby 7s

Abstract

Introduction: Current technology enables the capture of a player’s location on the field using computer vision technologies, providing insight into the complex, non-linear interactions between individuals that are responsible for team-level formations and collective movement that emerge without apparent central control. Self-organisation and emergent behaviour has been widely studied in biological systems and can be observed in the impressive displays of ‘flocking’ and ‘shoaling’/’schooling’ behaviour seen in species such as fish and birds. Studies of pedestrian movement have also demonstrated emergent collective behaviour and this provides a basis to study the development of collective behaviours that are embedded within the movement of individuals in team-based sport. This study performs a preliminary analysis, using techniques previously applied to biological systems, to study collective movement in Rugby 7s. Methods: Two segments of gameplay from a game of Rugby sevens were selected for analysis; one in which a try was scored (Play 1) and one which ended in a foul (Play 2). The player positions were recorded semi-autonomously using the Channel and Spatial Reliability Tracker (CSRT) algorithm implemented in OpenCV, and then transformed to a metre based coordinate system. Location was recorded every 3 frames from 30 fps video resulting in 222 records (~22 seconds) for play 1 and 624 records (~62 seconds) for play 2. Two collective motion order parameters, Polarisation (Op) and Rotation (Or) were calculated. Polarisation measures the consensus in movement direction between all team members, whereas rotation measures the tendency of the team to rotate about the team centre in a consistent sense (clockwise or anticlockwise). Results: Order parameters follow consistent trends in response to game events, such as passing through multiple players and tackling. Play 2 provides multiple examples of successful tackles which differ in group order from the unsuccessful attempt that allowed a break in Play 1, resulting in a try. Discussion: By using tools from biological systems we can quantify the coordination of team members and compare this to performance outcomes. The plays studied demonstrated different collective movement patterns, which resulted in either a positive or negative outcome. Rigorous quantitative analysis of collective team behaviour will help practitioners in tactical and technical training design for improved performance outcomes. Conflict of interest statement: My co-authors and I acknowledge that we have no conflict of interest of relevance to the submission of this abstract.