Match performance variables influential in selection into the English national football team

• Main Author: He, Qixiang
• Other Authors: Kee Ying Hwa
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Science::General
• Online Access: http://hdl.handle.net/10356/76783
• Institution: Nanyang Technological University

Background: In recent years, talent identification has become an imperative tool in football, given the highly lucrative nature of the modern-day game. However, despite attempts to utilize multidimensional assessments for talent identification, there is little evidence to propose a set of factors that can reliably predict future footballing success. Purpose: To examine match actions of players in the four major playing positions in football (goalkeepers, defenders, midfielders, forwards) to see if they can distinguish between players selected or not selected for the national team. Method: Match actions data of English players playing in the English Premier League (EPL) from 2012-2018 were collected, with analysis done on 828 season-long observations. Discriminant analysis was run to identify performance variables that best distinguished those selected or not selected for the English national team for each position. The discriminant functions were tested on their ability to successfully classify the players into the national team or non-national team categories using the leave- one-out method. Results: Four unique sets of critical performance variables were identified using discriminant analysis. Multiple critical performance variables were identified for each position but the performance variable best able to distinguish the two selection groups were: saves made (goalkeepers), clean sheets kept (defenders), tackles made (midfielders), and shots on target (forwards). The four discriminant functions provided high overall percentages of successful classification (>77%). Conclusion: The usage of match performance variables profile analysis is a viable addition to existing multidimensional methods of football talent identification.