Predicting match outcome according to the quality of opponent in the English premier league using situational variables and team performance indicators

ABSTRACT

The purpose of this research is to investigate the situational variables and performance indicators that significantly affect the match outcome (win, loss or draw) based on the quality of opposition. The data consisted of the situational variables and performance indicators of the matches in the English Premier League for the 2017–2018 season. One-way ANOVA, Tukey HSD, k-means clustering and decision tree approaches were implemented in the analyses. Scoring first was found as the most influential on match outcome in each decision tree, while the effects of clearances, shots, shots on target, possession percentage and match location on the match outcome varied according to the quality of opponent. An average of 2.43, 0.53 and 0.97 goals were scored by the teams that won, lost and drawn, respectively and teams that scored first won 67% of the matches. The decision trees based on the quality of opponent correctly predicted 67.9, 73.9 and 78.4% of the results in the games played against balanced, stronger and weaker opponents, respectively, while in all games (regardless of the quality of opponent) this rate is only 64.8%, implying the importance of considering the quality of opponent in the analyses. Coaches and managers can use these findings to create targets for players and teams during training and matches, and also can be prepared for these different competitive scenarios.

KEYWORDS:

• Decision tree
• clustering
• machine learning
• football
• scoring first
• quality of opposition

Disclosure statement

No potential conflict of interest was reported by the authors.