Detecting Anomalous Matches: An Empirical Study From National Basketball Association

Abstract

Match fixing and anomalous sports events have increasingly threatened the integrity of professional sports, prompting concerns about existing detection methods. This study addresses prior research limitations in match fixing detection, improving the identification of potential fraudulent matches by incorporating advanced anomaly detection techniques. This study develops a novel method to identify anomalous matches and player performances by examining series of matches, such as playoffs. Our findings reveal anomalous series in the 2022 NBA playoffs, with the Phoenix Suns vs Dallas Mavericks series having the lowest natural occurrence probability. This study contributes by developing a new approach to detect anomalous matches and assisting investigators in identifying responsible parties.

Keywords: Anomaly Match Detection, Match Fixing, Match Outcome Forecasting, Problematic Players Identification, Sports Betting.