Detecting anomalous matches in professional sports: a novel approach using advanced anomaly detection techniques
Dr Dulani Jayasuriya, Business and Economy, Business School
Abstract
This research focuses on improving the detection of match fixing and anomalous events in professional sports by developing a novel method that incorporates advanced anomaly detection techniques.
Our approach examines series of matches, like playoffs, to identify anomalous matches and player performances. The study also explores the potential impact on bettors’ profits and investigates factors behind unusual player performances.
Introduction
The integrity of professional sports is increasingly threatened by match fixing, necessitating the development of robust detection methods. Traditional research in this field has primarily utilized sports betting market data, often overlooking other potential motives behind match fixing, such as influencing ticket sales or deliberately losing matches. Our study aims to address these gaps by employing outlier detection theories and identifying anomalous matches and player performances that deviate significantly from the norm.
Research goals
Our primary objectives are to leverage outlier detection theories for better identifying potential fraudulent activities in sports data and to ascertain if it is possible to identify stakeholders or players associated with such anomalous matches. The study’s innovative methods contribute to the literature by refining and building upon existing approaches, assisting law enforcement and sports governing bodies in identifying individuals responsible for anomaly matches.
Methodology and results
The study employs several machine learning algorithms, including Ordinary Least Squares (OLS), Stochastic Gradient Descent (SGD), and Decision Tree Regression (DTR), to forecast player performances and, subsequently, match outcomes. The data were segregated into bench players and non-bench players based on their distinct performance correlations. Our approach involves a comparative analysis between our predictions and actual outcomes to identify anomalous matches. Here is the link to the paoer https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4493430
Conclusion
This research introduces a new methodological framework for detecting anomalous matches and player performances in professional sports. By employing advanced anomaly detection techniques and machine learning algorithms, we have identified several anomaly series in the 2022 NBA playoffs and players with exceptionally abnormal performances. Our findings suggest potential factors behind these anomalies, such as team financial difficulties and executive mismanagement, contributing to the broader understanding of match fixing in sports.