The Development of Mathematical, Statistical and Machine Learning Models of Sports Betting, with a View to Detection of Irregular or Corrupt Activities

Irregular or corrupt activities in Sport have made the News over recent years, with allegations and even legal convictions of players performing badly, “throwing” matches or doing very usual things due to the influence of gambling syndicates or corrupt bookmakers. Examples include test level cricketers deliberately bowling “no balls” at specific points in a match specified by corrupt bookmakers, to gambling on a minor team’s rather overweight reserve goalkeeper being caught eating a pie during a match.

Use of Statistical and/or Machine Learning based approaches to modelling “regular” behaviour in gambling patterns during matches, and hence detecting anomalous activities, should be able to help detect such illicit behaviour. The aim of this project is to use gambling data and changes in match or spot odds from real sports matches to develop mathematical and computational models to detect such irregularities for one or more sports. This will assist sporting and legal authorities detect result fixing and illegal gambling.