Abstract

According to the Fédération Internationale de Football Association (FIFA), while the sports industry itself generated revenue totalling USD 300 billion in 2011, sports betting was worth USD 350–400 billion.  The primary objective of this case is to analyze the past data from the English Premier League (EPL) to develop winning strategies for in-play betting. In-play betting allows bettors to place their bets while the match is in progress. In most instances, the bettors can place bets any time during the match and can bet on the final score, match result, next goal scorer, and so on. Unlike the other bets, in-play betting odds are updated in real-time by the betting companies based on the events in the match. In-play betting provides an opportunity to the bettors to mitigate the risk on their other bets. For example, consider a bettor who has placed a bet on a particular team to win before the start of the play. During the match, if the other team leads, he/she can mitigate the risk of losing money by placing another in-play bet. The objective of the case is to predict the outcome of football matches using classification techniques such as multinomial logistic regression and chi-squared automatic interaction detection (CHAID).  The accompanying data in the Excel spread-sheet (IMB403EPL.XLS) may be used by the instructor to demonstrate application of several classification tools.

Learning Objectives

The case can be used in advanced statistics or business analytics courses of MBA or executive MBA programmes. The case is suitable for teaching classification techniques such as binomial and multinomial logistic regression as well as chi-squared automatic interaction detection (CHAID).  The instructor may use the case to teach prediction of the classification probabilities, likelihood ratio tests, Wald’s test, classification table, classification plots, and so on.