Prediction of Season-End Point for Football using Pythagorean Expectation

Authors : Sezer Baysal, Engin Yildiztepe

Pages : 63-73

View : 23 | Download : 13

Publication Date : 2019-03-01

Article Type : Research

Abstract :The use of data collected on players, teams, and games for performance evaluation, player selection, score-outcome estimation, and strategy development using data mining tools and techniques are defined as sports data mining. Performance measures, unlike the common statistical methods, developed for each sport branch have an important role in sports data mining processes. Performance measures calculated for team sports can be used to predict the expectation of winning. The Pythagorean expectation developed for this objective was originally used in baseball games. The Pythagorean Expectation has also been adapted for other team sports with two results, such as basketball. However, the studies using Pythagorean Expectation for sports which have three possible outcomes are very limited. In this study, a suggestion for the calculation of Pythagorean Expectation for football is presented. In the application section, end-season rankings and points for the 2017/2018 season of  the selected fifteen European football leagues are predicted by using the suggested method. The data of the past five seasons of the selected European football leagues is used as the training dataset. All calculations are performed in R.
Keywords : Sports data mining, Pythagorean Expectation, Point prediction, Soccer, Football