Pythagorean expectation in EuroLeague data

The central subject of this study is to examine the applicability of Pythagorean Expectation (PE) formula (introduced by Bill James) on basketball EuroLeague data. A secondary aim is to identify the optimal value of the main turning parameter of this method based on EuroLeague data. The data of our study are composed of boxscore statistics of EuroLeague games for seasons from 2016–2017 to 2019–2020 (four seasons in total). By applying the PE model on EuroLeague data, the main conclusion is that the method predicts the final standings for 2019–20 season with accuracy comparable with other, more sophisticative, predictive methods such as logistic regression models using box-score statistics evaluated at the end of each game. Bootstrap was used to obtain confidence intervals of the PE main parameter. The estimated exponent parameter was found to range between 9.98 and 12.32 with 95% confidence (point estimate equal to 11.19). This is the first study which provides accurate (point and interval) estimation of the PE using EuroLeague Basketball data.