Financial performance forecasting of European football clubs: Machine learning model designs

The ability of professional football clubs to create sustainable financial structures is directly related not only to sporting success but also to effective financial management. In this context, correctly analyzing the financial performance of clubs and creating future forecasts is of critical importance in terms of both managerial decision-making processes and compliance with external financial regulations. In this study, the predictability of the financial performance of clubs was investigated using financial data from 1996 to 2022 belonging to 26 professional football clubs operating in Europe's top-level football leagues. Considering this situation, two basic performance indicators, return on equity (ROE) and return on assets (ROA), were tried to be predicted with the help of 19 different financial ratios. In the study, a total of 10 different regression algorithms based on both statistical and machine learning were applied and the prediction performances of the models were evaluated comparatively through various statistical measures such as MAE, RMSE, R². The findings confirmed that financial ratios exhibit structural variability over time and contain nonlinear relationships; as a result of the analyses, it was observed that nonlinear models such as Additive Model Tree (AMT), SMOreg and Gaussian Process exhibited more successful performance in estimating ROE and ROA. In contrast, linear models such as Elastic Net, Pace Regression and Linear Regression were limited and showed relatively low prediction success, especially in the face of complex financial structures. These results reveal the importance and effectiveness of nonlinear machine learning techniques in the financial performance analysis of football clubs. The findings show that machine learning-based models can be used effectively in the financial analysis of football clubs, going beyond traditional methods. It has also been concluded that such models can make significant contributions in terms of early diagnosis of financial risks of clubs and development of strategic decision support mechanisms.