Evaluating the efficiency of the association football transfer market using regression based player ratings

Abstract

In recent times, the use of quantitative methods to improve decisions within sports has increased. In association football, large amounts of match data has become available. This work first shows how simple match data describing the players on pitch and the time for events such as goals and red cards, can be used to derive an objective player rating. The rating is based on solving a large linear regression model. The resulting player ratings are in turn used as input to a regression model for analyzing transfer fees. It is shown that the performance of players, as reflected in the player ratings, is an important predictor of transfer fees. At the same time, several other important factors that determine the size of transfer fees are identified.