Are NBA players getting paid according to their performance on court?

It is customary for researchers and practitioners to fit linear models in order to predict NBA player's salary based on the players' performance on court. However, in reality the behaviour of the association is non-linear and in this article we take this into account. We first select the most important determinants or statistics (years of experience in the league, games played, etc.) using the LASSO penalised regression and then utilise those determinants to predict the player salaries by employing the non linear Random Forest algorithm. We externally evaluate our salary predictions, via cross-validation, thus we avoid the phenomenon of over-fitting observed in most papers. Overall, using data from three distinct periods, 2017-2019 we identify the important factors that achieve very satisfactory salary predictions and we draw useful conclusions.