A data-driven approach to predicting the most valuable player in a game

Abstract

The identification of outstanding behaviors is a matter of essential importance in sports analytics. However, analyzing how human experts select each match's most valuable player (MVP) according to objective and subjective factors is a great challenge. This article proposes a data-driven approach for sports team performance based on the weighted aggregation of statistical indicators. The proposal is divided into two approaches: The first conducts a principal component analysis to examine the relationship between each game's statistical indicators. The other addresses a meta-heuristic analysis to weight the attributes and choose the MVPs optimally. Finally, we apply the proposed approach to the 2018 European Men's Handball Championship and take the “Player of the Match” of each game as an example to illustrate its usefulness and efficacy. We perform multiple analyses, including a comparison with a fuzzy multi-criteria decision-making method that show that the data-driven approach can predict the “Player of the Match” in most matches. It also allows us to estimate and quantify the expert evaluations, which are often difficult to obtain in a disaggregated form.