Rally-Based Performance Evaluation Model for Highly Competitive Volleyball

Abstract

We propose a simple stochastic model to evaluate the effect of different complexes’ performance on the probability of winning a rally. The model uses as input the probabilities of success and failure in various complexes, which can be extracted from standard match reports. Our model reproduces well-established results; for example, we found that a team that starts the rally with a serve is more likely to obtain a point in the phase of complex k2 than in the phases associated with complexes k0 and k1. Conversely, if a team starts the rally receiving, it is more likely to win the rally in complex k1. The proposed model also provides a new approach to quantify a team’s performance in a rally and diagnose performance issues in different complexes. As a case study, we analyze the performance of a top South American team in the CSV Men’s Tokyo Volleyball Qualification 2020. Although our model can be applied to various individual actions, our performance analysis focuses on one pivotal game action: the serve. It is found that only power jump serves that decrease the attacking efficiency in k1 of the rival team have the potential to be more effective than jump float serves. The proposed model makes it possible to determine when one player’s serve is more effective than another’s, not only based on the number of direct points scored for or against but also on their influence on the probability of winning the rally.