A series of plays in volleyball is modeled using a Markov chain, by defining the main states such as serve, receive, spike and spike-receive for both teams. In order to find optimal tactics in each state, we apply the method of Markov decision processes, and we make a formulation for obtaining the expected number of points gained in a series of plays beginning with a service. This approach makes it possible not only to calculate the expected number of points gained, but also to find the optimal tactics in each state, which maximize the expected number of points, by solving a set of value determination equations. Using a simple example, we demonstrate how this approach may help to identify the optimal tactics for a serve or spike, and how much the expected number of pointes increases if the optimal tactics are followed. This method will be applied to a real-time game analysis system.
© Copyright 2004 The Engineering of Sport 5, Volume 2. Veröffentlicht von International Sports Engineering Association. Alle Rechte vorbehalten.