Bayesian Prediction of Volleyball Sets Using the Truncated Skellam and the Ordered Multinomial Models
share
In this work, we focus on building Bayesian models to analyze the outcome of a volleyball game as recorded by the difference of the winning sets for the Greek A1 men's League of the regular season 2016/17. More specifically, the first and foremost challenge is to find appropriate models for the response outcome which cannot be based on the usual Poisson or binomial assumptions. Here we will use two major approaches: a) an ordinal multinomial logistic regression model and b) a model based on a truncated version of the Skellam distribution. For the first model, we consider the set difference as an ordinal response variable within the framework of multinomial logistic regression models. Concerning the second model, we adjust the Skellam distribution in order to take into account for the volleyball rules. We fit and compare both models with the same covariate structure as in Karlis Ntzoufras (2003). Both models are fitted, illustrated and compared using data from the Greek Volleball League for 2016/17.
READ FULL TEXT
