|
Abstract:
|
We present a state-of-the-art Bayesian statistical analysis for studying large soccer data that has recently become available. Our data was obtained from OPTA in the case of the English Premier League and from the official data providers of the German Bundesliga and the World Cup 2014. The aim is to identify factors that play a major role in the goal creation. The major factors fall in three broad categories, namely the speed, the precision, and the discipline. Our approach is twofold. First, in the aggregate analysis section and the tracking section, we focus on the goal differential of the two teams, the difference between the home team and the away team scores. This approach is arguably more precise than focusing on probabilities of winning or losing the game as the corresponding models require non-linear transformation of data (in terms of logit or probit regressions). It also removes a high correlation of the tracking variables corresponding to the two teams. Second, we employ a Poisson regression when estimating the impact of precision (focusing on open crossing) on the scoring of an individual team. We study both the score differential and the absolute number of goals.
|
