Modelling the 2004 Super 12 Rugby Union Competition

This paper uses Bayesian methods via WinBUGS to model round robin play in the 2004 Super 12 Rugby Union competition in order to explore home advantage and how that impacts the outcome of the competition. The scores from the games are decomposed into counts of converted and unconverted tries, penalties and drop goals and are modelled as Poisson random variables with a log link. The dependent variables are the offensive and defensive capabilities of the teams along with terms for home advantage. The model is used to ascertain the effects of home advantage on the standings of the teams in the competition and, from that, how fairness in the competition could be improved.