The Modelling and Estimation of the On-Field Production Process in Rugby League: The Case of the National Rugby League

In recent years, the modelling of production in sport has emerged as a growing sub-set of production economics. There have been numerous studies concerned with the production processes within sport, with studies on football, basketball, baseball, ice hockey, cricket, and rugby league. This paper is an initial foray into the generation of game outcomes in the Australian-based National Rugby League (NRL) competition. In light of recent research identifying the need for thorough diagnostic testing and examination of data properties, examination of a cross-sectional time-series panel of data identifies evidence of both heteroskedasticity and autocorrelation. After modelling the data as a pooled generalised least squares (GLS) regression model, it is found that having an accurate goal-kicker, being in the lead at halftime, and being the 'stronger' team overall have statistically significant impacts on the final game outcome. It is also found that there is evidence of a home ground advantage for six teams upon modelling the data as a fixed-effects pooled GLS regression model.