Portal de Periódicos da CAPES

THE RUGBY LEAGUE PREDICTION MODEL: USING AN ELO-BASED APPROACH TO PREDICT THE OUTCOME OF NATIONAL RUGBY LEAGUE (NRL) MATCHES

2016; Volume: 2; Issue: 5 Linguagem: Inglês

2455-295X

Joel Carbone, Tony Corke, Frank Moisiadis,

Sports Performance and Training

We introduce and discuss our model, which is used for rating teams in the National Rugby League (NRL) competition. It is an ELO-style model, modified to integrate several match variables into its calculations. Match data obtained from the 1999-2014 NRL seasons were used to build the model, and matches from the 2015 NRL season were used as a test sample. An automated computer system, employing an iterative approach, and writtenin R script, was used to optimise the parameters in the underlying ELO model. The model correctly predicted a total 63.0% of results, with a mean absolute error (MAE) of 13.7 points per game during the build process. The test phase saw a slightly better MAE – of 13.0 points per game – produced, and a lower accuracy of 55.7%. Secondary experiments regarding the model’s ability to predict the top 8 at the end of the home-and-away season were also conducted.