ABSTRACT
Historical match data contains lots of useful information that can be used to predict how soccer teams will perform in the future. This chapter focuses on techniques for predicting the outcome of individual soccer league matches using historical data. In particular, the chapter introduces two popular match prediction techniques, Poisson regression and the Dixon–Coles model, and shows how these can be implemented in R using historical match data from the English Premier League. In addition, random forest and conditional inference tree models are introduced, with the reader shown how these can be applied using the match odds produced by bookmakers to predict the outcome of individual soccer matches.
