Utilizing logistic regression to apply the ELO system in forecasting Premier League odds

Abstract: This thesis provides insights into the creation of a model for predicting odds in the Premier League. It illustrates how the ELO system and historical odds, in combination with Monte Carlo simulations, can be implemented through logistic regression to predict odds in an unbiased way. The findings are that the model performs generally well, but significantly worse at the beginning and end of the Premier League seasons. For further improvements, it is most likely necessary to factor in variables not available in the current model. Such factors could for example be incentives, injuries, or changes in the squad, all not being accounted for by the model in this case.