The prediction of outcomes in the National Basketball Association

This dissertation derived statistical models for predicting the outcomes of NBA Basketball matches based on past data. The models were of two distinct types: a pre-game model and an in-play model. The pre-game model used a rating system with an exponential smoothing algorithm. The in-play model utilised probability distributions and simulations for predicting final scores in real time. The application of these models to the sports betting market was investigated and evaluated.