Machine learning, classification and prediction models have applications across a range of fields. Sport analytics is an increasingly popular application, but most existing work is focused on automated refereeing in mainstream sports and injury prevention. Research on other sports, such as table tennis, has only recently started gaining more traction. This paper proposes the use of machine learning to predict the outcome of table tennis single matches. We use player and match statistics as features and evaluate their relative importance in an ablation study. In terms of models, a number of popular models were explored. We found that 5-fold cross-validation and hyperparameter tuning was crucial to improve model performance. We investigated different feature aggregation strategies in our ablation study to demonstrate the robustness of the models. Different models performed comparably, with the accuracy of the results (61-70%) matching state-of-the-art models in comparable sports, such as tennis. The results can serve as a baseline for future table tennis prediction models, and can feed back to prediction research in similar ball sports.