Tennis Betting Strategies Based on Neural Networks

Abstract

The aim of this paper is to explore betting strategies for tennis matches using neural networks. We used public data and we implemented a neural network prediction model, with an accuracy of 85% on the validation and testing sets. Based on the predictive model we tested several investment strategies which incorporates investor's risk profile. The optimal strategy is to place bets on games where our model indicates a high likelihood of victory and an appropriate bookmaker's odds.

Keywords: Sport betting, Tennis, Machine Learning, Neural Networks, Statistical Models

JEL Classification: G01