This project is built in Python and it uses machine learning to create a prediction model for tennis matches (between 2013-2020).
The project is based on realizing a prediction algorithm to “beat bookmakers” and gaining an advantage in making choices while betting on matches. We choose tennis as a sport where betting in for some reasons:
- Variables in tennis are easier analyzable than other sports since we have found the elements of a good betting-strategy and this ones are more distinguishable and “evident” than other sports (for example the surface, the winning streak) ;
- Tennis is a zero-sum-game and this is a prerogative for applying powerful features that we have decided to use such as Elo rating for example;
- The datasets on the internet that we have found are well-made.
In this project we will first of all import the dataset and clean the data, then we will proceed by adding new features to improve the accuracy of the predictions. After a fine tune of the best classificator, we will conclude with a study on how to maximize the return on investment for a betting strategy.
Just open and run the Python Notebook file "main.ipynb"