This work's goal is to predict soccer matches results and compare them to betting odds.
The initial goal is to predict only the match result based on recent team's forms and also player attributes extracted from FIFA game.
The following commands set the environment ready for anaconda users
conda create env --name <myEnv>
conda activate <myEnv>
pip install -r requirements.txt
The first results are not satisfactory yet. Some of the improvements ideas to be implemented in the future are listed below:
-
Scrapping Football Manager players attributes, as they are a better source of information, especially in minor leagues
-
Predicting not only the match result, but also the ammount of goals and other metrics
-
Creating a trigger advising if a match odd differs from the model's predicted results (profit opportunities)
-
Creating a backtest simulator to analyse different betting strategies.
2022/11/09 - Implemented scraper to get today matches data.
2022/11/16 - Implemented the old matches scraper, which provides several amount of data in according to the user needs.
** 2023/03/01 -** Implemented scraper for player attributes, PrepareDf and first model.