This repository provides a machine learning AI for predicting NBA game winners and over/under totals, targeting sports bettors seeking data-driven insights. It aims to improve betting accuracy and identify value bets by leveraging historical game data and betting odds.

How It Works

The project utilizes a neural network and XGBoost to predict game outcomes, trained on NBA team data from the 2007-08 season to the present, matched with game odds. It outputs expected value for money lines and suggests bet sizes using the Kelly Criterion, offering a ~69% accuracy on money lines and ~55% on over/unders.

Quick Start & Requirements

• Install dependencies: pip3 install -r requirements.txt
• Run prediction: python3 main.py -xgb -odds=fanduel
• Requires Python 3.11.
• Odds data can be fetched automatically from specified sportsbooks (e.g., FanDuel, DraftKings) or entered manually.
• A Flask web app is included for data visualization.

Highlighted Details

• Achieves ~69% accuracy on money lines and ~55% on over/unders.
• Outputs expected value for money lines.
• Implements Kelly Criterion for bankroll management.
• Supports automatic odds fetching from multiple sportsbooks.

Maintenance & Community

Contributions are welcomed. No specific community channels or maintainer information are provided in the README.

Licensing & Compatibility

The repository does not explicitly state a license. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The project is presented as a tool for personal use and does not guarantee profits. The accuracy figures are self-reported and may vary. The README does not detail the specific neural network architecture or training methodologies.