NBA Sports Betting Models

This is the source code for an NBA player prop model. Regular updates on my twitter account @bendominguez011

A writeup for the reasoning behind each of these models will be located here (eventually).

File Structure

exports/
    player_props/
    sim_results/
spreadsheets/
logs/
    model.log
notebooks/
model/
    models/
        threes.py
    odds_api/
        __init__.py
        config.py
    utils/
        betting_math.py
    __init__.py
    nba_api_helpers.py
config.py
main.py

Project Requirements:

You'll need to install nba.api seperately, and when you do, use the command:

pip install nba.api --no-deps

Python version: <= 3.10.5 Numpy version: <= 1.21.5

To install Python dependencies: pip install -r requirements.txt

Environment variables

There's a couple environment variables that need to be set to be able to run the program.

The Odds API

ODDS_API_KEY: Can retrieve your API key from this website here: https://the-odds-api.com/. There is a free version of the API, but it is limited to 500 requests/month.

DEBUG: Should be set to 1 or 0. Controls logging output.

Running the model

So far we have a three point model which is located under model/models/threes.py

To import all of the data at the same time and run the threes model, use the main.py file after setting your environment variables and installing dependencies:

python main.py

Options:

    -n, --n-simulated-games           Number of games to simulate / simulation. Suggested to set to 50,000 to 200,000.
    -b, --bootstrap_samples           Number of bootstrap resamples to use in simulations. Suggested to just use 100_000 unless you really want to change it.

Alternatively, you could also run each of the models seperately, and simply import them from the model.models module.