Machine Learning Model Predictions
One of my projects this offseason will be getting the betting model up and running again! It outperformed everything in the industry when it was running a few years ago, but I haven’t had time to maintain it recently. Well…while it’s not all done yet…this is what it’s currently spitting out for this weeks games!
Some notes:
- It can’t take injuries into account. So if a key player is out, that hasn’t missed any time yet, then the model doesn’t know that.
- The Eagles are getting a slight downgrade because Minshew was their QB for a few weeks to close the season. They would have graded out stronger if Hurts hadn’t been injured.
- I’ll have a video next season that walks through the model and explains more about what it knows and doesn’t know. That way you’ll have a better understanding of what it’s using to get it’s predictions, and can therefore think about what sort of information it might be missing (ie. the Mahomes and Hurts info above).
- Yes, even in it’s current form the model outperformed Vegas closing lines for the 2022 season.