If [538’s NBA RAPTOR model](https://fivethirtyeight.com/features/introducing-raptor-our-new-metric-for-the-modern-nba/) is more accurate than sports books at predicting outcomes, then it would be possible to extract value using the [Kelly Criterion.](https://en.wikipedia.org/wiki/Kelly_criterion) Namely sport books often adjust odds to encourage or discourage action to ensure their books remain balanced, and not because of any change in the prior win probabilities.
I made a spreadsheet that will collect the data throughout the season to test this hypothesis. I just need to input the score, the RAPTOR Predictions and the odds. Here’s what the first two days of the season looks like.
​
https://preview.redd.it/zfj3y2l1lhu91.png?width=1856&format=png&auto=webp&s=999f5cee7241afb473f92b857741cc9220ac4005
by fingerbangchicknwang
1 Comment
My kind of money ball!