Beta_Rank is an advanced college football model built on drive level data from every game. It is an interlocking series of hierarchical models that provide deep insights into team performance.

 

In season Beta_Rank doesn't contain any preseason data after week 6 because it doesn't need it. Beta_Rank predicts 73.7% of college football game winners correctly (not counting in games vs. FCS because that is cheating) without using preseason data or spreads.

​

​