An Explanation Of The Three Pro Football Ratings
The PiRate Pro Ratings
The NFL version of the PiRate Ratings was not the same as the collegiate version from 1979 to 2008. The NFL version strictly used a statistical formula that could be reproduced by anybody who knew the equations used to devise the formula. No subjective data was used. Starting in 2009, a new PiRate formula was created and backtested to 1999. The results produced a more accurate rating. This new formula is a close cousin to the formula used for years in the college PiRate formula.
The new formula begins with the previous season ending ratings with adjustments for experience, personnel aging, new coaching, injuries, and a couple dozen intangibles. As the season progresses, the previous season’s results have less of an ipact to where it has little effect by mid-October.
The Mean Ratings
Just like the PiRate Ratings, the NFL Mean Ratings are not the same as the collegiate version. The NFL Mean Ratings consist of a dozen different calculations. Three calculations consist of different ways to look at point differential and strength of schedule. Five calculations look at yards gained and allowed rushing and passing and special teams play with the strength of the opponents’ rushing and passing. Point values are assigned based on each set of data. The remaining four ratings are my old four pro ratings from the 1970’s and 1980’s. The 12 ratings are given equal weight, and then I take the average (mean) to get the rating.
The Bias Ratings
The Bias Ratings consist of five of the components of The Mean Ratings. The five ratings are not given equal weight. The five ratings are weighted at 37.5%, 25%, 12.5%, 12.5%, and 12.5%. I have back tested these ratings and found that this weighting gives the rating its best predictive percentage.