Updated DPPI

I took the time since I have had a few days off to tweak my DPPI algorithm. My No. 3 son has decided this is a pretty fun challenge so he is writing his own algorithm, albeit, based on my methodology. His program will be far superior to mine since my programming skills are pretty basic. He is a lot more proficient on the code side of things. The new improved version of the DPPI reduces the standard deviation by more than 10%. In layman's terms, at least for the already played games it is more accurate at predicting the scores by more than 10%. You may say that I should be able to predict games exactly since they are already played but teams do not play perfectly consistently so the challenge is trying to find variables that can be predictive. Anyway, below is the latest DPPI (at least the top 58 teams since that is what fit on my screen and this is a screen shot). You will see that West Virginia is way out front in first. They have literally crushed just about everyone. And even though they have played a very weak schedule (SOS=312th of 351 teams), they that is all considered in the algorithm. Kansas is currently 7th and only 2.5 points or so from North Carolina, the 2nd place team. A couple really strong games and KU could rise pretty quickly. You can also see the Big 12 is very well represented with 8 Big 12 teams in the top 43 and 7 in the top 29. There are 6 undefeated teams but USC has played a weak schedule and should be crushed by UCLA and will be an underdog to Arizona at home and on the road.

Thanks, Corpus! It will certainly be interesting to see how this all plays out. There are so many variables, as you pointed out. it just seems to me that SOS would play a pretty big part, even in blow outs. What you don't know is if and when the bench was emptied when the game was out of reach, or if the coach was determined to work on some things later into the contest.

Will we get to see your son's work, eventually?