The long-awaited pre-season ratings are here. There’s a write-up on the rating algorithm and some notable teams in College Basketball Prospectus, so you should check that out when it becomes available shortly. I’m not going to repeat that stuff here. Of course, you will get a lot of other great content in it as well.
One change this season involved the computation of the Pythagorean winning percentage. Calibration of last year’s predictions revealed that 10.25 is the best exponent to use for the purpose of the game probabilities. This should provide a more realistic forecast of end-of-season records now and especially as the games start and real data goes into the system. This does not affect the calculations of any other statistics.
Right now this results in a lot of conference records bunching up near .500. This will prompt your message board friends to say something like “There’s no way every team in the Pac-12 finishes with between 6 and 12 wins. LOL.” Substitute your own conference of choice in the quote. Of course that’s not how the teams will finish, but this is setting an expectation for each team on an individual basis, and not trying to predict the record of the conference champ, whoever that may be. Considering that the last preseason number one to go unbeaten in conference play was Duke in 1999, chances should be fairly low for even the best teams to run their conference’s table. Though I’ll admit, chances are unrealistically low for some teams. That should change fast as games are played.