Well, this will be the next-to-last update on the combined rankings from Massey’s site. After the seedings come out, I hope to do one more and finally conclude whether there is any value in piling everyone’s guesses, votes, and calculations together…and then drawing conclusions from the biggest piles.
As usual, here is the frequency graph and the table:
If you’re feeling especially masochistic today, go back and look at the table from the 2/21 update.
Jerry Palm has reported that the AP Top-25 is a better predictor of seeding than the RPI formula. However, the NCAA Selection Committee used to make super-secret “adjustments” to the RPI formula before selecting the bubble teams and seeding everyone. In 2005, the RPI formula was changed and there are not supposed to be any more secret adjustments to the RPI calculation. (Though that certainly doesn’t mean that the Selection Committee blindly follows the RPI ranking.)
So, I looked at just 2005 to see how the new RPI formula compared to the AP Top-25….and the RPI calculations win out. For the top six seeds, 22 of the 24 teams were in the RPI Top-25….while the AP Top-25 had 20 teams. Here is a table comparing how the RPI Top-25 and the AP Top-25 were seeded compared to what you would predict from their ranking:
As you can see, the RPI Top-25 placed more teams within one seeding position than the AP Top-25 did. Of course, one year’s results don’t prove anything…but this comparison looks interesting enough to follow in the future. Hopefully, I will be able to update this comparison early next week while we are waiting for the opening round games.
