Yesterday, I went through and ranked the teams by how many wins they have above or below what an average at-large candidate would have with average luck against the same schedule. I believe it is an improvement over the RPI, but still looks at wins and losses, and not point margins. It tries to put into context and compare different types of wins and losses, and account for game location and various opponent quality. The RPI seems to overvalue whether a schedule is okay at the bottom, or bad, which has little to do with how well a tourney type team should perform.
Today, I put together a bracket using that wins above average system. I followed it for the most part, with some minor modifications when teams were close together in the ratings. I put Virginia in the field based on how they have played since the middle of November, since two of their bad losses were before Thanksgiving. Kentucky would be barely in, but based on the showing at Tennessee, slips out.
So, here is my field, or at least the Top 12 seed lines that would include all at-larges.
My teams that missed the cut: Kentucky, Baylor, Temple, California, Villanova, Boise State, Charlotte, Arizona State
[photo via USA Today Sports Images]