Since I’m the only bracketologist with Louisville as a #1 seed, I’d like to explain why I have them there. It really boils down to 2 factors — 1) Public perception matters much more than profile when it comes to the top seed lines, and 2) The need for contingency plans to find an alternative #1 seed this season.
I started building my math model to predict which teams would be chosen to the NCAA basketball tournament and where they would be seeded in 2003. The first few years, I made some progress getting the seeding more in line with what the committee was doing, though obviously getting it nailed down continues to be tricky to this day. However, I consistently got the top seed lines wrong, which was frustrating because it seemed like those should be the easiest to predict. Following a strict by-the-profile view did not accurately predict which teams got the #1 seeds. The breaking point for me was 2008, when Tennessee finished the regular season #2 in RPI and a profile easily worthy of a #1 seed, but ended up with a #2. I realized that public perception of the top teams had a substantial impact on the top seed lines, and the perception that season was that Tennessee was a step below the top teams in the land. The only way I felt I could measure this perception in an analytical way was to use the AP and ESPN polls as stand-ins for public perception. The following year I incorporated those rankings into my model and have had much better success in predicting the top seeds ever since. For example, last season, Gonzaga did not have the profile of a typical #1 seed, given the lack of overall schedule strength. Some bracketologists predicted a #2 seed for Gonzaga, but given their #1 ranking in both polls last season, I had them as a #1, which turned out to be an accurate prediction. This season, with Louisville at #5 in both polls at the start of the week and with Villanova losing (a team that also does not have the strongest case for #1 seed based on profile alone), my model has Louisville moving up to #4 in both polls by the end of the week and grabbing a #1 seed.
Furthermore, the case for making another team a #1 seed is complicated by the fact that most, if not all, of the other potential #1 seeded teams would be playing on Sunday. For at least the last 2 seasons, the committee has completed the bracket on Saturday night, with contingency plans created based on possible results of Sunday games. With Michigan, Wisconsin, Duke, and possibly Syracuse and UVA as potential #1 seeds, the committee would have to create a number of contingency plans based on the different combinations of possible results of Sunday’s games. I don’t think it’s likely that a committee of 10 people will be able to come to an agreement by Saturday on these contingency plans. I think it is much more likely that the committee would grab another team as a #1 and seed the rest of the highly seeded teams that play Sunday as #2 or #3 seeds based on their profiles as of Saturday night. Last season, the committee did not even have a contingency plan in place for Ole Miss winning the SEC tournament, which led to last minute scrambling to remove them from the first-round games in Dayton.
So given these factors, I think the committee will make Louisville a #1 seed come Sunday if they are able to come out of the AAC tournament as champions.