Louisville is the defending National Champion. They are 25-5, and have outscored their conference opponents by a whopping 281 points in league play. They are ranked 9th in the USA Today Coaches Poll and 11th in the AP Poll, and poised to move up further (Duke and Syracuse both had dreadful losses this week). Why then, do most bracket projections (I looked at 12 different ones) have them as a 4 or 5 seed? The RPI, which is in the high 20′s right now.
Similarly, Virginia clinched the ACC regular season, is also 25-5, and is ranked 5th in both polls. If they win at Maryland, they will certainly enter the tournament week ranked in the top 4 in both. The RPI, though, has them down at 8, so most bracket projections have them as a 3 seed (a handful have moved them up to a 2).
In these cases, though, I think reliance on the RPI is misguided. Do you remember what Indiana’s RPI was last year, when they won the Big Ten and got a #1 seed? It was 12. I could cite numerous examples, but while we rail on the RPI, that more relates to the bubble. At the top of the tournament, the committee mirrors the polls heavily. That transitions toward the second half of the polls to more of a mix between the RPI and rankings, and then for teams outside the rankings, the RPI becomes paramount.
Over the previous five years, there have been 37 teams that were ranked in the final AP Poll, and had a significant difference between where they would be seeded based on the polls, and based on the RPI. (For these purposes, that difference is at least two seed lines apart, for example a team ranked 6th-a 2 seed-while being rated 16th in the RPI-a 4 seed.)
For those ranked in the top 12, they were seeded by the selection committee closer to their AP rank on 7 occasions, were split evenly between the two 4 times, and only once were seeded closer to the RPI ranking. That one was a true exception from a one-bid league: Murray State in 2012, seeded as a 6 seed and ranked #12 in the final poll. I doubt Louisville gets the same treatment from the committee.
In the second half of the rankings, it moved between the two, though the AP ranking was more predictive. 4 teams were seeded closer to the RPI rank, 11 were closer to the AP ranking, and 9 were split between the two rankings. Oh, and there is one outlier–Oregon from last year, seeded as a 12 while slipping into 25 in the final poll and finishing at 36 in the RPI. I’m going to chalk that up to being as bizarre a decision now as it was then, probably fueled by long days and not wanting to re-do the brackets after an Oregon run in the Pac-12 tourney.
After that, last year, every remaining unranked team in the RPI top 50 made it, except Southern Miss (no quality wins) and Connecticut (ineligible). They were replaced by the next two major conference schools in the RPI rankings, California and Villanova.
Kansas is seen as the favorite for the final #1 seed, I think in large part because they have the #1 ranking in the RPI and the best strength of schedule. I’ll just point out that they are #8 in both polls, still behind both Virginia and Villanova. The last time a team was selected as a #1 seed while being ranked outside the top 5 in the final AP poll was in 2005, when the committee snubbed 4th-ranked Louisville and dropped them all the way to a 4 seed, while making 8th-ranked Washington the #1 seed out West (Louisville would crush Washington in the Sweet Sixteen).
For this reason, I have actually moved Virginia up to the final #1 seed for now. Both Louisville and Virginia are top five in Pomeroy’s rankings, and have a good chance of adding a tournament title. Louisville may not be able to get to a 1 seed, but I’m not ruling out a 2 seed if they beat Connecticut, win the American tournament, and finish at 29-5. For now, I’ve got them as a 3, but as the highest 3 on the board.
Here is the projection (one bid leagues, the highest rated team by Pomeroy is used as current auto bid)