Jay Bilas said last night that it was “who you lost to” that mattered more to the committee than who you beat. I actually disagree with this. I think the story of the selection committee can be told through the lens of a group that operated within a very tight window in using the RPI, and who actually did make some adjustments based on the best wins for teams. The story of why some teams (among several equals) were selected and some were not is really not related to who they lost to. It’s actually related to who they beat at the bottom of the schedule.
In other words, it’s the story of how the Mountain West successfully bypasses the RPI system, and how the ACC might want to learn a thing or two about things that should be essentially meaningless when it comes to rating teams to play in a tournament, but which matter greatly to the RPI and the committee (even if they don’t know it).
First, let’s talk about the RPI and seeding. I got 68 of 68 teams right this year. Big whoop. I really went 2 for 2, correctly having La Salle and Middle Tennessee in my field, instead of Tennessee, or Kentucky, or Virginia, or Maryland. As I noted last year, a monkey throwing darts could be trained to get 65 of 68 right, since you are given the answer on 31 of them.
If you just went by the RPI, you would have correctly called 36 of the 37 at-larges, and would have needed to only switch Southern Miss out for California to get the field right. Southern Miss was held out because of who they did not beat, namely no tournament teams, and certainly not because of bad losses.
If you blindly assigned a seed to a team based on the RPI rank relative to the 68 teams in the field (22 as a 6-seed, 31 as an 8-seed, etc), you would have gotten within one seed line of what the committee did on 71% of the teams. Only ten teams were seeded three or more seed lines away from where a blind RPI seeding would have placed them. In each of those cases, it was actually big wins (or lack thereof) that caused the moves.
The biggest improvers in seed, compared to their RPI: Wisconsin, Illinois, Pittsburgh, Villanova.
Two Big Ten and two Big East teams. Wisconsin beat five teams that appear in the top four seeds. Villanova beat four. Illinois knocked off both Indiana and Gonzaga. Pittsburgh beat Syracuse, and Georgetown on the road.
The biggest decliners in seed, compared to their RPI: Belmont, Middle Tennessee, Memphis, North Carolina, St. Mary’s, and Colorado State.
Belmont’s best wins were Middle Tennessee and South Dakota State. Middle Tennessee beat Ole Miss and no other teams in the tourney. Memphis beat Southern Miss three times. North Carolina went winless against Duke and Miami, and had a best win of UNLV. St. Mary’s only beat Creighton. Colorado State is probably the only one with a complaint if we view it from a “RPI, adjusted slightly for who you beat scenario”. They dropped below San Diego State and were seeded four spots lower than UNLV.
[RELATED: NCAA Tournament 2013 Bracket Breakdown: East]
It actually wasn’t bad losses to teams that cost seeding. It was bad wins. The RPI cares about your opponent’s schedule, and the opponent’s opponent’s schedule. A couple of really bad teams can really submarine your RPI.
Mark Turgeon had this to say earlier today: “There was so much talk about the Mountain West all year. A 9-7 team [San Diego State and Boise State] got in from that league. I just think that’s the way it works.”
You can call it sour grapes, but Boise State was in (and bracketologists who look at the RPI had them in), and Maryland was out, because of the funny math at the bottom of the RPI.
If you look at the teams ranked outside the Top 200 in the RPI, the Mountain West appeared to play a much tougher schedule. Their nine members combined to play 37 such teams (4.1 per team). The ACC, meanwhile, played 62 teams outside the RPI top 200 (5.17 games against weak opponents). Appeared, because the RPI doesn’t include games against non-Division I schools. The ACC only played one such excluded game (and that was North Carolina vs. Chaminade in the tournament in Hawaii). The Mountain West played fourteen!
If you add those games in, then the Mountain West teams averaged 5.67 games against teams outside the RPI top 200, and the ACC averaged 5.25.
If you think that is just funny math and will have no impact, think again. Let’s compare Maryland to Boise State, and also throw in Oregon. There is some thought that Oregon got screwed (yes, they did), but it wasn’t because of an East Coast Bias. It was a RPI bias. Oregon got seeded where the RPI dictated.
If you go by the RPI formula, Boise State played four weak non-conference opponents, and Maryland and Oregon each played six. Neither Maryland or Oregon played any non-Division I opponents. Boise State also played Corban and Walla Walla. I kid you not. They beat them by 56 and 67. Those results did not sink Boise State (nor should they, because that would make no sense).
Maryland playing Maryland Eastern Shore and South Carolina State, though, sunk Maryland. Oregon playing Idaho State and Portland State nearly did them in, and apparently meant they needed to win the Pac-12.
How big of an impact was Walla Walla and Corban wins not counting for Boise, and those bottom feeders counting against Maryland and Oregon?
The first column is the non-conference strength of schedule per the RPI, followed by the rank in SOS. The third column is the non-conference strength of schedule if we just remove the two worst teams (all wins by each). And no, I did not look up how Corban and Walla Walla did in the NAIA or wherever they play. Turns out, if we just lop off the last two games, which the RPI did for Boise but not for Oregon and Maryland, Boise had the easiest non-conference schedule. Maryland was near average, Oregon above average, and Boise slightly below average.
However, the RPI viewed it differently. If you want to know why you are out right now, Maryland fans, well, it’s because you played Maryland Eastern Shore rather than Walla Walla. Now, does that make sense? I suggest Mark Turgeon give Bluefield College a call next year when Maryland needs an extra game to sell popcorn and t-shirts, and tell the MEAC teams to stay at home. It’s not who you lost to, it’s the bottom 10% of who you beat that is unfortunately driving a committee hidebound to follow the RPI.