What to Expect When Inspecting Creme's Bracketology -- lessons from 2014

[DobbsRover2]

There are always fans who think that ESPN's Bracketology expert Charlie Creme is an off base idiot and never gets anything right with his picks, but actually he generally has done a very good job of picking the field except maybe when the committee does something jaw-droppingly stupid. And even when he does have major gaffes, the results of the tourney often show that his picks were better than the committee's.

Last year Creme nailed all 64 teams in the Tourney. If you don't think that's really good, try putting together your own projections from scratch without just cribbing off Creme's analysis. Getting 64 right is not easy. And you know the committee is not going to hew that closely to his Creme's brackets because it might put them out of business if everyone started saying, "Creme already told us what the brackets are."

Getting the seeds and placements is of course far far harder than naming all the teams, and Creme will always provide a lot more ammunition there for calling him an idiot. But again, I was floored last year by how well he did considering all of the difficulties of analyzing what will come out of a committee that seems to blow in different directions with each year's breeze.

Last year Creme nailed 12 of the 64 teams exactly in place, with three of them being the easier top #1 seeds and two of them in the back group of #16 seeds. 18.7% of exact picks may not sound great, but again, try doing better yourself. He got 39 teams (60.3%) on the right seed line, 13 teams underseeded (20.3%), and 12 teams overseeded (19.4%). He put together 5 first round pairings correctly, but interestingly only one of them (UTenn vs. Northwestern State) also got the region right. Altogether his total difference from the seed line from the errant picks added up to 32 (16 over and 16 under), or 1 spot per every two picks. Note that seeding line bumps for balancing a bracket can also lead to errant seed placements.

Biggest errors for Creme in 2014 were: 1) making Stanford a #1 seed instead of South Carolina, but of course Stanford went to the FF and USCar lost in the S16; 2) he was 3 spots over for BYU with a #9 seed instead of the #12 they got, but BYU went to the S16 game against UConn; 3) he was either over or under by 2 spots for five teams, all three teams he underseeded by 2 did lose, while one of the two teams he overseeded by 2 (James Madison) won a game, making him seem more prescient than the committee.

What last year's results means for this year is hard to say except that you can expect that if Creme does really well he will get at most a quarter of the teams set exactly in their bracket spot and about two-thirds on the right seed line. In a more average year, he will do a little less than that.

The biggest question mark I would have in his picks this year is his 9-seed for FGCU, which is ranked 20/21 in AP/Coaches polls, rated #20 in Sagarin and #12 in Massey, and even brain-dead RPI has them at #15, so a #9 seed sounds crazy. FGCU did get only a #12 seed last year with a 26-7 record, but this year they are 30-2 and on a 25 game winning streak, so I would expect them to get a much better seed than a #9.

 

[JoePgh]

I agree that saying he is an "idiot" (or any similar comment) based on his being unable to mind-read the Committee and its compromises, is completely off-base. I think you should take his Bracketology as what the Committee "should" do if it followed the official rules and its past precedents. But that can easily be quite different from what the Committee actually does.

This year, I think the most likely and obvious difference that the Committee will enact vs. his brackets is to put South Carolina in Greenville and Notre Dame in Oklahoma City. I think he is right that Oregon State goes to Spokane as a 3-seed.

The interesting question, of course, is where Tennessee goes as the top #2 seed. The one sure thing is that they can't be in the same bracket as South Carolina, so that leaves Spokane (Maryland), Albany (UConn) and either Greensboro or Oklahoma City (wherever Notre Dame goes). Spokane might be the "fairest" (and a Tennessee-Maryland match in the Elite 8 might well yield a Maryland victory just as it did last year), but as I've said before, Albany is clearly the best choice from a marketing/TV-ratings perspective.

I see that Geno advocates that the Committee hold strictly to the S-curve (implying Tennessee to Spokane). I assume that is because he does not want to play them in the Elite 8 and endure the media circus that would attend to that. And maybe he hopes that Maryland or someone else will stick a fork in them so that once again he does not have to play them in the Final Four. As a fan, I hope that does NOT happen.

From my perspective, the only thing better than a UConn-Tennessee game in the Elite 8 would be for the Orange Ones to be placed in the same bracket as South Florida (a #7 seed), and lose their #2-#7 match to the Bulls (in Knoxville, no less). That could plausibly happen, and it would quiet down the complaints from SEC partisans about the weakness of the AAC.