Another year of Bracketology prognostication is over, and the big controversy is West Virginia vs. North Carolina. I had both of them out, so in my book West Virginia can't complain they were robbed, but North Carolina really caught a break. But let's start from the top:
At-lg %Change Total Seed Rank Team Conf Rec %Odds -1dy -1wk 1seed %Odds Proj Actual 1. Auburn SEC 28-5 100.0 0.0 0.0 99.0 100.0 1 1 2. Florida SEC 30-4 100.0 0.0 0.0 98.9 100.0 1 1 3. Duke ACC 31-3 100.0 0.0 0.0 98.2 100.0 1 1 4. Tennessee SEC 27-7 100.0 0.0 0.0 39.6 100.0 1 2 5. Alabama SEC 25-8 100.0 0.0 0.0 33.2 100.0 2 2 6. Houston B12 30-4 100.0 0.0 0.0 31.1 100.0 2 1 7. Michigan St. B10 27-6 100.0 0.0 0.0 0.0 100.0 2 2 8. Kentucky SEC 22-11 100.0 0.0 0.0 0.0 100.0 2 3 9. Missouri SEC 22-11 100.0 0.0 0.0 0.0 100.0 3 6
Auburn was our overall #1 seed and perhaps surprisingly, the Selection Committee stuck with them. I was pretty sure they'd go with Duke or Florida, and if Florida had beaten Tennessee by 20 points I think the Dance Chance would have put them #1. The top three were very close and then the next three were, too, but despite winning the Big Twelve tournament Houston couldn't climb over the SEC teams, and Tennessee remained #4 despite their loss Sunday. It seems like every year the Dance Chance overvalues a few teams from the dominant conference while the Committee likes to spread things out. Hence we see Missouri getting our highest 3-seed while IRL the Tigers were just a 6-seed!
After that it was a litany of teams off by 1 seed line, generally. Formulas don't tend to do well in Bracketology scoring; humans are much better at "yes, but..." logic that the Selection Committee uses. AI would probably do well but the few attempts I've seen have been no better than the formula approach. The worst seeding job by the Dance Chance this year was Memphis:
At-lg %Change Total Seed Rank Team Conf Rec %Odds -1dy -1wk 1seed %Odds Proj Actual 34. Memphis Amer 28-5 93.6 +4.2 +11.8 0.0 100.0 9 5
The Tigers got a 5-seed, and that was questioned by many. The Bracket Matrix average was an 8-seed, so it wasn't just us.
Like I said, in terms of seedings, formulas don't do so great, but in terms of "who gets in" the formulas do pretty well, and the Dance Chance was 66 of 68, which is a good human score if you look at the Bracket Matrix. There weren't many surprises this year, but a big one was West Virginia being left out, and North Carolina getting in. The majority also thought Indiana would be in, instead of Texas.
As you can see below, the Dance Chance was right about Texas being in and West Virginia being left out, but missed on Indiana and North Carolina, as well as Baylor (in) and Boise State (out).
At-lg %Change Total Seed Rank Team Conf Rec %Odds -1dy -1wk 1seed %Odds Proj Actual 40. UC San Diego BW 30-4 73.5 +2.2 +8.4 0.0 100.0 10 12 41. Xavier BE 21-11 64.9 +1.5 +5.7 0.0 64.9 11 11p 42. VCU A10 28-6 61.9 +3.4 +10.0 0.0 100.0 11 11 43. San Diego St. MWC 21-9 59.8 +3.1 -1.0 0.0 59.8 11p 11p 44. Colorado St. MWC 25-9 58.1 +3.5 +13.9 0.0 100.0 11 12 45. Texas SEC 19-15 55.6 +2.5 +21.3 0.0 55.6 11p 11p 46. Indiana B10 19-13 52.7 +1.2 -10.9 0.0 52.7 12p xx 47. Boise St. MWC 24-10 50.9 +1.0 +2.8 0.0 50.9 12p xx 48. Baylor B12 19-14 49.1 +1.3 +2.0 0.0 49.1 9 49. North Carolina ACC 22-13 48.0 +0.7 +2.7 0.0 48.0 11p 50. West Virginia B12 19-13 44.1 +1.2 -11.1 0.0 44.1 51. Drake MVC 30-3 43.1 +1.1 -5.7 0.0 100.0 12 11
Not long before Selection Sunday I noted that the Dance Chance had Baylor and West Virginia—both from the Big Twelve—as the First Teams Out. I looked up the NET ranking for both and saw that Baylor was running around #30 while West Virginia was below #50. In the Bracket Matrix Baylor was a solid 9-seed while West Virginia was a low 10-seed. I figured I was going to miss on Baylor—it happens—but I felt West Virginia was more on the bubble than most people thought. In the end, despite 243 of 249 Bracketologists vouching for them, the Mountaineers were left out.
The governor of WVA has hilariously started an "investigation" of the Selection Committee over North Carolina getting in, but both were bubble teams and our formula—as unbiased as it gets when simple numbers pick the teams—had the Tar Heels ahead. Baylor ended up a 9-seed meaning they were solidly in, and we had Boise State just a few percentage points ahead or we would have had 67 right. There's no shame in missing on your Last Two In and First Two Out, but Baylor was a big miss. Part of it was their record—the Bears are just 18-14 vs. Division I opponents.
I can see the case against North Carolina, and my method didn't have them in either. But Indiana was probably a better choice than West Virginia. The NCAA should stop saying who the First Team Out was, that's the real solution. It was a bad look to have the Committee chaired by a school that was Last Team In, and they didn't help things with a bad excuse about West Virginia missing a player that was easily refuted. The Dance Chance knows nothing about missing players and still left the Mountaineers out, so there was other logic they could have used—like the #51 NET ranking. Of course then you could bring up 5-seed Memphis's #49 NET ranking. And so it goes.
There were two other teams that would have been "misses" if they hadn't won their conferences. UC San Diego only got a 12-seed instead of the 10-seed where we had them, and Drake got the 11-seed instead of the 12 we gave them. Luckily we didn't have to choose between the two. The Mid-Majors are the hardest teams to seed, hence the confusion around Memphis, etc. The one-bid leagues are spaced out enough to seed them pretty easily, but when teams play an easy schedule and have a gaudy record, the NET kind of goes out the door and conference rank and yes, a team's name has an influence.
Comments