Wednesday, January 3, 2018

Bracket Projection: 1/3

This is a terrible bracket. I mean, some parts of it are fine, I guess. So why do I even bother posting it? Let's take a gander at what makes it so terrible, and then spend the rest of this post discussing the shortcomings of the selection committee.

1 Duke Xavier Villanova North Carolina
2 Texas A&M Arizona St. Kansas Oklahoma
3 Purdue West Virginia Arizona Michigan St.
4 TCU Virginia Arkansas Seton Hall
5 Tennessee Kentucky Wichita St. Miami FL
6 Gonzaga Florida St. Texas Tech Clemson
7 Butler Alabama Cincinnati Creighton
8 Auburn Rhode Island Syracuse Nevada
9 Texas Louisville Missouri Michigan
10 Baylor SMU Notre Dame Saint Mary's
11 St. Bonaventure UCLA Minnesota USC
11 Maryland Florida
12 Lipscomb Missouri St. New Mexico St. Middle Tennessee
13 Vermont South Dakota Louisiana Lafayette Murray St.
14 College of Charleston Ball St. East Tennessee St. UC Santa Barbara
15 Bucknell Stephen F. Austin Princeton Montana
16 UNC Asheville Texas Southern Iona Northern Kentucky
16 Hampton St. Francis PA

First of all, I made this while watching the games yesterday, and by the time I had finished it, Texas Tech had beaten Kansas, Auburn had beaten Tennessee, Mississippi State had beaten Arkansas, Florida had beaten Texas A&M, and Vanderbilt had beaten Alabama. All of these games could have a significant impact on seeding, and they aren't accounted for. But that's not really my biggest complaint. As you gaze upon this garbage bracket for a few seconds you'll see what I'm talking about. What's that? The 3 seed in the Midwest region? The Spartans of Michigan State? That's right, the number 1 team in the AP Poll, the Sagarin ratings, and the number 2 team in Kenpom is like, 12th in my system, which has nothing to do with my own personal biases and everything to do with our good friend, the Rating Percentage Index. Let's dive a little deeper:

An RPI refresher: The formula for RPI is 

0.25*(winning %) + 0.5*(opponents' winning %) + 0.25*(opponents' opponents' winning %)

If that seems a bit silly, you are correct. 75% of your RPI is strength of schedule, and 25% of it is completely out of your control, because you can't control who your opponents schedule in their nonconference season. Strength of schedule is very important of course, but there's a certain category of teams that will absolutely kill your RPI no matter how badly you beat them, and the Big Ten loves to schedule them for some reason.

On December 17th, Michigan State had the #3 RPI in the country. Since then, they've won 4 games by an average score of 107-59. This has caused their RPI to fall to 28th. Seriously. After beating #312 Houston Baptist, they fell to 5th. #142 Long Beach State knocked them to 9th. #340 Cleveland State dropped them to 13th, and the final win over #172 Savannah State caused them to fall all the way to 27th. That is insane, and completely contradicts the product we've seen on the court. If any of this reminds you of this hilariously incorrect Graham Couch article, it's because last season Michigan State's RPI was actually the only thing keeping them in the field for a while, and the schedule he proposes in the article would have completely bombed them out of the tournament. 

So what does it mean? For Michigan State? Probably nothing. They're still a 1 seed on Bracket Matrix, but more recent projections have included more 2s and 3s, including obviously the one you see here. But every year this kills the Big Ten in seeding, with the notable exception of last year's Minnesota. The Gophers expertly scheduled teams projected to have winning records in below-average-but-not-atrocious conferences, such as Louisiana Lafayette, UT Arlington, Georgia Southern, and Arkansas State. This paid off for them in seeding, where they got a totally undeserved 5 seed that they used to embarrass themselves against Middle Tennessee. The previous year, Michigan State got screwed over by Oregon, who used the same scheduling tactics as Minnesota to sneak their way into a 1 seed. The Spartans were stuck with a totally undeserved 2 seed that they used to...also embarrass themselves against Middle Tennessee.

John Beilein has been asked about this, as he is one of the biggest offenders of terrible scheduling strategy. He claims (in more diplomatic terms) that he'd rather play a game that's an automatic win and use that time work on his own scheme, rather than running the risk of being upset by a mid-level team. That's totally fair, but stressful when you have anything resembling a bubble team, and annoying when it prevents your great team from getting the seed it deserves. The only solution would be for the selection committee to rely less heavily on such an archaic and clearly flawed ratings system, but a bracketologist can only dream. 

No comments:

Post a Comment