The 2012 Stat Geek Selection Committee Mock Bracket: The Brain Trust Speaks Notes from the Sports Nerds

The 2012 Stat Geek Selection Committee Mock Bracket: The Brain Trust Speaks

13 comments - NCAA Basketball, NCAA Tournament - posted by David Hess on February 18, 2012

Invited members of the media met Thursday and Friday in Indianapolis to simulate the duties of the NCAA selection committee. While the bracket they came up with is reasonable, it suffers from a fatal flaw — the RPI plays a huge role in the process.

To try to improve on their efforts, we gathered a dozen numbers-oriented bloggers for a rival Stat Geek Selection Committee (we outlined the details Thursday).

We finished our voting and consensus building today, and the result is this 2012 NCAA Tournament mock bracket, which uses the exact same simulated conference tournament winners as those from the media bracket to allow for a more direct comparison.

2012 Stat Geek Selection Committee Mock Bracket

(click to enlarge)

Later this holiday weekend, we’ll post detailed analysis of this bracket, and compare it to the media’s bracket. But for now, we just wanted to release this puppy into the wild.

If you’ve got questions about our bracket, feel free to ask them in the comment section below, or Tweet at any of the committee members directly:

Stat Geek Mock Selection Committee Members

Drew Cannon of Basketball Prospectus (@DrewCannon1)
Andy Cox of Crashing The Dance (@crashthedance)
David Hess of TeamRankings (@AudacityOfHoops)
Evan Jacoby of Rush The Court (@evanjacoby)
Monte McNair of Outside the Hashes (@OTH_blog)
CJ Moore of Need I Say Moore? and Basketball Prospectus (@cjmoore4)
Daniel M of D Stats (@DSMok1)
Michael Portscheller of Big Ten Geeks (@bigtengeeks)
Corey Schmidt of Basketball Prospectus (@cjscmhidt1)
John Templon of Big Apple Buckets (@nybuckets)
Nathan Walker of The Basketball Distribution (@bbstats)

http://twitter.com/brian72975 Brian Edmison

Saint Louis can’t play in St. Louis.
Richard Stacey

Murray 11- St mary 7????
Brianalisch24

Based on that bracket, my top first round upsets would be: Long Beach State over Notre Dame, Uconn over Memphis, Murray State over Temple … Duke probably wouldn’t like seeing LBS lurking as a possible 2nd round matchup. Long Beach is type of team that matches up well against Duke; long, athletic, can defend, good fast break team, inside presence with a dynamic outside score (ware). They have played tough comp. this year, Duke also plays down to comp.
http://www.teamrankings.com/ David Hess

Doh! I figured we’d make a a random error or two like that. I believe (hope) that the selection committee has a computerized list of conflicts like that, and doesn’t have to scrub the bracket by hand like we did.
http://www.teamrankings.com/ David Hess

Richard — Keep in mind that this bracket was created before Murray State beat St. Mary’s, so that head-to-head result (Murray State won by 14) didn’t play any role in this. On top of that, Murray State didn’t get one of the mock automatic bids, which means they had one more bad conference loss on their resume in this exercise.

Murray State was the top 11 seed, so our S-Curve ranks fro these two were #27 for St. Mary’s and #41 for Murray State. Compare that to #29 and #27 respectively for the media mock, and you can see the main difference was our evaluation of Murray State.

Our goal here was to follow the NCAA’s own guidelines for selecting teams, which say to select the “BEST” 37 at larges, not the most deserving. Every committee member was free to interpret that however they like, but I know many leaned heavily on various predictive power ratings, which generally rated Murray State much lower than the public consensus.

For example, our predictive power ratings from February 16 (when we started the selection process) had Murray State down at #37: http://www.teamrankings.com/ncaa-basketball/ranking/predictive-power-ranking-by-team?rating_date=2012-02-16. For comparison, our Wins Rating, which considers ONLY wins and losses, and not margin of victory, had them all the way up at #11: http://www.teamrankings.com/ncaa-basketball/ranking/wins-power-ranking-by-team

I think that was a reasonable seed. Against an absolute joke of an Ohio Valley schedule (ranked #299 by our ratings, like playing every game against Stetson College), Murray State has struggled more than their single loss would indicate. They’ve been forced to come back from large deficits on multiple occasions. In the RPI, a win is a win. But if you’re really trying to find the BEST teams, scraping by OVC foes is not a great sign.

At any rate, the great part about the tournament is that you can play your way out of a bad see. If Murray State beats Temple in the first round of our bracket, all of a sudden their path is the same one they would have gotten if they were a #6 seed.
http://www.teamrankings.com/ David Hess

Brian — Hard to disagree with your choices there. Long Beach played a ridiculous non-conference schedule (#7 according to our ratings:
http://www.teamrankings.com/ncaa-basketball/ranking/non-conference-sos-by-other) and went 7-6 against it.
AD Esteban

The PAC-12 even in this down year will get at least 2, if not 3 in the dance. Cal will not win the PAC 12 tourney so UW, AZ or Oregon will get the auto bid.
http://www.teamrankings.com/ David Hess

AD — We used the same tourney winners that the media used in their mock bracket, to make comparison easier. We agree that if Cal loses in the tourney, they’ll probably still get in.

Also, our goal here is not to project what WILL happen. It’s to pick the best at large teams. Those are very different things, due to the real committee’s reliance on the RPI.
Dan

How’d you get UW-Milwaukee coming away with a Horizon League championship? Valpo is first in the conference, and if things stay that way, will host much of the league tourney. Valparaiso’s home court advantage is strong this year (almost 10 points per game) and I see them having a significant chance at winning the championship at home, while Cleveland State is the stronger team and has a very good chance at winning as well. But Milwaukee? They’re fine at home but not a neutral or opposing court. Your Horizon League Tourney page says that your computer uses home court: “First Round – higher seed. Quarterfinals – home court of WI-Milwkee. Semifinals – home court of WI-Milwkee.” That’s not likely at all, and it really screws up the league tournament predictions.
Dan

Hmmm…I see now that you copied the media’s conference tournament predictions, so my question is somewhat moot. I’d still like an answer on why your Horizon League page has UW-Milwaukee as the tournament host, however.
http://www.teamrankings.com/ David Hess

Dan — Thanks for the heads up on the conference tourney location. We’ll look into this and get it fixed.
Davidperl

hey I am definitely going to use team rankings as a resource more often… the descriptions you guys have of each statistic are reasonable and make sense. if you guys could write an article about how to deal with teams you might be biased towards that would be great. for example, I am a big fan of michigan georgetown and marquette. so, it is hard for me to judge actually how good they are because i root for them every time
http://www.teamrankings.com/ David Hess

David — Thanks, we’re glad you’re finding the site useful. I definitely have that same problem staying objective about Kansas, as I grew up rooting for them. I’m a pessimist, and never believe they’re as good as they are. But generally, if I want an objective view, I’ll check out our new predictive ratings that take into account our preseason projections:
http://www.teamrankings.com/ncaa-basketball/ranking/predictive-by-other

If you look at that page, it shows each team’s record against teams rated in the 1-25 range, 26-50 range, etc. Those are useful for trying to figure out if the team’s rating is based on blowing out creampuffs, or beating really good teams.