Tightwad Hill

GUIDE TO THE NCAA BRACKET

Many of you are looking at your brackets and wondering which upsets to pick. It's useful to look at historical trends, and also to look at a statistic called Adjusted Scoring Margin. ASM measures the margin of victory (and defeat) for tournament teams compared with the average scoring margins for their opponents. It's one tool for evaluating teams who come from disparate backgrounds - i.e., mid-majors and big-conference teams. ASM is very useful for figuring out which teams will make deep runs in the Tournament; it's less useful in picking individual games.

Joe Lunardi makes a big deal about ASM; so does a guy named Nate from a blog called Nate's Random Stuff. He has helpfully done the research on ASM and made it available online for your review. Nate also has a formula that he says is more accurate called CRACK - don't ask, it's rather complicated but you can find it elsewhere on the site. We'll reference Nate's ASM rankings in this guide to bracket matchups.

1-16: As you all know, it's never happened - though there have been some close calls. Georgetown and Oklahoma in '89 won by a single point; more recently Purdue beat Western Carolina by a deuce in 1996. Resist the temptation, though.

2-15: It's happened four times, most recently in 2001 when Hampton beat Iowa State 58-57. I, along with lots of other bandwagoners, took Winthrop over Tennessee last year - resist the temptation in '07.

3-14: 17% of 14-seeds win in the opening round. That's a pretty amazing statistic when you think about it - almost one in five. It's happened in each of the last two tournaments - last year Northwestern State got Iowa and in 2005 Bucknell upended Kansas. There aren't a lot of good candidates this year:

Texas A&M-Penn; Oregon-Miami OH; Pitt-Wright State; and Washington St-Oral Roberts. Penn isn't going to beat A&M, and I doubt Oregon will be tripped up by Miami OH. Pitt should whip Wright State. That leaves Wazzu against Oral Roberts. ORU won at Kansas this season, and their ASM (#77) is nearly as high as Wazzu's (#53). I'd hate to see the Cougs fall victim to the upset, but it's a very reasonable pick.

4-13: 20% of 13 seeds win in the opening round. It's happened in five of the last six Tournaments. It probably behooves you to pick one (but not more than one) of these in your bracket.

Virginia-Albany; Texas-New Mexico State; SIU-Holy Cross; and Maryland-Davidson. The latter two games are by far the best candidates. SIU has a very low ASM for a four seed, and Ralph Willard is a tournament-tested coach at Holy Cross. Maryland is a good team, but Davidson's ASM is higher than the majority of teams in the Tourney. Your choice on which to pick; I lean toward Holy Cross for no good reason.

5-12: The magic number. One third of 12 seeds win their games. Starting in 2001, every year has had at least one 12-5 upset, and most years have had more than one. In 2002 all but one of the 12 seeds won; last year both Montana (over Nevada) and Texas A&M (over Syracuse) scored the upset. Interestingly, it's a relatively recent phenomenon - in the early days of the 64-team field 12-5 upsets weren't all that common. You need to pick at least one and probably two of these if the games look right.

Tennessee-Long Beach State; USC-Arkansas; Virginia Tech-Illinois; Butler-Old Dominion. Of these, ASM points you to pick Arkansas over USC. I just don't see it. Instead I'll take the Illini to drop an inconsistent Virginia Tech team, and that's it. If I had to pick another one, it would be LBSU over Tennessee.

6-11: Fewer 11 seeds than 12 seeds win in the first round (31%). Still it's likely to happen at least once in '07.

Louisville-Stanford; Vandy-George Washington; Duke-VCU; Notre Dame-Winthrop. Vandy is terribly overseeded, so consider them as a victim here. Winthrop is the trendy pick to drop the Irish, but Notre Dame is a pretty good team (12th in the nation in ASM) who almost knocked off Georgetown in the Big East semis. I'd probably lean toward GW here.

7-10: Hardly an upset anymore; 40% of 10-seeds win. The same rule applies - pick one or two if the matchups look right.

Nevada-Creighton; Boston College-Texas Tech; Indiana-Gonzaga; UNLV-Georgia Tech. ASM suggests that you consider the Blue Jays to knock off Nevada, but I don't see it. Georgia Tech over Vegas makes much more sense to me, given the historical poor performance of the MWC in the NCAAs. Gonzaga is intriguing, but Indiana is a pretty good 7 seed. I don't see Texas Tech doing much with Boston College, but that's more of a match-up problem since no one on the Raiders can guard Jared Dudley.

8-9: 9's win more frequently than 8's (53% of the time). Pick at least two.

BYU-Xavier; Marquette-Michigan State; Kentucky-Villanova; Arizona-Purdue. Xavier really doesn't belong based on their resume, but I was shocked to discover that the Musketeers are 14th in the country in ASM. I'm going to pick three "upsets" here, with Arizona the only 8 seed to survive.

posted by Tightwad @ 4:15 PM