Bracket Analysis: Picking Apart the Good and Bad
by Adam Wodon/Managing Editor
Now that the brackets are set for this year's NCAA Tournament ("hey, ESPN made only 11 gaffes on the Selection Show this year"), we can get to our next annual ritual — picking the brackets apart.
For many, this still means uninformed criticism. Read some of the chatter out there, and it'll make you weep. A vast amount of people still have never heard of the Pairwise, nor this site, apparently. It humbles those who have spent 15 years educating.
We've complained ourselves in the past, poked holes in the Pairwise and outlined many issues. But when it comes down to it, the completely objective system is preferable to a system that involves a lot of human intervention. We can argue the legitimate flaws of the Pairwise all day — it doesn't rate strength of schedule highly enough, there's the "TUC Cliff," there's not enough penalty for losing early in your conference tournament, etc... — but at least everyone knows what the criteria are and there's no secrets.
But there are legitimate questions about the process, and it's better to discuss those than get distracted by arguments of those who don't even know what the process is.
So assuming we take the Pairwise as is, with the strict 1-16 chart that we (most of us) have become accustomed to, then the issues come more down to seeding and placement. Every year there's some argument or another that comes to the forefront, and this year is no exception.
After mentioning this just about every year, it's still worth mentioning again — the committee's rigid dependency on the Pairwise for seeding is something that should be reassessed. There is too much dependency on numbers that are too imprecise to begin with. It's understandable, as mentioned above, to adhere strictly to the numbers when determining who is in the tournament — at least it avoids controversy over that, which is much worse. But for seeding, the human element should be given more leeway.
This year, once again, the concept that Regional hosts schools are placed in that Region, has caused problems. The "nightmare scenario" of New Hampshire, the Northeast Region host, dropping to a No. 4 seed, came to fruition. That means that Boston College couldn't be placed in that region without violating the other sacrosanct rule — avoiding first-round matchups between teams from the same conference.
Some have said the committee should just let the BC-New Hampshire take place. After all, they say, BC would rather have that than get "hosed" by being sent out West.
It's hard to say BC really got that "hosed." BC fans have never traveled that extensively to Manchester. And the travel burdens are not that great. BC is hardly the first No. 1 seed that has to fly for a Regional. WCHA schools do it almost every year. In the pantheon of NCAA Tournament injustices, this one barely measures a blip.
So I don't have a problem with the committee's two sacrosanct rules, and don't have a problem with BC going out West. But if the committee wanted to keep Boston College in Manchester, it could've done something less drastic — flipping No. 13 New Hampshire and No. 12 Nebraska-Omaha. In other words, moving UNH to the 12 slot (and thus a No. 3 seed), and UNO to the 13 slot (and thus a No. 4 seed).
The committee won't do that, though, because, according to another rule of theirs, teams are stuck in 4-team "bands." And the 12 slot is in the 3-band, while the 13 slot is in the 4-band. As a result, it's easier for the committee to flip 9 and 12, than 13 and 12.
That, of course, seems silly. However, I don't have a problem with the committee standing pat in this case anyway. If New Hampshire was moved to a 3 seed, it would've forced Merrimack elsewhere, and Merrimack will bring more fans to Manchester than BC will anyway.
Also, at least the "band" thing is a rule.
There is another place, however, where the committee could've made some common sense changes without violating any rules: Moving Denver away from North Dakota's bracket.
This is something I've harped on every year since the tournament went to 16 teams in 2003. That year, the committee essentially shifted to this strict 1-16 way of doing things, and threw out the window the concept of trying to prevent second-round matchups from teams in the same conference.
Consequently, many times over the years since then, teams have been in this very position, forced to play each other in the second round.
This circumstance is particularly unfortunate when the two teams in question just got done playing each other in the conference tournament championship game — as Denver and North Dakota did last night. It's not fair to North Dakota that it should have to beat Denver again just to make the Frozen Four. And it cheapens the conference tournament.
If it's a potential second-round intra-conference matchup with the 1 seed and 3 seed, then it's OK. Or, if the 1 and 2 seed in a bracket didn't just play each other the weekend before — such as with Yale and Union in the East — then you can also let it slide.
But the committee should do everything it can to avoid a kind of bracket where North Dakota and Denver are matched up again, assuming the higher seeds win.
And in this case, it had plenty of easy solutions. It could've switched Denver with Union, for example. That would've created a conflict with Denver/Minnesota-Duluth, so then UMD could've been flipped to the Northeast to play Merrimack, with Notre Dame coming to Bridgeport to play Denver. Union would then be out West playing Western Michigan, and the problem is solved.
Maybe the committee doesn't want to have to go through those hoops just to avoid that second-round matchup — but that's only if you assume a rigid adherence to the Pairwise to begin with. If you see the list, 1-16, in the Pairwise and think "that's the way we ought to do it as close as possible," then your approach will be to have as little flipping and shifting as possible. But if you look at the list as merely a guideline, and don't worry so much about sticking so precisely to imprecise numbers, then certain matchups tend to make a lot more sense.
Finally, it's always interesting to see how the field would like if the committee used KRACH, the much better computational tool for ranking teams.
This would be the bracket using KRACH:
1. North Dakota vs. 16. Air Force
8. Notre Dame vs. 9. Merrimack
2. Boston College vs. 15. Minnesota
7. Minnesota-Duluth vs. 12. New Hampshire
3. Yale vs. 13. Colorado College
6. Michigan vs. 11. Union
4. Denver vs. 14. Western Michigan
5. Miami vs. 10. Nebraska-Omaha
The only difference in the field is that Minnesota would replace Rensselaer. All other committee rules were adhered to, including the Air Force autobid.
This bracket somehow feels more right than the Pairwise one, but alas.