March 18, 2007 PRINT Bookmark and Share

Final 2007 Bracket Analysis

by Adam Wodon/Managing Editor

Selection Sunday came with very few surprises once again, which annoys a minority of people who like suspense, but is otherwise universally accepted by college hockey fans as a tradeoff for having a reasonably fair, objective-based system.

(See here for the completed official bracket)

That said, there are things that are still a mystery to many of even the most ardent followers. And some of those things were hopefully cleared up as a result of this year's selections.


Tragically, we made some missteps in the last month while doing our Bracket ABCs look-aheads to where things would go. In particular, an oversight on some of the potential fluctuations of the "Teams Under Consideration" list (a TUC is a team that finishes in the Top 25 of RPI) caused us to make some statements that were more definitive in nature than they should've been — particularly when it came to St. Cloud being a No. 1 seed and Denver being in the tournament.

So, that said, it's only appropriate now that we get to toot our horn about being correct, right?

(See here for Saturday night's CHN bracket projection and compare it to the final result)

Out in the digital college hockey blab-o-sphere, there was a misread of the Pairwise taking place last night when it came to breaking the tie in the 11th spot (see the final Pairwise list) — which is understandable given how things go. Many sites, including ours, that show the Pairwise — the committee's system of objectively selecting the field (see here for detailed explanation of how it works) — list the teams in order of the most "comparison wins." And, including ours, when there's a tie in the number of overall comparison wins, the teams are displayed in order by comparing the individual comparison among the tied teams.

There is no question that a tie within an individual team-to-team comparison is broken via the RPI criterion. But in the case of a tie in overall comparison wins on the Pairwise list, most sites — again, including ours — still list it wrong ... or not "wrong," per se, but just according to an older method.

In the really old method, the committee never even saw a Pairwise chart. It only saw the list of team-to-team comparisons. They just eyeballed it, and it was obvious which teams were winning most of them. But when it came to comparing one team to the next in microcosm, it just used those team's individual comparison to each other.

Eventually, life followed art, and the list was used wholesale. So breaking ties become trickier. So in recent years, the committee has shown time and again that it would break these kinds of ties simply by comparing the RPIs of the respective teams.

So when it came down to a tie at the No. 11 spot between Massachusetts, St. Lawrence and Maine — we, correctly, ignored the way it was listed in the Pairwise and compared the individual RPIs. In that case, St. Lawrence became No. 11, Maine was No. 12 and Massachusetts No. 13.

As a result, we got the bracket completely correct ... except in one spot ...

Air Force is Grounded

The committee decided to put the No. 15 overall seed — Air Force — against the No. 1 overall seed, Minnesota, instead of keeping it 1-16, 2-15. So Alabama-Huntsville will play Notre Dame in the Midwest Regional instead.

Without yet hearing a specfic rationale from the committee, it can only be speculated on the reasons — either to avoid two teams from having to fly, by keeping Air Force within 60 miles of home; or to boost the atmosphere in the West Regional at Denver, which suffered a blow when the Denver Pioneers failed to make the tournament.

Actually selling tickets to that regional is not necessarily a factor, because Denver fans were forced to buy those tickets as part of a Frozen Four/season ticket package. But those other factors could be. (How ironic, by the way, that Air Force will now not have to fly — it's the Air Force for heaven's sake, can't they fly themselves? ... And didn't Air Force join Atlantic Hockey. They didn't care about flying then! we kid...)

Colleague Mike Machnik pointed out Saturday night that this was a strong possibility, but it seemed like the committee would keep that 1-16, 2-15 setup. He was right.

What Else?

So, given those two factors above, the committee went strictly (other than the Air Force switch) by a serpentine order — which means, ideally, 1-16/8-9 ... 2-15/7-10 ... 3-14/6-11 ... 4-13/5-12.

This is exactly the way the committee has operated since going to a 16-team field in 2003. The committee has completely ignored the potential for second-round matchups against teams from the same conference, and ignored attendance factors — two things that were very much considered before 2003.

However, the committee did have to stray from this in two instances because of the sacrosanct policy of avoiding first-round matchups against teams from the same conference.

As a result, to avoid the 8-9 matchup between Michigan and Michigan State, the committee moved No. 10 North Dakota over to play No. 8 Michigan ... and that set up a 7-9 matchup between Boston University and Michigan State in Grand Rapids. The committee could just as easily have switched BU and Michigan, which would've put the 7-9 BU/Michigan State matchup in the West Regional instead. This would've made Minnesota fans very happy — since some of them are peeved that Minnesota may have to play North Dakota in the second round. But the inclination of the committee has been to preserve the No. 1 vs. No. 2 seed matchup in the second round, and thus it tends to flip-flop the teams in the No. 3 seed band.

The same was done with Maine and St. Lawrence. To avoid a 5-12 matchup between Boston College and Maine, the committee switched Maine to play No. 6 St. Cloud State in Rochester ... and made a 5-11 matchup between Boston College and St. Lawrence.

A Flight is a Flight

The committee, technically, is supposed to take the four No. 1 seeds, and place them in the regionals "closest to home," going in order from the No. 1 overall seed, to No. 2, etc... Technically, then, Grand Rapids was the closest of the four to Minneapolis.

But doing it that way, would then have pushed Notre Dame to Rochester, and then Clarkson to Denver. Ugh! That made no common sense.

So the committee did the right thing, realized that "a flight is a flight," i.e. it doesn't matter where Minnesota flies to, and kept the other two teams close to their home area.

For a while, we had been speculating that the committee would "protect" Minnesota, if it was the top overall seed, by keeping the Gophers away from the Pioneers in the Denver Regional. This was when it looked like St. Cloud would be a No. 1 seed. So we projected Minnesota to Grand Rapids, Notre Dame to Rochester, and SCSU to Denver.

As it turns out, Denver didn't make the tournament, so there was no need to protect the Gophers in that sense. So it didn't matter. And so we'll never know what would've happened.

The committee can do many things if it decides to care about attendance, but since switching to a 16-team tournament in 2003, the committee has gone Straight Pairwise every season, without fail, other than just switching to avoid first-round intra-conference matchups. The attendance issue has not been used as a factor whatsoever.

The Bubble

Clarkson's big rally against Quinnipiac in the ECAC title game meant a big sigh of relief to Massachusetts and Miami, which were the last two teams on the bubble. If Quinnipiac would've won, one of those teams would've been bumped out — depending on the Minnesota-North Dakota outcome. With Clarkson's win, it didn't matter, and both got in.

The last teams out of the tournament were Denver and Wisconsin, which would've gotten in had KRACH been used. In fact, Colorado College and Michigan Tech would've gotten in, too, which would've meant seven WCHA teams in the tournament, which is probably why they'll never switch to KRACH.

It should be noted that the RPI "bonus" for "good wins" on the road — which is estimated to be around .003 — is what got Maine in the tournament, thanks to two early-season wins at North Dakota. With a smaller bonus number, those games wouldn't have mattered as much, and Denver would've gotten in. And, oddly, with a slightly higher bonus number (around .005), Denver would've gotten a bigger boost over St. Lawrence and gotten in that way, too.

Gotta love it.

Overally, the committee did a great job again. Hat's off. No complaints this year, folks.

OK, that's enough geek talk for this year. It's time to play the games. Enjoy.

Mike Machnik contributed to this article.

Bookmark and Share PRINT

Comment on this Article

Send Feedback | Privacy Policy | Terms and Conditions

©2018 Adam Wodon. All Rights Reserved.