Bracket ABCs: February Edition
Breaking Down the Pairwise; What We Know Now
Back we are for another season of scouring the Pairwise — the objective system the NCAA uses to select and seed the NCAA tournament — with our Analysis, Breakdown and Comparisons (ABCs).
I've been writing about this for 20 years now. When I wrote my first article explaining the process, very few people got it or even believed it. It took over a decade to educate most of even the mainstream media.
But here we are. It's gotten so routine it's practically boring.
As noted the last few years, the machinations of the Pairwise are so slimmed down these days that there are fewer arcane nuances to decipher. There are only three criteria, and the RPI is the dominant consideration because it's also the tiebreaker. And one of the criteria is head-to-head, but not all teams play head-to-head, of course. The RPI has a lot of nuanced pieces to it behind the scenes, but the final number dominates the whole system.
Between that and the knowledge that most fans have these days, there is not as much to pick apart. As a result, what you see is kind of what you get. Take a look at the chart. Win games, you'll go up. Lose games, you'll go down. The bigger the gap in RPI between one team and another, the less likely they are to move, and vice-versa. There used to be more nuances, or what some people called "flaws," which is why they're no longer there.
The Pairwise has been tweaked many times over the years. It's not perfect, and never can be. But it's in a pretty good place right now. As it stands, KRACH — which is a better system we've long advocated, though not perfect either — and Pairwise are pretty close.
That said, there are things to be on the lookout for, particularly when it comes to seeding, and that's why we're here. As I've said many times over the years, demonstrating a bracket as if the season ended today is rather pointless, because the season does not end today. If you could prognosticate the bracket based upon what you think the rest of the season's results would be, that would be more useful, though would require a crystal ball, of course.
CHN does have some tools to help, however. Soon, we will bring back our Pairwise Probability Matrix, which does exactly that — extrapolates out the remaining results for the season, more than 20,000 times, and gives you the chances of each team making the NCAAs. There are too many games left for that, however, at the moment. Then there's the return of You Are The Committee, which allows you to plug in results of conference tournament games and see what the final Pairwise would look like if that scenario happened. That will have to wait as well.
(Last year, College Hockey News also rolled out the Pairwise Player. This is a tool for users to make it easier to see the Pairwise, and its various related details, on mobile devices.)
In the mean time, here are some things to look out for, and thoughts on the hopefuls.
1-2. Minnesota-Duluth, Denver
These two have been the best teams in college hockey all year, and there's every reason to believe that will continue. In fact, more than likely, what you're looking at here is a fight to see which team will avoid having to be in the Fargo Regional.
The top 4 seeds overall are typically placed in the four regions, by the Committee, "closest to home," with the top seed receiving first priority. But North Dakota is hosting the Fargo Regional, so it must be there by rule, in front of its rabid fans. If that's the case, the Committee will have a decision to make. North Dakota is currently a 3 seed, so there typically wouldn't be an issue, since the Committee usually doesn't look beyond the first round. But will the Committee choose to "protect" the top seed by keeping it out of the Fargo Regional, even if North Dakota is a 3 seed? I can see an argument for doing so.
If they do, then a simple switch of Minnesota-Duluth/Denver would be fine. There is another wrinkle, though. For Denver, it's a flight either way. But for UMD, it's just 242 miles across the state of Minnesota to Fargo, so that's a bus trip. The Committee may be very reluctant to go so far to protect UMD from a second-round game against North Dakota when it's the difference between bus and plane.
If Denver is the No. 1 overall seed, that switch is more likely. Though again, the Committee doesn't usually make a switch to avoid second-round matchups like that. But it could. And maybe should.
But there is another major wrinkle here. If North Dakota is a 4 seed, it would theoretically be matched up in the first round against UMD or Denver. But that is a matchup of two teams from the same conference, and that is also not allowed, by rule. In that case, neither UMD or Denver is going to Fargo, even if they are No. 3 or 4 overall seeds.
Believe it or not, it may only get wackier. Read on.
The East Regionals are in Providence and Manchester, N.H., this year, which are essentially equidistant spots from Boston. The same applies here for both Harvard and Boston University, which flipped spots recently and could do so again. Here's the crazy part:
If Harvard is the 3rd overall seed, and if a Hockey East team is the 13th or 14th overall seed, and if North Dakota is the 13th or 14th overall seed, the Committee is going to have their hands full. (Same goes for BU being 3rd overall and a Hockey East team like BC being the 13th or 14th).
Harvard cannot play that ECAC team in the first round (say, Cornell). And the top two seeds should be "rewarded" by playing the bottom two seeds, which will be the Atlantic Hockey and WCHA conference tournament champs.
The Committee is allowed to freely move the groups of four teams within their seed bands. Theoretically. But because the disparity between the 14th seed, and the autobid 15/16 seeds is usually much greater than the gap between any other two teams, out of fairness, the higher seeds are usually protected in this manner and "rewarded" with the lowest overall seeds.
However, in the scenario outlined above, that would be impossible ... unless Harvard was sent to Fargo, to face North Dakota.
That would be a cruel twist of fate.
4-5-6-7. Minnesota, Boston University, Penn State, Western Michigan
There's a big dropoff in RPI from Harvard at 3 to Minnesota at 4. And then there's another gap after Western Michigan at 7. So this looks like a good grouping to concentrate on.
One of these teams has a good shot of taking the fourth No. 1 seed. Where it lands depends so much on all of the shenanigans explained above that it's hard to say right now. But in a perfect world, this is the team that goes to Fargo. The thought of Minnesota playing North Dakota in the first round of the NCAAs, or even the second round, is a pretty exciting thought right now.
The comparison between Minnesota and BU tells the tale of why everything but RPI in the Pairwise makeup is all but meaningless these days. BU has a better Record vs. Common Opponents than Minnesota, but Minnesota has the edge in RPI. The only other criterion is Head-to-Head, and the teams haven't played, so that's out. So the comparison score is 1-1, with RPI as the tiebreaker. So the Common Opponents means nothing either.
Neverminding the Pairwise, the main thing people are watching here is whether Penn State can hang on and get the program's first NCAA tournament bid. It was No. 1 in the Pairwise a few weeks ago, but because its schedule hasn't been great and it only has .0009 of quality win bonus points (the lowest of any team in the top 25), things bottomed out quickly once the Nittany Lions lost a few games in a row, dropping to 11. Two wins over Wisconsin on the road this past weekend were huge, and Penn State looks very safe now, but since Wisconsin dropped out of the Top 20 of RPI with the losses, those aren't "quality wins" for Penn State at this point either.
8-13. Providence, Mass.-Lowell, North Dakota, Union, St. Cloud State, Cornell
This is a big pile of mush, and where these teams end up is anyone's guess. Don't believe anyone who claims they know. St. Cloud State got itself into the mix with a big sweep of Miami to put it one game over .500. You must be .500 or better to qualify. So that's something to keep an eye on.
Because you have two Hockey East teams in this pack, and two ECAC teams, there might be a whole lot of shuffling going around to avoid first-round intra-conference matchups. That just adds to the "beats me" nature of prognosticating this bunch at the moment.
14-15. Ohio State, Boston College
These teams are teetering on the edge. Boston College is more interesting because it is in first place in Hockey East, yet may not make the NCAAs, which is wacky. If you look at the Pairwise grid, Boston College owns the only comparison win against a team above it overall among all the teams in the top 30. That again just goes to show you how RPI rules nowadays. You used to see all sorts of different individual comparison scenarios.
16-21. Vermont, Nebraska-Omaha, Notre Dame, St. Lawrence, Air Force, Wisconsin
To me, the bubble looks like 20, realistically, at this point. Wisconsin is No. 21. It has its work cut out. Vermont, though, is within striking distance. In fact, the gap between 14 and 20 is smaller than the gap between 13 and 14.
Air Force is interesting because it could theoretically finish in the top 14, lose in the Atlantic Hockey tournament, and two teams from Atlantic Hockey would get in. That is unlikely, but not impossible.
If I had to guess, right now, this scenario seems likely:
1. Minnesota-Duluth plays WCHA champ in Fargo.
2. Denver plays Atlantic Hockey champ in Cincinnati
3. Harvard plays BC or Ohio State in Manchester.
4. Minnesota plays Cornell in Providence.
So much can change, but that seems plausible. Then you'd have some poor Eastern team, as the higher seed, going to Fargo to play North Dakota, for the right to face UMD in the Regional Final.