Bracket ABCs: Championship Weekend Edition
by Adam Wodon/Managing Editor
Hard to believe it's that time of year again. Time flies when you're having fun. Or getting old.
This weekend, will be the last of what the NCAA considers "regular-season" games — the championships of each of the six conferences. When that is done, the Pairwise, the mathematical system used to rank the teams, will be set for this season.
You can, of course, see the Pairwise as it stands right now. But thanks to CHN's own You Are The Committee tool, you can also punch in your prognostications for this weekend's games, and see what the final Pairwise would be based upon your results. There are still over 3 million different combinations of results (3,145,728 to be precise) — enough to keep you busy for a while. But, at this point, we can give you a pretty good idea.
For the as-yet-unintiated, here is our Pairwise Primer, a detailed but simple explanation of what the Pairwise is and how it works.
Remember to stay with CHN throughout the weekend for updates as the results roll in, and follow along with our annual Pairwise Live Blog.
At the bottom of this article, you can see a discussion about how the brackets are shaping up and what the main issues are. Remember, the Committee will take the teams, 1-16, and attempt to order them in a serpentine fashion. i.e. 1 vs. 16, 2 vs. 15, 3 vs. 14 ... and also put the 1-8, 2-7, 3-6 and 4-5 in the same region. But they can make adjustments for certain factors, some they must do, and others that are more subjective.
As we've mentioned in past Bracket ABCs articles, this year, there is less volatility in the Pairwise. There is less potential for craziness. That's because the Record vs. Teams Under Consideration criteria was discarded, along with the whole notion of what a TUC is. Instead, the Quality Win Bonus and home-road weightings have been rolled right into the RPI itself, and the amount of criteria used to compare each team to each other has been reduced to three, with RPI as the tiebreaker. Things like the "TUC Cliff" no longer exist, and so it's a much more straightforward "who wins, who loses" type of thing this weekend.
We've been tinkering, internally, with a computer program that calculates results, and spits out the odds of each team placing in each spot. However, it's not ready for prime time, so we haven't made it public. But internally, it gives us a pretty good idea. It runs what are called "monte carlo simulations," basically playing out the schedule, with the results of each game determined randomly, weighted by each team's KRACH rankings (KRACH is a whole other story, FYI). This is then done 1 million times, and the percentage chances are figured out for each team. The drawback of this method is that it doesn't necessarily factor in every possibility — though by doing the simulation 1 million times, it's essentially the same.
Our good friend Jim Dahl of College Hockey Ranked actually does run through all 3 million-plus combinations, and has calculated the expected finish of those results. You can see that here. The only caveat there is, not all combinations of results are as likely as others. For example, it's more likely Michigan wins than Penn State wins, but the combination-matrix treats them both equally. Again, this is basically splitting hairs, and each method will give you very good insight.
So, enough with the geeky stuff ... let's try to break down each team's chances into layman's terms, and also try to look ahead a bit at a bracket.
The Gophers have been locked into the top two for about six weeks, and locked into No. 1, for all intents and purposes, since Notre Dame defeated BC in the regular-season finale. Minnesota will play the Atlantic Hockey tournament champion in St. Paul.
2-3. Boston College, Union
It's rare that so many No. 1 seeds are already locked in, but these two, along with Minnesota, already have their spot. There is a minuscule chance that Union could move up to No. 2, but really, it doesn't even matter. BC will be the top seed in Worcester, while Union will be the top seed in Bridgeport, the place where it won the Regional two years ago.
4. Ferris State
The Bulldogs will most likely grab the final No. 1 seed if they win the WCHA tournament. But even if they do, Wisconsin could grab this slot with a win at the Big Ten tournament. Whatever team it is, will be the No. 1 seed in Cincinnati.
The Badgers are also a lock for the NCAAs, and have a shot at a No. 1 seed with a Big Ten tournament championship. The most likely scenario is a No. 2 seed for the Badgers, and now we start getting into dicey questions with the Committee, in terms of where it will place teams. Will attendance concerns be a big factor, like last year, or only a secondary factor? There's plenty of wiggle room there. If Ferris/Wisconsin remain 4-5, it would make sense just to have Wisconsin in Cincinnati. Madison to Cincinnati is 442 driving miles according to Google Maps, whereas it's just under 300 miles to St. Paul. That means St. Paul is driving distance, while Cincinnati is flying distance, according to NCAA criteria. That could come into play, if the committee so chooses.
Quinnipiac has a chance to get a No. 1 seed by winning the ECAC tournament, but even if it does, it needs other things to break right — like Ferris, Wisconsin and Lowell all losing. It is highly likely Quinnpiac will be a No. 2 seed, and so, here again, we wonder what will happen — whether the Committee follows its serpentine ordering (1-16, 2-15, 3-14, etc...) or leans more on attendance factors. There's no doubt Quinnipiac would pack it in at Bridgeport. That could set up another potential Union-Quinnipiac Regional final, like last year in Providence. It would also work out if Quinnipiac remains 6 while Union is 3 — then you have your "pure" 3-6 pairing. But Quinnipiac could just as well go to Worcester and draw just about as well. The issue is if Quinnipiac finishes 5 or 8 and must go West — that will be the test of the attendance issue.
Again, a team with a chance for a top seed, but a most likely No. 2 seed, with all the same questions on attendance as Quinnipiac.
8. Notre Dame
We're still in "absolute lock" territory here, and who would've believed that just a couple weeks ago, when the Irish were clinging to their NCAA hopes. Three wins in four meetings against Boston College sure did solidify things in a hurry.
This is the first team on the list not a 100 percent lock, according to Jim Dahl's calculations. But it would take a practical miracle for the Friars to miss out. Even a semifinal loss at the Hockey East tournament would likely only drop Providence to around 11. The interesting thing to watch out for here is that, you have three — and potentially four (Vermont) — Hockey East teams in the 5-12 range. Since the Committee is mandated to avoid first-round matchups between teams from the same conference, that could make for some interesting juggling.
10. St. Cloud State
In the same boat as Providence in terms of being "almost locks" but not quite 100 percent. This is the first team on the list that isn't playing this weekend, so there is less chance for wide swings. If all the higher seeds win this weekend, this is the exact spot the Huskies will find themselves in. Going by pure serpentine order, that would put them in a 7-10 matchup in the bracket with the 2nd overall seed (Boston College), which means Worcester. Putting the Huskies in St. Paul will be enticing, but the Gophers alone should be plenty powerful of a draw there, and favors aren't usually given to No. 3 seeds. The only way to naturally pair with Minnesota would be to get to the 8-9 game, which has a decent shot if Providence or Notre Dame loses Friday. However (see below), there is another way, if a swap is needed with a Hockey East team.
This is where, if you check Jim Dahl's calculation, is the one place where the figures are most deceiving. Michigan is listed as having a 58.8 percent chance of an NCAA bid, but that is based upon the possibility of losing to Penn State in the Big Ten first round game. When you run all 3 million combinations, every outcome is treated equally, so that makes the Michigan-Penn State game a 50/50 proposition. But realistically, it's not that — Michigan's chances are much better, its two losses this year to Penn State notwithstanding. And if Michigan wins that game, it's all but assured of a bid. So in reality, I'd put the Wolverines' chances at returning to the NCAAs at about 75 percent. Odds are good Michigan will be at one of the eastern Regionals.
12. Minnesota State
A win and a loss at the WCHA tournament, and things could get dicey for the Mavericks, if there are upsets elsewhere. ("Upsets" at this time of year means, teams that wouldn't ordinarily get an at-large NCAA bid win their conference tournament, thus getting an automatic bid, and thus knocking out another team.) Certainly losing the first game, to Bowling Green, will make things very difficult for Minnesota State, and open the door to the likes of Northeastern to sneak through. If they get in, I'd put Cincinnati as the most likely location for the Mavericks.
13. North Dakota
While everyone else has dropped the consolation game at this point, the NCHC, in its inaugural tournament, has decided to have one. This could have a profound effect. That means if North Dakota loses its semifinal game to Miami, it still has a good chance to get in with a win in the consolation game. Hard to believe a team as hot as it was in mid-season is this close to the bubble, but a typically poor start was compounded by the fact that the entire league did poorly out of conference early on. With one win, UND will definitely be in the Top 15, and then just has to make sure there aren't upsets elsewhere. Thing is, if it doesn't win the NCHC tournament, any of the other three teams (Denver, Miami, Western Michigan) would be considered an "upset" and take away a spot — meaning North Dakota needs to finish in the Top 14 to make it.
Vermont is not playing this weekend, but still has a very good chance of getting in — better than 50/50. It's rooting for teams like Union and North Dakota to win their tournaments. You could easily see Vermont wind up in Bridgeport, the place where it won the Regional in 2009 to make the Frozen Four.
15-16. Colgate, Cornell
Both of these teams had some very nice wins this season, but stumbled just enough at times to jeopardize what seemed like strong NCAA chances. It's possible both get in, if they both win their semifinal games. But that's no easy task, with Union and Quinnipiac the opponents. There is no ECAC consolation game this year, which hurts these teams' chances of getting in. If both lose in the semifinals, both would probably be out. Whereas, in years past, the winner of the consolation game would have a very good chance of getting in anyway, giving the ECAC three teams. This is the exact opposite of what the ECAC hoped when it eliminated the game — although you could say eliminating it is a good idea anyway, if you think it's the kind of thing that just outlived its usefulness as a concept. Let the debate rage.
17. New Hampshire
Despite being this close, one win followed by a loss in the finals doesn't move the needle on the Wildcats' RPI to move them up. So UNH has to win the Hockey East tournament to get in. If it does sneak in that way, it's looking at going to Bridgeport, since it can't play Boston College in the first round.
Northeastern, sitting this weekend out, has plenty of ways of sneaking into the Top 15, but some of those ways entail upsets elsewhere, which means it would have to finish Top 14. But Northeastern does have a fighting chance. Like UNH, that would probably mean going to Bridgeport.
Everyone else still playing must win its conference tournament to get in. That group includes Western Michigan, Ohio State, Bowling Green, Alaska-Anchorage, Denver, Miami, Michigan State, Penn State and all of the Atlantic Hockey schools. None of these schools can finish higher than 15th in the Pairwise, which means all the non-Atlantic teams would be slotted at No. 15 and play Boston College in Worcester.
All, that is (as reader Tim pointed out in the comments), except Miami, which is hosting the Cincinnati Regional. This could actually create a very dicey situation, akin to 2003, if the old-timers remember. If Miami is in Cincinnati, and the Atlantic champ is in St. Paul, and two Hockey East teams — say, Vermont and Northeastern — are in slots 13 and 14, then what team would Boston College play? It cannot play Vermont and Northeastern by rule. It would be unfair for the top seed not to get the lowest seed, given the discrepancy between 15 and 16. And it cannot play Miami — at least not in Worcester. Furthermore, you can't switch 15 and 16, because 15 and 1 are both hosting, so can't move.
So ... would you move Boston College to Cincinnati? Probably not. So, now what? You take the Atlantic champ and send it to Worcester, and Minnesota then gets either Vermont or Minnesota in the first round.
This is the same thing that happened to Cornell in 2003, the first year of the 16-team tournament, when instead of the Atlantic champ, it got Minnesota State in the first round. Cornell was up in arms, briefly, but clobbered the Mavericks anyway.
This is where I've advocated in the past — many times — for the Committee to simply swap 12 and 13 instead of swapping 13 and 16. But, according to rule, the teams are in "bands" of four, and must be kept in those bands. In other words, teams 9-12 are 3rd seeds, and can only be swapped with other 3rd seeds — they cannot be swapped with 13 because 13 is a 4th seed. This is silly, since 12/13 is a closer swap than 13/16 — but the Committee doesn't afford itself that opportunity.
The Committee might have an escape clause, though, because it has given itself the wiggle room when one conference has five or more teams in the tournament. In that case, first-round intra-conference matchups are allowed. So, BC could face one of those two teams.
Otherwise, the Atlantic Hockey champ will play Minnesota in St. Paul.
There is very little chance it happens exactly like this, but for instructional purposes, we can see what a bracket could look like if all the higher seeds win every game this weekend. First, let's place them in strict 1-16 serpentine order.
1. Minnesota vs. 16. Atlantic champ
8. Notre Dame vs. 9. Providence
2. Boston College vs. 15. Minnesota State
7. Quinnipiac vs. 10. St. Cloud State
3. Union vs. 14. Michigan
6. Wisconsin vs. 11. North Dakota
4. Ferris State vs. 13. Vermont
5. Mass.-Lowell vs. 12. Colgate
The first thing that must be changed is Notre Dame and Providence. As members of the same conference, they can't play each other. Given the attendance factor, this would be a no-brainer switch — Providence and St. Cloud State swap so that Providence plays in Worcester against Quinnipiac, and St. Cloud State plays in St. Paul against Notre Dame.
Otherwise, things are fine — though I could see the Committee doing one other swap: moving the Wisconsin-North Dakota game to Cincinnati, and the Lowell-Colgate game to Bridgeport. This, again, would maintain decent "bracket integrity" while also helping those teams' fans more easily make the games.
Keep checking in with CHN to see what happens, and our analysis of it.