Understanding The Process
by Adam Wodon/Managing Editor
Over 100 letters have already passed through the desk of St. Lawrence coach Joe Marsh. And in various ways, they all say the same thing: how lousy a job Marsh did as chairman of the NCAA Division I men's ice hockey tournament committee.
Many have come from Minnesota, and most haven't been quite that diplomatic.
"I'm pretty much guaranteed I can't run for governor of Minnesota," Marsh said. "I've got a stack of stuff that [USA Today's Jon] Barkan sent me from the Internet. I've been up to my ears."
Gopher fans — and even their head coach, Doug Woog — are upset because Minnesota, despite being fourth overall in the criteria that ranks the teams, is set up to play No. 1 Michigan in the second round.
Minnesotans are crying foul; some go so far as to claim conspiracy, or at least an anti-Minnesota bias.
None of it comes close to the truth.
Minnesota is in its current position for one simple reason: the overriding philosophy that tells the committee to avoid first- and second-round matchups against teams from the same conference. This mandate was passed down from the coaches, after discussions over the years at the American Hockey Coaches Association conventions.
By the same token, Cornell fans are up in arms: "How can we be out West?" they say. Or, "We're a sixth seed? We get no respect." But again, the process was followed as an exact science, and Cornell's seeding has nothing to do with respect.
In the case of Minnesota, and everyone else, it comes down to simple numbers. The mistake that many make is in believing that the seedings are a wholly subjective process, or in being concerned too much with what number seed a team gets. That number is often irrelevant, because of the flip-flopping necessary to avoid those intra-conference matchups.
That philosophy was instituted after coaches came to the conclusion that it was unfair to bye teams. For example, why should No. 2 North Dakota, a team that just defeated Minnesota in St. Paul to earn the bye, have to play a conference rival again to make the Final Four? The same philosophy holds no matter which teams are involved.
There's also been a misconception, even among media and fans trying to be knowledgeable, about the exact method used by the committee.
"We had our team doctor telling me how it works," Marsh said. "I said, 'Doc, with all due respect, I'm on the committee. I've been on it for four years.' So these are our own people that think I'm an idiot. It's so frustrating."
One thing that upset Gophers seem to be forgetting is that the team they are playing, Michigan State, is the lowest-seeded team in the tournament. The fact that the Spartans are team with a history doesn't mean anything. By the numbers, Minnesota is getting the easiest first-round draw. Perhaps it can be argued that the Spartans are actually better than other teams, but that's just an opinion, and opinion doesn't come into play for the committee. Not one bit.
"I'm trying to explain that to Doug," said Marsh. "They're a traditionally strong team — they're Michigan State. [But] what if we threw Wisconsin in there, and, '[Whoa] it's Wisconsin.' We've got to just take the numbers.
"I mean, I think it's safe to say Michigan's probably the best team; you don't have to go out on a limb there. But at the same time, if we're going to protect not having the conference matchups, that's the way we've got to do it. That was what coaches expressed they wanted. That's the reason Minnesota is there. The only two things that are subjective (are) that, and the protection of the gate."
To get a better idea of how the process works, let's break down this year's draw:
When the committee gets into the room, they are presented with printouts showing every team compared to every other team based on five criteria:
Ratings Percentage Index (RPI) Record vs. Teams Under Consideration Record in the last 20 games Record vs. Common Opponents Head-to-Head record between the two teams
The RPI is a weighted winning percentage that helps factor in strength of schedule. It's comprised of 35 percent won-loss percentage, 50 percent opponents' won-loss percentage, and 15 percent opponents' opponents won-loss percentage. A "Team Under Consideration" is anyone .500 or better.
So the first step is to determine the 12 teams in the tournament. Five were automatics due to winning conference tournaments, regular-season championships, or both: Boston University, Clarkson, Michigan, North Dakota and Cornell.
The other teams are determined by looking at the Teams Under Consideration, and comparing them to each other using the criteria outlined above. Schools like Minnesota, Miami, New Hampshire and Vermont were obvious choices once the committee could easily see those teams were beating all the rest under consideration.
The committee then took all the "bubble teams," like Michigan State, St. Cloud, Denver, RPI, Princeton and Colorado College, and compared them all to each other. In doing so, Michigan State, Colorado College and Denver rose above the rest.
So, with the field of 12 set, the teams broke down like this:
Colorado College Boston University
Michigan New Hampshire
Michigan State Vermont
Next comes the process of seeding the teams. By virtue of winning the conference and regular-season title, Michigan, North Dakota and Boston University received automatic byes. The coaches voted to implement this rule a couple of years ago.
The final bye went to Clarkson, which clearly beat out the remaining teams in the East. The committee won't give a team which is geographically outside a region a bye unless there is an overwhelmingly compelling reason do so, and in this case, there wasn't.
Now, however, you have uneven regions, because seven Western teams qualified. Therefore, the lowest-seeded team in the West — in this case Michigan State — was moved East to make it six-and-six.
Policy then dictates that two non-bye teams from each region be switched to the other. The committee first attempts to take the bottom two from each region, but if that would negatively impact attendance in a severe way, or create intra-conference matchups, different teams can be moved. This occurred last year, when Mass-Lowell was moved West.
In this case, moving the bottom two worked out perfectly, especially since doing so sent Michigan State back West, where the committee was going to ensure it was anyway.
"It just worked out that way, but also it's sort of serendipitous in that you would want Michigan State in the West anyway to protect the gate [attendance in Grand Rapids, Mich., site of the West Regional]," Marsh said.
There was some debate in the East, however. Some members of the committee thought it was a no-brainer to keep Vermont and New Hampshire in the East, no matter the numbers, because of their huge draws. But Marsh didn't want to be too hasty. He said he owed it to Cornell to check the numbers, because the Big Red bring a large contingent in their own right.
"I wasn't comfortable as the chairman to say, 'We'll just do it carte blanche,' because they're the two biggest draws," Marsh said. "I said I really would rather us go with the numbers, because Cornell has a great following, very loyal alums, a lot of alums that will go [to the tournament games].
"I said if it was another team that doesn't have the following, then maybe, but I thought Cornell deserved the benefit of the doubt in terms of their ability to draw. So I said, 'Let's look at the numbers.' And we really spent a lot of time with this."
Cornell edged Vermont in the criteria, while Vermont defeated New Hampshire and UNH defeated Cornell. So with the teams even in that sense, the RPI was used to break the tie, and Cornell was the odd team out.
Another conspiracy theory foiled.
In the case of CC and Denver, they were the bottom two teams [after moving MSU], they weren't going to be big draws in Grand Rapids, and moving them avoided a number of possible intra-conference scenarios. It was a no-brainer in every way.
So, taking into account the switch in regions, and then comparing each team to every other, we get these seedings in each region:
1. Michigan 1. Clarkson
2. North Dakota 2. Boston University
3. Minnesota 3. Vermont
4. Miami 4. New Hampshire
5. Cornell 5. Colorado College
6. Michigan St. 6. Denver
Here is where some people got confused, especially those who have become familiar with U.S. College Hockey Online's Pairwise Rankings. Some of the above is a slight contradiction to the straight numbers presented by the PWR. While some believed that the committee would use the Pairwise Rankings exactly in determining the seedings (even if it was called something else), that's not entirely true.
For example, the Pairwise Rankings take every team and compare them to every other team, using the five criteria up above, and count how many "wins" a team gets. In this case, that would have placed New Hampshire ahead of Vermont, Denver ahead of CC and Cornell ahead of Miami.
But this isn't exactly what the committee does. While, in essence, it is totalling up wins to see which teams are grouped near which other teams, when it comes down to seeding two (or more) teams, the head-to-head comparison is used. In each of the above three cases, the team with fewer overall Pairwise wins won the individual comparison.
"When you get down to where you're taking one team over the other, it's that team versus the other team," Marsh said.
"The critical factor here is, it's not that alone. If you just look at the Pairwise Ranking and say UNH is ahead of Vermont and that's it — well, you haven't gone far enough. You have to take New Hampshire versus Vermont individually, five criteria, plug it in and the winner goes.
"The Pairwise thing is fine, it's accurate, but it's not being applied fully. You have to take just those two teams."
Those pesky intra-conference matchups
That cleared up, then comes the fun part, a source of great controversy for many. Once the teams are in the proper regions, now they have to be seeded in a way that avoids intra-conference games at the Regionals as much as possible.
The East is set — no conflicts there. In the West, however, the potential second-round matchup between North Dakota and Minnesota was staring the committee in the face.
It should be noted that the possibility of just one conflict is pretty lucky. In the past, a lot more juggling was needed to avoid those matchups. It should also be noted that the policy of avoiding intra-conference matchups is NOT written in the championships manual, but it is something that is discussed at the coaches' convention, and the consensus has been to try and avoid then.
"We've kicked that around a lot," Marsh said. "I don't remember it being any kind of an edict. We talked about how we better formalize this sort of thing. But it was definitely discussed on the floor of the coaches' convention, and the coaches are the ones that gave us this input."
That being clear, the committee's first order of business was to flip-flop No. 3 Minnesota with No. 4 Miami. The result, which would force Minnesota to face No. 1 Michigan in the second round, may be unfortunate, but not part of a scheme to get the Gophers.
"They're deeming that they're being punished. [But] they're playing the lowest-seeded tournament team," Marsh said.
Nevertheless, even rational fans will complain about the absurdity of trying so hard to avoid those intra-conference matchups, accusing the committee of being clueless, but not realizing that the mandate to avoid those matchups comes from the coaches themselves. The committee, in essence, is following orders.
"The coaches know this," Marsh said.
The idea behind avoiding intra-conference matchups is a protection to the bye team. Why should North Dakota, the argument goes, have to play Minnesota again, a team it just beat out in the regular season, and defeated practically on the road, in overtime, to take the conference tournament?
Some say the margin between the Sioux and Minnesota was so small, that it shouldn't hurt Minnesota so badly. It's an irrelevant point, however. The other side can say that Minnesota is the only non-bye team in the region not to gain at least a tie against Michigan this season (losing in overtime in their only meeting), but that's an equally irrelevant point.
The philosophy of avoiding intra-conference games applies no matter which schools are involved. If that's a policy that a lot of fans don't agree with, or even a lot of coaches don't agree with, then they can get together and decide to change it. But for now, the committee is following established guidelines long known to be used.
"I talked to [an upset Woog on Tuesday]," said Marsh. "That's how we explained it. He would've liked to have gone to the East, but the numbers don't suggest that. We can't just arbitrarily say, 'Who would you rather play, Doug?'
"He's saying, 'CC's got a much easier draw.' Well, the numbers don't say that at all."
So, how much consideration was given to moving Minnesota to the East?
"None," said Marsh.
"Why couldn't you say, well, New Hampshire, they've gotta play CC, wouldn't they rather play Michigan State? I don't know. We're just doing it by the numbers and the criteria, and boy, it seemed to be cut and dried.
"We're just using the numbers next to the school's name, not the school's name."
Back to business
Getting back to the regional pairings, after flip-flopping Miami and Minnesota, you're left with a Michigan State-Miami first-round matchup. That's even worse than the previous situation.
To avoid that, the committee simply did another flip-flop, switching No. 6 Michigan State with No. 5 Cornell. Though it wasn't necessarily intentional, that restored the original Minnesota-Michigan State matchup and produced the following final draw:
1. Michigan 1. Clarkson
2. North Dakota 2. Boston University
3. Miami 3. Vermont
4. Minnesota 4. New Hampshire
5. Michigan State 5. Colorado College
6. Cornell 6. Denver
Of course, Cornell fans then began to complain, charging the committee with a lack of respect for the team. "A sixth seed?" they said.
But, now, with the process laid out, one can clearly see how Cornell came to be a sixth seed, despite being ranked better than four other teams in the tournament on the Pairwise Rankings. In the end, the number six is somewhat irrelevant, and should not be considered a slam at the Red. It made no difference whether Cornell was seeded fifth or sixth; it would have played the same team.
Cornell was moved West because of losing the tiebreaker with New Hampshire and Vermont; it was given a five seed originally because it lost the head-to-head comparison with Miami; and it was dropped to sixth so Michigan State wouldn't face Miami in the first round.
It's that simple. No conspiracy, no lack of respect, no witchcraft, no doubt.
"It just fell into place," said Marsh, the longtime coach at St. Lawrence.
"Did I have any preconceived or hidden agendas? Geez, I'd just like to be playing in [the tournament] once in a while. And I'm talking to people just pissed off that they're there, and I'd love to play. We're a million miles away from playing."