It’s time once again for what we like to call Bracketology — college hockey style. It’s a weekly look at how the NCAA tournament might look if the season ended today.
More than that, it’s a look into the thought process behind selecting and seeding the NCAA tournament teams.
This is a special installment of Bracketology, and we’ll be bringing you a new one every week, until we make our final picks just before the field is announced.
Here are the facts:
• Sixteen teams are selected to participate in the national tournament.
• There are four regional sites (East – Worcester, Massachusetts, Northeast – Amherst, Massachusetts, Midwest – Grand Rapids, Mich., West – Minneapolis, Minn.)
• A host institution which is invited to the tournament plays in the regional for which it is the host, and cannot be moved.
• Seedings will not be switched, as opposed to years past. To avoid undesirable first-round matchups, including intraconference games (see below), teams will be moved among regionals, not reseeded.
Here are the NCAA’s guidelines on the matter, per a meeting of the Championship Committee:
In setting up the tournament, the committee begins with a list of priorities to ensure a successful tournament on all fronts including competitive equity, financial success and likelihood of playoff-type atmosphere at each regional site. For the model, the following is a basic set of priorities:
• The top four teams as ranked by the committee are the four No. 1 seeds and will be placed in the bracket so that if all four teams advance to the Men’s Frozen Four, the No. 1 seed will play the No. 4 seed and the No. 2 seed will play the No. 3 seed in the semifinals.
• Host institutions that qualify will be placed at home.
• No. 1 seeds are placed as close to home as possible in order of their ranking 1-4.
• Conference matchups in first round are avoided, unless five or more teams from one conference are selected, then the integrity of the bracket will be preserved.
• Once the six automatic qualifiers and 10 at-large teams are selected, the next step is to develop four groups from the committee’s ranking of 1-16. The top four teams are the No. 1 seeds. The next four are targeted as No. 2 seeds. The next four are No. 3 seeds and the last four are No. 4 seeds. These groupings will be referred to as “bands.”
Additionally, the NCAA recently clarified its selection criteria to include a bonus factor for “good” nonconference wins, which are wins against non-league opponents in the top 15 of the Ratings Percentage Index.
Given these facts, here are the top 16 of the current PairWise Rankings (PWR), and all conference leaders, based on winning percentage (Quinnipiac, Michigan, Bemidji State, Cornell, Boston College and Denver) (through all games of Friday, March 11, 2005):
1t Colorado College
1t Denver
3 Boston College
4t Minnesota
4t Cornell
6 Michigan
7 Harvard
8 New Hampshire
9 Ohio State
10t Boston University
10t North Dakota
10t Dartmouth
13 Wisconsin
14t Maine
14t Colgate
16t Massachusetts-Lowell
16t Northern Michigan
26 Bemidji State
— Quinnipiac
Step One
From the committee’s report, choose the 16 teams in the tournament.
We break ties in the PWR by looking at the individual comparisons among the tied teams, and add all of the conference leaders, based on winning percentage.
From there, we can start looking at the bubble in a more detailed fashion.
Breaking ties in the PWR using head-to-head comparisons among the tied teams, the 16 teams in the tournament, in rank order, are:
1 Colorado College
2 Denver
3 Boston College
4 Minnesota
5 Cornell
6 Michigan
7 Harvard
8 New Hampshire
9 Ohio State
10 Boston University
11 North Dakota
12 Dartmouth
13 Wisconsin
14 Maine
15 Bemidji State
16 Quinnipiac
All ties were broken because of individual comparison wins.
So, after Friday’s games, Maine is in and Northern Michigan is out. We also have Minnesota moving back into the top band of seeds, as we predicted Monday, with its win over Minnesota State. And UNH made a bold move with two wins over TUC Northeastern. The loser Friday? Wisconsin, despite a win over UAA.
Step Two
Now it’s time to assign the seeds.
No. 1 Seeds — Colorado College, Denver, Boston College, Minnesota
No. 2 Seeds — Michigan, Cornell, Harvard, New Hampshire
No. 3 Seeds — Ohio State, Boston University, North Dakota, Dartmouth
No. 4 Seeds — Wisconsin, Maine, Bemidji State, Quinnipiac
Step Three
Place the No. 1 seeds in regionals.
We place host schools first and then place the other No. 1 seeds based on proximity to the regional sites.
No. 4 Minnesota is placed in the West Regional in Minneapolis as the host school.
No. 1 Colorado College is placed in the Midwest Regional in Grand Rapids.
No. 2 Denver is placed in the East Regional in Worcester.
No. 3 Boston College is placed in the Northeast Regional in Amherst.
Step Four
Now we place the other 12 teams so as to avoid intraconference matchups if possible.
Begin by filling in each bracket by banding groups. Remember that teams are not assigned to the regional closest to their campus sites by ranking order within the banding (unless you are a host school, in which case you must be assigned to your home regional).
If this is the case, as it was last year, then the committee should seed so that the quarterfinals are seeded such that the four regional championships are played by No. 1 v. No. 8, No. 2 v. No. 7, No. 3 v. No. 6 and No. 4 v. No. 5.
So therefore:
No. 2 Seeds
No. 5 Cornell is placed in No. 4 Minnesota’s Regional, the West.
No. 6 Michigan is placed in No. 3 Boston College’s Regional, the Northeast.
No. 7 Harvard is placed in No. 2 Denver’s Regional, the East.
No. 8 New Hampshire is placed in No. 1 Colorado College’s Regional, the Midwest.
No. 3 Seeds
Our bracketing system has one Regional containing seeds 1, 8, 9, and 16, another with 2, 7, 10, 15, another with 3, 6, 11, 14 and another with 4, 5, 12 and 13.
In this case with the No. 2 seeds being displaced, we’re trying to get the 8-9, 7-10, 6-11, and 5-12 matchups as close as possible.
Therefore:
No. 10 Boston University is place in No. 7 Harvard’s Regional, the East, as the host.
No. 9 Ohio State is placed in No. 8 New Hampshire’s Regional, the Midwest.
No. 11 North Dakota is placed in No. 6 Michigan’s Regional, the Northeast.
No. 12 Dartmouth is placed in No. 5 Cornell’s Regional, the West.
No. 4 Seeds
One more time, and this time we’re going back to taking No. 16 v. No. 1, No. 15 v. No. 2, etc.
No. 16 Quinnipiac is sent to Colorado College’s Regional, the Midwest.
No. 15 Bemidji State is sent to Denver’s Regional, the East.
No. 14 Maine is sent to Boston College’s Regional, the Northeast.
No. 13 Wisconsin is sent to Minnesota’s Regional, the West.
The brackets as we have set them up:
West Regional:
13 Wisconsin vs. 4 Minnesota
12 Dartmouth vs. 5 Cornell
Midwest Regional:
16 Quinnipiac vs. 1 Colorado College
9 Ohio State vs. 8 New Hampshire
East Regional:
15 Bemidji State vs. 2 Denver
10 Boston University vs. 7 Harvard
Northeast Regional:
14 Maine vs. 3 Boston College
11 North Dakota vs. 6 Michigan
Our first concern is avoiding intraconference matchups. We have three.
If we switch Maine and Wisconsin, we take care of two of these matchups. If we switch Dartmouth with North Dakota, we take care of the other one.
So we have our new brackets:
West Regional:
14 Maine vs. 4 Minnesota
11 North Dakota vs.5 Cornell
Midwest Regional:
16 Quinnipiac vs. 1 Colorado College
9 Ohio State vs. 8 New Hampshire
East Regional:
15 Bemidji State vs. 2 Denver
10 Boston University vs. 7 Harvard
Northeast Regional:
13 Wisconsin vs. 3 Boston College
12 Dartmouth vs. 6 Michigan
Bracketing the Frozen Four, if all four number-one seeds advance, then the top overall seed plays the No. 4 overall, and No. 2 plays No. 3. Therefore, the winners of the Midwest and West Regionals face each other in one semifinal (Colorado College and Minnesota’s brackets), while the winners of the East and Northeast Regionals (Denver and Boston College’s brackets) play the other semifinal.
But…
Bonus Time
We know there is a bonus component to the criteria, the NCAA’s tweak to the system which rewards “good” nonconference wins.
Without official word on the size of the bonuses, we take these numbers: .003 for a good road win, .002 for a good neutral win and .001 for a good home win.
Now remember, non-conference wins against conference opponents do not count as a good win. Therefore when UAA defeated Minnesota in the Nye Frontier Classic, that doesn’t count as a good win.
Our seedings are now:
1 Boston College
2 Colorado College
3 Denver
4 Minnesota
5 Michigan
6 Cornell
7 Harvard
8 New Hampshire
9 North Dakota
10 Dartmouth
11 Ohio State
12 Boston University
13 Wisconsin
14 Maine
15 Bemidji State
16 Quinnipiac
There are some slight differences in the middle of the pack, but the big one is that Boston College is the overall No. 1 seed here. It makes a substantial difference, as we’ll see later.
So, our new brackets, using bracket-filling as above, are as follows:
West Regional:
13 Wisconsin vs. 4 Minnesota
11 Ohio State vs. 6 Cornell
Midwest Regional:
15 Bemidji State vs. 2 Colorado College
9 North Dakota vs. 7 Harvard
East Regional:
16 Quinnipiac vs. 1 Boston College
12 Boston University vs. 8 New Hampshire
Northeast Regional:
14 Maine vs. 3 Denver
10 Dartmouth vs. 5 Michigan
Ugh. Too many intraconference matchups here.
Let’s start in the No. 4 band. We have to move Wisconsin. There’s only one place that the Badgers can go, and that’s to play No. 1 Boston College. BC won’t be happy about that.
Okay, let’s wait a minute, and go back to this rule:
Conference matchups in first round are avoided, unless five or more teams from one conference are selected, then the integrity of the bracket will be preserved.
I’ve hesitated to use this rule until now. I am invoking it because of the fact that the overall No. 1 seed should not play the overall No. 13 seed in the first round. So I am going to leave the fourth band alone.
Now let’s turn our attention to Boston University-New Hampshire. We can’t switch BU because the Terriers are the host. So we have to move UNH. The closest seed to UNH is Harvard, so we switch.
So our new brackets are:
West Regional:
13 Wisconsin vs. 4 Minnesota
9 Ohio State vs. 6 Cornell
Midwest Regional:
15 Bemidji State vs. 2 Colorado College
9 North Dakota vs. 8 New Hampshire
East Regional:
16 Quinnipiac vs. 1 Boston College
12 Boston University vs. 7 Harvard
Northeast Regional:
14 Maine vs. 3 Denver
10 Dartmouth vs. 5 Michigan
So there is our bracket this week. Bracket integrity is preserved, competitive equity is a little askew, but it’s okay because all the one seeds are in the proper slots, all the two seeds are in the proper spots, the three seeds are matched as well as possible with the two seeds, and the four seeds are matched exactly with the one seeds.
It’s not a perfect 1, 8, 9, 16 or 2, 7, 10, 15, but that’s within the rules.
Let’s now take a look at it another way. What if we put Boston College in Amherst as the No. 1 seed? We certainly can see if bracket integrity is better than in the previous situation, where competitive equity was good, but bracket integrity was a little askew.
If we do this, then we have these brackets:
West Regional:
13 Wisconsin vs. 4 Minnesota
11 Ohio State vs. 6 Cornell
Midwest Regional:
15 Bemidji State vs. 2 Colorado College
10 Dartmouth vs. 7 Harvard
East Regional:
14 Maine vs. 3 Denver
12 Boston University vs. 5 Michigan
Northeast Regional:
16 Quinnipiac vs. 1 Boston College
9 North Dakota vs. 8 New Hampshire
Again, we invoke the five-team rule to keep the No. 4 band intact.
Now we have to switch Dartmouth out, and that means switching Dartmouth with North Dakota.
Our new brackets:
West Regional:
13 Wisconsin vs. 4 Minnesota
11 Ohio State vs. 6 Cornell
Midwest Regional:
15 Bemidji State vs. 2 Colorado College
9 North Dakota vs. 7 Harvard
East Regional:
14 Maine vs. 3 Denver
12 Boston University vs. 5 Michigan
Northeast Regional:
16 Quinnipiac vs. 1 Boston College
10 Dartmouth vs. 8 New Hampshire
The competitive equity here is much better than that in the previous bracket, and the bracket integrity remains intact.
This is my bracket.
What if we took these numbers: .005 for a good road win, .003 for a good neutral win and .001 for a good home win.
Does anything change? Nope, so our bracket will stay the same.
What does Saturday night bring?
Can BC lose the overall No. 1 seed, assuming it beats Massachusetts?
No. Even if Denver wins, it cannot pass BC in the RPI, which is the only factor that can change in that comparison. Against CC, BC has already lost the RPI criterion, but wins TUC and COP. That cannot go against BC even if CC wins. Bottom line: if BC wins on Saturday night, it will remain the No. 1 overall seed.
Can Cornell move up to a No. 1 seed this weekend?
No. Cornell needs two additional PWR wins in order to move past Minnesota. Minnesota wins the head-to-head comparison with Cornell, even if Minnesota drops two games to Minnesota State and Cornell wins on Saturday. So Cornell needs to turn two of three comparisons with BC/CC/Denver.
The Big Red can turn the Denver comparison, since it hinges on RPI. But Cornell won’t catch CC on RPI, so that comparison cannot be turned. And the gap in RPI with BC is too large as well. The Big Red will have to wait a week to try to gain a No. 1 seed.
Is UML really out if it loses to Maine on Saturday night?
Yes, for sure. In order for UML to move up, it would have to win some additional comparisons. Without an opportunity to play, can the River Hawks watch as comparisons turn? Most likely not, but let’s look at the numbers — specifically, at comparisons with the teams closest to the River Hawks, ones they can hope to win back.
Michigan State: The Spartans can only lose three more this year, meaning that their TUC record will be, at worst, 9-12-3, since losing only once to Miami will leave Miami out of being a TUC. If UML loses to Maine, it’s TUC record will be 7-10-4, which would not be btter than MSU. MSU’s RPI will be higher than UML’s too. So UML can not change that comparison.
Maine: No chance to turn the comparison.
Wisconsin: Same scenario as Michigan State.
You can continue up the line and look at the comparisons, but if UML loses to Maine, it can’t turn any more comparisons its way.
We’ll be back tomorrow to see what Saturday’s results have done.
