Where does your team stand with less than one month left in regular season?

27 Jan 17: The Saint Cloud State University Huskies play against the Bemidji State University Beavers in a quarterfinal game of the North Star College Cup at the Xcel Energy Center in St. Paul, MN. (Jim Rosvold)
Where would Bemidji State and St. Cloud State start the NCAA tournament if the event began today? (photo: Jim Rosvold)

We’re at that time of the year where one thing is on everyone’s minds.

Will my team make the NCAA tournament? Where does it sit in the PairWise Rankings (PWR)?

Those of you that are veterans of the college hockey scene know that it is all about the PairWise Rankings. This is USCHO’s numerical approach that simulates the way the NCAA Division I men’s ice hockey committee chooses the teams that make the NCAA tournament.

Since USCHO began the PairWise Rankings, we have correctly identified all of the teams that have been selected to the NCAA tournament.

Five of the last six years, I am the only prognosticator to have correctly predicted the exact brackets for the NCAA tournament, meaning that I have predicted how the committee thought when putting together the brackets.

With that in mind, it’s time once again to do what we like to call Bracketology, college hockey style. It’s our weekly look at how I believe the NCAA tournament might look like come selection time, using what we know now.

It’s a look into the possible thought processes behind selecting and seeding the NCAA tournament teams.

This is not a be-all, end-all analysis of the bracket. I am trying to give you, the reader, an idea of what the committee might be thinking and not exactly what they are thinking.

This is the next installment of Bracketology for 2017, and we’ll be bringing you a new one every week until we make our final picks before the field is announced on March 19.

If you want to skip the inner workings and get to the results of the analysis, then click here.

Here are the facts:

• Sixteen teams are selected to participate in the national tournament.

• There are four regional sites (East – Providence, R.I.; Northeast – Manchester, N.H.; Midwest – Cincinnati, Ohio; West – Fargo, N.D.).

• A host institution that is invited to the tournament plays in the regional for which it is the host and cannot be moved. The host institutions this year: Brown in Providence, New Hampshire in Manchester, Miami in Cincinnati and North Dakota in Fargo.
• Seedings will not be switched. To avoid undesirable first-round matchups, including intra-conference games (see below), teams will be moved among regionals, not reseeded.

Here are the NCAA’s guidelines on the matter, from the 2015 pre-championship manual:

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 the likelihood of a playoff-type atmosphere at each regional site. For this model, the following is a basic set of priorities:

1. Once the six automatic qualifiers and 10 at-large teams are selected, the next step is to develop four groups from the committee’s rankings of 1-16. The top four teams are 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. The next four are targeted as No. 2 seeds. The next four are No. 3 seeds and the last four are No. 4 seeds.

2. Step two is to place the home teams. Host institutions that qualify will be placed at home.

3. Step three is to fill in the bracket so that first-round conference matchups are avoided, unless it corrupts the integrity of the bracket. If five or more teams from one conference are selected to the championship, then the integrity of the bracket will be protected (i.e., maintaining the pairing process according to seed will take priority over avoidance of first-round conference matchups). To complete each regional, the committee assigns one team from each of the remaining seeded groups so there is a No. 1, No. 2, No. 3 and No. 4 seed at each regional site.

Given these facts, here is the top 16 of the current PairWise Rankings (PWR), and the conference leaders through all games of February 14:

1 Minnesota-Duluth
2 Denver
3 Harvard
4 Minnesota
5 Boston University
6 Penn State
7 Western Michigan
8t Providence
8t Massachusetts-Lowell
10 North Dakota
11 Union
12 St. Cloud State
13t Cornell
13t Boston College
15 Ohio State
16 Vermont
20 Air Force
24 Bemidji State

Current conference leaders based on winning percentage:

Atlantic Hockey: Air Force
Big Ten: Minnesota
ECAC Hockey: Union
Hockey East: Boston College
NCHC: Denver
WCHA: Bemidji State

Notes

• Bracketology assumes that the season has ended and there are no more games to be played — i.e., the NCAA tournament starts tomorrow.

• Because there are an uneven amount of games played inside each conference, I will be using winning percentage, not points accumulated, to determine the current leader in each conference. This team is my assumed conference tournament champion after applying the tiebreakers.

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 in any current league leaders that are not currently in the top 16. The only teams that are not are Air Force and Bemidji State.

From there, we can start looking at the ties and bubbles in a more detailed fashion.

The ties and bubbles consist of none this week.

We break all of our ties based upon the RPI.

Therefore, the 16 teams in the tournament, in rank order, are:

1 Minnesota-Duluth
2 Denver
3 Harvard
4 Minnesota
5 Boston University
6 Penn State
7 Western Michigan
8 Providence
9 Massachusetts-Lowell
10 North Dakota
11 Union
12 St. Cloud State
13 Cornell
14 Boston College
15 Air Force
16 Bemidji State

Step two

Now it’s time to assign the seeds.

No. 1 seeds: Minnesota-Duluth, Denver, Harvard, Minnesota

No. 2 seeds: Boston University, Penn State, Western Michigan, Providence

No. 3 seeds: Massachusetts-Lowell, North Dakota, Union, St. Cloud State

No. 4 seeds: Cornell, Boston College, Air Force, Bemidji State

Step three

Place the No. 1 seeds in regionals.

No. 1 Minnesota-Duluth is placed in the West Regional in Fargo
No. 2 Denver is placed in the Midwest Regional in Cincinnati.
No. 3 Harvard is placed in the East Regional in Providence.
No. 4 Minnesota is placed in the Northeast Regional in Manchester.

Step four

Now we place the other 12 teams so as to avoid intra-conference 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 would be played by No. 1 vs. No. 8, No. 2 vs. No. 7, No. 3 vs. No. 6 and No. 4 vs. No. 5.

So therefore:

No. 2 seeds

No. 8 Providence is placed in No. 1 Minnesota-Duluth’s regional, the West Regional.
No. 7 Western Michigan is placed in No. 2 Denver’s regional, the Midwest Regional.
No. 6 Penn State is placed in No. 3 Harvard’s regional, the East Regional.
No. 5 Boston University is placed in No. 4 Minnesota’s regional, the Northeast Regional.

No. 3 seeds

Our bracketing system has one regional containing seeds 1, 8, 9, and 16; another with 2, 7, 10 and 15; another with 3, 6, 11 and 14; and another with 4, 5, 12 and 13.

North Dakota is a host, therefore they are placed first in this pod:

No. 10 North Dakota is placed in No. 8 Providence’s regional, the West Regional.
No. 9 Massachusetts-Lowell is placed in No. 7 Western Michigan’s regional, the Midwest Regional.
No. 11 Union is placed in No. 6 Penn State’s regional, the East Regional.
No. 12 St. Cloud State is placed in No. 5 Boston University’s regional, the Northeast Regional.

No. 4 seeds

One more time, taking No. 16 vs. No. 1, No. 15 vs. No. 2, etc.

No. 16 Bemidji State is sent to No. 1 Minnesota-Duluth’s regional, the West Regional.
No. 15 Air Force is sent to No. 2 Denver’s regional, the Midwest Regional.
No. 14 Boston College is sent to No. 3 Harvard’s regional, the East Regional.
No. 13 Cornell is sent to No. 4 Minnesota’s regional, the Northeast Regional.

The brackets as we have set them up:

East Regional (Providence):
14 Boston College vs. 3 Harvard
11 Union vs. 6 Penn State

Northeast Regional (Manchester):
13 Cornell vs. 4 Minnesota
12 St. Cloud State vs. 5 Boston University

Midwest Regional (Cincinnati):
15 Air Force vs. 2 Denver
9 Massachusetts-Lowell vs. 7 Western Michigan

West Regional (Fargo):
16 Bemidji State vs. 1 Minnesota-Duluth
10 North Dakota vs. 8 Providence

Our first concern is avoiding intraconference matchups. We have none this week.

Now let’s look at maximizing attendance.

I can see on swap that we can make here. Providence with Penn State.

East Regional (Providence):
14 Boston College vs. 3 Harvard
11 Union vs. 8 Providence

Northeast Regional (Manchester):
13 Cornell vs. 4 Minnesota
12 St. Cloud State vs. 5 Boston University

Midwest Regional (Cincinnati):
15 Air Force vs. 2 Denver
9 Massachusetts-Lowell vs. 7 Western Michigan

West Regional (Fargo):
16 Bemidji State vs. 1 Minnesota-Duluth
10 North Dakota vs. 6 Penn State

And I think we have it.

Can we approach the bracket another way?

There is always that school of thought that you should not put the number one overall seed in the same regional as a host school. We are of course referring to the fact that UMD is currently in the same regional as North Dakota.

Let’s change it up a bit then and see what happens.

Here’s how we place the number one seeds:

UMD in Cincinnati, Denver in Manchester, Harvard in Providence and Minnesota in Fargo.

That means our bracket falls out like this:

East Regional (Providence):
14 Boston College vs. 3 Harvard
12 St. Cloud State vs. 6 Penn State

Northeast Regional (Manchester):
15 Air Force vs. 2 Denver
11 Union vs. 7 Western Michigan

Midwest Regional (Cincinnati):
16 Bemidji State vs. 1 Minnesota-Duluth
9 Massachusetts-Lowell vs. 8 Providence

West Regional (Fargo):
13 Cornell vs. 4 Minnesota
10 North Dakota vs. 5 Boston University

We solve Massachusetts-Lowell vs. Providence. We want Providence in Providence so we move it instead of Massachusetts-Lowell.

East Regional (Providence):
14 Boston College vs. 3 Harvard
12 St. Cloud State vs. 8 Providence

Northeast Regional (Manchester):
15 Air Force vs. 2 Denver
11 Union vs. 7 Western Michigan

Midwest Regional (Cincinnati):
16 Bemidji State vs. 1 Minnesota-Duluth
9 Massachusetts-Lowell vs. 6 Penn State

West Regional (Fargo):
13 Cornell vs. 4 Minnesota
10 North Dakota vs. 5 Boston University

Now attendance items. We swap BU and Western Michigan.

East Regional (Providence):
14 Boston College vs. 3 Harvard
12 St. Cloud State vs. 8 Providence

Northeast Regional (Manchester):
15 Air Force vs. 2 Denver
11 Union vs. 5 Boston University

Midwest Regional (Cincinnati):
16 Bemidji State vs. 1 Minnesota-Duluth
9 Massachusetts-Lowell vs. 6 Penn State

West Regional (Fargo):
13 Cornell vs. 4 Minnesota
10 North Dakota vs. 7 Western Michigan

Is there anything else?

Not that I can see.

Now choose a bracket.

I still have to go with the original one based solely on attendance issues and go away from the theory on the second one.

See you here in a few weeks for the next Bracketology.

Here’s a summary of everything that we have covered.

This week’s brackets

East Regional (Providence):
14 Boston College vs. 3 Harvard
11 Union vs. 8 Providence

Northeast Regional (Manchester):
13 Cornell vs. 4 Minnesota
12 St. Cloud State vs. 5 Boston University

Midwest Regional (Cincinnati):
15 Air Force vs. 2 Denver
9 Massachusetts-Lowell vs. 7 Western Michigan

West Regional (Fargo):
16 Bemidji State vs. 1 Minnesota-Duluth
10 North Dakota vs. 6 Penn State

Conference breakdowns

NCHC — 5
Hockey East — 4
ECAC Hockey — 3
Big Ten — 2
WCHA — 1
Atlantic Hockey – 1

Movement

In: St. Cloud State
Out: Vermont

Last Week’s Bracket

East Regional (Providence):
14 Vermont vs. 4 Minnesota
10 Boston College vs. 7 Union

Northeast Regional (Manchester):
13 Cornell vs. 3 Boston University
12 Massachusetts-Lowell vs. 5 Harvard

Midwest Regional (Cincinnati):
15 Air Force vs. 2 Denver
11 Penn State vs. 6 Western Michigan

West Regional (Fargo):
16 Bemidji State vs. 1 Minnesota-Duluth
9 North Dakota vs. 8 Providence

149 COMMENTS

    • Not sure how Jayson missed that. Additionally, he keeps flipping teams into Fargo “to improve attendance” as if the regional isn’t going to be a sell out with UND there. The committee may very well do that, but it’s pointless. Focus on the other regionals if you’re looking at attendance-Fargo will be more than fine no matter who the other 3 teams are.

    • Not sure how Jayson missed that. Additionally, he keeps flipping teams into Fargo “to improve attendance” as if the regional isn’t going to be a sell out with UND there. The committee may very well do that, but it’s pointless. Focus on the other regionals if you’re looking at attendance-Fargo will be more than fine no matter who the other 3 teams are.

  1. I figured St Cloud would eventually get in, giving the NCHC 5 teams. Sure enough, they finally got a sweep and it happened. They have improved greatly since the first half of the season, especially in goal and on defense. Omaha is at 17 and if they put together a sweep soon they could sneak in too.
    Penn State being in Fargo instead of Cincy is a bad idea. I know the Nittany Lions are bonkers about hockey in Happy Valley, but Cincy isn’t that far away so they could possibly improve attendance there.

    • Take a look at Jim Dahl’s new feature at his web site collegehockeyranked.com called “Pairwise By Wins”

      It’s pretty cool but this tool only goes through the end of the regular season and the post season tourney is going to be important for cliffhanging teams like Omaha, who are probably going to need a win or two or three in their respective conference tourneys to get to the NCAA’s, and that is IF such teams put themselves in a good spot between now and regular season’s end.

      For example, in the Omaha case that you cite, they are probably going to have to win 2 of their final 4 regular season games to be “well positioned” to make the NCAA tourney on the back of whatever they do in the conference tourney. Those 4 games are against the teams currently ranked 2 and 10 in Pairwise.

      2 more regular season wins gives them a best possible Pairwise finish of 12th and a worst of 21st. 3 wins would be better and gives them a best possible of 9th and a worst of 18th. Bottom line is, those conference tourney games are going to matter in their case.

      They are a unique case among the cliff dwelling teams because of their SOS due to the conference they play in, who is left on their current regular season schedule, as well as which one of the three likely opponents they are going to face in their conference tourney. Those being one the teams that are currently ranked 1, 2, or 7 in Pairwise.

  2. I figured St Cloud would eventually get in, giving the NCHC 5 teams. Sure enough, they finally got a sweep and it happened. They have improved greatly since the first half of the season, especially in goal and on defense. Omaha is at 17 and if they put together a sweep soon they could sneak in too.
    Penn State being in Fargo instead of Cincy is a bad idea. I know the Nittany Lions are bonkers about hockey in Happy Valley, but Cincy isn’t that far away so they could possibly improve attendance there.

    • Take a look at Jim Dahl’s new feature at his web site collegehockeyranked.com called “Pairwise By Wins”

      It’s pretty cool but this tool only goes through the end of the regular season and the post season tourney is going to be important for cliffhanging teams like Omaha, who are probably going to need a win or two or three in their respective conference tourneys to get to the NCAA’s, and that is IF such teams put themselves in a good spot between now and regular season’s end.

      For example, in the Omaha case that you cite, they are probably going to have to win 2 of their final 4 regular season games to be “well positioned” to make the NCAA tourney on the back of whatever they do in the conference tourney. Those 4 games are against the teams currently ranked 2 and 10 in Pairwise.

      2 more regular season wins gives them a best possible Pairwise finish of 12th and a worst of 21st. 3 wins would be better and gives them a best possible of 9th and a worst of 18th. Bottom line is, those conference tourney games are going to matter in their case.

      They are a unique case among the cliff dwelling teams because of their SOS due to the conference they play in, who is left on their current regular season schedule, as well as which one of the three likely opponents they are going to face in their conference tourney. Those being one the teams that are currently ranked 1, 2, or 7 in Pairwise.

  3. Even though I would not be able to get a ticket in Fargo, I would prefer the second scenario. And I would bet a lot of the Fighting Sioux/Hawks fans would too. It would be a very interesting bracket.

  4. Even though I would not be able to get a ticket in Fargo, I would prefer the second scenario. And I would bet a lot of the Fighting Sioux/Hawks fans would too. It would be a very interesting bracket.

  5. Sick and tired of seeing teams hovering around .500 ( St. Cloud ) getting in the tournament due to a strong schedule. Strength of schedule is important, BUT you need to win those games at a reasonable clip. The computer should get the first 10 teams and let a committee pick the final 6. Why have any polls with humans involved if the Pairwise is the ONLY tool to get the NCAA teams into the tourney..

    • Teams that low in win pct shouldn’t be in the tourney, but I also think it’s silly that the 30th best team, but best in their conference deserves a spot over the 15th, 16th best.

      • I agree with Bob and Ray. The problem is that it is so hard to plot SOS vs overall win percentage. Both are important, I have no idea how fair things are in CH right now..??

    • Are you kidding me — strength of schedule is the primary statistic of importance. Win or lose — if your teams plays good teams, they should be in the tournament. Don’t be surprised when NCHC feel good team Colorado College crashes the party, and gives a whole new reason to yet another fan base to travel well.

    • How do you propose to fix it then? Remove the autobids and then just take the top 16 overall? St Cloud has played the second hardest schedule in the country to this point. They’ve also beaten Bemidji State, who, based on Jayson’s projection, would be the autobid from the WCHA. I had to go look at St Cloud’s schedule to see this. It appears that by seasons end they will have played Duluth and Denver a total of 9 times, plus more if they end up paired against each other in the playoffs. So literally, 1/3 of their schedule could be against the #1 and #2 ranked teams in the country.

    • Ouch, “let a committee pick the final 6”. What could go wrong there? How about teams 11-16 being picked from ECAC and HE? My suggestion is to do away with “committees” altogether. Maybe teams should schedule tougher OOC schedule and not complain about teams hovering around .500 getting in the tournament. At least the Pairwise is unbiased, though far from perfect.

      • They don’t even have to schedule tougher OOC, they just have to win all of the games they schedule against the cupcake teams.

    • There is no other way to do it. If you play the best teams, you are going to be rewarded for doing so. When teams like Penn State stack their schedule against teams who will not make the tournament, it shouldn’t propel them in just because they beat low life teams. Saint Cloud should be rewarded for beating good teams like they have.

    • I think that is more of an argument to get rid of polls than it is to go back to selection committees. You are basically arguing to not have standard criteria to judge teams by. That’s all the Pairwise is.

      • Agreed.

        The polls are utterly ridiculous since they have zero meaning in college hockey. A ridiculous beauty contest that serves no purpose.

        Pairwise IS the poll in my book.

        • The polls have more meaning in the first half because the pairwise doesn’t really have enough comparisons until about New Years. But I get your point.

  6. Sick and tired of seeing teams hovering around .500 ( St. Cloud ) getting in the tournament due to a strong schedule. Strength of schedule is important, BUT you need to win those games at a reasonable clip. The computer should get the first 10 teams and let a committee pick the final 6. Why have any polls with humans involved if the Pairwise is the ONLY tool to get the NCAA teams into the tourney..

    • Teams that low in win pct shouldn’t be in the tourney, but I also think it’s silly that the 30th best team, but best in their conference deserves a spot over the 15th, 16th best.

      • I agree with Bob and Ray. The problem is that it is so hard to plot SOS vs overall win percentage. Both are important, I have no idea how fair things are in CH right now..??

    • Are you kidding me — strength of schedule is the primary statistic of importance. Win or lose — if your teams plays good teams, they should be in the tournament. Don’t be surprised when NCHC feel good team Colorado College crashes the party, and gives a whole new reason to yet another fan base to travel well.

    • How do you propose to fix it then? Remove the autobids and then just take the top 16 overall? St Cloud has played the second hardest schedule in the country to this point. They’ve also beaten Bemidji State, who, based on Jayson’s projection, would be the autobid from the WCHA. I had to go look at St Cloud’s schedule to see this. It appears that by seasons end they will have played Duluth and Denver a total of 9 times, plus more if they end up paired against each other in the playoffs. So literally, 1/3 of their schedule could be against the #1 and #2 ranked teams in the country.

    • Ouch, “let a committee pick the final 6”. What could go wrong there? How about teams 11-16 being picked from ECAC and HE? My suggestion is to do away with “committees” altogether. Maybe teams should schedule tougher OOC schedule and not complain about teams hovering around .500 getting in the tournament. At least the Pairwise is unbiased, though far from perfect.

      • They don’t even have to schedule tougher OOC, they just have to win all of the games they schedule against the cupcake teams.

    • There is no other way to do it. If you play the best teams, you are going to be rewarded for doing so. When teams like Penn State stack their schedule against teams who will not make the tournament, it shouldn’t propel them in just because they beat low life teams. Saint Cloud should be rewarded for beating good teams like they have.

    • I think that is more of an argument to get rid of polls than it is to go back to selection committees. You are basically arguing to not have standard criteria to judge teams by. That’s all the Pairwise is.

      • Agreed.

        The polls are utterly ridiculous since they have zero meaning in college hockey. A ridiculous beauty contest that serves no purpose.

        Pairwise IS the poll in my book.

        • The polls have more meaning in the first half because the pairwise doesn’t really have enough comparisons until about New Years. But I get your point.

    • No — position for quality shot metric analyst. Wisco & I are moving forward on that, but based on his input, I think we to include Vinnie Vega as well.

    • No — position for quality shot metric analyst. Wisco & I are moving forward on that, but based on his input, I think we to include Vinnie Vega as well.

LEAVE A REPLY

Please enter your comment!
Please enter your name here