If the season ended today, 1999 February 9

© 1999, Joe Schlobotnik (archives)

URL for this frameset: http://slack.net/~whelan/tbrw/tbrw.cgi?1999/pairwise.990209.shtml

Remember, you can go through this process interactively using the "You Are The Committee" script.

Back for the 1999 season, this is an occasional look at how we might expect the NCAA Tournament selection procedure to play itself out given the current records of the teams involved. Using the results from the Division I Composite Schedule on US College Hockey Online after the conclusion of the Beanpot Tournament, we've calculate the various selection criteria and performed pairwise comparisons among the 25 tournament-eligible teams with overall records at or above .500, which are known as teams under consideration, with the following results:

    Team         PWR  RPI                  Comparisons Won
 1 North Dakota   24 .648   MeNHMSQnBCCCCkPnDUOSCtMiNtSLNMRPNiPvCgMLFSCrAAHC
 2 Maine          23 .627 __  NHMSQnBCCCCkPnDUOSCtMiNtSLNMRPNiPvCgMLFSCrAAHC
 3 New Hampshire  22 .624 ____  MSQnBCCCCkPnDUOSCtMiNtSLNMRPNiPvCgMLFSCrAAHC
 4 Mich State     21 .596 ______  QnBCCCCkPnDUOSCtMiNtSLNMRPNiPvCgMLFSCrAAHC
 5 Quinnipiac     20 .556 ________  BCCCCkPnDUOSCtMiNtSLNMRPNiPvCgMLFSCrAAHC
 6 Boston Coll    19 .575 __________  CCCkPnDUOSCtMiNtSLNMRPNiPvCgMLFSCrAAHC
 7 CO College     17 .593 ____________  Ck__DUOSCtMiNtSLNMRPNiPvCgMLFSCrAAHC
 8 Clarkson       17 .569 ______________  PnDUOSCtMiNtSLNMRPNiPvCgMLFSCrAAHC
 9 Princeton      16 .550 ____________CC__  DUOSCtMi__SLNMRPNiPvCgMLFSCrAAHC
10 Denver U       13 .546 __________________  OSCt____SLNMRPNiPvCgMLFSCrAAHC
11 Ohio State     13 .543 ____________________  CtMiNt__NMRPNiPvCgMLFSCrAAHC
12 Connecticut    12 .545 ______________________  MiNt__NMRPNiPvCgMLFSCrAAHC
13 Michigan       12 .564 __________________DU____  NtSLNMRP__PvCgMLFSCrAAHC
14 Notre Dame     12 .562 ________________PnDU______  SLNMRP__PvCgMLFSCrAAHC
15 St Lawrence    12 .546 ____________________OSCt____  NMRPNiPvCgMLFSCrAAHC
16 Northern Mich   8 .538 ______________________________  RP__PvCgMLFSCrAAHC
17 RPI             8 .531 ________________________________  NiPvCgMLFSCrAAHC
18 Niagara         8 .487 ________________________MiNt__NM__  __Cg__FSCrAAHC
19 Providence      7 .522 __________________________________Ni  CgMLFSCrAAHC
20 Colgate         5 .521 ______________________________________  MLFSCrAAHC
21 Mass-Lowell     5 .514 __________________________________Ni____  FSCrAAHC
22 Ferris State    3 .498 __________________________________________  CrAAHC
23 Cornell         2 .506 ____________________________________________  AAHC
24 AK-Anchorage    1 .461 ______________________________________________  HC
25 Holy Cross      0 .422 ________________________________________________  
    

Automatic bids are given to the four regular season and playoff champions of the estalished conferences. For the sake of this excercise, we assign those to the current leaders in each conference: North Dakota, Michigan State, Maine and St. Lawrence. SLU is tied for the conference lead with Clarkson, but according to the ECAC Tiebreaker System, the Saints rank ahead based on a head-to-head goal differential with the Knights of +1. This is currently very important, since the Saints' pairwise comparisons are insufficient to snare them an at-large bid.

There would ordinarily be up to four more guaranteed bids for winners of the conference tournaments, but rather than conjecture as to those, let's leave them out and move on to filling eight at-large slots. The next complication comes from the issue of relative conference strength. The six Division I teams in the Metro-Atlantic Athletic conference play no games against any teams from established conferences, which makes it difficult to gauge their strength based on tradition pairwise comparisons. The selection committee anticipated this by including language about relative conference strength in its guidelines for this season. This question must be addressed since a team from the MAAC, Quinnipiac, is solidly in the top twelve according to the pairwise rankings. But since their only non-conference game was a win over Division I independent Army (the second-worst team in Division I according to the Ratings Percentage Index), it's safe to say that Quinnipiac should only be judged to be as strong as the MAAC as a whole. Looking at the average RPI of the eligible teams in each of the five conferences, as well as their performance against the four eligible Division I independents (the only bridge between the MAAC and the big four):

                vs Indies    vs Army    vs Niagara   vs AFA      vs Mankato
      Avg RPI   PF-PA  Pct  PF-PA  Pct  PF-PA  Pct  PF-PA Pct    PF-PA Pct
 HE    .528     10- 2 .833   8-0 1.000   0- 2 .000   2-0  1.000   0-0  .---
 WCHA  .506     28- 8 .778   0-0  .---   0- 0 .---  10-0  1.000  18-8  .692
 CCHA  .504      5- 5 .500   0-0  .---   2- 4 .333   0-0   .---   3-1  .750
 ECAC  .496     28-10 .737  10-0 1.000  10-10 .500   2-0  1.000   6-0 1.000
 MAAC  .449      5-17 .227   4-4  .500   0- 2 .000   1-3   .250   0-8  .000

I think it's safe to say that the committee would decide the MAAC has not reached "competetive equity", and exclude Quinnipiac, Connecticut and Holy Cross from tournament consideration.

With that out of the way, we proceed to assigning the at-large bids. UNH and Boston College win their comparisons with all of the teams below them, and with the exception of the Princeton-Notre Dame comparison, so do Colorado College, Clarkson and Princeton. So those five should get bids easily. The remaining three bids need to be given based on pairwise comparisons among the "bubble teams". Where life gets really complicated is in the definition of the bubble, because Division I Independent Niagara wins so many comparisons against team higher in the PWR and loses comparisons against those farther down. To illustrate, consider the broadest possible set of teams that could end up on the bubble, leaving out only teams which have already been granted at-large or automatic bids, plus the bottom five teams in the PWR, who lose all but one of the comparisons with teams above them. That leaves the following eight teams in contention for the last three bids:

   Team         lPWR RPI Comps Won
 1 Ohio State     6 .543   Mi__NtNiNMRPPv
 2 Michigan       5 .564 __  DUNt__NMRPPv
 3 Denver U       5 .546 OS__  __NiNMRPPv
    
 4 Notre Dame     4 .562 ____DU  __NMRPPv
 5 Niagara        3 .487 __Mi__Nt  NM____
 6 Northern Mich  2 .538 __________  RPPv
 7 RPI            2 .531 ________Ni__  Pv
 8 Providence     1 .522 ________Ni____
    

It would seem like Ohio State, Michigan and DU are the teams to take, but if we drop the bottom three teams (who are clearly not going to make the cut), give a bid to top team OSU, and instead define a four-team bubble, we get the following:

    Team         lPWR RPI Comps Won
 1 Niagara         2 .487   MiNt__
 2 Michigan        2 .564 __  NtDU

 3 Notre Dame      1 .562 ____  DU
 4 Denver U        1 .546 Ni____
    

In which case Niagara makes the field in place of Denver. I don't think we can predict what the committee would do here. The definition of the bubble is the only vague aspect of selecting the field, and while that vagueness has never mattered before, it could be crucial if Niagara's comparisons stay about the way they are. It's even possible that Notre Dame would receive the final bid rather than Niagara or Denver. They do have the highest RPI of the three, and win the pairwise comparison with Denver.

Let's guess that Denver would receive the bid in question, which would leave each of the four major conferences with three teams in the tournament:

      West                                  East
1 North Dakota   5 .648 MSCCMiDUOS | 1 Maine          5 .627 NHBCCkPnSL
2 Mich State     4 .596   CCMiDUOS | 2 New Hampshire  4 .624   BCCkPnSL
3 CO College     3 .593 __  MiDUOS | 3 Boston Coll    3 .575 __  CkPnSL
4 Michigan       1 .564 ____  DU__ | 4 Clarkson       2 .569 ____  PnSL
5 Denver U       1 .546 ______  OS | 5 Princeton      1 .550 ______  SL
6 Ohio State     1 .543 ____Mi__   | 6 St Lawrence    0 .546 ________  

First-round byes are assigned to the top two teams in each region, which are clearly North Dakota and Michigan State in the West and Maine and New Hampshire in the East. We need to shift two teams from each region to the opposite one. In the East, the pairwise comparisons indicate that those should be Princeton and St. Lawrence. There is the drawback that leaving Boston College in the East will neccessarily set up an undesireable second-round matchup between them and another Hockey East team. However, with the East Regional being held in Worcester, Massachusetts, it is basically inconceivable that a strong potential draw like BC would be moved out of their region in contradiction of the pairwise comparisons. Likewise, Clarkson not only wins comparisons with the other two ECAC teams, they are also the best choice to maximize attendance in the East.

Out West, the bottom three teams are in a rock-scissors-paper standoff, but Michigan has the highest RPI of the three (not to mention attendance considerations) and so should stay in the West. Swapping Denver and OSU for Princeton and SLU gives the following regions:

      West                                  East
1 North Dakota (W)   1 .648 MS     | 1 Maine (H)          1 .627 NH     
2 Mich State (C)     0 .596        | 2 New Hampshire (H)  0 .624      
          
3 Princeton (E)      3 .550 CCMiSL | 3 Boston Coll (H)    3 .575 CkDUOS
4 CO College (W)     2 .593   MiSL | 4 Clarkson (E)       2 .569   DUOS
5 Michigan (C)       1 .564 __  SL | 5 Denver U (W)       1 .546 __  OS
6 St Lawrence (E)    0 .546 ____   | 6 Ohio State (C)     0 .543 ____
    

The natural seedings give two potential second-round intraconference games: BC-UNH is unavoidable with the Eagles in the East, but we can get rid of CC-North Dakota by swapping Princeton with Colorado College. This also removes the Princeton-SLU first-round matchup, which is unequivocally forbidden.

5W Michigan (C)                    6E Ohio State (C)    
4W Princeton (E)                   3E Boston Coll (H)   
     1W North Dakota (W) --+--2E New Hampshire (H) 
                           | 
     2W Mich State (C)   --+--1E Maine (H)         
3W CO College (W)                  4E Clarkson (E)      
6W St Lawrence (E)                 5E Denver U (W)     

The Gory Details

You also can look at the detailed description of each of the pairwise comparisons, if you're so inclined.


Last Modified: 2011 October 9

Joe Schlobotnik / joe@amurgsval.org

HTML 4.0 compliant CSS2 compliant