Sunday's NCAA women's hockey bracket announcement left many shaking their heads. Based on the Pairwise rankings and the selection criteria in the NCAA handbook, most expected Princeton, the 7th ranked team in the Pairwise, to be sent to Boston College, with Northeastern, at #6, being sent to Minnesota.
Here is the selection criteria as set forth in the women's hockey handbook:
The Women’s Ice Hockey Committee will seed the selected participants as follows:
1. The top four teams according to the selection criteria will be seeded 1-4 at the time of the selection call. The remaining four teams will be placed in the bracket based on relative strength as long as these pairings do not result in additional flights. These teams will not be reseeded and the committee will not change the bracket once the tournament has begun.
2. Assuming it meets the committee’s hosting criteria, the highest seeded team will be given the opportunity to host the quarterfinal game.
Pairings in the quarterfinal round shall be based primarily on the teams’ geographical proximity to one another, regardless of their region, in order to avoid air travel in quarterfinal-round games whenever possible. Teams’ relative strength, according to the committee’s selection criteria, shall be considered when establishing pairings if such pairings do not result in air travel that otherwise could be avoided.
Based on this, and the committee's decision to send Princeton to Minneapolis, it appeared on its face that the committee was putting extra emphasis on the "geographic proximity" portion of the criteria, making the bus trip to Boston significantly shorter and even removing a hotel stay since the Huskies can just stay on campus.
As it turns out, that wasn't the case at all.
BC Interruption reached out to Sarah Fraser, chair of the NCAA women's hockey selection committee, for clarification on the methodology used, and it's not what anyone expected at all:
Committee deliberations included a review of each team's full body of work while considering all of the selection criteria. Strength of schedule and win percentage are included among those criteria. It must also be considered that the committee has the responsibility to reduce flights when possible.
Princeton and NU were neck-and-neck in regional and national committee deliberations all winter and in the final discussion Princeton was deemed to be a stronger team. Strength of schedule played into that discussion.
So it wasn't a travel consideration at all, but rather a decision, in the committee's opinion, to recognize Princeton as a stronger team than the Huskies.
We asked Mrs. Fraser if that meant that the Pairwise was not an accurate model for determining how the committee ranked the field.
I wouldn't use the term "model," but I will say that the women's committee selection criteria does allow the committee to use its discretion and reserves the right to weight criteria differently based on relative team performance. The men's situation is different where they straight by the numbers, so that could confuse people.
I don't know if USCHO created the Pairwise, they very well might have, but somebody did, and that's not the same system or the same numbers that the NCAA looks at. Mathematically it could be the same, but we don't use the Pairwise ever.
So it was more the committee thought Princeton was a stronger team than it was shortening a bus trip from Princeton to Boston rather than Northeastern to Boston?
Yes, in this case it was that Princeton, we felt, their body of work was ranked higher than Northeastern.
First, a big thank you to Mrs. Fraser for taking the time to clarify some of this for us.
What we learned, though, casts even more of a shadow over the women's selection criteria than there already was. While the criteria itself left a lot to be desired (namely with saving money the primary consideration over producing the best possible bracket), one of the saving graces of the system was that there was almost no ambiguity or "smoke filled room" aspect. The Pairwise was a mathematical formula (if a bad one) so there was very limited subjectivity as to the ranking of teams, and there was a clear set of rules (flights + bracket integrity) to determine who would be sent where.
As it turns out, there appears to be significantly more subjectivity in the ranking of teams than anyone realized. There is not a single ranking system that ranks Princeton above Northeastern -- not WCHODR, not Rutter, not KRACH, and obviously not the Pairwise. Even the completely subjective USCHO poll has Princeton ranked significantly lower than Northeastern.
What makes this even more crazy is that despite the reasoning specifically cited by the committee, that strength of schedule played a factor in choosing to rank Princeton higher than Northeastern, both RPI and KRACH rank Northeastern's schedule tougher than Princeton's.
But, it doesn't matter -- the bracket is the bracket, and BC, Princeton, Minnesota, and Northeastern can only play the hands they were dealt. The selection criteria was tweaked this past offseason, so we can only hope that this new information will cause the coaches to push for a more transparent and objective system.