D-I Women’s Bracketology:
March 2, 2005

This week’s Women’s D-I* Bracketology column begins with a confession — USCHO can’t claim to predict exactly what working principles the NCAA Women’s Hockey Committee will use when choosing tournament pairings and host sites. There’s no shame in admitting that. In fact, the committee itself might not know exactly what it would do if the season ended today either. After all, this is the inaugural year of the eight-team tournament, and everyone faces a learning curve.

Because this column is not omniscient, does that mean it’s useless? Of course not. What we can do is use all information sources available to make the best possible bracket prediction given the current results to date. Such information sources include the women’s tournament’s handbook, comment from women’s hockey’s NCAA staff liaison Troy Arthur, and the pairing decisions from all past NCAA championships across all team sports.

This information suggests that the following working principles are the most probable: teams will be chosen under criteria approximated by the USCHO Pairwise Rankings (with one slight tweak in the RPI), and that once those teams are chosen, the NCAA will officially seed 25 percent of the field (two teams) and place them in the bracket so that they don’t face each other in the final. From there, teams will be paired based first on minimizing airline travel, and second on avoiding intraconference matchups and maximizing bracket integrity.

Given this information, we believe the most likely outcome of the selection process is that the NCAA will choose a bracket arrangement that requires just one flight. Because we expect three WCHA teams to make the tournament, and there are no other teams within driving distance of them, we expect that two WCHA teams will play each other in the NCAA quarterfinals.

We do recognize, however, that it is within the realm of possibility that an 11th hour source of funds could appear that allows the women’s hockey committee to fly three teams for the quarterfinals, although we do not believe such an outcome is at all likely. Nevertheless, we will provide the optimal bracket pairing involving three flights to account for this remote possibility.

That’s the disclaimer–on to the selection process.

Picking the Field

The first step is to rank the teams using the NCAA’s criteria, which is very close to the USCHO’s Pairwise Rankings. The only difference is that we need to drop certain games against weak opponents from each team’s RPI calculation in the manner set out in the NCAA handbook.

The following table lists the adjusted RPI, which approximates the RPI used by the NCAA selection committee. For the sake of comparison, the table also lists the unadjusted rank — the rank using the RPI calculated by USCHO’s Pairwise Rankings.

    Team Name   Adj.RPI  Unadj.Rank
1   Minnesota     .693       1
2   UMD           .661       2
3   Dartmouth     .656       3
4   Harvard       .623       5
5   Wisconsin     .619       4
6   St. Lawrence  .616       6
7   Mercyhurst    .601       7
8   UNH           .589       8
9   Princeton     .566       10
10  Providence    .559       9
11  Brown         .533       11
12  Colgate       .519       13
13  Ohio State    .518       12
14  Yale          .511       15
15  UConn         .509       14

The follow table ranks the teams using the NCAA’s criteria, which is the same as the Pairwise Rankings except it uses the adjusted RPI as one of five criteria instead of the RPI listed on USCHO. For the sake of comparison, each team’s Pairwise Ranking is also listed.

    Team Name      PWR
1   Minnesota      1
2   UMD            2
3   Dartmouth      3
4   Harvard        5
5   Wisconsin      4
6   St. Lawrence   6
7   Mercyhurst     7
8   UNH            8
9   Princeton      10
10  Providence     9
11  Brown          11
12  Colgate        13
13  Ohio State     12
14  Yale           15
15  UConn          14

The only difference resulting from using the adjusted RPI instead of the unadjusted RPI in the top eight is that Harvard jumps Wisconsin in the rankings.

The NCAA awards automatic bids to the champions of the WCHA, ECACHL, and Hockey East and five at-large bids. Each conference has a team in the top eight of the above rankings, so we’ll assume those top eight teams make the NCAA tournament in our bracket predictions.

Bracket Prediction With One Flight (The Likely Outcome)

The NCAA can only seed two of the eight teams. If the season ended today, Minnesota would be No. 1 and UMD would be No. 2. Teams are then paired geographical in order to minimize air travel with an eye towards bracket integrity. Because there are only three teams from the West Region, two must be paired with each other to avoid three flights because there exists a bracket pairing that requires only one flight. Therefore, Wisconsin is paired with UMD. Because New Hampshire and St. Lawrence are within driving distance of both Dartmouth and Harvard, and Mercyhurst is more than 300 miles away from everyone, Mercyhurst must be paired with Minnesota. Superior bracket integrity results from pairing New Hampshire with Dartmouth and St. Lawrence with Harvard. Under these pairings, each half of the bracket is balanced with the Minnesota site winner playing the Harvard site winner, and the Dartmouth site winner playing the UMD site winner.

If the season ended today, the expected bracket would be:

Mercyhurst at No. 1 Minnesota
St. Lawrence at Harvard
New Hampshire at Dartmouth
Wisconsin at No. 2 Minnesota-Duluth

Bracket Prediction With Three Flights (The Unlikely Outcome)

The bracket with perfect integrity (1 vs. 8, 2 vs. 7, etc.) can be done with just three flights. However, there is one intraconference matchup (St. Lawrence vs. Dartmouth) that we need to avoid. Since Mercyhurst is not within driving distance of Dartmouth, and that would require four flights, Dartmouth would have to host New Hampshire, leaving Minnesota against with Mercyhurst.Note: previous editions of this column incorrectly assumed Mercyhurst was within driving distance of Dartmouth.

Thus, the expected bracketed would be:

Mercyhurst at No. 1 Minnesota
Wisconsin at Harvard
New Hampshire at Dartmouth
St. Lawrence at No. 2 Minnesota-Duluth

Bracket Prediction With Strict Travel (Hopefully Not Likely)

(Addendum, 3/2/05 7 p.m. ET) According to NCAA travel policies, any team over 300 miles from its destination can choose air transportation over ground transportation. The above bracket with one flight assumed that the [email protected] and St. [email protected] matchups would not require air transportation. After all, these teams do routinely make these trips by ground during their conference regular seasons. However, the visiting teams in each case would have the option of travelling by air since they are 300-400 miles away from their destination. So what if the NCAA choose to minimize the number of pairings more than 300 miles apart to avoid any possibility the visiting teams would choose air travel?

Dartmouth is the only qualifying school within 300 miles of St. Lawrence, and Minnesota is the only school within 300 miles of Wisconsin, so those schools are paired together. Of the remaining schools, only New Hampshire and Harvard are within 300 miles of each other, so they are paired together. That leaves Mercyhurst vs. Minnesota-Duluth for the final pairing.

Under such a scenario, regular season results are meaningless within the top eight teams, aside from assuring that Minnesota and UMD do not play each other in the quarterfinals. While such pairings seem unorthodox for a D-I tournament, this kind of thing happens in D-III tournaments all the time. So the following bracket is what the committee would choose if it’s willing to allow its D-I tournament to feel like a D-III tournament:

Wisconsin at No. 1 Minnesota
New Hampshire at Harvard
St. Lawrence at Dartmouth
Mercyhurst at No. 2 Minnesota-Duluth

Until next time.

*Note: Technically the women’s hockey championship is called a National Collegiate championship because in theory, it includes schools classified in both D-I and D-II. In reality, all schools in contention play up to the D-I level, so we refer to the championship as a D-I championship throughout the column.