This is pretty long, so I've organized it into parts.

0. Preamble
1. Comments related to the seeds themselves
1.1 Placement of non-bye teams in regionals
1.2 Placement of bye teams in regionals
1.2.1 A plea to avoid awkward situations in a 16-team tournament
1.3 Michigan and St. Cloud: PWR vs individual comparisons
2. Comments related to Jack MacDonald's statements in the interview
2.1 Rematches from conference tournaments
2.2 Harvard vs Quinnipiac and MAAC PWCs
2.3 Automatic qualifiers
3. Conclusions

0. Preamble

Ever since reading Jack MacDonald's explanation of the 2002 NCAA
Tournament Seedings in his interview with Jayson Moy
<http://uscho.com/news/2002/03/18_004304.php>, I've been concerned
about the decrease in the transparency of the procedure and meaning to
write some sort of comment.  The news that the tournament will
definitely be switching to a 16-team format next season makes it that
much more important that the new selection procedure be clearly laid
out ahead of time.

1. Comments related to the seeds themselves

1.1 Placement of non-bye teams in regionals

The issue which has received the most attention in discussions of this
year's brackets is the regional nature, with all six Eastern teams in
the East Regional and all six Western teams in the West Regional.
Because of the nature of the 12-team selection procedure, a
distinction needs to be made between the placement of the top four
teams nationally and the placement of the remaining eight teams.  It
was well publicized before the tournament that an NCAA policy enacted
in response to the terrorist attacks of September 11, 2001 constrained
the placement of the eight non-bye teams, which meant that the four
Eastern and four Western teams in that group all had to play in their
own regions.  Not only was that well known, but since the NCAA
Championships Handbook at
<http://www.ncaa.org/library/handbooks/iceHockey/2002/2002_d1_m_icehockey.pdf>
is not specific about the placement of non-bye teams, the new policy
was completely consistent with the published procedures.

1.2 Placement of bye teams in regionals

As MacDonald states in the interview, the NCAA travel policy did not
apply to the placement of the bye seeds.  In the interview MacDonald
said "We were allowed to take the top four seeds and place them where
we saw fit."  I would have thought that the placement of those seeds
was to be done according to the procedure laid out in the
championships handbook, which contains two relevant instructions.  One
is point five under Seedings and Pairings on page 12, which says "The
top four-seeded teams 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," and the other is the bracket on page 30 of the handbook,
which shows that the #1 seed in the East plays the #2 seed in the West
in the national semifinal, while the #1 seed in the West plays the #2
seed in the East.  Given that the ranking of the top four teams
according to the selection criteria was 1. New Hampshire 2. Denver
3. Minnesota 4. Boston University, the championships handbook
indicated that New Hampshire had to be bracketed to play Boston
University in the semifinals, which meant, given the bracket, that one
of them had to be sent to the West Regional.  However, the committee
chose--and was allowed--to disregard the bracket contained in the
championships handbook, placing New Hampshire and Boston University in
the Eastern Regional and re-drawing the brackets so they would meet in
the semifinals anyway.

1.2.1 A plea to avoid awkward situations in a 16-team tournament

Some would say overruling the instructions in the championships manual
was an application of "common sense" on the committee's part.  But
this implies that the authors of the manual and the instructions it
contains didn't anticipate the situation where they would be in effect
requiring a committee to send and Eastern and a Western bye team to
the opposite regionals to make the bracketing work out right.  If so,
that was a fairly major oversight, given that there are only a few
possible combinations of regions for the top four teams, and a number
of us in the college hockey fan community looked at the handbook and
foresaw just the sort of situation that came up this season.  This is
where I hope the NCAA will learn from the experience of this season,
and carefully consider the implications of the guidelines they set for
seeding a 16-team tournament.  For example (assuming the new format
involves four four-team regionals at neutral sites), if the guidelines
dictate that the top four teams have to be bracketed 1/4 and 2/3 in
the semifinals, either the guidelines will need to allow the committee
to fill out four four-team regionals and then place them into an
overall bracket based on the relative rankings of the #1 seeds, or the
NCAA should be willing to accept #1 seeds being shipped to "strange"
regionals to ensure they're properly bracketed.

1.3 Michigan and St. Cloud: PWR vs individual comparisons

One other (minor) surprise which came out of the seeds, which hasn't
received much mention, was that Michigan was seeded 4th and St. Cloud
5th in the Western Regional.  This is surprising because St. Cloud won
the individual pairwise comparison with Michigan.  Michigan was ranked
ahead of St. Cloud in the overall PWR, because they won comparisons
with Boston University, Maine, and Cornell, which St. Cloud lost, but
the NCAA has repeatedly stressed that they do not use the PWR to seed
the tournament field, but instead look at the individual pairwise
comparisons.  (In fact, Cornell was seeded below Miami in 1997 for
exactly this reason, as then selection committee chair Joe Marsh
explained to Adam Wodon in an interview archived at
<http://www.uscho.com/news/1997/03/20_000770.php>.)  So the question
here has to be: Was this a mistake, or has there been a change in
Selection Committee policy on this issue?

2. Comments related to Jack MacDonald's statements in the interview

2.1 Rematches from conference tournaments

"Most of the people that go to Worcester are not there to watch
Colorado College and Minnesota. They're there to see their team play
the team they lost to last weekend, or that they beat last weekend."

This reflects a very different sentiment than was expressed by
selection committee members in past years.  The attempts to avoid
intraconference games in the early rounds were supported in part by a
desire not to have the regionals become a replay of the conference
tournaments, and on one occasion a second-round conference matchup was
described as less of a problem because those two teams had not played
since the middle of the season.

2.2 Harvard vs Quinnipiac and MAAC PWCs

"Based purely on the numbers, and you know the Pairwise as well as I
do, probably either Harvard or Quinnipiac would have gone west.
Pairwise, who was better? I'm not telling you who was the better team.
Harvard goes west based on the math. I'll tell you what I told the
committee: Who has the better program? Harvard, without doubt. Who's
the better team? Probably Harvard. But that's not the criteria given
to the committee."

Based purely on the numbers, Quinnipiac would have received at-large
bids in 1999 and 2000 and Niagara would have been seeded 3rd in the
East rather than 6th in the West in 2000.  Since the MAAC began play,
the NCAA Selection Committee has stressed its discretionary power to
consider the strength of a team's conference when evaluating their
selection criteria.  Considering the MAAC's overall 3-44-1 record in
non-conference games this season, it seems like the conference's
"competitive equity" with the rest of the NCAA is very much in doubt.
I hope that, had the question actually arisen, MacDonald, as athletic
director of Quinnipiac, would have abstained from the committee's
subjective decision of how to evaluate Quinnipiac's pairwise
comparison win over Harvard.

2.3 Automatic qualifiers

"I hope that the visionaries of college hockey realize that we have 60
institutions playing Division I college hockey for only five automatic
bids. The math is not in anybody's favor. We have the opportunity to
create more automatics in a very, very easy way. [...]  There are two
more automatics right under our nose and no one is talking about
it. The Ivy League and Big Ten. [...] those automatics are there and
we're wasting them. The NCAA recommends that half of the field be
automatic bids. [...] when we go to 16, we'll have six, we could have
eight.  My point is that by having more automatics, we allow
everybody, we could now have 8 automatics for 60 schools, that gives
every school a better shot."

Since there are no Division I independents, and since the champions of
all six Division I conferences will receive automatic bids starting
next season, every team in the NCAA will have the opportunity to
receive an automatic bid via their conference tournament.  Why is that
situation unsatisfactory?  MacDonald's comments seem to imply that by
splitting into more, smaller conferences, hockey could make more
tournament bids magically appear in the form of automatic qualifiers.
But of course, those bids would be at the expense of at-large bids,
which are based on selection criteria evaluating a team's performance
throughout the season.  This is a zero-sum game, and if teams which go
on a hot streak during their conference playoffs, and teams in smaller
conferences, have a better shot at making the NCAAs, then teams which
play well all season but have the misfortune to play in strong
conferences, or get upset in the playoffs, have a worse shot.

Also, conferences like the ECAC, CCHA, and WCHA (who would be broken
up to form an independent Ivy League and Big Ten) have functioned very
well in their current form, and do not need a "visionary" to come
along and re-structure simply to increase the number of automatic bids
in the tournament, especially when it's far from apparent that that
itself is always a good thing.

3. Conclusions

Several years ago, the NCAA undertook a conscious effort to educate
both the coaches and the general public about the selection and
seeding of the tournament.  In recent years, the efficiency of that
communication has slipped somewhat.  Statements from the selection
committee chair such as "So, put down your microscopes and enjoy
this. The people involved in the bracketology of the whole thing can
be too critical of something that could be the best weekend college
hockey has ever seen" come across as insulting to the segment of the
community which is trying to understand the details of the seeding
procedure so we can help disseminate the information.

With the structure of the NCAA tournament bracket about to change to
accommodate sixteen teams, there is a great opportunity to increase
the transparency of the system.  Lay down the rules in the new
championships handbook carefully (checking with us "bracketologists"
to avoid painting the committee into a corner in a foreseeable
hypothetical situation), then be responsive to requests for
clarifications, and have the committee follow the prescribed
procedure.  If the NCAA siezes this opportunity, it will put college
hockey fans in a position to focus on anticipation of the tournament
to come rather than questions over how that tournament was seeded.

                                          John Whelan, Cornell '91
                                                 [log in to unmask]
                                     http://www.amurgsval.org/joe/

Enjoy the latest Hockey Geek tools at slack.net/hockey