> Can anyone be more specific regarding where at the U.S. College Hockey
> Online website that these guidelines are published?  Are they official NCAA
> guidelines, or are they U.S. College Hockey Online's
> interpretations/opinions?
 
The principal document is the Tournament Selection FAQ that Adam Wodon
wrote at
 
http://www.uscollegehockey.com/rankings/19992000/selectionfaq.html
 
Unfortunately, the NCAA doesn't have its own comprehensive description
of the procedure, so we have to go by what gets reported in media
conference calls and NCAA News articles, and answers from the NCAA
officials in response to our questions.  The watershed in this regard
was an interview that Adam did with then-selection committee chair Joe
Marsh after the 1997 tournament, where he really spelled out the
details.
 
http://www.uscollegehockey.com/tournament/032097.html
 
> Also, can anyone provide John Whelan's web address and/or a brief
> explanation of who John Whelan is.
 
I guess it's been a while since I plugged this.  I'm just a fan with
enough knowledge of computer programming and mathematics to work out
the selection criteria defined by the NCAA (Charlie Shub, USCHO and I
all do this independently, although we start from the same game
results).  I've also peppered the NCAA with a number of questions
about the process over the past few years, and the answers are
synthesized into a description of the process, to the best of my
understanding (before the seeding of this year's tournament; it has
not yet been revised to include the new information that Niagara is
definitely considered an Eastern team, and as apparent change in how
teams are distributed between the regionals when there are more teams
from one region than the other), at
 
http://www.slack.net/~whelan/cgi-bin/tbrw.cgi?pairwise
 
That also page contains a bunch of links to USCHO and NCAA News
articles relevent to tournament selection.
 
Adam and I differ on some of the finer details of the precise
algorithm, but we agree that the committee considers the following
factors, and only the following factors, when choosing the field,
deciding who plays in what regionals, and seeding the regionals
themselves:
 
1. Regular season and tournament championships in the WCHA, CCHA, ECAC
and Hockey East (these are only used for determining automatic bids
and automatic byes, and the committee has no discretion to overrule
them)
 
2. Pairwise *comparisons* determinined using games between
tournament-eligible Division I teams according to five criteria using
an algorithm described for example at
http://www.uscollegehockey.com/rankings/19992000/explanation.html
 
3. Relative strengths of the conferences
 
4. Which teams are hosting the regionals (the regional host is
required to play in its own regional, and the information is not used
otherwise)
 
5. Potential attendance at the regionals (only considered in placing
the teams in regionals)
 
6. The desire to avoid potential intraconference matchups, especially
rematches of playoff games, in the regionals (only used in placement
and seeding, not selection)
 
The original point of all this was to pin down the procedure enough to
make a CGI script, "You Are the Committee", at
 
http://www.slack.net/~whelan/cgi-bin/tbrw.cgi?tourney
 
which takes one interactively through the process and lets one make
the decisions the committee does based on the same data they have at
their disposal.
 
> Please understand that I'm not
> DEBATING any point here.  I'm just an average sportsman who's always
> believed (maybe naively) that competition takes place on the court, ice,
> field, etc.
 
By the same token, a number of us fans believe that when seeding a
tournament, it's more fair to analyze the results of that competition
using an objective statistical measure than the whims of a bunch of
people in a smoke-filled room.  After all, conference playoffs are
seeded by the final standings and a specified tie-breaking algorithm,
not by the subjective decisions of a committee.
 
>  I'm just now trying to educate myself on the "purely
> mathematical process", why there is a selection committee if "there are no
> subjective decisions", and now how Q failed to receive a tournament
> invitation if "there are no subjective decisions".
 
In the past, the selection committee was at liberty to do as it
pleased when choosing and seeding the tournament field, which led to
some bad blood.  Now it has a somewhat proscribed set of guidelines it
follows.  The implementation of factors 1 and 4 above is completely
automatic; the calculation of factor 2 is likewise hands-off, but the
committee's role comes in deciding what to do with it.
 
Factor 3 is a big question mark, since the most precise statement
we've so far had from the NCAA (right before the MAAC's first season)
is "the committee noted that it reserves the right to evaluate each
team based on the relative strength of their respective conference
using the overall conference ratings percentage index (RPI) in
determining competitive equity."  None of us know what the committee
means by "overall conference RPI", especially given that one of the
shortcomings of the system is that teams in a weak conference have
RPIs that are "too high".  Given the way it's been used in the past,
it's clear that it's targeted at teams from new conferences which the
committee believes should be seeded lower than their pairwise
comparisons indicate.  A number of us believe that if the criteria
were improved, there would be no need for factor 3.
 
The committee's judgement comes into play as follows:
 
In selecting the tournament field, deciding whether factor 3
(conference strength) should cause them to overrule factor 2 (PWCs).
 
In placing teams in regionals, deciding how much weight to give
factors 2, 5 (attendance) and 6 (intraconference matchups), as well as
factor 3.
 
In seeding teams within regionals, deciding whether factor 3 should
cause them to overrule factor 2.
 
Otherwise the procedure is more or less automatic, although some of
the finer details of how things work when the comparisons (factor 2)
are non-transitive (i.e., where a lot of situations exist where team A
wins the comparison with team B and team B with team C, but team A
loses the comparison with team C) are not completely spelled out to
us, and possibly also not to the committee.
 
> Who knows?  If you educate me (and others who may be observing this
> thread),
 
The idea of asking the NCAA all those questions and putting together
the web pages on the subject was to educate both myself and the
general public.  Presumably the same goes for the USCHO folks (and
Keith Instone).  I think we've all benefitted from the conscious
effort the NCAA made a couple of years ago to educate the coaches and
the public on the selection procedure.
 
> you may be able to convince me why Niagara's PWR ranking should be
> ignored, their tournament performance should be ignored, and why the
> Niagara athletes shouldn't think of themselves as a Top 10 team in 2000 D1
> hockey.
 
I think it's important to distinguish among three separate questions:
 
        1. Where and whether a team should be seeded in the tournament
according to the NCAA's current procedure
        2. Whether or not that procedure seeded a team fairly given
the data available when the tournament was seeded
        3. How a team's season should be evaluated after the
tournament has been played
 
As for question 1, I think it's clear that the committee was acting
within the bounds of its procedure both to select Niagara and to sent
them West and seed them sixth, contrary to the pairwise comparisons.
However, the apparent subjectivity of the conference strength
consideration means that the committe would also have been acting
within those limits to give an at-large bid to Mankato in Niagara's
place.
 
Question 2 is what most of us have been talking about this week.  The
point is that the conference strength escape clause used to exclude
Quinnipiac is basically a band-aid for the flaws in the criteria used
to define the pairwise comparisons.  Using broken criteria and an
inexact fix effectively took the objectivity out of the process as
regards Niagara.  According to the objective measure I came up with
last year, which tries to operate as much in the spirit of the current
criteria as possible while taking better account of strength of
schedule, Mankato was more "deserving" of the bid, but if Niagara had
for example beaten Canisius and taken the fourth point from UNO or
Brown, they would have been in line for the bid.  (In contrast, in
either situation I would have said the committee was within their own
bounds to choose either Niagara or Mankato under the current system.)
My system is not the only modification possible (although it's the
only one I've seen); the point is that it would be IMHO fairer to use
*some* system that didn't require a semi-subjective judgement of a
team based on the strength of their conference.
 
As for question 3, Niagara can say without ambiguity that they were
National Quarterfinalists when all was said and done.  Whether that
makes them one of the eight best teams in the country is really a moot
point.  The winner of the NCAA tournament is the national champion,
and the loser of the championship is the runner-up.  Most people
(myself included) regard that as more important than being the most
highly-rated team in the country before the tournament begins.  What I
would object to would be the claim that the national champions should
have been the #1 seeds going into the tournament.  Likewise the
conclusions that if the eventual champion was seeded lower, the
seeding must have been unfair, or that if the eventual champion was
seeded #1 that means the system used to seed them must automatically
have been fair.
                                          John Whelan, Cornell '91
                                                 [log in to unmask]
                                     http://www.amurgsval.org/joe/
 
HOCKEY-L is for discussion of college ice hockey;  send information to
[log in to unmask], The College Hockey Information List.