Sorry, Dick, but your argument is based on two unwarranted
assumptions:
 
* Assumption #1: that one game can tell us anything about who should be
in the tournament.
 
The question at hand is not whether Quinnipiac would stand a
snowball's chance in hell in the NCAAs, but whether they deserve to be
there given their performance this season.  Mankato, judging from
their performance before the weekend, overcame 8-to-1 odds in beating
North Dakota, but that doesn't mean that a 17-to-1 longshot like
Quinnipiac should get a free ride to the tournament in the face of
other considerations.
 
* Assumption #2: that the ratings percentage index is a good indicator
of a team's strength even in the presence of disconnected schedules.
 
RPI is an ad-hoc attempt to gauge a team's performance and strength of
schedule by combining their winning percentage and those of their
opponents and opponents' opponents'.  It becomes inaccurate quickly
when teams' schedules are "balkanized" since opponents' winning
percentage becomes a poor measure of schedule strength.  The NCAA
recognized this weakness and this stressed the committee's prerogative
to overrule the RPI and pairwise based on conference strength.  Yes,
there is a process in place, and it includes judging whether a group
of teams who predominantly only play each other has an unreasonably
high RPI due to a lack of competitive equity.
 
I would prefer that the NCAA in the future use a criterion which is
robust enough not to need overruling in these situations.  Ken
Butler's KRACH rating system seems to do what RPI was intended to do,
and do it better, by determining a list of relative strengths which
would predict each teams' winning percentage given its schedule.  With
no preconceptions of who the "establishment" is or even who is in
which conference, it does automatically what the committee will have
to do by hand this year: compare the MAAC teams' performances against
the independents with those of the rest of Division I and judge each
team's success against their schedule accordingly.  According to that
measure, Niagara, who have the worst RPI of any team without a losing
record, are 4/5 as good as Mankato.  Mankato are twice as good as
Quinnipiac, three times as good as UConn and almost five times as good
as Holy Cross.  And Mankato themselves are less than half as good as
the teams on the bubble.  (All of this is excluding games against
Nebraska-Omaha and before last night's game, which of course improved
Mankato's ranking noticeably.)  The moral is that by a more robust
statistical measure Quinnipiac and UConn are a lot worse than the #9
and #18 figures you quote by blindly following the RPI.
 
Finally, your MAAC-vs-"establishment" comparison further pads its
numbers by comparing the MAAC teams (and Niagara) to UMD, Merrimack,
Lake State and Brown, teams whose only chance at even qualifying as
"teams under consideration" for the NCAA tournament is to win their
conference tournaments.  Those four teams place a lot lower in the RPI
than the MAAC teams (although according to the KRACH Duluth is around
the level of Quinnipiac and the other three are comparable to
Niagara), but to make the tournament each needs to advance through a
conference playoff system which is seeded against them and probably
get past some of the top 10 team in the country.  In comparison you
are proposing that Quinnipiac be awarded a bid if they win three games
against teams ranked even lower than themselves.  All of these
opponents have KRACH ratings less than half that of Northeastern, who
did not even qualify for the Hockey East tournament.
 
Before you respond that KRACH is not the system used by the NCAA to
judge the performance of teams, let me re-iterate my main point that
while the NCAA uses RPI, they are also aware of its weaknesses in the
face of the MAAC schedule, and have provided themselves with the
leeway to override the RPI.  I am appealing to the KRACH, which
appears not to suffer from those weaknesses, to demonstrate that using
the "MAAC escape clause" would indeed be reasonable and fair.  In
fact, I think it would be unfair not to.
 
                                          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.