>>It could be the fact that no member of the MAAC has been anywhere near
>>the top 25, let alone the top 12, in any rating system that adequately
>>addresses strength of schedule.

> Yes, let's make the decision by use of mathematical formula rather than by
> seeing who puts more pucks in the net.

You would think the HOCKEY-L community would be immune to this sort of
anti-intellectual myth that rating systems exist in some magical
netherworld.  What do you think goes into these formulas?  It's the
number of wins, losses, and ties each team has against each other
team.

You want to compare who puts more pucks in the net, but you can't
ignore the question of what sort of opposition they faced.  For
instance, when Mercyhurst was not going 0-4-1 against non-MAAC
Division I opponents, they were 22-6-1 within their own league.
Winning two-thirds of their total games looks impressive until you
consider that the Division I members of the other five conferences,
including CHA, went 32-6-4 against the MAAC.  Mercyhurst lost as many
games against their MAAC comrades as the rest of Division I combined,
and in less than 3/4 as many games.

> With strength of schedule, we're talking about a relative number that
> penalizes not only teams but entire leagues, and and not just the MAAC and
> CHA teams -- it's a common problem for the ECAC, too. Going by KRACH, the
> highest-ranked SOS for an ECAC team was Harvard at #32.

This year the ECAC had a very bad season, and the WCHA a very good
one, in interconference play.  In the end, how can you judge the
strength of teams in one conference relative to those in another if
not by their performance in interconference games?

> Not one Hockey East team had an SOS better than 12th overall (BU) or lower
> than 26th (Lowell); the others fall in betweem with #13 BC, #14
> Northeastern, #15 Maine, #17 Providence, #19 UNH, and #20 UMass.

> I guess their SOS ratings all suffer because they're too busy beating up on
> each other.

You've got it backwards.  Suppose two teams have identical KRACH
ratings based on all their other results, and then play ten games
against each other, splitting them 5-5.  In that case their KRACH
ratings *will* *not* *change* *at* *all*.  The fundamental definition
of the KRACH ratings is that the winning percentage they predict for
each team given their schedule is the same as their actual winning
percentage.  In games against a team with the same rating, you're
predicted to go .500, so if you actually do, the previous set of
ratings continues to predict your overall winning percentage
perfectly.  This division into won-lost ratio and strength of schedule
is just a pedagogical aid.  If the two teams in this example are
tournament-caliber, their own KRACHes are presumably higher than those
of their average opponents, so their strength of schedule actually
goes up by virtue of beating up on each other.  But at the same time,
splitting a bunch of games brings their winning percentage down.  The
definition of KRACH ensures that these two effects cancel out, as I
would assume most of us would consider sensible.  With RPI, winning
percentage and strength of schedule are just added together in an ad
hoc manner which guarantees nothing.

> KRACH and all the others are fun to look at from time to time, but they
> serve no purpose when it comes to the tournament. There is only one rating
> system that matters, and that is the NCAA's PWR.

This statement is just plain disingenuous.  We all know that the NCAA
doesn't trust their own selection criteria, or else they would have
given Quinnipiac an at-large bid the last two seasons, and given
Niagara a better seed last year.  The NCAA's current procedure is to
use a system they know overrates MAAC teams, and then deny them
at-large bids by a judgement call.

Looking at KRACH (or RHEAL, or CHODR/CCHP) gives us some objective
standard on which to base the consideration of the MAAC's performance.
And since the issue is whether Mercyhurst was worthy of the bid to
which they were entitled, that seems as valid a starting point as any.

> In the PWR, the MAAC has three teams in the top 25: #13 Mercyhurst, #17
> Quinnipiac, and #18 Canisius. Considering that the NCAA chooses 12 teams
> for it's tournament, I think it's quite just that the MAAC had one
> representative there.

It must then have been very unjust that the MAAC went unrepresented
back when they had not only three or more teams in the top 25, but
also a team in the top 10 in PWR in 1999 and 2000.

It's one thing to say the MAAC is entitled to a bid for satisfying the
requirements of a Division I conference for two years straight, but to
point to the criteria which the NCAA has overlooked as flawed in the
MAAC's case, and which haven't gotten appreciably better over the
conference's three-year existence, is pretty faulty logic.

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

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