It would be interesting to pursue the reasons for discrepancies in
CHODR, TCHCR, and RPICH ratings.  Here's a couple of quick observations:
 
Suppose that we assume that the East (particularly the ECAC) has a couple
of weak teams.  How should this affect the ratings of the "good" teams
that play them?  Contrary to an earlier post, playing a weak team will
not necessarily help one's CHODR rating, since the "good" team would
already be predicted to win by a considerable margin.  They can only
improve their rating by exceeding that margin. For example, given the
current CHODR ratings, if Michigan were to visit Yale, they would need
to score more than 10 goals to improve their offensive rating. Anything
less than 10 and their CHODR offensive rating would decrease.  On the other
hand, if Michigan were to add a two game set against Yale at this point in
the season, regardless of the outcomes of those games, I believe we
would see Michigan's TCHCR and RPICH ratings go down.
 
It is true that a couple of extraordinary outings (like Harvard's 12-1 win
over Yale or UMass-Lowell's two recent losses) can have a lasting effect
on one's CHODR rating which is not nearly so dominant in TCHCR or RPICH.
I wouldn't characterize either of the UMass-Lowell games as involving
any weak teams.  I wonder if style of play in the East makes such games
more likely.
 
Certainly, winning a number of close games (Northeastern is probably a
good example here) will help one's TCHCR and RPICH ratings much more than
one's CHODR ratings.  My sense is that the ECAC has suffered a pretty
dismal out-of-conference record this year, but many of the games have been
fairly close (lots of one-goal or OT losses to HE) - something good for
Mike M. to dig up some stats on(:>)? This greatly affects TCHCR and RPICH
strength of schedule values throughout the ECAC, but has less impact on
CHODR ratings.
 
Finally, it has seemed to me that teams which play more games tend to
fare better in TCHCR and RPICH than they do in CHODR.  It's not clear
whether this is a function of the systems or just confounded with the other
east/west distinctions.
 
Robin Lock
[log in to unmask]