> Really?  Why, then, do we find the 10th place ECAC team at #42 in the
> RPI, below all other conference affiliates?  Or 4 ECAC teams at #33,
> #34, #36, and #42?
 
As an ECAC fan, I think this is a valid question.  After all, the ECAC has
indeed placed several teams in the lower tier of the RPI for several years.
 Here are some *possible* reasons which would have nothing to do with
comparative league quality, and I'm sure cleverer lasses and lads than I
can think of more.
 
1.  There are statistical consequences of having more teams in the league
which tend to pull the bottom ECAC teams lower in the RPI.
2.  There are statistical consequences of having fewer games in the season
which tend to pull the bottom ECAC teams lower in the RPI.
3.  The ECAC teams have a historical record of / have strategically adopted
/ are stuck playing non-conference games which tend to pull the bottom ECAC
teams lower in the RPI.
4.  The RPI reflects such slight interleague play, that comparisons between
leagues have no statistical significance whatever.
 
 
I think the PWR (with its strong RPI compoment) is superior to former, less
formal mechanisms for selection.  There may be mathematical ghost effects
which hide biases -- I don't really want to dredge up the math necessary to
find out, but if there are any truly stellar statistical geeks out there
who eat heteroskedasticity for breakfast...
 
HOCKEY-L is for discussion of college ice hockey;  send information to
[log in to unmask], The College Hockey Information List.