Although I do not believe that the RPI system with the modifications
that have been used (if difference is < .01 then ..., etc.) is the
best system that could be developed, it is IMHO much better than what
I think is the main alternative, namely a selection committee meeting
in a smoke-filled room and starting from scratch.
 
Years ago, when the ECAC consisted of between 15 and 17 teams and only
8 made the playoffs, some of the strangest choices seemed to have been
made.  Because there was no Hockey-L at the time, all that I found out
was what I read in the local press and the RPI student newspaper and
what I heard on local radio and TV stations and WRPI, clearly all
biased to RPI.  But in any case, it all seemed quite arbitrary.  One
of the problems was that it was quite sometime before mandatory
scheduling was in place and unbalanced scheduling made everything
quite subjective.
 
Personally, I feel that RPI (the rating, not the school) suffers
mainly because it has no apparent mathematical basis -- the .25, .5,
.25 weights are quite arbitrary, but it certainly does combine three
quantities that I would agree indicate which teams are better than
other teams.
 
Ralph Baer
RPI '68, '70, '74