(For HOCKEY-L people, this comes out of a discussion on hockey3 dealing
with Chris Lerch's DivII-III rating, which he is considering switching
this season to match the RPICH rating done by Erik Biever for DivI
teams (and which closely approximates the RPI used in selecting teams
for the DivI tourney).)
 
I'm curious about the fact that RPI seems to be preferred to SOS.  I'm
not saying that I disagree, but it is interesting that people are
expressing this view at the same time that the use of RPI has been
lambasted in DivI.
 
What are the reasons for preferring RPI?  Is it that some teams who
were perceived to have been better were ranked higher in RPI than SOS?
Then perhaps we have to ask what those perceptions are based on.  What
about cases in which RPI ranked teams lower?
 
I have said this before, but my problem with the use of RPI in the
selection process is that (in DivI), I think it is weighted too
heavily.  First, I'm not convinced that the weights given to the
individual components are based on anything that makes sense.  Should
winning pct really be worth only 25%?
 
And second, RPI should be used to generate an initial ranking of teams
and then eliminated from the process if it is deemed that two or more
teams have RPIs that are too close.  Last year, if any two or more
teams had RPIs that were within a certain threshold (again, a value
that seemed arbitrary to me), the committee looked at "other factors".
 
These factors included head to head, record against teams under
consideration, record in last 20 games, record against common
opponents...and RPI.  In a head to head comparison between two teams,
when looking at these factors, you received one "point" if you were
better in a certain factor (i.e. common opp) than the other team.
(You could receive > 1 point in head to head if, say, you had beaten
a team 3x and lost to them once - difference is +2 points.)  In
some cases, RPI was a deciding factor - say if from those five
factors, one team had +2, the other had +2, and one team had a slightly
higher RPI.  That team ended up with +3.
 
But this was after the committee had already determined that the RPI
difference was small enough that it effectively did not tell which
team deserved to be ranked higher!  That's why we went to these other
factors in the first place.
 
That, IMO, is one of the major problems with the process.  RPI should
NOT be one of the factors considered in head to head comparisons when
two teams' RPIs are within the threshold.
 
Just something to consider if the DivII-III people push to have RPI
used in the selection process.
 
BTW, in response to another question, the SUNYAC and MIAC winners have
both received DivIII automatic bids.  I believe that is related to the
fact that all teams in those conferences are (or have been) DivIII,
and so it is not possible for a non-DivIII team to win the conference
tourney (say UConn in the ECAC East) and lay claim to the conference's
automatic bid.
---                                                                   ---
Mike Machnik                                            [log in to unmask]
Cabletron Systems, Inc.                                    *HMM* 11/13/93
<<<<< Color Voice of the Merrimack Warriors (station TBA for 94-95) >>>>>