Thanks very much for the explanations concerning *some* of the
operation of this extremely interesting ranking system.  I look
forward to it each week.
 
One comment, on a part I was sure I could understand:  the re-rating
of a team's opponents after each of THEIR games and the subsequent
re-rating of the team in question is an interesting approach.  It
could be argued that your rating, relative to someone you've already
played, should not be changed by the LATER performance of that
someone (opponent).  On the other hand, maybe you should do this
Anyone have comments on this, out there?
 
My former institution, Eastern Washington University was quite an
NAIA basketball power in the Pacific Northwest during the 1970's
and early 1980's (and then pretty good at the NCAA II levelin the
mid-80's, before going all the way up to Div I status about 1988).
Anyway, despite the team's overall success we could never seem to
beat our arch rival, Central Washington University (in Ellensburg),
at playoff time, and this seemed due in large part to Centrals some-
how always being awarded the home court advantage in the finals (which
were best-of-three in those NAIA days).  Our coach, a pretty cerebral
guy who was frustrated as hell by his lack of success in big games
with CWU, devised a simplified version of the TCHRP to rank the teams
in the northwest division of NAIA Div I (the Northwest).  It gave the
most points for beating a winning team on the road, the least for
losing to a losing team on your home court, etc.--which I gather is
similar to the approach TCHRP takes.  HOWEVER, "winning" and "losing"
definitions of teams depended upon their won-loss record AT THE MOMENT
THAT YOU PLAYED THEM.  If they were at .500 or above when you played
them, they were "winners."  Even this system didn't do for us what our
coach had hoped.  One time, late in the season, we were locked with
Central in a tight race for first place in the final standings, prior
to the playoffs.  Central played a common opponent of ours on a Saturday
night.  The third team had a record of .500 at that point.  Central
beat them, and thus got credit for beating a "winning" team (the game
was at home for Central, so they got, I believe, 4 points).  The next
night WE played the same team.  Of course it was now one game below
.500, so although we beat them (and, I think, by more points than
Central had the night before), we got only 3 points!  As one of my
friends said--"It's hopeless!  We built the system and we still can't
beat (out) those guys!"  Needless to say, as usual CWU had the home
court advantage that year again.  (In all fairness I must admit that
the one time in all those years Eastern did get the home court advantage
as fate would have it we won the opening game on THEIR court, only to
have CWU come back and beat us two straight on ours!  Arrghh!)
 
Anyway--I hope no one minds this apparent digression in to basketball.
My point is the various arguments for weighting opponents' records
in the figuring of different teams' ratings.  What do others have to
say about this?
                  Steve Christopher, NMU (Go 'Cats!)