For the record, here is the letter I sent to Bob Croce. I did not plan
on sending a copy to HOCKEY-L, but now it seems appropriate. You can
see that I "borrowed" many points that were brought up on HOCKEY-L.
Thanks!
My work and home phone numbers were included at the top of the page. I
added the 1/11 TCHCR as a second page.
Keith [log in to unmask]
=======================================================================
1/12/92
Bob,
I called the Albany Times-Union several times this week, but I could
never catch you there to talk to you regarding your article about me
and my rating. I am sending you this fax instead.
I must thank you for the publicity, but I am disappointed because you
made several glaring errors in your reporting of TCHCR. And it upsets
me that you never called to get your facts straight before printing
your article in such a well-circulated newspaper. Specifically, I have
the following points to make regarding your story.
o I was disturbed by your overstatement of the RPI players' reaction
to TCHCR. After calling Buddy Powers and hearing from Tony Hejna, I
now realize that words like "fury" and "infamy" are just your attempt
to spice up your story and are not a true indication of the reaction
there at RPI.
o Your analysis that "there are only 11 teams...worse than the Engineers"
is incorrect. TCHCR IS NOT A POLL. You tell your voters to rank the
top 15 teams in the country. But TCHCR is merely a set of mathematical
calculations based on how each team has performed relative to the
other teams in Division I so far this season. A better claim would
have been that 11 teams have done worse than RPI so far, but the best
statement is just to say that RPI is ranked 34th by TCHCR.
o Is there any basis for your claims about Union? Did you talk to any
Union players or coaches?
o "Strength of conference" is not a factor in TCHCR! TCHCR has no
concept of who is in what conference, and I have never stated that it
does. However, strength of SCHEDULE is important. Your assumption
that TCHCR ranks the conferences probably stems from the table of
inter-league numbers that I included (you quoted them, after all). If
you had read the disclaimer before the table, however, you would have
found that TCHCR does not use those numbers. I compiled them
separately and included them to help explain what is going on in
TCHCR.
o Regarding Maine's drop to 6th, you missed the boat by so much I need
sub-topics.
--If you look at the heading for the 1/4 printout, it says that
"last week" actually stretched back to Christmas. You then must
consider Maine's performance at the Badger Showdown, where they
lost to Wisconsin. After seeing Maine's drop myself, I went
back and calculated TCHCR for 12/29--the Black Bears had
already fallen to 5th by then. The results of the first weekend
in January (including the Dexter Shoe) dropped them 1 more spot
to sixth.
--In the 1/4 printout, I already explained the effect that the
Maine-Army game had. Please re-read it. (Also, the score was
15-5.)
--There is so much more to consider than just Maine's wins!
TCHCR builds a graph off ALL of the Division I results and
considers everything simultaneously. The ramification is that
Maine's rating is also determined by Maine's opponents (because
Maine's opponents determine Maine's schedule strength). For
example, UMass-Lowell's two losses at North Dakota hurt Maine:
UML is not rated as high, making Maine's schedule easier.
--Also, you have to look at the other teams: Michigan, Lake
Superior, Minnesota and Michigan State were also winning, but
against higher-rated opponents. Maine's fall to sixth was in
part due to other teams out-performing them.
o Your comparison to ATU's rankings are not very relevant. TCHCR is
not supposed to be like any poll. In fact, it is quite the opposite:
polls are opinions and TCHCR is math. The reason I started the rating
was to give people a different point of view; this does not make it
"flawed."
I feel that these mistakes not only give your readers the wrong
impression of The College Hockey Computer Rating, but also false
information about it. I would love to see another article in the
Times-Union about TCHCR, but without the errors.
I will gladly answer your questions about TCHCR. Feel free to call me
at work or at home. I can send you more information about it, if you
are interested. Hopefully, the 1/11 TCHCR included will help you gain
an understanding of how it works.
Sincerely,
Keith Instone
|