HOCKEY-L Archives

- Hockey-L - The College Hockey Discussion List

Hockey-L@LISTS.MAINE.EDU

Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Condense Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Sender:
College Hockey discussion list <[log in to unmask]>
Subject:
From:
Keith Instone <[log in to unmask]>
Date:
Mon, 1 Mar 1993 12:15:06 EST
Reply-To:
Keith Instone <[log in to unmask]>
Parts/Attachments:
text/plain (37 lines)
(We've been over this before, but since we have been getting a lot
of new members, let's review.)
 
>This has been bothering me about the ratings, including TCHCR, RPICH,
>and CHODR.  How do you average the ratings of a team's opponents?
 
For RPICH, you don't. Erik's "Opp Win %" is more like the average *winning
percentage* of your opponents.
 
>Do you use the ratings from the previous week?
 
No
 
>I guess my question is, what order do you do your operations in?
 
SIMULTANEOUSLY!
 
The basic algorithm is the same for TCHCR and CHODR and involves solving
a system of equations (linear algebra).
 
This basic algorithm is described "so a hockey coach would understand" (^;
in my paper "Inside The College Hockey Computer Rating". "Inside TCHCR"
is available online: if you have anon ftp access, connect to andy.bgsu.edu
and get the file pub/Hockey/InsideTCHCR.ps (Postscript) or pub/Hockey/
InsideTCHCR.sit.hqx (Word 4 on the Mac). If you want to receive it by
email, send a note to me directly. I can also fax it or snail mail it, but
I prefer not to.
 
If after reading "Inside TCHCR" you still have questions, I encourage you to
email me directly (there is enough stuff broadcasted over H-L already).
 
Although I do not speak for Erik or Tim/Robin, I also suggest you email them
directly if you have questions about their rating systems. Their email
addresses appear at the top and bottom of every one of their posts.
 
Keith [log in to unmask]

ATOM RSS1 RSS2