Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: ICC rating study

Author: James T. Walker

Date: 14:44:36 04/21/00

Go up one level in this thread


On April 21, 2000 at 16:58:08, Jeremiah Penery wrote:

>On April 21, 2000 at 08:56:55, James T. Walker wrote:
>
>>On April 21, 2000 at 03:08:38, Angrim wrote:
>>
>>>On April 20, 2000 at 08:34:05, James T. Walker wrote:
>>>
>>>>Hello Bob,
>>>>Interesting as it is, I think that to get a real feel for Crafty's strength on
>>>>ICC you need to change Crafty's formula to "Formula=rated".
>
>>>another point, if "Formula=rated" here are the results I predict:
>
>>>game 1. crafty is matched to a game of 3 0 wild17(losers chess) and of
>>>  course loses it.
>>>game 2. crafty is matched to a game of 3 3 blitz by a 1200 and wins 0 pts
>>>game 3-50 same as game 2.
>>>game 51. crafty is matched to a game of 3 0 crazyhouse and loses by flag
>>>since it doesn't understand the rules.
>>>game 52. crafty is matched to a game of 9999 999 by a 1400 who plays the
>>>first few moves of a game, then when crafty leaves book and starts
>>>thinking leaves on vacation with his account logged in(cable modem).
>>>Two days later Bob gets ICC admins to abort the game, even though the
>>>1400 still has plenty of time left on his clock.
>>>
>>>Angrim - Whose comp has 9 lines of formula on FICS
>>
>I was of course
>>talking about "CHESS".  What the hell does aborting a game have to do with
>>anything???   Who mentioned anything about 9 lines of formula??  Where did you
>>learn to read ??
>
>Hey, take it easy there.  You said:
>
>Interesting as it is, I think that to get a real feel for Crafty's strength on
>ICC you need to change Crafty's formula to "Formula=rated".
>
>And he pointed out a few things that happen if "Formula = Rated".  Ignoring the
>variants, you will still get idiots do things like this:
>
>1) A 1000 rated player will match Crafty to a 999 999 game.  He'll make a few
>moves until Crafty is out of book.  When Crafty comes out of book, it will start
>to think forever, and the player will leave.  He will never come back.  I have
>had similar things happen.  So until Bob come back to see what's happening, no
>games are getting played.
>
>2) A 1000 rated player will match Crafty to a million games in a row, at some
>fast time control.  Of course he'll get slaughtered every time, but that doesn't
>stop him.  This is absolutely not useful for the rating, because Crafty doesn't
>gain points and the player loses no points for it.  But Crafty just wasted a
>very long time in this series.
>
>3) A 1000 rated player will play a bunch of games and lose them all.  He'll get
>frustrated, so he decides to cheat using a computer.  He'll eventually win a
>game, then he might quit.  Crafty's rating just went down 32 points, yet that
>loss has nothing to do with what a true rating of its strength would be.
>
>4) There are a million other possible scenarios, just as absurd.  But they
>really happen.  This is the reason people set such long and convoluted formulas.
> Some use it to "protect" their rating, while some use it just to make sure
>decent chess is getting played.  These are reasons why Formula=Rated just
>doesn't work to provide a true rating of strength.
>
>Jeremiah

Hello Jeremiah,
Thanks for your insight into the problems of ICC chess.  I welcome any
reasonable rebuttle of anything I post here but some people just go off on the
deep end for no reason.  Your points are well taken.
Regards,
Jim Walker



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.