Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: A Sober Estimate of the Strength Of Modern Programs

Author: Derrick Williams

Date: 18:48:32 03/13/00

Go up one level in this thread


On March 13, 2000 at 17:25:45, Dann Corbit wrote:

>On March 13, 2000 at 16:48:39, Derrick Williams wrote:
>>
>> How many agree with these estimates for the strength of Modern programs on
>>various platforms?
>>
>>1. Deep Junior Quad Xeon    2590 fide
>>2. Fritz6 k6-450  2480-2530 fide     k6-233  2450 fide
>>3. Rebel century k6-450    2500-2560 fide     k6-233  2450 fide
>
>Mathematically, your estimates are pointless.  As expressed, they are
>meaningless.  What is the time control?  What is the standard deviation for the
>figures.  With the current number of games played with exactly this setup, I
>should think the standard deviation will be absurd.
>
>For all of these programs, eventually we should get some good numbers.  Probably
>not on that hardware, though.



  I meant quad xeon 400 thinking of the machine Crafty runs on at Icc. I was not
attempting to be scientific or mathematical, I was only interested in expressing
a guess based on intuition and expierence playing the machines. It may be
meaningless to guess, but it is extremely interesting. Regardless of whatever or
not there is enough hard evidence to base an opinion, people like myself want to
get some kind of ideal (even if that Ideal is vague), of what the playing
strength of their silicon opponents are.  In many circumstances Theory must
always come before fact, my theory is computers are GM's, based on intuition and
a Good amount of 40/2 games.
>
>"Quad Xeon" does not mean anything by itself.  There are many different CPU
>speeds, memory configurations, etc. for the Xeon versions of Intel processors.
>
>I don't know why people like to spend so much time saying that computers are
>GM's or computers are not GM's or computers are this or that ELO.  Eventually,
>we will have enough data for an answer.  It appears (from the Rebel matches)
>that Rebel on a 600 MHz machine is just under 2500 ELO [IIRC] but that number
>should become more refined over time.



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.