Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Which Program Would do the Best on ICC?

Author: Robert Hyatt

Date: 20:09:54 06/28/99

Go up one level in this thread


On June 28, 1999 at 16:38:15, James T. Walker wrote:

>On June 28, 1999 at 11:00:40, Robert Hyatt wrote:
>
>>On June 28, 1999 at 09:57:20, Bruce Moreland wrote:
>>
>>>
>>>On June 28, 1999 at 01:27:53, blass uri wrote:
>>>
>>>>
>>>>On June 27, 1999 at 16:43:59, Mark Young wrote:
>>>>
>>>>>On June 27, 1999 at 15:49:53, odell hall wrote:
>>>>>
>>>>>>Hi
>>>>>>
>>>>>>
>>>>>>  I am thinking about setting up a computer account on ICC? Does anyone know
>>>>>>Which Program would  do the best? If the choice was between Genius 5 , hiarcs6,
>>>>>>Mchess 8 or Genius 3? I have heard that Genius 5 is the Best a blitz chess
>>>>>>(game/15) is this still true. What program would tear the crafty clone Data to
>>>>>>shreds!!  (incredibly strong data is!) Or should I just go out and purchase
>>>>>>Chessmaster 6000 to do the Job?
>>>>>
>>>>>The best results I have had is with Hiarcs 7.01. I have not had a chance to play
>>>>>7.32 online because of the testing I am running now. I most likely will only
>>>>>test Hiarcs 7.32 against humans online, after testing Hiarcs 7.01 online against
>>>>>other computers its left a bad taste in my mouth. I try to test against the best
>>>>>programs online and play a set amount of games, but with Hiarcs 7 this became
>>>>>almost impossible because of the high winning percentage Hiarcs 7 was
>>>>>generating. The automatic programs would no play me, or change the time formula
>>>>>to only play 0 inc games in the middle of my testing, and some of the other
>>>>>manuale computer operators were no better. Rating protection in online play is
>>>>>only getting worse, and thats to bad because it was a nice tool to use in the
>>>>>evaluation of chess programs.
>>>>
>>>>I do not understand it.
>>>>I think that programmers should prefer to play against strong opponents because
>>>>they can learn from the games that they lose to improve their programs.
>>>>
>>>>Uri
>>>
>>>He is not talking about programmers.
>>>
>>>bruce
>>
>>
>>Bruce, myself, Stanback, Ban, plus the usual cast of the "under 2600 club" here
>>play on ICC and do just what you suggest.  But we are a tiny bit of the total
>>number of computers playing there.  There are dozens of Fritz, Hiarcs, shredder,
>>and so forth (not to mention crafty, comet, tcb, etc) that are run by _others_.
>>And _they_ are not working on improving their programs.  Many work only to
>>improve their 'rating' which is a point I don't understand.
>
>He may be talking about programmers.  If the shoe fits, wear it.  Programmers
>are not immune from the action he is talking about.  There seems to be a rating
>point contest on ICC and programmers seem to be caught up in this action too.
>It seems like programmers can't separate their egos from their programs either.
>There are many childish "games" being played on ICC with computer accounts.  I
>run a computer account on ICC as a hobby.  I enjoy watching the different
>programs play each other.  It is a long time hobby of mine that goes back to my
>first two chess computers (Chess Challenger 10 vs Sargon 2.5).  I can play Fritz
>vs Crafty at home but I don't have a "Quad Xeon" so I don't get the same quality
>of play I would get if I get a chance to play Fritz/Hiarcs vs Crafty. So I enjoy
>the games on the Internet.  And If I get a chance to watch a GM / IM play
>Fritz/Hiarcs it's a real treat for me!  That's what I get out of running a
>computer account.  And Bob, I got to tell you, your words ring a little hollow
>on this point.  I'm still on the Crafty "Noplay" list and my only "Crime" is
>that I played 5 games vs Crafty.  This after I explained, that there was nothing
>in your notes about this and it was just a misunderstanding.  Is this your way
>of protecting your rating points?

Silly question.  If I wanted to protect my rating, first program I would +noplay
would be ferret,  yet we play regularly when Bruce is there.  I would also dump
the other fast programs (machinehead, shredder 3 on a 500+ megahertz machine.

I _never_ +noplay for rating.  I do have a custom interface that catches the
fact that I am playing a computer, and when it does, after 4 games it will
automatically +noplay you, although it will -noplay you 4 hours later.  In
your case, it is possible that either ICC went down, or I logged crafty out
before the automatic -noplay.  However, my notes have _always_ had the 4 game
limit.  I apparently deleted it recently...  which was my fault.  But this has
been in crafty's notes for at _least_ 2 years, so it isn't 'news'...




>I, like the above poster, was under the impression that programmers would like
>to test their programs against top level competition to look for flaws.  I would
>think you would learn more from one loss vs Fritz than winning 16 of 17 vs a FM.
> I thought it would be better if the progams were actually playing chess rather
>than just sitting idle waiting for a GM / IM to come along.  Apparently I was
>wrong.
>Jim Walker


I am _much_ more interested in playing humans.  And when I log on and get a
complaint from a GM about "I was playing a match and a computer jumped in
before I could match again" that computer becomes +noplayed and it _stays_
there.  Because I can play computer games whenever I want here off-line, but
on ICC I get the chance to play 50 games vs GM players _every_ day.  And I'd
much rather sit idle for an hour waiting for them to come along.  I don't pay
much attention to games vs computers because I (a) don't know much about the
hardware; (b) don't know what influence the operator exerts.  If rating was
all I was interested in, however, I could absolutely get to 3300+ and _stay_
there.  No computers.  No increments. And it is easy...

That isn't my goal, however...  solving 'anti-computer' _is_ and I am not going
to do that playing computers 24 hours a day...



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.