Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Someone Better Stop Tiger or This GM Debate is Going to be over Quic

Author: Chessfun

Date: 11:41:03 06/26/01

Go up one level in this thread


On June 26, 2001 at 14:31:18, Robert Hyatt wrote:

>On June 26, 2001 at 14:13:20, Uri Blass wrote:
>
>>On June 26, 2001 at 12:12:21, Robert Hyatt wrote:
>>
>>>On June 26, 2001 at 06:48:46, Martin Schubert wrote:
>>>
>>>>On June 26, 2001 at 06:44:44, Mark Young wrote:
>>>>
>>>>>On June 26, 2001 at 06:38:30, Martin Schubert wrote:
>>>>>
>>>>>>On June 26, 2001 at 06:30:58, Mark Young wrote:
>>>>>>
>>>>>>>On June 26, 2001 at 06:23:07, Gian-Carlo Pascutto wrote:
>>>>>>>
>>>>>>>>On June 26, 2001 at 05:15:56, Mark Young wrote:
>>>>>>>>
>>>>>>>>>I don?t think this that arguments passes the laugh test, most players? trainee
>>>>>>>>>with some kind of computer aid, chess bases, programs etc. They have too, to
>>>>>>>>>compete in today?s chess tournaments. Programs have been around for many years,
>>>>>>>>>and I would doubt this is any players first time seeing a chess computer.
>>>>>>>>
>>>>>>>>Let me tell you about some of my real life experiences with this.
>>>>>>>>
>>>>>>>>When Mat(h) (Belgian program that competed in Leiden) was first
>>>>>>>>publicly presented, an exhibition match was played vs about 34
>>>>>>>>volunteers.
>>>>>>>>
>>>>>>>>It was quite well publisided so players up to 2250 ELO joined (this
>>>>>>>>is much for Belgium).
>>>>>>>>
>>>>>>>>In the games, most opponents were crushed. Although the program is
>>>>>>>>not very strong at all (look at the Leiden results), even the strongest
>>>>>>>>players lost without much chances, or had to content themselves with
>>>>>>>>a draw.
>>>>>>>>
>>>>>>>>There was one, at that time 1300 rated player, that did quite the
>>>>>>>>opposite. He crushed it. Rather easily even. (I knew this player and
>>>>>>>>had been watching the match closely)
>>>>>>>>
>>>>>>>>How come this 1300 did so well even though the 2200's had trouble?
>>>>>>>>
>>>>>>>>This player sometimes played on chess severs, and was familiar with
>>>>>>>>chessprograms, and had played them before. He knew where they were
>>>>>>>>weak. In the Mat(h) game he castled on opposite sides and stormed
>>>>>>>>Maths kingside. The program had no clue what was going on and lost.
>>>>>>>>
>>>>>>>>Most of the 1800-2200 players knew Fritz quite well. They used it to
>>>>>>>>look up games. They used it for analysis. They sometimes played in
>>>>>>>>sparring or friend mode. But they never played it seriously full-
>>>>>>>>strength. They had no idea how to beat the program. Most of them
>>>>>>>>played totally open postions. I saw nearly no closed manoeuvering
>>>>>>>>games. They were just playing like they played other humans. And it
>>>>>>>>didn't work.
>>>>>>>
>>>>>>>Is this what is happening with Chesstiger, the IMs and GMs are just Ignorant
>>>>>>>on how to cope with ChessTiger.
>>>>>>>
>>>>>>>I guess there is no way to prove that, so your safe if that is what you believe.
>>>>>>
>>>>>>Oh yes, there is a way to prove that. Just let someone play who is experienced
>>>>>>with playing computers. I know someone how offered to play against a computer
>>>>>>and he wanted to show he is better than a GM. But nobody was interested in such
>>>>>>a match. You know why? Because it's not PC. The only thing you can learn from
>>>>>>that is that computers are not that good as everyone tells. But that doesn't fit
>>>>>>in the "Computers have super-super-GM-strength"-hype.
>>>>>>Just look at some games posted here, how to beat programs with 2.Na3 for
>>>>>>example.
>>>>>>A few month ago Eduard asked to give him any opening and any program. He wanted
>>>>>>to beat that program in that opening. He did. Very soon. Do you think that would
>>>>>>be possible against a GM?
>>>>>
>>>>>Look at what you are saying. Yeah anyone can sit and find a win this way at home
>>>>>against one line of play. Yes it is easy to exploit the computer programs under
>>>>>those conditions.....SO WHAT! My grandma could do that and she is dead.
>>>>>
>>>>>The trick is being able to win under tournament conditions where you can not
>>>>>control the lines of play.
>>>>>
>>>>
>>>>It's not a problem of lines! It's a problem of understanding what is important
>>>>in anti-computer-chess comparing with what is important in usual chess.
>>>>Do you think 1. e4 c5 2. Na3 or 1. e4 c5 2. e5 is a problem of lines, you can't
>>>>get in tournaments? Of course you need ideas how to beat computers not only in
>>>>the sicilian. But it's impossible to put all these things into an opening book.
>>>>
>>>>
>>>
>>>
>>>You are wasting your breath.  Some believe that (a) you can learn how to beat
>>>a program by doing it over and over, but that doesn't mean it isn't a GM player.
>>>They don't realize that they could play a GM under the same circumstances, over
>>>and over and _never_ win a single game.
>>
>>I doubt if they never win a single game against the GM.
>>
>>The difference is that the program is available to play every time they want
>>when the GM is not available to play 24 hours in a day so of course humans can
>>play more against a program and not against a GM.
>>
>>I know about an 2100 and 2000 players who beated  GM's at tournament time
>>control so it is not impossible.
>>
>>Uri
>
>
>You aren't going to beat the GM by playing the same opening over and over,
>and finding improvements in your play until you win.  The GM will react too,
>and _he_ will also change moves here and there to improve his position even
>more, even though he won the previous 20 games in a row...
>
>that is an important distinction between how humans learn and the _ridiculously_
>primitive forms of learning we do in the chess engines.


Exactly correct.

Sarah.



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.