Author: stuart taylor
Date: 17:07:48 02/15/00
Go up one level in this thread
On February 15, 2000 at 12:42:05, blass uri wrote:
>On February 15, 2000 at 12:12:56, stuart taylor wrote:
>
>>On February 15, 2000 at 11:49:41, Dan Ellwein wrote:
>>
>>>On February 15, 2000 at 10:02:40, blass uri wrote:
>>>
>>>>On February 15, 2000 at 09:25:02, Côme wrote:
>>>>
>>>>>On February 15, 2000 at 08:53:27, blass uri wrote:
>>>>>
>>>>>>On February 15, 2000 at 07:16:12, Côme wrote:
>>>>>>
>>>>>>>On February 15, 2000 at 06:16:27, blass uri wrote:
>>>>>>>
>>>>>>>>On February 15, 2000 at 05:51:31, stuart taylor wrote:
>>>>>>>>
>>>>>>>>> I wonder if it will eventually be discovered that chess is not absolute,
>>>>>>>>>and that a human will therefore always be able to beat a machine by playing
>>>>>>>>>exactly against the weaker points of that particular machines style-everything
>>>>>>>>>else being to perfection?
>>>>>>>>> Maybe chess isn't an exacting art-absolutely?
>>>>>>>>> Stuart Taylor
>>>>>>>>
>>>>>>>>If the machine has no weak points then it is impossible to play exactly against
>>>>>>>>the weaker points of the machine and there is no reason to assume that it is
>>>>>>>>impossible to do a machine with no weak points.
>>>>>>>>
>>>>>>>>Today there is no machine with no weaks points and there is no machine that can
>>>>>>>>pass the turing test(every machine can do positional mistakes that I do not
>>>>>>>>expect humans even with 1800 elo rating to do) but it is practically impossible
>>>>>>>>for most of the humans to play exactly against the weaker points of the machine
>>>>>>>>because you cannot go practically to the positions that the machine does not
>>>>>>>>understand.
>>>>>>>>
>>>>>>>>Uri
>>>>>>>
>>>>>>>Hello Uri,
>>>>>>>I don't agree with you Uri !
>>>>>>>It's not so hard to play against weaks points of machine !
>>>>>>>Best Regards
>>>>>>>Alexandre Côme
>>>>>>
>>>>>>Hello Alexandre,
>>>>>>If it is not so hard then what is the reason that these machines can beat more
>>>>>>than 99% of the humans with rating above 1600?
>>>>>>
>>>>>>There are positions when the machines are stupid and humans knows about them(for
>>>>>>example programs do not understand fortress positions when one side has a big
>>>>>>material advantage) but you usually cannot go for these position in
>>>>>>a practical game.
>>>>>>
>>>>>>Uri
>>>>>
>>>>>Hello Uri,
>>>>>Ok you must be stronger than 1600 to play against weaks points of machines. :-)
>>>>>There are really a lot of positions when the machines are stupid and IMHO
>>>>>machines will not be never stronger than humans.
>>>>>Best Regards
>>>>>Alexandre Côme
>>>>
>>>>I am stronger than 1600 and my rating is 2025 but even players with 2300 get
>>>>less than 50% against the machines.
>>>>
>>>>If it was easy to play against the weak points of the machines they could win
>>>>the machines.
>>>>
>>>>I agree that there are a lot of positions when the machines are stupid but
>>>>I do not agree that machines will never be stronger than humans in these
>>>>positions.
>>>>
>>>>Uri
>>>
>>>chess is 'absolute' in the sense that you have a finite amount of space - 64
>>>squares...
>>>
>>>and a finite amount of material - 32 pieces...
>>>
>>>finite times finite (usually) equals finite...
>>>
>>>computers have not (yet) reached this finite-ness...
>>
>> What I was wondering was something else. Perhaps there is no such thing as
>>perfect chess play in the sense that it cannot possibly be beaten by someone who
>>has prepared himself perfectly to play against the other persons style.
>>We don't know that yet, but I don't know if many people have ever thought
>>of it either.
>
>We know that there is a strategy to get at least 50% against everyone.
>We do not know the stratefy but we know that there is a strategy to do it.
>
>people thought about the problem of games in mathematics.
>
>
>> Why for example does the (possibly) greatest ever chess playing entity
>>Kasparov have to prepare so much for every opponent? Can't he just play perfect
>>chess?
>
>He cannot because he does not know the best move in every position.
>He also wants to win and sometimes for practical reasons perfect chess does not
>give the best chance to win.
>
>There are also many possibilities to play perfect chess and the probability to
>win is not the same in all of them.
>
>> Could it not be possible that chess is only 99% or 99.999% exact, and
>>ocassionaly the most perfect chess player-who has not yet existed-can also
>>lose a game since there might be weak aspects to any style-even at perfection
>>level?
>> S.Taylor
>
>No
>Chess is 100% exact.
>There are clear rules and every position has a clear result(a draw or a win for
>white or a win for black) if both sides do not do mistakes.
>
>Uri
O.k. then! Maybe! So the aim of machines is to play perfect chess. It has to,
so that no opponent will ever beat it, what ever he attempts. Whereas the aim of
human chess is to outwit the opponent-since even Kasparov cannot aim for higher
(although Fischer claimed that he could-but maybe fooled himself).
Therefore machines are slated to overtake humans by never losing a game,
even if the games may not be all that spectacular-atleast not romantic-as
perfect chess doesn't usualy allow much romanticism in equal positions.
And in a few years time it will only be the one or two masochists-top human
players-gluttons for punishment, trying to see if there are any holes left in
computers by which to lead them astray-usually becoming a cropper in the
proccess. How charming!
S.Taylor
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.