Author: Christophe Theron
Date: 12:47:59 10/11/99
Go up one level in this thread
On October 11, 1999 at 09:29:25, blass uri wrote:
>On October 10, 1999 at 15:49:09, Thorsten Czub wrote:
>
>>On October 10, 1999 at 15:35:53, Enrique Irazoqui wrote:
>>
>>>It is not very active, it likes to play cat and mouse without doing much until
>>>it sees something in the search, but it is very efficient.
>>
>>I do have a different opinion. the games i get are different from yours.
>>tiger plays in all games very active and having initiative.
>>it forces the wins. it makes the game.-
>>do you test a different program ?
>>older tiger versions were passive. not 11.2 and not 12.0.
>>
>>
>>
>>>I start thinking of
>>>it as the opposite of CST in every way.
>>
>>??
>>
>>
>>>I guess that if someone likes CST or
>>>Mchess won't like Tiger so much, and viceversa. Do you agree? Now that I think
>>>of it, Thorsten likes both, CST and Tiger, so I may be wrong.
>>
>>I like the games. it is not important HOW a program gets a beautiful
>>game, it is important THAT it plays beautiful chess games.
>>i am not interested in games that are dump, produced by dump programs.
>>by boring games between junior-fritz or nimzo-fritz or whatever.
>>i am interested in planful games . hiarcs and mchess and ctiger and rebel
>>and virtual and cstal and and and produce those interesting games.
>>this is what i like.
>
>I believe that all the programs are dumb and have no plan.
>programs also calculate a tree when both sides move and I know no program that
>calculates things like I am going to play Be7 (in order not to let the opponent
>to play Bh6-f8) after it Kh8 after it Bg8, after it Bh7 after it Bg6 and after
>it I control the important diagnol (h7-b1)so I can go forward with the king and
>play Kh7 and win the game because the bishop at h6 i trapped.
What you are talking about here is what programs do when they use recursive null
move or related search techniques.
You begin by "what can I achieve if my opponent does not move?", and then when
you have found something interesting you ask youself: "is it possible to do it
even if my opponent moves and tries to prevent me from doing it?". Null movers
do the same.
Nothing new.
After all, maybe programs think a little bit more like humans than you seem to
believe.
>They may discover ideas like this by search but if the plan is long enough then
>they have no practical chance to discover it.
>
>Some programs have better evaluation function but I believe that all of them are
>based only on search of lines when both sides play and this is the reason that
>humans are sometimes superior relative to programs (mainly in the endgame) and I
>see sometimes stupid mistakes by all programs (including Tiger and Hiarcs) in
>the endgame.
>
>Uri
No, that's not the reason why human are (maybe) still superior, because many
programs do that already. Or almost the same in a different way.
Christophe
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.