Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: more examples for search-based stupidity

Author: blass uri

Date: 15:11:03 06/12/98

Go up one level in this thread



On June 12, 1998 at 16:33:00, Thorsten Czub wrote:

>On June 12, 1998 at 15:16:50, blass uri wrote:
>
>>the programs were unable to handle the topic because they do not know
>>what lines to search.
>>knowing what lines to search is tactic.
>
>???? I think this is nonsense. If not knowing which line to search is
>tactics, than - per definition - anything is tactics.
>Thats silly.
>If your definition cannot differenciate between tactics and positional,
>than your definition sucks.

evaluation function is positional but I think that a computer program
with the same evaluation function who knows better which line to
search can find that it should not play h6xg5.

there are cases when computer program fails in evaluating the position
for more than 20 plies(there is no forced variations)
and in these cases to avoid the mistake it
must have a better evaluation function(knowing what to search
will not help practically because if you want to prove by a tree and by
the
the evaluation function the computer has that the mistake is a mistake
then the tree is too large (for example look at the first game kasparov
won
deeper blue)

>>I do not agree we measure how fast their searches are
>
>And why not ? You will see that programs like fritz and ferret will
>always make a good job in those suites.
>
>>there are test positions when fritz5 is faster than Junior4.6
>>not because it is a faster searcher(more than 30 times faster)
>>and there are test positions when the opposite happens
>>because in 1 positon fritz5 understand better what lines to search
>>and in the other position the opposite happens.
>>
>>Uri
>
>And who is the better tactician, when not finding "the right lines" is
>tactis, or here - NO tactics ?
>Your definition of TACTICS is as stupid as the definition of GOD or
>RANDOM.
>It says nothing.
>Saying: Anything is GOD is the same as saying anything is random or
>anything is tactics. Same level of argumentation.
>
>I don't think you will be able to increase the strength of a chess
>program by calling anything tactics/god/random.



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.