Author: Christophe Theron
Date: 18:56:52 10/09/03
Go up one level in this thread
On October 09, 2003 at 02:37:24, Thorsten Czub wrote:
>On October 09, 2003 at 02:16:37, Dana Turnmire wrote:
>
>>Thanks for the explanation though I'm not sure who the original poster was of
>>this. It just didn't make sense that an evaluation function that was supposedly
>>inherently defective could propel Richard Langs programs to ten world titles.
>>Regards.
>
>the method worked to come deeper in the tree because you did not have to
>compute ALL moves in your own plies.
Thorsten I really think that it is not what Richard did.
Not looking at all of your moves in all the search tree (or a significant part
of it) does not work. It provides only a marginal speedup and has huge
drawbacks.
Genius has an asymetrical search only in the last few plies of the search. It
does not look like what you are describing.
In the position posted in this thread, search asymetry is not the reason why
Genius could not find the key move. The real reason in this case is asymetry in
the evaluation.
Christophe
>In the moment many programs used NULLMOVE the method was not
>effective anymore because the programs with nullmove came deeper than
>genius.
>
>from this time, genius did not win anything anymore.
>
>thats the reason the success it had over years stopped.
>the others simply outsearched it.
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.