Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Making a computer play more like a human

Author: David Blackman

Date: 00:45:32 09/30/99

Go up one level in this thread


On September 29, 1999 at 20:52:37, Michael Fuhrmann wrote:

>          From my limited understanding of computer chess, I gather that
>programs choose moves based on the assumption that their opponents always make
>the best possible moves. This may not happen -- especially in the case of a
>human opponent. Which leads to the paradox that computers might win more games
>if they sometimes played weaker moves.
[ cut ]
>             So (a) Is any computer today capable of choosing this kind of
>technically unsound, counter-attacking move (in positions where it will probably
>lose anyway given best play) instead of the "correct" move according to its
>eval?
>             And (b) Does anyone else agree that if a computer could do this, it
>would add a human-like, "psychological" dimension to its arsenal?

Various versions of my program Desperado have tried to play traps and swindles.
The most recent version has been tuned up to play in a computer only tournament
however, and testing suggested this is not a good way to play against other
computers, so it has been turned off.

I tried a few trapping methods, but the one that seemed most effective was when
doing a N ply search for the computer, only to try opponents replies in the
search if they looked ok at N-2 ply. (Then you would try them again as N-1 ply
as normal.) This was done at several levels in the tree.

This produced some spectacular crushes against players of about expert strength,
but a game by IM Guy West showed that strong human opponents could keep the game
under control, at least in some cases. Without the trapping code, it plays very
boring chess and puts no pressure on the opponent unless they blunder.

There have been rumours of some better known programs also being programmed to
play traps. Aren't there "Anti-GM" and "Anti-patzer" modes in Rebel that are
supposed to do this? Also i suspect Deep Blue did this in at least some of the
games against Kasparov. And if it didn't, it should have.



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.