Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: How Does HIARCS Work?

Author: Ratko V Tomic

Date: 10:10:05 09/30/99

Go up one level in this thread


> You should take a look at Chessmaster 6000, it plays for increasing
> its mobility and restricting its opponents... I notice this with
> the modified personality , CM6666.

I didn't notice it with regular CM (I noticed it plays odd looking moves, like
putting its bishop on d6 before d7 pawn was moved, and its positions often look
weird). I looked at the Shep's site for CM 6666 personality (few days ago), but
it doesn't seem to have a link for download. Have those files come back online
now?

All programs use in some form the mobility/control type evaluations. But these
evaluations span only the search depth, i.e. at most they'll pick something that
improves their mobility (or reduces opponents) within the search horizon,
perhaps 8-12 plies ahead (or wherever the end of capture/check extensions was).
That kind of short term positional plan is not very hard to see and work
against. But with Hiarcs I often find my position squeezed out of any
counter-play, with nothing good left to move, but without ever noticing how did
it happen, when did it start. So it is more subtle than usual short term contest
for mobility as you see in other programs. Hiarcs somehow considers the
potential long term paths of pieces and makes moves which cut or undermine your
piece paths well ahead the time they become interesting in a short term tactical
calculations. If I get that CM 6666 setting, I'll check it out just to see
whether the regular evaluation parameters adjustments in a regular alpha-beta
searcher can simulate such behavior.





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.