Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: What is horizon effect?

Author: Robert Hyatt

Date: 08:52:53 01/30/06

Go up one level in this thread


On January 30, 2006 at 03:59:07, Renze Steenhuisen wrote:

>On January 29, 2006 at 23:10:30, masros tukiran wrote:
>
>>What is  phenomenon called "horizon effect" in chess computer?.I don't
>>understand what it is.Can somebody give me an example to desribe it.Thank you.
>
>You can only make a decision based on what you know, see, or so.
>It is not different with computers.
>When something is unknown to you but would have changed your decision if you
>knew, then going from the situation that you don't know  to the state that you
>do know is called horizon effect.
>
>Or sometimes people call the horizon effect just the former situation, that a
>wrong decission is made because the engine is uninformed, or at least
>under-informed.
>
>Cheers,
>
>Renze


Not quite correct.  The "horizon effect" is something caused by a fixed search
horizon that the search can't penetrate.  If the search can therefore force
something out beyond the horizon, in effect "delaying the problem for a while"
that delaying can appear to be a "permanent solution" to the problem, since if
the problem is pushed beyond the horizon, so far as the search is concerned, the
problem no longer exists.

search extensions are one way to mitigate this, but it is impossible to
eliminate.

These "big searches" are quite good at pushing things off far enough that they
appear to disappear, only to come back on the next move...



This page took 0.01 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.