Author: Christophe Theron
Date: 01:16:37 02/27/04
Go up one level in this thread
On February 26, 2004 at 20:37:11, Russell Reagan wrote:
>I have a question for chess programmers.
>
>I think most engines have a maximum search depth (most that I have seen anyway).
>The only one I've heard of that doesn't have a maximum depth is Steven Edwards
>program Symbolic.
>
>I was wondering if there would be any advantage to having no maximum search
>depth. If you could, for instance, dynamically allocate more space to continue
>searching deeper, and it would cost you very little (mostly pre-allocated in
>non-time-critical parts, in between moves, or whenever), would you even want to
>do that? Or is there really no advantage to be gained there? It doesn't seem
>like you would miss much if you cut off your qsearch at 64 plies.
>
>On the other hand, are there any practical disadvantages to having no search
>depth limit? I could see a potential situation where the qsearch explodes.
>
>My guess is that it's probably not a big deal to have a limit, since you might
>only cut off some very, very deep qsearch nodes, but a hard cutoff like that
>seems to be a little hackish, and not entirely correct to me.
You should not count on the maximum ply depth to avoid qsearch explosion anyway.
So in theory there is no need to have a maximum ply depth (you must have some
other limits in order to avoid search or qsearch explosion).
In practice it's probably pointless to go deeper than 64 plies deep, unless you
are trying some special check-check mate in 40 or some weird endgame study.
There would not be any measurable strength difference between a program that
limits the search to 64 plies and one that would limit it at 256 plies.
Christophe
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.