Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Depth vs Time

Author: Aaron Tay

Date: 18:24:18 06/26/02

Go up one level in this thread


On June 26, 2002 at 14:10:17, Steve Coladonato wrote:

>On June 26, 2002 at 12:51:05, Daniel Clausen wrote:
>
>>On June 26, 2002 at 07:26:10, Steve Coladonato wrote:
>>
>>[snip]
>>
>>> And say I also want to have the analysis go to 11 ply.
>>
>>And that's exactly the point where you go wrong. It makes not much sense to want
>>to have the analysis go to 11 ply. Maybe what you want is to have an analysis
>>where the engine sees all tactical combinations within 6 moves or something like
>>that. But a 11 ply search doesn't necessarily do that - it depends on the
>>engine. It also depends on the position, and what not.
>>
>>If you want to analyze a game and you don't want to spend more than X seconds
>>per position, do exactly that - limit the time the engine should take for a
>>position accordingly.
>>
>>Sargon
>
>There is a threshold where the engines take a major factor to get to the next
>depth in their analysis.  If I know that an engine can do an 11 ply search in N
>minutes then I can set my time accordingly.  However, if I just limit the time
>to say 3 minutes, the engine may only be capable of getting to ply 8 in that
>time.  So I've wasted an analysis because the depth is too shallow.  I'm looking
>for the threshold where getting to the next depth is just not worth the
>exorbitant amount of time to get there.


You want something like an "estimated time to reach next depth" ?
I wonder how accurate such estimates if possible would be..

How about estimated possibility of changing it's mind when it finishes this ply?








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.