Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: diminishing returns

Author: Christophe Theron

Date: 19:42:45 01/30/04

Go up one level in this thread


On January 30, 2004 at 18:34:07, GuyHaworth wrote:

>
>Using the SSDF data to compare same_prog/different_hw is an interesting idea.
>
>I must correct my recollection of Heinz' results, having looked them up now.
>The 12-11 match showed an advantage of 84 ELO.
>
>I reckon one ply extra requires 4x the computing power as effective branch-out
>with all the good techniques is ~3, quite a difference from ~30.



An extra ply should rather be in the 2.5-3 range with state of the art
techniques.




>So halving the time is probably worth about 40 ELO,



If you consider that an extra ply takes 4x, but that's not the case.



    Christophe



> less given that the machine
>- in a man-machine game - would have the opponent's time to think in, and the
>opponent would also have a lower ELO because they only have half the opponent's
>time to think in.
>
>g



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.