Author: Uri Blass
Date: 12:23:04 03/01/04
Go up one level in this thread
On March 01, 2004 at 14:27:04, Robert Hyatt wrote: >On March 01, 2004 at 14:15:45, Leen Ammeraal wrote: > >>I am currently playing a match of my engine with pondering on >>against a special version of it which does not ponder even if it >>is allowed to do so. Apart from this, the two engine versions >>are identical. I do this experiment just to see if pondering >>works correctly, but I have no idea by how much points I >>should expect the pondering version to be stronger. >>Are there any other engine authors who have done this >>experiment for their engines? >>Leen > > >from 30 to 100 points. At best, the pondering program will be 2x faster than >the non-pondering program, if it ponders correctly every time. At worst, it >will be only a little faster as it is probably impossible to never ponder >correctly in comp vs comp. > >2x faster is anywhere between 50-100 rating points depending on the program, >roughly. If the program has bugs being twice faster may be less than 50 elo better and even a negative number. I suspect based on what I read that it is the case for at least one free program when the time per move is long enough(RDchess). Uri
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.