Author: Robert Hyatt
Date: 07:41:34 09/21/05
Go up one level in this thread
On September 21, 2005 at 09:48:01, ALI MIRAFZALI wrote: >On September 20, 2005 at 13:22:16, Robert Hyatt wrote: > >>On September 19, 2005 at 09:38:10, ALI MIRAFZALI wrote: >> >>>On September 18, 2005 at 14:19:47, Chan Rasjid wrote: >>> >>>>I forget something :- >>>> >>>>You cannot match a GNU chess with some very basic fundamental flaw (may be >>>>undiscoverd) at 1000 x nps with a mature good program. 1000x nps that chooses >>>>a losing line will still lose some games. >>>> >>>>So the only proper test must be w/o bugs, ie. >>>>material + mate vs material + mate + best eval() on planet earth. >>>> >>>>Rasjid >>>I will say this again and again:There was nothing flawed about GNU chess 4 or >>>the experiment!!! >> >> >>you can say it all you want. But it won't make it true... >> >>That's the point you are overlooking... >Wont make what true? You are a master of minipulating words and the things you >have said in the past.Are you saying that you do not believe these 4 games are >possible.Or maybe you just cannot accept defeat.Sad but true;essentially Gnu >chess 4 is a normal program WHICH GOES PRECISELY AGAINST your claim in the past >that 1000X is not possible between reasonable programs.How many people in this >forum can beat GNU 4? You are certainly not one of them. I have beat it many times in fact. It was not particularly strong when I did play it often back in the late 90's. I doubt I would beat it today on modern hardware, since I don't play tournament chess much any more. And sadly, gnu 4 is _not_ a "normal program" by any measure. It was not as strong as almost any program in the time when it was being developed. It certainly is not as strong as anything today. Try your experiment with palm tiger vs something current. Even Crafty. Then report back when you have some real data...
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.