Author: Robert Hyatt
Date: 09:01:00 09/19/05
Go up one level in this thread
If you were to ask anyone here, at the beginning of a WCCC-type event, "who is the favorite?" the most common answer by far would be Shredder. Probably followed by Deep Junior and/or Deep Fritz (if they played). Why is that? Because Shredder has won several events in recent years? Because everyone watches shredder games on the chess servers and notice how well it plays in most all positions? Those are the _same_ criteria I would use in guessing how DB would perform today, based on how it performed when it was active, and how its predecessor performed under the same conditions... It really was a remarkable machine. All the more so back in 1996/1997. It might not have nearly the same edge over us today that it had back in 1997, but if we are going to speculate, then we at least have to speculate that development continued on DB. In 1996 I was doing about 80K nodes per second. Today I can do about 100X more. So Deep Blue should get at _least_ that same improvement, if not more. 20 billion nodes per second or more is just unimaginable... If they had continued software improvements as well, they would be doing beyond 20 ply searches, assuming null-move was added, probably some forward pruning since they did some of this in the hardware anyway, etc... It would be remarkably strong. Probably impossibly strong.
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.