Author: Mark Young
Date: 02:05:47 12/09/98
Go up one level in this thread
On December 09, 1998 at 03:04:13, Dave Gomboc wrote: >On December 08, 1998 at 23:15:09, Mark Young wrote: > >>Not at all, I just don't underestimate the depth of chess. And I know how all >>computer programs find chess moves. And 1 Billion position a second is not going >>to cut it at that time control. Not when each new ply is exponentially larger. >>No program yet can select the few good lines and just search just them, like a >>strong human play can do given 1 move per day. The program would be badly out >>searched by Kasparov. One billion positions a second is not even a drop in the >>bucket, given that level of play. > >My _belief_ is that Deep Blue is really good at searching the shit out of the >really important variations, as compared to the fluff. I don't think it is that much better then some of the other programs. If it were I think it would have had a much higher level of play on the hardware it was on. It beat Kasparov at 3 min a move in a short match, but not convincingly. And it will only lose ground to Kasparov as the time controls get longer. I suppose it will take >more than my belief to convince you of that, though. :-) And justifiably so. To bad we can not see a match between Junior 5 (5 days per move) vs Deep blue( 3 min a move). I think that match would be telling. > >I was interested by a comment Amir made some time ago, something like that he >_knew_ that Junior extended more aggressively than Deep Blue did. That isn't a >direct quote, and I hope my memory hasn't retained a sentence meaning something >different from what he actually said. In any event, I imagine it would be easy >enough for him to know if Junior extends moreso than the team did back around >1991, when they published articles in the ICCA Journal. What I wasn't sure >about was how he would know that Junior extends more than their current >software. Did the printouts from Kasparov - Deep Blue lend credence to this >assertion? > >Dave Gomboc
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.