Author: Chris Carson
Date: 04:33:53 05/31/01
Go up one level in this thread
On May 31, 2001 at 02:56:10, Joshua Lee wrote: >As everyone has heard before Computers are not GM's as they can't get the title >but they are at a minimum Master strength and not only in the rarest of >Positional Games which few Players out there will be able to pull off anyhow. >Remember Deep Blue 1 if you ignore the first game as being a tactical one and >try to figure out the performance of DB you get approximately 107points below >it's 2642 Rating for the match 2535. I am sure somebody will argue that DB is >alot stronger than this and that is fine but the point is is if a computer with >192 processors and some of the best programmers in the field (not counting who >knows how many bugs as it still did play well) only plays 2535 at best and 2588 >as a rough average and compared to various commercial software >Deep Blue how fast compared to 1ghz PIII??? 2642 2535 2588 14Ply >Deep Junior 8x 700 or about 4.130Ghz 2702 2594 2648 ? > >As you can tell it is hard to compare available programs to ones that aren't >available. Also the Depth of search is different. This goes back to if Crafty >Shredder Ferret ETC searched 100Mnodes a sec will the depth be the same. >The answer is no, so my question is >1 at what Depth are various programs playing at GM strength >2 at which depth are they comparable to the Deep Blues ...... >3 What Programs are best with Positional Play and is their play as good as deep >blue or better, if not what ply depth is needed. Interesting questions. There have been discussed here at length, but not lately. It will be interesting to see the posts. I have my own opinion and researched data. :) If you take a look at the last Monty Newborn book on the Deep Blue(96) vs Kasp. match, you will find some analysis on this. Levy and Newborn also wrote an earlier book with additional information. :) I have done some research (published here) in this area using nodes and depth based on LCTII results, different programs, SSDF and HvC ratings compared. I have published these before, much debate followed. Usually resulting in name calling about how stupid such comparisons are. I do not think research is stupid, but some people do. I quit posting when the discussion stops and the name calling begins. :) I asked about test positions for Deep Blue to compare nodes, depth, HvC ratings data, but no test position can be found for Deep Blue for comparison, just the published 100M NPS and 200M NPS (DBII) average and the 12-14 ply from game logs. More name calling on this subject, so I quit posting, maybe different this time and this group can just talk about this. Some smart people on both sides of the arguments. :) Best Regards, Chris Carson
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.