Author: stuart taylor
Date: 23:11:53 01/22/00
Go up one level in this thread
On January 21, 2000 at 09:47:11, Robert Hyatt wrote: >On January 21, 2000 at 02:10:26, stuart taylor wrote: > >>On January 20, 2000 at 22:20:14, Robert Hyatt wrote: >> >>>On January 20, 2000 at 22:09:37, Robert Hyatt wrote: >>> >>>>On January 20, 2000 at 18:53:15, Chris Carson wrote: >>>> >>>>>On January 20, 2000 at 17:55:06, Robert Hyatt wrote: >>>>> >>>>>>Doesn't matter. What appears strong to a class B player doesn't necessarily >>>>>>appear strong to a master. DB pretty well proved itself rating and performance >>>>>>wise. Both against humans and computers... >>>>> >>>>>Dr. Hyatt, >>>>> >>>>>I have always treated you with respect. I hope you will >>>>>take back this tasteless comment. >>>>> >>>>>Best Regards, >>>> >>>> >>>>>Chris Carson >>>> >>>>What is "tasteless". The point being that to "us" a computer might look >>>>impossibly strong. To a GM, it might look impossibly weak. And _both_ would >>>>be right. Although I doubt DB looks "weak" to any GM really looking... >>> >>> >>>That was worded backward. To a weaker player, DB might look pretty close to >>>equal with PC programs. To a GM, it doesn't. >> >> >>Mr. Hyatt, >> Many positions and moves have already been compared, and DB is not always >>more correct than software. > >That happens. But no computer has beaten Kasparov in a match. No program >has beaten the old Deep Thought more than one game, period. I have posted >DT's results over a 10 year period, which included 2 losses to microcomputer >programs... and a bunch of wins... > > > >> And regarding what weak players can see (however you put it) are you >>personally a stronger player than us? If you don't know the answer to that >>question, then how can you make such comments? > >I make the comment from two perspectives: (1) Kasparov commented that DB >was something "new" after game one, a computer with a whole new "view" of >the game. He hasn't said that about _any_ micro. Other GMs have said >similar things. So the GM opinion is that DB is very strong. Maybe all don't >think it is as strong as the top 10. But those same GMs don't think the micros >are as strong as the top 200 either. > >(2) I am a decent chess player. And I am a more-than-decent chess programmer >with a lot of experience. And I have had the opportunity to _actually see_ >the old deep thought program/machine in action. And I understand what the >difference between 1M nodes per second and 200M nodes per second really is. >And finally, I understand how much time and hardware Hsu invested in DB's >"supposedly inferior to micro programs" evaluation. And I know how DB using >one chess processor at 1/10th its normal speed crushed several top micro >programs in 1995-1996. > >Taking all of that into account, I see no way, with my knowledge of the above, >to conclude that "micros are not very far behind..." > > > > > >> At any rate when people like like myself write something don't you presume we >>have enough >>sense to base what we write on clear evidence and reasoning based on clear >>facts? Not to write clutter? > > >I generally don't write "clutter". But I fail to see how _anyone_ can claim >to write intelligently about what deep blue can and can't do, without ever >having talked to Hsu, attended one of his talks, or watched his machine during >a chess game. I have done all three of the above, which does give me some >information to base an educated guess of an opinion. Others simply jump on the >micro's side with "DBs eval sucks".. "DBs search depth sucks, it only gets to >10-11 plies" without having _any_ understanding of how they count plies. The >list goes on and on... > > > > > >>Regards >>S.Taylor From what you write here and later, you certainly know much more than I could have possibly known about what's going on RE. DB etc. And I am thankful to you for taking the trouble of enlightening us a bit more in this complex and fascinating subject. It is a massive job trying to computerize human intelligence, and Chess might be just on the very border between human intuition and what is possible to compute. It may JUST make it-and then dominate chess-. Stuart Taylor
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.