Author: Chris Carson
Date: 08:14:41 07/12/00
Go up one level in this thread
On July 12, 2000 at 11:05:57, Ed Schröder wrote: >On July 12, 2000 at 10:52:44, Jeroen Noomen wrote: > >>On July 12, 2000 at 10:23:40, blass uri wrote: >> >>Hi Uri, >> >>You are wrong: There is only one thing to come up with better >>moves in blocked positions: CHESS KNOWLEDGE. >> >>Give DJ a 1000 times faster machine and no extra knowledge and >>it still will play moves like c4?? It doesn't KNOW what to do >>in a blocked position and that is fatal. >> >>Van Wely-Fritz from the Dutch championship and Kramnik-DJ prove >>that it still very easy for a strong human to make a computer >>program look silly. >> >>But that is the paradox in computer chess: More knowledge would >>avoid such losses, but slow the program down. And then it would >>score worse against other computer programs, making people think >>it is weaker. > >>How to solve this!? > >Give up comp-comp competition... it would force programmers to focus on >things that really matters as now (and this has been true since day one >in CC) they are forced to do well in comp-comp as this is the measuring >rod a program is judged. > >I frankly believe that if there was no comp-comp competition nowadays >programs would be more intelligent. But wait... comp-comp is fun too, >don't abolish that! > >Ed You kill me Ed. :) Rebel is great against other computers and against GM's. You have done a great job and produced a great product. I say, keep both and put pressure on you and all the other programmers/products to improve. Chess is Chess. :) Best Regards, Chris Carson > > >>Jeroen >> >>>It does not prove it because it is possible that Deep Junior could find better >>>moves with faster computer and better search techniques. >>> >>>The question is what are the mistakes of Junior and if it can avoid them in >>>slower time control. >>> >>>Uri
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.