Author: Jonas Cohonas
Date: 09:41:10 01/31/03
Go up one level in this thread
On January 31, 2003 at 11:58:04, George Sobala wrote: >On January 31, 2003 at 09:26:42, Jeff Lischer wrote: > >>Continuing a thread from yesterday... >> >>On January 30, 2003 at 07:06:40, Mike Hood wrote: >> >>>On January 30, 2003 at 02:49:07, Jouni Uski wrote: >>> >>>>http://www.chessbase.com/columns/column.asp?pid=160 >>>> >>>>The top humans are significantly stronger than the top programs at classical >>>>time controls. Hmm?!?! Bareev? >>>> >>>>Of course You must be careful when reading Chessbase report, bit this is OK. >>>> >>>>Jouni >>> >>>Jouni, although Chessbase can often be accused of a lot of spin, which is >>>understandable, I can accept the statement made by Mig as an honest evaluation. >>>To quote a statement later in the article, which qualifies your quote: >>> >>>"We surmise that today's top programs play consistently at a 2500-2600 level of >>>chess quality. The difference is that they instantly and mercilessly punish >>>every human mistake and almost never let a winning position slip. This >>>near-elimination of the margin for error pushes their practical performance up >>>toward the 2800 level." >> >>When I first read Mig's statement, I agreed with it as well. However, now I'm >>not so sure. I have these visions in my head... >> >>Humans talking about the way computers play chess: >> Positionally, they consistently play at a 2500-2600 level, but tactically >>they play at more like 3000! In practice, this combination pushes them up >>towards 2800. >> >>Computers talking about the way humans play chess: >> Tactically, they consistently play at a 2500-2600 level, but positionally >>they play at more like 3000! In practice, this combination pushes them up >>towards 2800. >> >>What's the big difference what combination of skills makes up the ultimate >>playing strength? Isn't strength strength? >> >>On re-reading Mig's article, I have to say it is very fair overall. I think >>there's an asymmetrical balance now in human/computer chess that makes these >>matches great events to watch! > >I completely disagree that it is a question of computers-do-tactics v. >humans-do-postional. E.g. the top 10 or 20 GMs find profound tactical tricks >that the current programs cannot. The top GMs, at their optimum performance, >simply play and understand chess much better than the best programs. But humans >cannot maintain optimum performance. Difference is that computers play at their best, tactically speaking, all the time and humans don't. The fact that humans can find "profound tactical tricks that the current programs cannot" does not mean that they can find them in standard 120/40 games (those which the computer won't find that is). That's another difference, computers seem to need much less time to find tactical blows than the human GM's. BTW do you have any examples at hand, where GM's find these "profound tactical tricks that the current programs cannot" ? Regards Jonas Regards Jonas
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.