Author: Robert Henry Durrett
Date: 07:28:10 07/05/02
Go up one level in this thread
On July 05, 2002 at 08:06:17, Marc van Hal wrote: >On July 04, 2002 at 12:35:49, Sune Fischer wrote: > >>On July 04, 2002 at 12:08:37, stuart taylor wrote: >> >>>On July 04, 2002 at 11:34:13, Sune Fischer wrote: >>> >>>>On July 04, 2002 at 10:40:31, Thorsten Czub wrote: >>>> >>>>>they lack creating a plan. and choosing the right opening for the right >>>>>opponent. >>>> >>>>One could also say that they investigate all possible plans. >>>> >>>>Some of the great moves in history, eg. where Fischer seemingly sacrifices a >>>>rook for no reason at all in an "equal position", doesn't take computers more >>>>than a few seconds to find. >>> >>>I don't think they'ii easily find that rook (exchange) sacrifice Petrosian made >>>against (i think) Reshevsky. > > > >>I'm not saying they can find all great moves ever played, only that they have >>less of a problem with the unintuitive moves. Many of those exclamation mark >>moves made by grandmasters are simply piece of cake for the progs. >> >>-S. > > > >What about Kasparov bischop sacrefice in his game against chiburnatse in a >Gligoric Kings Indian >or his rook sacrefice for nothing but position against Karpov. Perhaps true positional sacrifices would be examples of moves found by humans but not by computers? Would need some sort of experimental verification to be sure. This raises a more general question: We have talked a lot about "computer moves" which are played by computers but not by humans. But there is another side to that coin. There may also be "human moves" which are played by human top GMs but never played by computers. Elaboration: In positions where there is only one "best" move in the position, then perhaps both the top GM and the computer might find that "best" move most of the time. Exceptions might be a form of "blindness." For example, the best move might be a true positional sacrifice and only the human would see it. Or, alternatively, the best move might be a move which humans simply cannot see, but the computer finds it. All this is speculation. Need concrete examples. Bob D. >in a Ruy Lopez >>This kind of moves will not be played by programs. >But something which is a bigger weaknes is finding simply positional moves to >keeping a grip on the position. >And in the mean time defendig it self from counter attacks. >resulting to more clear positions. >Attacking is fine but defending is just as important. > >Maybe it would be nice to make some epd's of defending moves. > >Only when you find a better balance of Defending and Attacking a program could >improve. >Though technicaly this is a dificult task. > >Or how to handle positions after f4,d5 e5 is some openingslines > >So there still is enough space to explore! > >Regards Marc van Hal
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.