Author: blass uri
Date: 02:58:23 06/07/00
Go up one level in this thread
On June 07, 2000 at 04:30:45, Dann Corbit wrote: >On June 07, 2000 at 04:16:50, blass uri wrote: >[snip] >>I am also not sure if e4 is the best move and it is possible that Nxf7 is the >>best move. > >Or something altogether different. > >Call to mind the evergreen game. Most people "ooh'ed and ah'ed" when they first >saw the ending sequence (including me). Computers actually avoid that series of >trades and don't 'fall for it.' >But the point I am making is that a big group of experts might look at that >position and conclude that a certain course was best. But at least one person >(his opponent) did not see what was coming. > >So, until checkmate is formally proved (or unless the move is literally forced) >I will never believe that any move is the best unless it is unmistakeably proved >to lead to checkmate. > >Programs can be fooled by null move problems and miss something brilliant. It >is rare, but it clearly does happen. The main problem is not null move(There are some programs that are not null movers) but cases when the combination is too deep for programs to see because they do not know to search the right lines. > >Humans can be fooled too. Look at all the busts in ECM, for an example of why >human analysis cannot always be trusted. The main reason that I do not trust humans analysis is the fact that in most of the cases it is not important for them not to do mistakes. This is the reason that in many cases humans do not use chess programs to help them in analysis. I trust more a team of human and program than only one of them. 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.