Author: Bruce Moreland
Date: 22:56:35 02/22/99
Go up one level in this thread
On February 22, 1999 at 18:06:15, Dann Corbit wrote: >If we define a sacrifice as a loss of material for positional advantage without >being able to see a tactical bonus, then it was a sacrifice *for him*. > >If we can see deeply enough, then all sacrifices become tactical manuvers, >unless they are unsound sacrifices. At which time, they become blunders. >;-) > >Crafty might make what it thinks is a positional sacrifice, but Deep Blue, at >the same time control says "Combination" because it see 'x' plies further to the >actual solution. > >It seems (from previous discussion) that a sacrifice is only a sacrifice has >"hoped for" benefits rather than directly calculated material benefits. We may >further differentiate and say that 'accidental' sacrifices do not count unless >your program made the move for positional reasons. I don't like these definitions, but I can't do much better. A real sacrifice is one whose outcome you can't determine. A sham sacrifice is one whose outcome you can determine (a risk-free sacrifice), and you get your material back or mate. Spielmann calls positional sacrifices sham sacrifices, which I find interesting, since in many cases it would seem to me that they involve risk. Averbakh, in his awesome "Chess Tactics for Advanced Player" states (page 9) that a combination "... refers to a co-ordinated combination of contacts, which necesssarily gives a player an advantage, or to the conversion of a certain combination of White and Black pieces into co-ordinated combination, which necessarily gives one side the advantage." This is hard to understand, which is fine, but it's clear from the book that a combination doesn't have to involve a sacrifice of any sort, which is also fine. A sacrifice can have a positive material outcome. Spielmann would call these sham sacrifices. Computers find many of these effortlessly. I haven't looked through his examples, but it is likely that some of his real sacrifices would be shown to be sham sacrifices by computer programs. When I am talking about computer chess, I use the term "real sacrifice' to refer to cases where a program sacrifices material and doesn't see return of the material. Real sacrifices by computers are quite common now, but many of them are mundane. Spielmann would probably refer to some of them as sham sacrifices. I think in game two of my program's match with Blatny last week featured an offered real sacrifice, which Blatny declined (he didn't play ... Qxb2). WChess will often make real sacrifices against the enemy king position, today I saw it do Nxg5 after ... h6 and ... g5. This was a real sacrifice because it never got the material back (and in fact lost). I've seen programs make real sacrifices of the most of the same types that I see in human games, but I've also seen them mis-judge either way, meaning that I've seen them make bad ones and fail to make obvious ones. bruce
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.