Author: James Robertson
Date: 21:56:25 10/23/99
Go up one level in this thread
On October 22, 1999 at 17:57:47, KarinsDad wrote: >On October 22, 1999 at 16:42:13, Robert Hyatt wrote: > >>On October 22, 1999 at 11:43:37, KarinsDad wrote: >> >>>Does any program currently do a sacrifice search? >>> >>>By this, I mean the program intuits that a sacrifice may be in order (within the >>>first 3 ply maybe), and then pretends that the sacrifice move is the PV and >>>searches down it's ply to a deeper depth or somesuch in order to find out it's >>>validity. >>> >>>Or, do all current programs only make a sacrifice if they perceive a definitive >>>gain within the same ply depth as most other moves? >>> >>>KarinsDad :) >> >> >>I find them via a normal search. The point being that you should see how >>many times a sacrifice works, how many times you 'think' it works, and then >>figure out how much effort you spend proving that most don't work. As >>opposed to spending that effort on a normal tree search... > >Hence the word "intuits" in my previous post. To make it worthwhile, you would >only want to do it when some set of rare conditions are detected (1 root >position out of thousands perhaps) which could indicate that a sacrifice might >be favorable. What those conditions might be is unknown to me and part of the >reason I posted the question. I had the feeling that most programs would only >find them via the normal search. > >KarinsDad :) It seems like a heck of a lot of work to decide what conditions are good for sacrifices.... Even seemingly obvious situations (e.g. three bishops, two knights, a rook and a queen all bearing down on the enemy castle) would be difficult.... James
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.