Author: Richard Pijl
Date: 02:51:18 12/20/05
Go up one level in this thread
>I wonder why other engines doesn't implement this knowledge which is rather
>simple:
>1.Stronger side has only one piece.
>2.Its king has no moves.
>It will help in 99% cases.
I'll experiment with this in the Baron. Looks feasible to add it, although I
doubt that it will have any effect on playing strength.
>Here are some more examples:
>
>[D]1NK1k2N/4p3/8/8/8/4p3/8/8 w - -
>
>1.Na6 e2 2.Nc7+ Kf8 3.Ne6+ Kg8 4.Nf7! e1=Q 5.Nfg5=
there is an alternative solution with:
1. Na6 e2 2.Nc7+ Kf8 3.Ng6+ Kg7 4.Nxe7 (TB draw) or
Kf7 4.Ne5+ Kxx 5.Nf3 and pawn is stopped.
This solution was found by the Baron _without_ using tablebases (although it did
not show an appropriate score)
If you want to find the solution with the locked away king, you'll have to
detect that it cannot go anywhere as it will still have one square it can move
to.
>It is funny too see some great engines struggling even on endpositions of some
>of these studies, but...
>Rybka shows us that endgame knowledge doesn't equal to playing strength.
As you probably also found out, endgame knowledge is not important to score well
against most engines as:
- Opponents are also bad at playing endgames
- Chances are that the games is already decided in the middlegame
Still, I hate it when the Baron cannot win a won endgame, or loses a drawish
one. So I will try to add and correct the endgame evaluation whenever this
happens!
Richard.
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.