🚀 go-pugleaf

<liam@ork.net> wrote in message news:B2bN6.1005$du2.96321@news.shore.net...
> In rec.games.chess.politics Robert Hyatt <hyatt@crafty.cis.uab.edu> wrote:
>
> > Here is a terrifying thought:  If you make the raw assumption that there
> > are 10^50 chess positions possible (this is based on many attempts to
> > find the smallest number of bits necessary to store any legal chess
position
> > plus enpassant status and castling status), then the tablebase generator
> > has an interesting task:
>
> You don't need to search them all. Once you have a won position, you don't
> need to analyze every position from there.
>
> For example, give me a "dead" position, a knight up, and it's a
> theoretical win, I can assure you, even if I might not win it against
> Kasparov.

Look a the material tossed out during the Evergreen game.  Down considerably
more than a knight at one point.

There are winning material sacrifices of a knight or more.

The only way to know you can trim a branch of the tree is to go all the way to
won/loss/draw -- just assuming a big material advantage will win won't cut it.
It works most of the time, but sometimes its dead wrong.  There are times when
it is obvious that the material advantage is temporary (the Evergreen game
being a rather clever combination, for instance).  But there are positional
moves that pay off surely, and only much later -- quite out of range of most
of today's chess programs.

The LCT II EPD test suite has some very, very difficult positional moves, for
instance.  Here's one example (analyzed by computer for over an hour without
correct solution):

1r2r1k1/4ppbp/B5p1/3P4/pp1qPB2/2n2Q1P/P4PP1/4RRK1 b - - acd 13; acn 109856709;
acs 4101; bm Nxa2; c0 "Barbero - Kouatly, Budapest 1987"; ce 52; id
"LCTII.POS.13"; pv Rbd8 Be3 Qf6 Bb7 a3 Bc6 Rf8 Bc5 Nxa2 e5 Qxf3 gxf3 Nc3 Bxe7;

Solution:
Barbero - Kouatly : ...Nxa2!!
(Budapest 1987)
Two passed pawns on the third rank are worth the quality.
--
C-FAQ: http://www.eskimo.com/~scs/C-faq/top.html
 "The C-FAQ Book" ISBN 0-201-84519-9
C.A.P. FAQ: ftp://cap.connx.com/pub/Chess%20Analysis%20Project%20FAQ.htm