Test Position

[Daniel]

Geschrieben von: / Posted by: Daniel at 11 January 2004 17:11:14:

6k1/8/7P/7P/7P/7P/4B2P/3B1BK1 w - - 0 1

It is a draw for white.
Hint: Your engine should have code for handling wrong bishop color and rook pawn,and bishops of same color.
regards
Daniel

[Mike Scheidl]

Geschrieben von: / Posted by: Mike Scheidl at 11 January 2004 19:42:55:
Als Antwort auf: / In reply to: Test Position geschrieben von: / posted by: Daniel at 11 January 2004 17:11:14:

6k1/8/7P/7P/7P/7P/4B2P/3B1BK1 w - - 0 1

It is a draw for white.
Hint: Your engine should have code for handling wrong bishop color and rook pawn,and bishops of same color.

I've tried several engines, and none recognized the draw in this position. But this doesn't mean that they don't have code for that motif (which is very common)! I think it has to do with the total material value. Some - if not all - engines activate some endgame knowledge only below a certain total material value, which makes sense most always.
Some engines may have problems to understand the motif with more than one bishop of that colour, or - more important - when additional pawns are in the h file. OTOH, without such additional pieces, the tablebases will provide the evaluation anyway...
Anyway, I think engines should have decent knowledge for that (earlier than in KBP-K), also because most of them cannot use tbs. access in the Q-search and would therefore risk to miss it, or see it too late.
For example, CB-Crafty 19.06 sees the draw in this position:
New game
6k1/8/7P/7P/7P/7P/7P/5BK1 w - - 0 1
Analysis by Crafty 19.06:
1.Bd3 Kh8 2.Bc4 Kh7 3.Kg2 Kxh6 4.Kf3 Kh7 5.h6 Kxh6 6.h5 Kh7 7.h6 Kxh6 8.Kg4
= (0.00) Depth: 1/3 00:00:00
= (0.00) Depth: 15/24 00:00:02 1351kN, tb=287

(But not i.e. with 2 white-coloured bishops and 1 pawn.)
Regards,
M.Scheidl