Author: Uri Blass
Date: 01:24:33 12/29/03
Go up one level in this thread
On December 29, 2003 at 01:53:43, martin fierz wrote: >On December 28, 2003 at 10:46:11, Robert Hyatt wrote: > >>On December 28, 2003 at 04:46:07, martin fierz wrote: >> >>>On December 27, 2003 at 00:51:11, Mike Byrne wrote: >>> >>>> >>>>[d]8/r7/8/5bk1/8/5B2/5RPP/6K1 b - - 0 1 >>> >>>for all programs that instantly saw the draw in mike's >>>position, this may be interesting too (of course only >>>without tablebases): white to move and win, a study by >>>J. Vancura 1922, which is also on the chessbase site. >>> >>>[d]8/8/8/8/8/2K4B/5k1P/8 w - - 0 1 >>> >>>the point is that although black's king is in the >>>square of the white pawn, he can never get through >>>to the h-file. programs which are too speculative about >>>"wrong bishop => draw" might show a draw score before >>>they see the win. >>> >>>cheers >>> martin >> >> >>Mine _never_ says draw in this position: > >nice. as i suspected, this can confuse programs, as the result of ruffian shows >- it says draw at shallow depths. you seem to have implemented the wrong bishop >thing better than per-ola ;-) > >cheers > martin Ruffian Evaluates Be6 Kg2 as a draw and I guess that the reason is that the king controls the square after the pawn. I guess that Ruffian probably evaluates positions when the king controls a square that the pawn can advance to it as a draw in KBP vs K type positions. Here is another example for Ruffian's wrong evaluation and it says 0.00 in small depthes. [D]8/8/5K1k/8/7P/1B5P/7P/8 w - - 0 1 Uri
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.