Author: Charles L. Williams
Date: 22:18:31 03/06/99
Go up one level in this thread
On March 06, 1999 at 19:48:30, Charles L. Williams wrote:
>These are the statistics for the first four plies of
>my program's move generation.
>
>Node count at Ply 1 is 20
>Node count at Ply 2 is 420
>Node count at Ply 3 is 9322
>Node count at Ply 4 is 206599
>
>The last count should be 206603.
>
>Ply 4 is the first ply where checks must be countered,
>which might be significant. There are no en passant or
>castling moves yet. Any ideas on what could cause four
>missing moves? Thanks.
>
>Chuck
I found the missing moves. The count came out right by either taking out my
white bishops, or my black rook pawns. It seems that if a bishop (or presumably
any piece) moves behind a pawn, the pawn thinks it's blocked and doesn't make
the next move. For instance, if a WP moves a2 to a4 and a BB moves to a3, the
pawn won't make the move to a5.
I used the Crafty perft command to compare the count.
WP a2 to a4
BB f8 to a3
BB a3 to f8
BB f8 to b4
WP a4 to a5
WP a5 to a4
BB b4 to f8
BB f8 to c5
WP a4 to a5
WP a5 to a4
BB c5 to f8
BB f8 to d6
WP a4 to a5
WP a5 to a4
.
.
.
I'll figure out why tomorrow. Nothing like a bug like this to make the weekend
go by too fast.
Chuck
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.