Author: Scott Gasch
Date: 10:22:21 01/08/03
Hi,
First I'd like to add my thanks to everyone else's for Ed taking the time to
write his "Programmer Stuff".
While reading the Evaluation section I was interested in the implementation of
attack tables. Attack tables are something I have always wanted to do but never
wanted to spend the cpu time on.
But I am a little unclear as to how the Rebel scheme works.
[D]4r3/4qnb1/2nk4/r6R/8/2BN1N2/4Q3/4K3 w KQkq - 0 1
If I understand, we get:
KQRMPccc
WB[e5] = 0y01110??? // the bits for piece type attacks set
BB[e5] = 0y11110??? // the bits for piece type attacks set
But what are the counts? Do they take into account the fact that there are more
than one minor piece attacking the e5 square for each side? If not how can you
accurately do a SEE function for a position where there are different numbers of
minors attacking the square from each army? It would also be possible to make a
position where one side had two rooks attacking a square and the other only one.
If there were no minors and no pawns attacking that square, the side with more
rooks controls the square and that affects the value of proposed moves ending on
that square.
Is this just rare enough of a situation that you accept some inaccuracy in the
interest of speed?
Thanks for the insight,
Scott
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.