Author: Robert Hyatt
Date: 11:16:16 01/07/02
Go up one level in this thread
On January 07, 2002 at 12:06:01, David Rasmussen wrote: >On January 07, 2002 at 10:20:03, Robert Hyatt wrote: > >> >>Make it something impossibly large. Then make sure that the king is the _last_ >>piece used to make a capture. All works fine that way... > >I know thats what you do :) > >But won't that just make the capture sequences that ends with a king capture, >look very good? No, because my SEE "minimaxes" the scores. If one side captures with the king and the other side then captures the king, the minimax in the SEE code will decide "OK, I won't make that king capture because then the opponent will capture my king for even more gain..." Works fine. :) > >If the value of the king is 100 pawns, the capture sequence I mentioned above, >would stop after the queen has captured the king, and this capture would then >get the value of 101 instead of 1 (disregarding that the king is actually mated. >Just pretend that the knight on d7 is gone). There might be other captures in >the position that wins a queen, and as such gets a value of 9, but this capture >winning only a pawn, will look better as it will have a value of 101. Isn't that >a problem? Couldn't it just be solved by saying "If we capture a king, the value >of this sequence is the SEE value two moves before the king is captured"? > >/David You should be doing that _anyway_ since _no_ capture is "forced" in chess, only in checkers. So at any point the side on move can choose to make a capture or "stand pat". If your SEE doesn't work like that, it will deliver broken scores and screw up move ordering beyond belief...
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.