On 2/22/06, Anand Buddhdev <[EMAIL PROTECTED]> wrote:
I've just built quackle from CVS, and it runs fine. I would like to use
it to simulate end-games, using positions that Pete Finley has posted
on uk-s recently. How can I setup the board with tiles and scores to
run the simulations?
 
At present the only way to set up a board position is by putting words on the board and committing them.  It would be some work to do it and get the score correct.  It wouldn't be difficult at all to have a real board-editor, but we haven't done it yet.
 
The version of Quackle in CVS does not have an endgame solver.  Its endgame plays are chosen with the same static evaluator used for the rest of the game, which needless to say is far from ideal.  I've written endgame solvers before, but I haven't been happy with any of them.  What I would really like is to use the B* algorithm that Maven has, but I haven't gotten my head around it yet.  If anybody thinks they can help with the endgame solver, I'd encourage them to take a stab at it.
 
The B* tree search algorithm : a best-first proof procedure
by Hans Berliner

Type: English :  Book
Publisher: Pittsburgh, Pa. : Carnegie-Mellon University, Dept. of Computer Science, 1978.
OCLC: 9204185 
 
I can send scans of the pages if anyone is interested.

John



SPONSORED LINKS
Collectible board game Cranium board game Chess board game
Scrabble board game Monopoly board game Milton bradley board games


YAHOO! GROUPS LINKS




Reply via email to