On 5/14/06, Andrew Fisher <[EMAIL PROTECTED]> wrote:

After a few hundred 3-ply sims, Quackle (0.92) likes
OWLIER then ZANIER (minus 19) then ZAIRE (minus a
further 18), WORE (minus 4), WON (minus 5) and then
CLAW (minus 5). Maven has WORE top by 15 over CLAW,
then WON (minus 3), ZAIRE (minus 1) and NEWEL (minus
6). No mention of OWLIER or ZANIER even after they
were added to the sim.

Andrew,

This is a very embarrassing error. If you click on the Details button, you'll see "Spread" and winning% values that make a lot more sense and agree with Maven. But somehow the "Valuation" numbers in the main window are screwed up, and they probably always are in pre-endgame positions. I haven't had time to figure out exactly what it's doing and why, but this will be fixed before the next version, which will hopefully be very soon. I don't know how we didn't notice this already.

Also, if you're this close to the end of the game, you should use "Many" iterations, because mixing iterations where the game ends and iterations where it doesn't is bad. That's not really the problem here, because I ran my sim with Many and the Valuations were still screwed up, but after we fix this bug, you're going to want to use Many, and when you do pre-endgame positions in Maven you should use 99 iterations.

John O'Laughlin




SPONSORED LINKS
Collectible board game Milton bradley board games Child board game
Family board game Cranium board game Parker brother board game


YAHOO! GROUPS LINKS




Reply via email to