Please send the gcg. I would guess it is the same as the issue I previously commented on - partial racks don't evaluate correctly. I can determine if that is the issue from the gcg fairly readily.
Matt --- In [email protected], "zev_kesselman" <ke...@...> wrote: > > John, > > I think Chris has the right idea. I have a similar issue, and I have > a .gcg for it. > > With one tile in the bag, Expert Player lists 15 possible moves, the > first of which has a 50% chance of winning - depending on which of > the 8 tiles is in the bag (3 of them are E's, simplifying the > scenario). Play number 5 has 0% chance. Both evaluations seem > correct. > > But when I do a sim of #1 against #5, #5 is reported as 100% win. > I've looked at the XML file, Quackle's moves don't add up to a win. > So Chris' concluding sentence seems appropriate. > > Zev > > > --- In [email protected], "John O'Laughlin" <olaughlin@> > wrote: > > > > Chris, > > > > Do you have a gcg for this game? > > > > John > > > > On Tue, Jan 6, 2009 at 7:16 PM, tendenc <tendenc@> wrote: > > > Quackle AAGINR 451 pts > > > Me NITROSO 368 pts > > > > > > New Player 1: Turn 12 > > > A B C D E F G H I J K L M N O > > > ------------------------------ > > > 1|B E E C h N U T P =| > > > 2| - " R E " U - | > > > 3| - G O X O N - | > > > 4|' - I D U T '| > > > 5| N O D R | > > > 6| " I F " I " | > > > 7| Q A T ' ' P E ' | > > > 8|= f I B R I L S L ' =| > > > 9| O ' A ' Y E ' | > > > 10| V T H E M E D " | > > > 11| E E - D | > > > 12|Z A G - S ' - W '| > > > 13|I L ' ' O Y H A J| > > > 14|T O U C A N W A V E R | > > > 15|S M ' A L K I E ' F E| > > > ------------------------------ > > > > > > The winning play is NITROSO at O5 for 74 points. This wins the > game. > > > > > > When I simulate the generated choices, NITROSO at M5 (62 points) > > > gives a 100% win, when in actuality loses the game immediately. > > > > > > I don't pretend to understand all the intricacies of simulation > but > > > it would seem that a math function to calculate the score after an > > > outplay to validate the win% would seem appropriate. > > > > > > Cheers! > > > > > > Chris ten Den > > > > > > > > >
