Scott,
 
My understanding is that Quackle's endgame player is still not able to look  
ahead more
than 2 turns, and even out-in-2 combos are not necessarily solved with 100%  
accuracy
yet.  Stuck tile endgames can take anywhere from one to four more  turns than 
that to 
play with greatest accuracy, and Quackle will not be able look ahead far  
enough to 
calculate the true spread value of sticking the tile when it has a choice  of 
sticking or 
not sticking with the first move, so it will get most positions of that  
nature laughably 
far from right.  Larry played a game against it a couple of weeks ago  in 
which he
took its suggestion for his play, and lost by 16, by declining to stick it  
with a Z,
and I looked at it for a few minutes, found the best of a handful of ways  to 
stick 
the Z, and wound up winning by 19 -- by eschewing a 6-turn endgame for an  
out-
in-2 combo, Quackle made a 35-point spread error.
 
Joel Sherman



************************************** See what's free at http://www.aol.com.

Reply via email to