coloin wrote:No i think your pattern calculations are good - as Serg verified it !
A quick and approximate value:
There are C(9,2)=36 ways to fill a box with two clues.
Then, before taking into account any automorphism, there are (36^9)/(2x6^8) = 30,233,088 "ed"-patterns.
This number has to be increased by a factor equal to the average number of automorphisms in such patterns.
A first and very quick Monte Carlo estimation gives a factor of 1.1
therefore, the approximate value would be : 1.1 x 30,233,088 = 33,256,397 which is in accordance with my 33,135,278
number and approved by Serg!
That 33,135,278 number was calculated by using the conjugaison classes of the group of transformations acting on that set of patterns and Burnside's theorem.
Actually, I tried to use an other way to find that number, but I found a bug in a program I've been using for a while...
JPF
