>>64
Let me think. If for a function that takes
n elements and outputs
n elements, with no collisions, and the output is suitably random, then the chance of encountering a disjoint cycle in dimension
N would be...
Taking a first random
x, chance to find one that is disjoint, is 1/n. Chance of finding a disjoint sum taking
x and two other random elements is n-1/((n-1)!)
With three elements its... what the hell is the equation for the number of all unique subsets of a set where each subset has
n elements?
Just apply this, but with the mod
n for the group. Due to the mod it'll go up for each dimension, otherwise the chance would diminish.
I'll work this out and see.