Single Comment

Thank you for finding my error. I'm wrong, so I take back my nyaah nyaaah.
But you're also wrong.

When you're really stuck with probabilities, you can use the laborious
case-by-case analysis. With a zFF going from e1 to e5, if e3 is occupied it
can't get there and f4/d4 deserves no two-path bonus. If e3 empty and both
f2/d2 occupied, no twopath. If e3 empty and one of f2/d2 occupied, no
two-path. If e3 empty and both f2/d2 empty, I was in error, twopath
applies. I haven't worked out the correct number, but it's higher than
mine, lower than yours, closer to mine than yours. You applied twopath in
too many of these cases.

I'm not sure to write this in a more general manner. (for example, zFF
going from e1 to e9 on a larger board).

I haven't worked out the number the laborious way, pending your agreement
to this.

You found an error, I think, but I think you were also wrong. 

I know you'll reply!