Single Comment

>My verbal description is saying that the choice between the two paths is
made
>only once, and therefore the two-path probability correction should be
made 
>only once in the calculation

This works fine if the first step forces you to make a choice, but
sometimes both directions are unblocked after the first step, so you
still
have a choice of which way to go when you get to the third step.  A piece
that moves 2 squares as a Crooked Bishop then started moving as a Rook
would be easier to block than a Crooked Bishop is, as it would only get
the
two-path correction once.  Likewise, a piece that made bigger zig-zags,
going to c3 or g3 instead of e3, would get the higher number.

Nyaah, nyaah! :)