Single Comment

Brad, I think your method favors positions with the king near the edge of the board, relative to the standard method. <p> This is because, since the king has to be between the two rooks, it's more likely to be closer to the center than to the edges. If the king is on the b1, there are only 6 combinations of squares the rooks can be on, while if it's on c1, there are 10 possible sets of squares for the rooks, and on d1, there are 12. (The fact that bishops must end up on opposite colors muddles things a bit too--it means that the king and two rooks are less likely to all three end up on the same color than they would be without the provision that bishops must be on opposite colors--but I can't think of a good way to explain this in detail without making this post much much longer than it already is.) <p> To look at it from a brute-force perspective, if you simply count up the king positions out of <a href='http://www.chessvariants.com/diffsetup.dir/fischer-random-fen.html'>all 960 possible setups</a>, you will find there are:<p> 108 setups with kings on b1 & b8<br> 168 setups with kings on c1 & c8<br> 204 setups with kings on d1 & d8<br> 204 setups with kings on e1 & e8<br> 168 setups with kings on f1 & f8<br> 108 setups with kings on g1 & g8<p> The standard method gives an equal probability to each of the 960 setups .