When there are an even number of swaps to solve the edges you can go seamlessly into the corners, but lately I haven't had an even number of edge swaps. In this latest scramble there are 11 edge swaps so the corner that starts at 2 is loaded at 3 when the edges are done. So during analysis I have to start with the corner at 2 instead of 3. In this scramble it swaps with the piece at 4 which happens to live at 3. So after the first corner, I have to start a new cycle. I was having trouble seeing it on the scrambled cube so solved the edges so I could get a clearer idea of what was happening. It still took me a while but I finally came up with the above solution and noticed something.
4 solves 4 and does not load another piece to solve.
doughnuts loads a piece into 3 that starts a 3-cycle, namely, 2 to A to I.
baked loads a piece into 3 that starts a swap, namely, 1 and B.
So 4 swaps are used to do a corner 3-cycle that could, in this case, be done with an 8-move commutator. hmmm...