This week's giveaway is in the EJB and other Java EE Technologies forum. We're giving away four copies of EJB 3 in Action and have Debu Panda, Reza Rahman, Ryan Cuprak, and Michael Remijan on-line! See this thread for details.
Since your method depends upon the story as a basis for what needs to be done, do you discuss how to clarify weak or incomplete stories? Or how you handle if the story needs to change as you go through the process. In my company, the developers are still working on creating good stories for those of us creating automated tests.
If you need to change the story (or Goal) you have to look at the impact of the change on your code base. In general, you can probably just assume the current state as a new starting position for the changed story. If you want or need to, and can, revert your partially implemented story you can do that.
If that happens often, maybe you should not try to find a rock solid process to handle it, but rather see if you can change how you work, to avoid it altogether? This is just a thought, and has little to do with the Mikado Method. :-)