So, er, I went a different direction. Came up with a different game that's a lot less simple, but possibly easier _for me_.
Ideas have been bombarding me for a day or two. I want to do something with machine learning. With maze autogen. And, er, ... maybe I have it figured out. But I spent the last five or six hours just getting the beginning of the framework together (and by that, I mean everything works except "start game" ;) )