Nice work! I love an "impossible" problem that falls to bounding estimates like this one does. There was a surprisingly lot of work done in protein folding that had similar sorts of techniques to eliminate structures that would either never happen or would self destruct if they did kinds of things.