I like you proposal but my personal preference is to run 50 different experiments at the state level until we get rules like this that work. Check out my proposal if you would. Certainly not as detailed, but I think it would stab at the heart of the problem.