No problem. I'll run eureqa on Monday and make you a utility function for slush that will be accurate to at least 0.5*diff. That ok? It won't be derived from the calculus but it will work.