At present, we use the Brett 2-piece method. We've thought about the Lancaster method, but haven't done anything. Using the Yale method would require us redistributing a dataset with the point-to-point map -- something we could do, but would be more effort. Is it clearly superior?