Now that we have the legal_encodings iterator, it is simpler to find an alternative branch encoding with a better range.