Why Best-Worst Method Works Well