ParentFull threadUhUhUhUh·I would think that one could find a function giving efficiency of gerrymandering as a function of the number of parties involved. Obviously, f(1) = 0.View on HN