Ranking multiple measures in a stack of stars/circles without labeled 'better / worse' margins seems like a pretty terrible data visualization.