In this paper we rank a finite set of agents based on their performance on ordered pairs. That is, we assume that the performance of any two agents could be different depending on the agent in the first position. We provide a characterization of the ranking using cooperative games and the well-known Shapley value and finally, we show an example of how this index could be applied in industrial organization, social sciences and the Mexican soccer league to look for the most popular team in the league.
