Harvard scientists partnered up with their colleagues at Emmanuel College to develop an algorithm that could predict cooperation. The algorithm could also determine whether changes done in these networks could lead to better social cooperation. Giant social networks may create connections by destroying the barriers of geography, nationality, and language, but to establish real social cooperation, scientists suggest something more intimate, like strong pairwise relationships.

The study published in Nature on claims that loose networks spread all over the world are not enough to create a social structure conducive to real cooperation. Program for Evolutionary Dy­namics director Martin Nowak, who is also the senior author of the study said the algorithm they make can calculate the benefit-to-cost ratio for cooperation and how well it works on a fixed population structure. He said that a social network with strong pairwise ties is most conducive for cooperation, and their mathematical argument is about measuring the stability of relationships among families and friends.

The main idea of the mathematical argument is that there is a downside to the global interconnectedness brought by the internet and its many social networks, Harvard Gazette reported. Program for Evolutionary Dynamics researcher Benjamin Allen, who is also the assistant professor of mathematics at Emmanuel College, said more connectivity doesn't necessarily mean people will be good to one another. It doesn't mean that global interconnectedness is a bad thing, but it simply can't substitute the relationship that small and strong local connections bring.

Nowak said that cooperators tend to influence their close friends to become cooperators too, which means this cluster will get positive payoff from one another. Meanwhile, people can't get cluster in a well-mixed population, which means cooperation is selected against. Newak and his team figured out that they should include weak selection in their mathematical argument to come up with the algorithm that can quantify the social cooperation among people in a more intimate network.