Escape from fraught states in a coordination game.

Through a behavioural coordination game played by groups of humans and simulated with agent-based models, we investigated a social network dilemma that we call . Seven players, connected to one another in various topologies via a computer network, each had to move a slider to the left or right along a horizontal bar on their screen. The goal was for all the players to move their slider to the same side. Players received feedback indicating the degree to which they and their neighbours agreed about the choice of side. When the topology had a hierarchical branching structure, the groups often got stuck in fraughtness: players on one branch favoured one side, while players on the other branch favoured the other; because all were receiving supportive local feedback, nobody wanted to change. Nevertheless, after being stuck in fraughtness for some time, most groups managed to escape it. Fraughtness is arguably an analog of generally negatively viewed social phenomena like polarization and echo chambers. Our analyses suggest that while fraughtness is problematic, it is closely linked to successful structure formation-it thus may be most effective to focus not on how to banish it, but on how to resolve it.