All the world – a game? 

We develop model agents whose cognitive capacity can be constrained and who act human-like in public goods games. Exploring their collective behavior in networked game geometries will pave the way to predictive simulations of social equilibria and instabilities. Initial results show remarkable universalities in the dependence of willingness to invest in public goods on the degree of connectedness of game groups. These correlate well with everyday experience, which renders the agent model promising for further exploration.