Theoretical models on network formation focus mostly on the stability and efficiency of equilibria, but they cannot deliver an understanding of why specific equilibrium networks are selected or whether they are all actually reachable from any starting network. To study factors affecting equilibrium selection, we designed a network formation experiment with multiple equilibria, which can be categorized in terms of the demand on players’ farsightedness and robustness to errors. In a second scenario, we increase the need for farsighted behavior by players, as well as the perceived riskiness of equilibria by adding a stage in which the network is disrupted. This setting allows us to analyze the interplay between the need for farsightedness and perceived risk of errors and its effect on network formation and equilibrium selection.