> The second irony is that the Nash equilibrium for rational-actor based models of deterrence is not peace, but random violence.

Do you have a citation for that? I'd be interested to read further about it.