The Lagrangian answer covers this a bit more obviously by stating a solution which results in a closed curve of states around some equilibrium point(s) for certain initial variables. I think that matches your understanding of this as two peaks in a closed circuit.