One common technique in solving nonlinear systems is to do a perturbation expansion about some small parameter. This note demonstrates that a conventional perturbation expansion might not be helpful when the system is chaotic because the perturbation solution could have no chaotic behavior.
I am pretty sure I know what happened here, but I welcome your speculations.
I will add your commentary to this page.