ENSO is the largest signal in the climate system on interannual scales. The extratropical response to ENSO is associated with changes not only in the ambient seasonal mean flow but also in the statistics of synoptic weather systems (storm tracks). The question is, to what extent do such altered storm tracks then feed back on the seasonal mean response ? It is generally thought that this synoptic-eddy feedback is large. However, we find it to be rather minor in our study performed with 100-member ensembles of NCAR atmospheric GCM integrations with strong ENSO-like tropical SST forcing. The large feedback claimed in previous studies is argued to be an artifact of inadequate sampling that confuses storm track noise with storm track signals. Indeed when we sub-sample our own runs in a manner similar to those in the previous studies, we obtain a similar spuriously large feedback. Our 100-member ensemble is both necessary and sufficiently large for an accurate estimation of the rather small eddy feedback. The extratropical seasonal mean response to ENSO can thus be described as mainly an adjustment of the quasi-stationary extratropical wave pattern in response to the anomalous upper tropospheric Rossby wave source associated with the ENSO-induced changes of tropical diabatic heating. This seasonal-mean response does appreciably alter the extratropical storm tracks, but that storm track response is passive, i.e. it does not feed back appreciably on the seasonal mean response.