I have discovered that the channel order does seem to be significant. Moving the lThigh_smooY down so it sits on top of the rThigh_smooY channel seems to cure the asymmetry, though I have not examined this in detail to confirm exact symmetry. Strangely, moving the rThigh_smooY up so it is directly underneath the lThigh_smooY did not cure the asymmetry.

Tentative ******rule of thumb,  place the smothScale channels directly above the OffsetA channels.