The whole way the downswing pivot works is that ground reaction forces work at diagonal angles towards each other, coming from the feet, in reaction to opposing forces that the feet generate. The left and right foot are producing force vectors that work opposite to one another. The GRF vectors would have to be heading towards one another as a result. It causes a force coupling event that allows rotation to occur, as well as the lateral translation of the pelvis.
It would only make sense that if you initiated a similar move to begin the backswing, those same forces could work to allow for greater rotational torque going back.
Guys are so concerned with centered pivots yet they KNOW that they can get a lot of rotation going with a lateral shift of the pelvis in the downswing. Would you not be generating the same type of rotational potential going back if you shifted weight?
Just a guess.