I thought you might be talking symmetrical, but since I'm an engineer, I didn't assume In any case, believe it or not, with everything you've stated, you can still be flying at a non-zero AOA (in fact, it's likely) depending on fuselage lift & drag, thrust line vs. CG, wing and stab ACs vs. CG, and other aerodynamic traits of the plane that may dictatate a need for the wing to be producing some lift, positive or negative, on a downline. You get the idea. The ideal aerobatic airplane, I guess, would be one that has all component ACs in line and all of its control surfaces aligned with their respective axis of revolution (i.e., no moment arms other than for the function of the surface).

On the AC vs. CG roll issue, by definition, the roll (all, really) axis goes through the CG. So more accurately stated, the line between the each side's AC should ideally pass through the CG's vertical location, intersecting the roll axis. That way, when the plane is rotating about the CG, the wings are not barrel-rolling around the CG, they are twisting around the CG. And adding differential does not affect this- it is a natrual trait of the aircraft. Adding differential would only help the separate, adverse yaw due to drag issue (by balancing the drag).

I didn't explain it well before, but in the longitudinal axis of the airplane, if the AC of the wing is well away from the CG, then aileron deflection would produce a slight off-axis moment around the CG, yielding a small wobble. I don't believe differential will cure this well, as the lift and drag forces would be on different axes.