But work (torque and travel) is what dictates (or limits) how our control linkage is set up; neither of those variables can be left out of the picture.

Let's consider the rudder example: We change the pushrod attachment point on the control horn to double its original distance from the hinge axis. Yes, IF we change nothing else, we've doubled resolution and cut holding torque in half, BUT we've also lost half of our control surface deflection. To restore the original range of motion, the attachment point on the servo arm needs to be changed to double its original lever arm length, which cuts resolution in half and doubles holding torque.

>>> If we consider the need for additional servo travel as a "secondary effect", there's no end to the amount of "improvement" we could rationalize, e.g. use a foot-long control horn, but don't worry about rudder deflection

So we halve a doubled resolution, or double a halved torque, either way the product is 1 (no change ). The only way to increase res and reduce holding torque for a given surface deflection would be to use a servo having greater total range of motion acting through a reduced linkage ratio.... which calls for a different servo! Sorry, another case of “Can't get there from here”

Diverging from FCC’s original topic, the post made by John B cites a good point that didn’t hit me until after building umpteen kits - By using the outer holes on servo arms and control horns (provided surface deflection reqmt's are met), linkage slop is minimized, both initially and due to wear. Reason is simple - a given distance (slop clearance) gets taken up by a smaller angle of surface deflection the farther out you go from the axis of rotation. Another advantage to using outer holes is that it reduces the load throughout the entire linkage, from servo to control horn, WITHOUT increasing the torque load anywhere, nor reducing resolution, again provided that the range of surface deflection isn't changed.