Very interesting conversation and the dime just dropped for me.

I now understood the notion of the servo still being effective at full throw if it's slightly offset out and the arm angles in at neutral. On the flip side, if completely parallel ay neutral, it will be investing very litle of it's motion to actually hold the aileron at full throw...

Actually, if I extrapolate, in order to identify the linkage best angle at rest, it should be parallel to the servo at full push throw. So one would need to move the arm to the best angle (between 45 and 60) and measure a straight line from the hole to determine the right horn location.

Furthermore, the arm at neutral should not be at 90 degrees, because if that is the case, during pull (aileron up), you will encounter an area of low yield and low hold, probably somewhere after 50 - 55 degrees of throw...