Rudolph Optics,
I downloaded the .pdf file, opened it then started reading about that "Box Test." When I got to the words:
This is a true test of repeatability of the riflescope's turret adjustments.
... I laughed for a couple minutes.
It cannot be a true test because it includes the repeatability errors of the rifle, ammo and person shooting them. Smart people know that several 2-shot groups have a spread across extreme spreads of 5X or more; and their centers are never in the same place.
A true test is best done on a scaled bench collimator which only measures the scope adjustment repeatability as well as their accuracy.
Clamping the scope in a solid mount they watching its reticule move across a yardstick 50 yards away is another very precise way. That also will show that several scopes of the same make and model do not have the same exact adjustment per click because all their lenses do not have exact focal lengths; a 1% or more spread across each one is normal. If the collimator is like the one I built, it can also show how much error there is in reticule reseating from recoil for each shot.
A very precise and inexpensive way is to put an optical collimator in the muzzle then zero the scope on it. Move the adjustments 20 or 40 clicks in each direction to box the scope reticule about the collimator reference then see if it ends up where it started. Look through the scope with a monocular and the image will be several times larger so you easily see a 1/16th MOA error. No human nor ammo nor rifle errors are included in the result. Boxing the scope reticule on the collimator 2 clicks in each direction is an excellent test of both repeatability and backlash. Adjustment backlash can't be easily resolved by the popular "box test."
Therefore, I don't know what the repeatabiity of that scope is. But your test did show a lot of the non-repeatability of the rifle, ammo and shooter lumped together.
