Thanks for your input. Real life data always trumps theory.
That said, I remember you posting a target with holes that looked to me like the bullets were tipping a wee bit, though you thought they looked fine.
Temperature, humidity, and velocity have a big impact on stability, as does any bullet deformation or casting flaw. Hence the general rule of thumb is to avoid operating on the ragged edge of stability. This is especially true from the mold maker's perspective if his customers expect his recommendations to work in many different loads and conditions.
Your loads are running at 1900 - 2000 fps while my recommendations for the 444 deliberately chose a more conservative 1800 fps, which reduces stability because the bullet does not spin as fast. Bottom line is that the Miller formula is just an estimate, intended to put us in the ballpark. It seems to be more accurate than the old Greenhill formula but it's still just an estimate.