In earlier conversations about thrown charges I have mentioned how I do things. I fill the hopper and throw at least ten charges to get things normalized, for lack of a better term. I don't tap the measure myself. I don't let the throwing motion tap at either end of the stroke. I grasp the whole handle in my hand, not just the part that sticks out for the fingertips. I use a slow, high torque motion. If grains want to get cut, then they get cut. They don't hang up my throw. Because I'm using "overwhelming torque" the troublesome grains don't cause a lot of vibration to upset the way the powder is settled in the measure. If I do feel a really rough one then it and the next throw are returned to the hopper.
I did enough testing early on to know I don't get variation over a few tenths either way. Because of that I quit checking, and that's why I say I don't know what it really is. The fundamental problem, and early on it was a problem for me too, is getting over the idea that a few tenths matter to what you'll see on target provided the load is well developed in the first place.
To be clear, I'm not talking about long range benchrest and the like. At those distances the variations in MV from charge variations that don't affect short-mid range groups will start to show up as vertical.
