My guess is that any truly meaningful test would require a statistically meaningful number of receivers of each type, a decent number of barrels, and swapping barrels between receivers of each type. I won't lie, I have no interest in spending my time swapping even a single barrel between one receiver of each type and certainly not taking the additional time to do what would be required to make the results even approach being something meaningful. I think it would be an interesting test. However, not interesting enough for me to dedicate my resources and limited time to it. With the warm weather I'd rather spend my free time on my boat, riding my motorcycles, riding my horses, or attending more training courses (I'm hoping to get a HSP pistol class and perhaps a Vickers class in this year). Any AK time I can muster is already spoken for with a pending suppressed SBR 7.62x39 project I have begun. Maybe one of the people who likes to spend their free time making videos for their youtube channels would take it up. As I said, it would be interesting. Sadly most people probably don't have the ability to pull and install barrels, even if they did have the other resources to do it right.
I'm content to say that I have never seen any milled gun shoot consistent groups that I have not seen equaled by stamped guns. That, among other reasons, is why I don't think that a milled receiver per se gives a more accurate gun.