There is at least one component that is adding to the "loosening" of take down screws in my theory.
Slickness of smooth anodized aluminum chassis against the slickness of fine surface finishes of the receiver. The friction on receiver and chassis when clamped down with torqued takedown screws is diminished greatly.
If the recoil lug is not fitting tightly in the mortise, and the receiver is moving, that would tend to "yield" those 2 takedown screws at some point during firing. After all, they need re-torquing, so those screws aren't rotating loose, they are getting "yielded". If one has to continually re-torque after shooting sessions, those screws cannot do the job by themselves.
Has anyone bedded the recoil lug/ mortise?