The last page or two has accelerated the discussion to thoughtful responses from several angles that I imagined/hoped for when starting the thread.
With a repeatable controlled test, one can compare that controlled data to random. You can still use the drop test, and perhaps even validate it, or find differences in outcomes and explore those as weak points for improvements.
I have in my mind a drop jig that I hang from my garage rafters onto my rifle bench vise, that would cost me $0 to build, deliver measurable, relatively precise blows, that anyone could take the same recipe and reproduce to exacting specifications down to the rifle system, the same test. I know, the force factors would vary from a drop test, hence, they are entirely different tests. We could examine/discover trends, ie: the weak point tends to be the windage turret, or this optic tends to lose zero at X newtons of impact (equivalent to a a fall onto a rock from X height….)
The “millions of dollars” claim is pretty outlandish. The biggest thing is the access to optics and time. There is an opportunity here as well. There is value in this data.
The other opportunity here is, to learn, improve processes, and obtain more, not necessarily better, data. Data I am currently missing when trying to navigate critical gear decisions. That is the intent of the thread, to think/discuss the nitty gritty nuances of our hobbies and find enjoyment in the process.