The trouble is though, isn't it, that although your thoughts about "Kinks" and "Bent Sticks" may well be perfectly valid, they won't go for "tasting it" as a proof so long as the "sheer chance" option is so wide open. Now, if you want to spend 25 years collating hundreds of examples....

But on the other hand, the wide measuring tolerances of Smithhills don't matter if the broad alignments are there in large numbers. The "errors" need assessing statistically, and the opinion of a statistician might be that the alignments are proven valid, beyond reasonable doubt. The same could be said to be needed for the whole alignment industry, which is presently merely the alignment art and open to all to sneer at... Until then, you're in for a lot more rejection slips I fear - not because you don't have a compelling case but because it's a "risky" case.
I suspect that posterity will listen hard to you though, if that's any consolation...