Do the same thing over many samples and report back, thanks. Cooincidence and confirmation bias is all it is.
That's the trouble, you can't do many samples because those arrays aren't available via Socrates......I ran into the same issue back in 2016. Only way to get those are to wait for him to post them online in his blog or take them out of a report. Find me more current ones and I'll be happy to run tests galore.