Something I’ve been wondering.
Imagine Shop A has 25 tests over two years:
99.1%
98.7%
99.4%
97.9%
99.2%
98.5%
Shop B also has 25 tests:
99.9%
99.8%
100.0%
99.9%
99.9%
100.0%
Without knowing anything else, which testing history would actually make you more comfortable?
My instinct is weirdly Shop A.
Not because lower purity is better, obviously, but because real-world manufacturing and sampling should have some variation. A messy-but-consistent dataset can sometimes feel more believable than a flawless highlight reel.
This raises a bigger question:
Should “believability of the testing history” be considered separately from the actual test scores?
Not accusing any particular shop of anything. I’m genuinely curious where people draw the line between “excellent consistency” and “this dataset looks curated.”
What would a testing history have to look like before perfection itself became a yellow flag?