| HN Mirror

Remember your Karnaugh maps in school? Those logic diagrams where you use boolean operators to cover as large a space as possible on the map, which results in as simple as possible a boolean expression? And how much you appreciated the "don't cares" because it made it so easy?

For similar reasons, a cheap binary test for X is inevitably going to be easy to convince to give positive or negative results [1] for a wide variety of substances not considered by the test creator, because you just can't make a cheap test that tests COVID-19 and exactly and only COVID-19. Basically, every presented substance other than a correctly-done test is a "don't care". The resulting test takes advantage of this, and it will map quite a lot of things to either true or false outside of the domain of interest. I don't know if such a test would even be possible to create strictly 100% accurate, but certainly for a lot more expense we could get much closer than we do today. But we wouldn't be shipping some swabs and some strips, we'd be shipping tests where mass spectrometers would be a key component, along with a very large instruction book.

But for situations where the test is being taken in good faith, who needs that?

And the expense it would take to stop kids faking the test itself is too silly to contemplate. Instead, while the news articles didn't mention it, I'm sure where this cheating was endemic they took the simple step of making the child do it in front of them. That only raises the bar, but it'd be enough for the vast majority of the cases.

The math makes this inevitable, by the fact that the set of inputs that could be presented to a test is effectively infinite. It can't hardly be false that of the thousands of substances you have ready access to, quite a few of them will produce false positives. This is also not specific to "COVID-19", it's true of any of the simple paper strip tests. In practice, though, this is just a curiosity, not a problem.

[1]: e.g., while I don't know if this would work, if you want to fake a false test, add in some bleach to your sample. I haven't tried it and won't but I wouldn't be surprised something caustic like that would simply destroy both COVID-19 and the test itself, in relatively small quantities. If not bleach, something else.