[DonaldFisk]

> wouldn't it be more likely that the test was botched?

That's what I want to find out. From what you just wrote, the probability the test result is correct is at least 96%. But, as the result is so surprising and the tests aren't 100% reliable, it should still be checked.

You shouldn't assume it's a false positive simply because it doesn't fit in with our current understanding. Maybe our current understanding is wrong.

[collyw]

"Then they ran tests on samples taken between January 2018 and December 2019 and found the presence of the virus genome in one of them, collected on March 12, 2019."

If they did 100 samples you would expect between 1 and 4 false positives. (It doesn't go into detail about how many, but it's clearly more than one).