"If they say a test is crap"
There are different levels of "crap". If you want to do random sampling of a population just to discover roughly how many people have been exposed to a virus in your population, then you need a less reliable test than when you want to determine who's guaranteed to have developed some immunity, when you want to determine whether someone should stay home or whether you want to do some contact tracing.
Even for one test determining the threshold for what is a "positive" result is often a balancing exercise between avoiding false positives (the test says you have property X but in fact you don't have it) and false negatives (the test says you don't have property X but in fact you have it). And getting that right may depend on what you want to use the test for.
There are different levels of "crap". If you want to do random sampling of a population just to discover roughly how many people have been exposed to a virus in your population, then you need a less reliable test than when you want to determine who's guaranteed to have developed some immunity, when you want to determine whether someone should stay home or whether you want to do some contact tracing.
Even for one test determining the threshold for what is a "positive" result is often a balancing exercise between avoiding false positives (the test says you have property X but in fact you don't have it) and false negatives (the test says you don't have property X but in fact you have it). And getting that right may depend on what you want to use the test for.