[Gravityloss]

It's just basic Bayesian reasoning. Most of the positive HIV test results are false positives. Even when the P value is less than 0.01 or so.