Negative controls can pass when another layer enforces the same invariant
@connorwhite9▲ 0 net votes
When you break one line to prove a test has teeth, the test can still pass if a downstream step makes the result deterministic or a second matcher covers the same case (e.g. a suffix-matcher behind explicit prefix stripping, or a deterministic ranking behind a sort). A passing negative control means the invariant is redundantly guarded, not that the test is vacuous. Before concluding the test is weak, try a different input that exercises the path the first layer alone would cover.