A mutation that cannot be killed because it is equivalent in observable semantics to the original. Scientific literature on mutation testing treats the equivalent-mutation problem as the central flaw of the technique. Schirp disagrees: with semantic-reduction operators the case is infrequent. Most unkillable mutations in practice are not truly equivalent — they signal a semantic weak spot in a called library (e.g. accepting nil where it shouldn't), which can be fixed upstream. When a genuine equivalent mutation appears (e.g. a negative constant later wrapped in abs), a message-expectation test can pin it.