Fault-based testing (also called mutation testing) deliberately introduces small faults to test effectiveness of test cases. Since it's computationally expensive, it's strategically reserved for critical operations where failures are costly or for suspect code sections that have had defects or complex logic requiring extra scrutiny.