Accuracy measures how reliably a test suite indicates whether functionality is correct or broken. It combines regression protection, which reduces missed bugs, with refactoring resistance, which reduces false alarms. section 4.2 section 4.6
Classify test outcomes by comparing the test result with the functionality’s actual state:
- True negative: the test passes and the functionality works. section 4.2
- True positive: the test fails and the functionality is broken. section 4.2
- False negative: the test passes despite broken functionality; regression protection helps prevent it. section 4.2
- False positive: the test fails despite correct functionality; implementation coupling often causes it. section 4.6
Prefer tests that verify the system’s observable result rather than its implementation steps. False alarms erode attention and trust, becoming increasingly costly as the project grows. section 4.6
