Do you know this situation, when you implemented some automated tests based on a requirement, but a test has been failing for ages because a bug in the implementation was never fixed?