The slingshot penetrating your tank armor is not a helpful failure. Except for telling you that your processes are so wildly off base that you need to start over.

People who claim this kind of failure is useful are clueless. Failures are interesting in exploratory processes and useful when occurring within a predicted failure regime (i.e testing to failure). Unexpected failures in predicted success regimes just indicate process weaknesses. Repeated and continuous failures in similar fashions do not indicate strength, they indicate structural process deficiencies despite what cybersecurity bozos would like you to believe.