The problem is both. We can't pretend that failed solutions with good intentions are not problems. It is not worth adding more problems on the stack because the

The problem is both. We can't pretend that failed solutions with good intentions are not problems. It is not worth adding more problems on the stack because the guess-and-check solution whims fail. Solutions can be measured and worked towards, not guessed at where they do such large changes that their failure is also large.