I've run into this situation a couple of times, where people give an "answer" that basically repeats the problem in the original question. (Not going to give examples here obviously, but it's happened.) So a hypothetical example might be a question like:
I've noticed that addition overflows just get wrapped around in x86. Why?
and a hypothetical non-answer might be something like:
x86 naturally wraps around addition overflows in numbers, just like MIPS and other architectures. You need to be careful when performing arithmetic; check the carry/overflow flags to see if something overflowed.
On the one hand, this doesn't answer the question. But on the other hand, it's well-written (no specific quality issues), and it seems to answer the question.
Should an answer like this be flagged? (Sometimes commenting doesn't help, since there's just no response.)