Generally, we like for text-based identification-request questions to contain as much information as possible. Things like names of characters, major plot points, and genre/atmosphere are very important, and less significant things like subplots, where/when the OP saw it, details about art style, etc are all very helpful for answerers. Questions which don't include enough information are usually bad questions, as they may have multiple possible answers and it's essentially a guessing game as to which is the correct one.
With that said, sometimes one has a reasonable guess what the right answer is, despite the lack of information. By reasonable, I don't just mean "this matches a couple of the OP's points," but "this matches everything the OP said, and it's very likely the correct answer." This is a subjective judgement, but I think it's something that users here can be trusted to make.
Normally, if a non-ID question is even remotely answerable, it can be edited to be a good question that fits our standards in some way. Either the question is clear, and can be edited and answered by anyone, or it's unclear and requires more input from the OP and should be closed. However, ID requests are different in that we can't even edit anything in. The OP is the only one who knows what series they're thinking of. Furthermore, they may not know any more information besides what they provided.
So, what's the proper policy when a low-quality ID request comes along that I think I know the answer to? Is it appropriate to post the answer despite the fact that the question has problems, and let the OP and the community decide on how useful it is? (After all, the problems were not significant enough to stop us from finding the answer.) Should I comment to request for clarification when I already think I know what the answer is? Should I vote to close as unclear? In addition, if I do manage to guess correctly, at that point should more details be edited in to the question?
I personally lean towards posting the answer--after all, all text-based ID request answers are guesses to some degree, and the community here is capable of judging whether the guess is at least reasonable. However, I can see arguments for other actions as well.