I share your concern; tagging questions into categories that are accurate but also not helpful is not the neatness we're looking for. As Galastel says, tags that are too specific aren't helpful.
As many other things in Stack Exchange, I hesitate to have a firm ruling -- it's hard to predict how the site may be used and what questions may arise; it's conceivable that we build up a serious body of questions that's specific to a subgenre, and then I'd support the tag.
But what I wouldn't do is try to "make it happen," to forcefully create subgenre tags. For example, the new dark-fantasy tag, at the moment, only has three questions. Two are "what are some common tropes for dark (fantasy/romance)" questions (one closed, and not even about dark fantasy). One is a question about authentically writing depictions of brutal social norms in historical/fantastical settings, specifically trying to avoid writing a "dark" story; this re-tagging seems outright incorrect to me, and I'll be amending it.
I bring this instance not as a personal criticism, but as an example of judging the appropriateness of a tag. Hypothetical questions that could merit a dark-fantasy tag, or another specific subgenre tag, might be things like:
- Dark Fantasy:
- How far can I go in a YA Dark Fantasy
- Will a happy ending to a dark fantasy disappoint readers
- Urban fantasy: Questions on customizing popular urban fantasy tropes (vampires; werewolves; witches; etc.) for your particular book/story
- Alternate history:
- Questions about getting across what the story's historical deviation has been
- Questions about making clear what elements are historical and which are not
- Questions about making a story accessible even to readers unfamiliar with the actual history the story deals with
So I could definitely see subgenre tags hypothetically being useful, but I think a lot of the trick is simply not to rush forward and create ever hypothetically-useful tag.