As random points out, questions are closed when they don't meet site standards or aren't a good fit, which means they generally don't belong here. However, there are levels of "needing to go".
Potential for improvement. There used to be a "very low quality" close reason, a criterion for which was that the question didn't even have the potential to be improved. Deletion decisions should lean in that direction. Even if the question doesn't get improved during the hold period or soon after closure, if it received significant upvotes, or good answers, that points to some raw potential.
Good answers answer something, and it ought to be possible to make the question worthy of the answers. Worst case for a thread that would face deletion, anyway, would be to rewrite the question to reflect how the question "should have been asked". I'm not talking about a case where someone writes a great answer to something that is different from what the question asks. We don't rewrite questions to fit the answer. The solution there might be to post an entirely new question and recycle the answer. I'm referring to the case where the answer answers what was asked, so rewriting the question doesn't change the meaning or intent.
This question is a case in point. Clean up would have been better before closure, but people were spurred to action by this Meta question. The question went from being deleted to being improved and reopened because the potential was there all along, as evidenced by the answers.
Community opinion. Not much about the site's standards are absolutely cut and dry or set in stone, and that's on purpose. Within limits, the community decides what should be here. Sometimes the voting looks like a lot of people were under the influence, but when the community has spoken, we should be extremely careful about overruling it.
Actual answer quality. When a crappy question is sitting there by its lonesome, we can speculate that it is likely to attract low quality answers. And if those low quality answers start coming in, we can know we were right. In either case, we can give serious thought to closure and deletion.
However, if good answers start coming in, we need to reassess the speculation. You can't act on the basis of "answers will tend to be almost entirely based on opinions" when you have the answers and clearly they are not.
Value of the answers. The purpose of the site is a knowledge base of quality answers; the answers are the "meat" of the site. The questions provide the framework for originally creating the answers, and serve as a way to anchor and organize the answers to help find them. But Google finds answers directly, the searches don't rely on the quality of the question.
Like the question that triggered this thread, a poor quality question might not get improved, at least in a timely way. bwDraco's answer to a related Meta question makes the point that high quality answers should never be discarded on the basis of the quality of the question. We shouldn't throw the baby out with the bath water because the bath water is dirty.
