At various times in the past 9 years or so, we have tried several different social conventions to resolve the dissatisfaction that arises with the existing opening and closing rules. This may not be a comprehensive list:
Person X with at least 3000 points would comment, "I vote to keep open", so that Person Y with at least 3000 points who wanted to close would have to comment, "I cancel X's vote" instead of actually voting to close. I think this failed because most people didn't bother to follow this convention, and other people got upset by this, etc.
After n reopens, the question stays open, presumably enforced by moderator fiat. I forgot how much n was, and I don't think we ever had a true consensus. On the few occasions we did try that, people definitely got upset.
One benefit of the open/close system as it currently stands is that no one with qualifying score gets denied the opportunity to make a vote at some relevant time. This sort of denial-of-voice seems to be the source of most people's frustration. The other main source of frustration seems to be people who have an answer they want to write, but can't submit to a closed question. However, there is a specific place where people can make reopen requests, and using it is much better than making lots of annoying minor edits to bump closed questions to the top in hopes of attracting 5 reopen votes.
For the most part, the questions where we encounter these problems are not precise research-level mathematics questions, which form the core of MathOverflow's mission. While I appreciate the temporal and emotional investment in crafting a question, or writing an answer to a question, I do not think questions on the frontier of appropriateness on MathOverflow need special rules.
@Asafis necessary. $\endgroup$