Closure wars are caused by asymmetrical close/reopen voting process.

The close/reopen voting system has an asymmetry that promotes closure wars on meta and is vulnerable to abuse.

Once the first vote to close is cast, the matter of a question's survival is raised for a vote. But it is a one-sided vote. Only those who support one side of the matter can express an opinion. Once a question is closed this of course biases the vote on re-opening, as some people who would have originally voted to retain the question (had that option been available), may not want to promote closure wars. Also, after closure, fewer people will see the question, especially those who saw it while open and did not know there is any vote in progress.

Several recent closed questions were reopened, after the matter of closure had consumed much time on the meta. This indicates that, had the voting options been "close" and "keep open", either keeping open would have won outright, or the number of close votes minus open votes may not have reached as high as five. Thus, several of the closure wars would have been avoided had the voting system been more expressive.

As the user population grows, a closure system that requires a small fixed number of votes to close a question can be abused through the presence of a sufficient number of hard-liners acting individually or in concert (e.g., at present, five grammar pedants would suffice to close any question with a syntactically incorrect title). This would be much more difficult in a bipartisan voting system that allowed both Close and Keep Open as options.

A bipartisan voting system would offer:

-more certainty (or at least, a higher chance) that outcomes reflect a consensus

-fewer reversals

-less conflict on meta

-less potential for abuse.

Offered for discussion before adding a [bug] tag.

• Looks like a great idea; maybe this should be in meta.stackoverflow? – J. M. is a poor mathematician Sep 26 '10 at 9:45
• yes this is a good point & certainly we've been witnessing this happening a lot recently. – anon Sep 26 '10 at 11:35
• Is there any example of a question on math.SE that was closed, opened, and then again closed? – Rasmus Sep 26 '10 at 15:16
• @Rasmus: I glanced through the closed questions. If there are any like you describe, they are currently open. – Larry Wang Sep 28 '10 at 7:09
• @T..: I think the feature-request tag would be better than bug. – Larry Wang Sep 28 '10 at 8:23
• Answer to Rasmus reposted from other thread: ''[the phrase] "endless closure wars on meta" meant that there is a steady stream of such contests on the meta, not that any individual question will have an endless cycle of closing and opening (though that could happen as the user base grows). The more common cycle is closing-metadiscussion-reopening.'' A bilateral voting process would prevent the cycle entirely in most cases that, under the current system, go to the meta. – T.. Oct 5 '10 at 21:48

Agree, this would be a nice feature to have, especially in the formative stages of any such site.

Unfortunately, this has already been declined: How about a "Vote not to close" option to counter the "Vote to close"?

As a workaround, I suppose you can always add a comment to the question saying why you think a question should not be closed.

As far as I can tell this is not something we have individual control over; the issue has also come up on meta.MO. Our provisional solution is that users with the appropriate amount of reputation can leave a comment to the effect that "I cancel out the next vote to close." Whether these comments will be honored is up to the community, but in lieu of a software solution it's not bad.

• Yes, in fact we could simply decide not to use the built-in close system and agree on our own. E.g. yes/no close votes could be comments and the post could be flagged when if/when it reaches enough close votes. Or there could be a CW answer that people edit to add their yes/no close votes and keep a running tally. Perhaps the tally could be kept in a (single?) meta thread instead of the question thread. – Bill Dubuque Sep 27 '10 at 4:33
• In fact we could use up/down votes to a "vote to close" question on meta to count yes/no close votes, except that would allows everyone to vote to close. – Bill Dubuque Sep 27 '10 at 4:38
• For posterity, here's a link to the relevant meta.mathoverflow.net thread: tea.mathoverflow.net/discussion/506/… – Rahul Sep 30 '10 at 5:17
• I've tried this occasionally, but it was never clear that it actually worked, until today, when it became clear that it didn't. How well does it work on MO? – MJD Mar 12 '13 at 15:22
• This question was recently closed despite two well-supported comments that requested that it not be closed. I am coming to the conclusion that this strategy simply does not work. – MJD Apr 26 '13 at 15:26
• @MJD just voted to reopen that question. Perhaps you might follow. – James S. Cook Sep 30 '13 at 0:54
• Thanks. I just tried to reopen that one myself, but it informed me I had already voted to reopen. – MJD Sep 30 '13 at 1:44