As a spinoff of my other Meta question:

In the case of a new question, which duplicates other, unanswered questions (or questions with no good or accepted answer), we should close the older questions as duplicates instead of the new question.


  1. The new question is more likely to attract new answerers, because a lot of us just trawl the "recent questions" feed.
  2. The new question asker is a lot more likely to accept an answer, completing the question, than the asker from a question that is years old (I've never seen a 2+ year old question accept an answer).

Therefore, if it's our goal to have good questions with good, accepted answers as a source of knowledge, it makes a lot more sense to close the old unanswered question than the new one.

Now, clearly "no good or accepted answer" is a judgement call, and it should be up to individual reviewers to make that call. But we can share this out as a pragma and start following it most of the time.


I disagree with the suggestion of favoring a new question over an old one when closing duplicates.

I think whether an answer is accepted or not is overrated - while it indicates that one specific user found a given answer (the most) helpful1, it’s not necessarily what a future reader with the same problem would consider as such. The community vote on the other hand is a much better indicator of quality. So we may (or may not) get more accepted posts, but the quality of the answers is not affected by a green check mark.

In an ideal Stack Exchange world, a duplicate would get closed before any answers are posted and the informed users would vote to close instead of answering. We know that this corner of the SE world is not so fast, and that’s ok. But the underlying principle of “oh, look, we already have a couple of answers for you, isn’t that cool” still persist so I think:

  • As a default, the older post takes precedence and newer Q/As get closed. Note that the new and potentially very good answers are not lost, they still remain on the site as does the question. There’s also the possibility of merging questions and answers, should the need arise (but it’s not a default procedure).

  • In the case where a newer question is significantly more generic and potentially more useful (think about how this question could be the reason even newer questions get closed), it may be better to close the older one as a subset of the newer. But I would expect to see that argument documented in the comments by the respective close voters.


1 And that may also be because this was the fastest post or determined by some other factors. Including the Fastest Gun in the West Problem.

  • Er ... did you miss the part where I'm talking about older questions that have NO good answers? – FuzzyChef Jan 8 at 0:53
  • No, I didn’t miss it. rumtscho explained in the corresponding Q/A that the potential of getting answers, good or otherwise, hinges on the community knowledge, not on the age of the post. – Stephie Jan 8 at 6:22
  • So ... assuming that we have an old question. Say 5 years old. What's the workflow for that question getting answered by anyone? How would someone who knows the answer, and recently joined SA, discover it? – FuzzyChef Jan 8 at 21:49

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .