The title is the same as in the corresponding topic on Meta.StackExchange.com

Unpinning the accepted answer from the top of the list of answers

Today we changed the way we sort answers on Stack Overflow. We no longer pin the accepted answer (with the green checkmark) to the top of the list of answers. By default, we now sort strictly by votes (descending order by highest score), and the accepted answer's order in the list is based on its score.

The history of feature requests to unpin the accepted answer on Stack Overflow dates back to 2013 and has been raised almost every year since. This year within the Outdated Answers project we did research that showed that we are good to proceed with the accepted answer unpinned.

We can change the way the engine sorts answers in site settings. We would like to hear from you all if it is something you want to see on your site. (Please let me acknowledge in advance that we will not be able to run a test on each site.)

I believe that we should discuss about this also here.

  • I made a pseudo-feature-request on Meta.SE, suggesting a per-site option for moderators to pin answers above the accepted answer based on community feedback. This is similar to our usual approach of raising a question on Meta.TeX.SE if we want to discuss question closures, re-openings, tagging and the like. Link to request.
    – Werner
    Oct 13 at 23:54
  • 1
    @Werner there is also a real feature request for this: meta.stackexchange.com/questions/369576/…
    – Marijn
    Oct 14 at 8:30
  • @Marijn: Thanks for that. That feature request is specific to meta sites only, not the main sites (unfortunately).
    – Werner

I ran the query posted on the main meta for our site (https://data.stackexchange.com/tex/query/edit/1462421) which shows there are currently 8958 questions where the accepted answer is not the highest scored answer and would therefore be impacted by this change, which is 4% of the total number of questions.

This was more than I had expected. I think unpinning the answer would therefore be a good idea on TeX.SE, as the higher scoring answers are usually better and therefore deserve to be shown first.

  • 7
    Looking at the results of the query, it's not very clear that the higher-scoring answers are usually better: in many cases they have more votes because of being more impressive, rather than being more relevant to the question. Sep 19 at 19:27
  • 1
    I am in favour of this change. I just ran a field test ;).
    – AlexG
    Sep 29 at 10:46
  • Have you tried the modified SEDE query mentioned here? Oct 4 at 1:00

My personal opinion is that some users are a bit fast in accepting answers: “I have a problem, ask it at TeX.SX, someone answers, problem solved, tick.”

While in many cases the first answer is good, in particular when the problem is simple, in other cases the first answer glosses over some aspects of the problem, proposing

  • a not very general solution, or
  • a solution that's suboptimal, or
  • a solution that works by chance, or
  • a solution that's altogether wrong.

Yes, the last case happens. Sometimes the later answers generalize a correct accepted answer, making the solution applicable to other situations. Pinning the accepted answer at the top can hide better answers, maybe given months or even years later when new tools have become available: nicematrix, tabularray or expl3 come to mind, but also package updates that provide the feature asked for in the question.



Upvote if you prefer to unpin

Downvote if you prefer to keep accepted answers at the top


I'll add my contrarian view here. While it's undoubtably true that there are accepted answers that are either no longer valid or not as good as other answers, I'm not sure that this fact alone is enough of an argument to unpin accepted answers. Surely the measure of whether it makes sense to unpin is if that number is sufficiently large, and other than the measure that @marjin gives in his answer, it's not obvious that we really know how many answers deserve to be unpinned.

On the other hand, if the vast majority of accepted answers are valid 'best' answers, then users lose out when all accepted answers are unpinned.

A better approach IMO (although presumably not one that the SE network would like to implement) would be some way to unpin particular accepted answers, perhaps via a voting system like the close voting system.

But absent such a system, I'm not convinced that unpinning all accepted answers isn't like throwing the baby out with the bathwater.

  • 4
    "if the vast majority of accepted answers are valid 'best' answers, then users lose out when all accepted answers are unpinned." Why? If the accepted answers are the best, they will be still at the top because they get most votes Oct 5 at 18:59
  • @samcarter_is_at_topanswers.xyz Maybe so. I guess it depends on whether votes always equals best answer. But you're right in the cases where they line up it shouldn't be a problem.
    – Alan Munn
    Oct 5 at 22:20
  • Of course one needs to trust that users who upvote a several years old post because they recently got an answer by the same author and even document this abuse of votes in a comment, are the exception and most users will use votes wisely. Oct 5 at 22:48

This site covers technologies that are actively under development, and it’s been around for a long time. It’s sometimes the case that an old question has an accepted or highly-voted answer that worked well at the time, but has become obsolete since then. We might want a newer, updated answer to be at the top for anyone trying to solve the same problem today.

  • 1
    Does that mean you are in favor of pinning the accepted answer (because the OP could change the accepted answer) or against it (because then voting can overtake the old accepted answer)?
    – Marijn
    Oct 12 at 9:22
  • @Marijn I think there should be a way to discuss and pin a new answer, in cases where an edit isn’t a good solution.
    – Davislor
    Oct 12 at 17:33

You must log in to answer this question.

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