248

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 This year within the Outdated Answers project we did research that showed that we are good to proceed with the accepted answer unpinned.

Would you like to have the accepted answer unpinned on your site?

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.)

Currently we are planning to move forward with one of two scenarios, based on your feedback:

  1. Unpin the accepted answer on all SE sites by default and pin it back on a few sites that ask us to do so.
  2. Keep the accepted answer pinned on all SE sites by default and unpin it on a few sites that ask us to do so.

Please let us know what you think will work best for your site! If you can discuss this question with your community it would be awesome. We are going to collect feedback before the end of September 19th.

Site settings are per-site and child metas are also considered a "site".

Update

As Anita Taylor shared:

The Public Platform team will go through these requests and determine what to prioritize, defer and decline. We ask for your patience -- key members of the team are on vacation over the next few weeks, so we won't be able to triage these requests until mid-to-late October.

Update 2

Thanks to everyone for the feedback. Seeing that some sites do not want to unpin the accepted answer, we decided to move forward with the status quo and not to change the default behaviour on existing sites. If you think unpinning the accepted answer on your site makes sense, please do the following:

  1. Start a meta discussion on your per-site meta.
  2. When the discussion is completed and the community agrees to change the default behaviour (agrees to unpin the accepted answer), please ask one of the site moderators to add the tag to the question. It will generate a ticket and let CMs know that we need to adjust the site setting for your site.
  3. We would appreciate it if you add the outcome of the discussion to the accepted answer in this question. This will allow all of us to see the full picture.

There's no need to do this on the sites that are currently on the list; we have already adjusted the site setting for them.

Improve this question
44
  • 48
    My upvote on this post reflects that I think it's great to solicit per-site feedback instead of copying SO changes to the network, not that I necessarily support the change made on SO.
    – Rand al'Thor
    Sep 8, 2021 at 16:47
  • 59
    This is just brilliant! Not only because I very much want to not have the accepted answer first (they're too often not the best answer), but also because of the way you are approaching this, asking for feedback before pushing network-wide. Thank you!
    – terdon
    Sep 8, 2021 at 17:07
  • 20
    I think it would be a mistake to make this different per site. I would much rather unpin it every where and have a link near the answer sort selection to navigate directly to the accepted answer. If there is an accepted answer, there's a nice visible green button or something to get there. That makes it easier to document accepted answers as well.
    – ColleenV
    Sep 8, 2021 at 17:12
  • 4
    What is the process to request this for a specific site? I'll have to check with the other mods, but this is something we specifically requested on Skeptics years ago as the concept of accepting an answer doesn't really make sense for us, and the pinning is harmful.
    – Mad Scientist
    Sep 8, 2021 at 17:58
  • 46
    @ColleenV: Speaking personally, not as a CM (I'm certainly not speaking on behalf of what the company will do): I like the idea of unpinning the accepted answer, but leaving some easy way to navigate from the question to the accepted answer (e.g. something like "This user has accepted an answer. [Click here to jump to the accepted answer.](link)."
    – V2Blast
    Sep 8, 2021 at 17:58
  • 6
    @MadScientist: Given that it hasn't yet been decided how this will be rolled out network-wide, I think that right now, the idea is to help decide whether the accepted answer should be unpinned or remain pinned on the rest of the network besides SO - and then, once that's settled, a process will be established for having the relevant setting be enabled or disabled on specific sites.
    – V2Blast
    Sep 8, 2021 at 18:01
  • 5
    It looks like Google lists the accepted answer first instead of highest scored Is that something SE controls? (For that particular question, that's the desirable outcome, but it may not always be)
    – ColleenV
    Sep 8, 2021 at 18:35
  • 7
    @Mari-LouA No? Votes still sorts by highest votes. Active is most recently active, meaning it's been added or edited recently, Oldest is sorted by date of post, oldest first. Neither of the other views is by score, only Votes. Making Active into Newest would just be a reversal of the Oldest sort, which, while potentially useful, is not the same as Active sort.
    – Catija
    Sep 8, 2021 at 19:44
  • 7
    Whatever you do, please don't roll this out to all sites before we have time to discuss it among ourselves :) A week at least!
    – bobble
    Sep 8, 2021 at 20:13
  • 13
    Has this proposal been communicated to the moderators across the network?
    – Mari-Lou A Слава Україні
    Sep 9, 2021 at 18:46
  • 5
    Can you clarify whether this change will affect local child metas as well, and whether the setting can be different on the main site and its meta (i.e. pin on meta but not on main or vice versa)?
    – ACuriousMind
    Sep 10, 2021 at 14:14
  • 5
    @CGCampbell I think the usefulness for the asker is not relevant anymore after the accept, (his/her/or else) problem is solved, what we need to focus then is the usefulness for new visitors (so, it's best to have a "colectively approved" solution at the top, sorted via vote instead).
    – Largato
    Sep 14, 2021 at 18:01
  • 6
    @CGCampell " we no longer care about the usefulness to the asker (?)" I think that nails it. Indeed we don't. The asker can of course kind of say that no existing answer was helpful to him, but what does it matter to anyone else? It may be a hint that people misunderstood the question or that it's still unsolved but it also may just mean that the asker moved on. In the end, the signal is not reliable. So better not put much weight on it regarding the sorting order
    – NoDataDumpNoContribution
    Sep 14, 2021 at 22:15
  • 4
    I wonder if the simple "accepted post is pinned second, if it's not the top-scoring answer" works (?).
    – Rebecca J. Stones
    Sep 16, 2021 at 3:33
  • 11
    Is it worth revisiting the decision to not allow us to delete accepted answers now? I have an answer that at the time was "there is no feature that supports what you are doing.", but there is now such a feature. It was accepted at the time, but there is no way for me to delete it now and the asker is unresponsive to moving the answer flag. Seems like the original limitation to prevent the accepted answer from being deleted might no longer be useful.
    – Matt
    Sep 20, 2021 at 0:34

39 Answers 39

Reset to default
1
2
4

Anti-

Do not reintroduce the accept vote into any ranking mechanism.

We've already removed some of the most insidious unbalance that the accept vote brings on the system. Please make sure none of it is introduced back. The accepted mark has too many benefits that we don't need to add more on top of it.

It gives extra rewards, effectively almost tripling the potential reward that a single user can do for free and that also benefits the user that do it (+15 to the answerer and +2 to the asker), which is unlike bounties. It also removes the question from the unanswered list, which makes sense since the asker doesn't have the privilege to cast an upvote to select their preferred answer (this have in some cases motivated me to upvote a good answer and accept other, just because the accepted one was the first I used).

It also would give unduly preference to a single user vote over the others if reintroduced to affect the order of answers. We don't need a tie-breaking mechanism that would make the selection more complex, as we've been able to live without it when the asker doesn't accept any answer.

Leaving ranking of answers to be only in the power of users that have demonstrated some modicum experience with the site would undoubtedly benefit the content that we are creating here, so that the best answer always stays to the top.

Improve this answer
3

I'll repeat something I suggested on the math meta post:

A soft option would be to make the tick ✅ count for some K≥1 upvotes, say K=5 . Then early on, the accepted answer will be up top, but if a newer answer can prove itself, it will beat the accepted answer.

I would have liked to spend more time researching its effects eg with SEDE but I had other commitments. Its similar to William Walker III's suggestion from politics.se to use a percentage modifier e.g. 125%, but it might be better suited for depreciating old answers (especially on popular posts) while still effectively pinning accepted answers before there are many votes.

Two basic examples comparing the two: if all answers have ≤3 votes, then 125% acts the same as a tie breaker. If the accepted answer has 100 votes, then 125% acts like my suggestion with the rather large K=25.

Improve this answer
3

I would suggest unpinning accepted answers, but with a certain UI/UX change.

There should be a checkmark under the question saying "this question has an accepted answer" that on hover/touch transiently shows overlay with accepted answer. This gets us best of both worlds - current best answer on top and quick access to the answer that solved the question upon request (simply by hovering).

Improve this answer
1
3

Unpinning makes sense for some questions and not others

Stack Exchange has three types of questions on the network:

  1. Ones with an objective answer that is correct, and a different answer may become more correct over time
  2. Ones with a combination of objective and subjective components to answers
  3. Ones with almost entirely (but not completely) subjective components to answers

All three of these types are answerable on stack because all three types allow for expert opinion to avoid wrong answers. It is possible for there not to be a correct answer to a problem while simultaneously having very definite wrong answers. For example, "How would I solve world hunger in the world I am creating for my novel? (parameters are x)" would be a fine question for Worldbuilding, it doesn't necessarily have one true "right" answer, it does, however have at least one very wrong answer "have [insert disaster] wipe out everyone on your world". Sure it solves the problem of world hunger, but leaves the querent without a world to build in. Expert opinion allows answers to avoid the wrong, and provide potential good solutions to the problem.

Whether or not a site should use the unpinning feature or not depends on what mix of these three types of questions there are on that site. Sites that have more of types 2 & 3 should retain pinning, sites that have more of type 1 should remove it.

What does pinning actually achieve?

Pinning an answer gives the querent an active opinion on which answer solved their problem. Unpinning changes that opinion to being more passive (readers have to scroll past the first answer to get to an unpinned accepted answer, which is the use case this change is attempting to solve).

This decision seems to be based on the assumption that the querent is the least expert person in the discussion. Making this assumption in all cases is a dangerous road to go down.

The experiment this change was based on, was run on the site which has a plethora of objectively answered questions (Stack Overflow). In that situation the querent will be the least expert person in the room.

Someone, on the other hand, asking "How should I handle situation x which is causing issues for reasons x & y" is not a situation where the querent is the least expert person in the discussion, it's a situation where they may be the most expert person on what constitutes the best answer.

The moment you bring a subjective topic into a question is the same moment the fundamental assumption that the SO experiment was investigating shifts dramatically. For example, the querent cannot possibly put down all of the history and nuance of the interpersonal relationships (because space), however they are uniquely qualified to spot what is the best answer for their situation once presented (in that situation).

Example Site Classifications

For sites where there is an objective answer to a question (Physics, Mathematics, Stack Overflow, etc.), the querent's opinion doesn't matter as much on what is the best answer, as the best answer for the site is "what is the highest quality and most correct answer", which is a criteria that can change over time.

For sites where instead there is frequently going to be a subjective answer (Role-playing Games, Parenting, Politics, etc.). Answers to these types of questions will (commonly) have two components, objective facts (e.g. game rules) and subjective opinion (how those game rules should be interpreted and applied). The querent's opinion on what solved their problem should hold more weight for these questions as it is a prominent and visible signal of what solved the subjective part of their problem.

There is a third set of questions which are (almost) totally subjective, those dealing with interpersonal relations. Given the focus of some sites (like Role-playing Games), we allow a subset of those (e.g., managing unruly players in my game, or x player's behaviour is unfun). For these types of questions the most correct answer is the one which solved the querents question the best, and this can only be indicated by pinning their accepted answer.

So which of the two approaches should we take?

Given the different mix of questions that sites can have, it should be an opt-in situation for a site to de-pin accepted answers, as opposed to an opt-out of the de-pinning. This would produce the least harm.

As a counter point, I think a better approach would be to try and classify most sites in the network based on the balance of these three buckets of questions on their site. Then provide an initial rollout of de-pinning to those sites with primarily questions of type 1, with an opt-out for these sites. It should then be opt-in for all other sites.

When a site is in beta, it should be a configurable criteria by the mods of that site, so they can experiment and figure out which approach works best for their fledgling site. The final decision should then be made once the site graduates out of beta.

Improve this answer
5
  • 5
    Some sites don't fit into these categories - for example, Puzzling has many questions with objectively correct answers which will never become outdated. But then Puzzling bucks the Q&A model in a lot of other ways as well.
    – bobble
    Sep 15, 2021 at 18:39
  • @bobble I would put puzzling as containing type 1 questions. The "may become outdated" isn't a mandate that they will, just a possibility. For puzzling that possibility is 0 (without breaking logic). For something like physics.SE the possibility is also close to 0 (for example it's exceptionally unlikely that an answer about relativity would become outdated based on our current understanding). Puzzling, however, does have other features which make it particularly unique, and thus potentially unsuitable to lose the pinning behaviour even though it's primarily populated by type 1 questions.
    – illustro
    Sep 16, 2021 at 4:33
  • For example, being first with an answer to a puzzle is an important criteria on puzzling.SE which the pinning behaviour rewards.
    – illustro
    Sep 16, 2021 at 4:36
  • 1
    There are quirks - for example, riddles would ideally have an objectively correct answer but sometimes the factor determining checkmarking is just whichever answer was intended if multiple are plausible. (We have a close reason for that, if too many answers are plausible). But even then, which answer is considered correct by the poster is important information for evaluating the puzzle and the other answers, so pinning is beneficial.
    – bobble
    Sep 16, 2021 at 4:40
  • "This decision seems to be based on the assumption that the querent is the least expert person in the discussion. Making this assumption in all cases is a dangerous road to go down." <-- my jaw dropped when I read this. It perfectly sums up the problem with automatically unpinning accepted answers.
    – Kenn Sebesta
    Sep 22, 2021 at 23:34
3

There seems to be a diverse number of answers ranging from all-in to all-out. So, here's another one to throw in the mix that aims at something already covered by the scope of each site:

pseudo-

Allow moderators the power to pin a specific answer above the accepted answer.

The motivation here is that the scope of per-site metas are specifically to discuss site-specific content. This includes questions about tagging, re-opening questions, why questions were closed, etc. Why not make questions about pinning a new (more appropriate) answer above the accepted answer part of that discussion. That is, someone would post a question, suggesting answer X should be pinned above the accepted answer Y and state the reasoning/motivation. The community can vote on it and a moderator takes action/not depending on result.

The accepted answer tick mark still represents "the answer that helped the OP most" (this was the original definition of the accept tick, if I recall). Not necessarily THE answer, but just the one they ended up using. A community discussion can opt for something more representative for the community to be ranked first, if that's the case.

Perhaps some sites can do with an all-in (top-voted answer is ranked first) approach, so this idea suggests a middle-ground that let the sites decide how they'd like to handle content on a per-question basis as and when needed.

The potential exists (although it's not necessary) for some larger reputation milestone for users who could achieve this new pinning privilege.

Improve this answer
2

Currently we are planning to move forward with one of two scenarios, based on your feedback:

  1. Unpin the accepted answer on all SE sites by default and pin it back on a few sites that ask us to do so.
  2. Keep the accepted answer pinned on all SE sites by default and unpin it on a few sites that ask us to do so.

I think the change should be applied network-wide.

This point formed part of the answer by ColleenV, but I think it's a key point that merits its own consideration independently of the UI suggestion about a button.

As well as being "better than forcing each site to come to a consensus about whether to pin or not to pin" (per ColleenV), it is a more consistent overall UX for the network as a whole (who uses just one site?). Admittedly, sites already have subtle UI variations, but I feel this one would be a step too far.

Improve this answer
-2

One thing people keep bringing up here, is that the asker of the question may not be the best qualified to judge which answer is the best. Could we make it so that the asker can earn the privilege to pin an answer to their own question, if they reach a high enough reputation, or earn for example, a gold (or platinum?) badge on one or more of the tags on the question? Since at least six years ago, many users have expressed support for adding more reputation levels anyway :)

Improve this answer
16
  • 9
    Reputation is not a measure of expertise in any topic other than “how to participate on the Stack Exchange network in a way that earns more reputation than it loses”.
    – ColleenV
    Sep 14, 2021 at 16:36
  • 4
    @ColleenV yet reputation is what is needed to earn the privilege to upvote, to downvote, to edit, and to do many other things.
    – user1271772
    Sep 14, 2021 at 16:37
  • Making pinning a privilege does not solve the problem of askers being unqualified to judge which answer is best. Expressing "This is the answer that was most helpful to me personally." is not a privilege that needs to be earned. Having to earn the right to control how an accepted answer is displayed doesn't seem to fix anything. There will still be users pinning answers that the community thinks are wrong. There will still be answers that were fine when they were pinned, but are now obsolete.
    – ColleenV
    Sep 14, 2021 at 17:04
  • 2
    For 13 years SE has allowed the asker to pin their preferred answer to the top. I'm proposing a less drastic change.
    – user1271772
    Sep 14, 2021 at 17:45
  • 7
    So, just to get this out of the way, I didn't downvote. I'm pointing out that I don't understand how this would be "better" instead of just "different". If you made that a little clearer, maybe I would be able to upvote it. I'm sorry if my comment came across as picking on your idea - that was not my intention.
    – ColleenV
    Sep 14, 2021 at 17:51
  • 2
    @ColleenV This answer had a positive score of +1 before your comment. Often a negative comment can be 10x worse than a downvote. My understanding is that you feel that high-rep users may not know enough about a site's topic to deserve this privilege. Could you give an example of a high-rep user that is not very competent in a site's topic?
    – user1271772
    Sep 14, 2021 at 18:13
  • 14
    Or maybe more than one person is having a negative reaction to your idea for the same reasons I articulated. Would you rather have downvotes with no comments explaining why? The topics a site covers are very broad. Show me a high reputation user on SO who is an expert in all of the top 50 tags. Also, remember you're talking about the person asking the question, who probably wouldn't be asking if they already knew everything about it. Reputation is not a measure of expertise. It's a measure of participation on the network and whether you can be trusted with various privileges.
    – ColleenV
    Sep 14, 2021 at 18:42
  • 5
    You do realize that people have accepted wrong (as in malfunctioning) answers, as well as answers that contain known vulnerabilities or explicit Easter eggs to indicate that they were copied from Stack Overflow, right? Allowing the OP to rank any answer gives the impression that they are an authority on what Right™ is, which leads to situations where people just copy bad code only because they blindly trust Stack Overflow. So uh, no, the asker should get the ability to indicate that this answer worked for them, but if it really is bad, then the community gets to say that it is.
    – Makoto
    Sep 16, 2021 at 17:54
  • StackOverflow is not the only site in this network. Furthermore: for 13 years, OP has been allowed to pin their preferred answer to the top. Please stop pretending that this was a crazy, non-sensical idea.
    – user1271772
    Sep 16, 2021 at 22:04
  • 1
    This. As an example bicycles.stackexchange.com/questions/49886/… the accepted answer is bad, unsafe, and flies in the face of common sense. But OP accepted it so its at the top and will be first-answer to be read. OP was looking for confirmation, rather than asking a serious question.
    – Criggie
    Sep 18, 2021 at 22:33
  • Excuse me, why are you writing that comment on my amswer? I'm suggesting that the user would need a lot of rep to pin answers to the top, and you've now shown an example of a user with 183 rep who you think made a bad decision on pinning an answer.
    – user1271772
    Sep 19, 2021 at 9:11
  • I can only go by my personal experience, which is that the best answers are most frequently the highest upvoted ones.
    – President James K. Polk
    Sep 22, 2021 at 11:37
  • 1
    @ColleenV stackoverflow.com/help/badges/15/generalist
    – Ekadh Singh - Reinstate Monica
    Sep 29, 2021 at 12:10
  • @EkadhSingh-ReinstateMonica If you're implying that getting a better than 15 score on an answer in a tag is a measure of expertise, I would disagree. Even if we forget for a moment that tagging is woefully inaccurate, a 15 score just means the person who is obviously the less informed of the pair liked the answer, not that it was correct or particularly erudite. That's just one of several reasons I can think of that makes generalist completely irrelevant to my point.
    – ColleenV
    Sep 29, 2021 at 13:06
  • 1
    The comment was addressed to me so I responded to it. If they didn't want a response, they shouldn't have pinged me. And yeah I noticed that you didn't.
    – ColleenV
    Sep 29, 2021 at 17:16
-11

It would be amazing if the OP could choose to pin the accepted answer or not! It would be better if there are choices.

Maybe, you could add a new privilege - Vote for pinning accepted answers for the high-reputation users and asker.

Or, you could just limit this privilege to the question-asker and the moderators

Improve this answer
1
  • 7
    I think it's an interesting idea, but I don't think it solves the issues that pinning an accepted answer causes. This feature would make it seem like the issues could be solved if the community just tried harder, when the entire idea of pinning a single answer is kind of broken no matter how it's done.
    – ColleenV
    Sep 14, 2021 at 13:11
-17

TL;DR: change the design of accepted answers so that it's more obvious an answer is accepted, now that it's not pinned. My idea is have a background color for it. Mockup:

mockup of background for accepted answer

Reasoning:

The unpinning pretty much renders the accepted answer mechanism as whole almost pointless.

People knew that an answer was the "best" (in terms of helping the original question author) when it was on top. It was simple. Now people will just see a weird green "V" icon to the left of an answer and might just wonder what it means.

It's not bad; it's just a bigger change than what most people think, i.e. not just moving cheese, but turning the cheese into something else.

So to "compensate" over this, at least make it more obvious an answer is accepted, so that it will not be totally lost.

Improve this answer
9
  • 2
    I'd instead want the preference idea in another answer. Should that be implemented, this would be moot.
    – Sonic the Anonymous Hedgehog
    Sep 8, 2021 at 16:52
  • @Sonic two different things. I don't talk about people who want to see it first, just a way to make it more clear "this answer is accepted, even though it is not on top".
    – Shadow Wizard Is Sad And Angry
    Sep 8, 2021 at 16:53
  • If someone's disabled such a preference, I'm certain they wouldn't appreciate the answer being highlighted.
    – Sonic the Anonymous Hedgehog
    Sep 8, 2021 at 16:53
  • 21
    I disagree with this. The whole point of ending the accepted answer pinning is to lower the "authority" that the check yields, to highlight it less. I think that, a lot of the time, the accepted answer shouldn't matter to most folks (though, to be fair, I'm speaking largely about SO here), and the new change reflects this. The way I see it, adding more highlighting to accepted answers would counteract what this change aims to achieve.
    – zcoop98
    Sep 8, 2021 at 16:59
  • @Sonic maybe people would like it, maybe not, we can't know. I'd like it, and keep in mind there are already background color used in several places: deleted posts, questions in lists which match favorite tags, etc. So the idea of background color for special occasions isn't new.
    – Shadow Wizard Is Sad And Angry
    Sep 8, 2021 at 16:59
  • 5
    As a sidenote, I'd posit that this would be an easy userscript to implement.
    – zcoop98
    Sep 8, 2021 at 17:01
  • @zcoop98 there's no way I'm installing a userscript on my workplace PC. Give me vanilla.
    – bad_coder
    Sep 8, 2021 at 17:03
  • 2
    The point of unpinning is to remove the special emphasis on that answer. The checkmark is staying, we don't need another indicator.
    – Nij
    Sep 9, 2021 at 0:10
  • 7
    "The unpinning pretty much renders the accepted answer mechanism as whole almost pointless" - emphasis on almost. The way I see it, you get 15 extra rep for actually solving the OP's original problem. That's it! From there it's up to the community to decide the order of the answers by votes. The questions and answers should serve a greater purpose than to the one asked the question, but you still get a little "bonus" for actually helping them
    – Tomerikoo
    Sep 9, 2021 at 19:30
1
2

You must log in to answer this question.

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