Check in on the accepted answer being unpinned; do we want to leave it as is or re-pin it to the top?

Question

Way back in September of last year we posted a meta asking if the community wanted the accepted answer to be unpinned from the top of posts. The vote was very close but an answer suggesting unpinning won and we've been running an experiment to see how people feel on the matter. Admittedly this experiment lasted longer than we wanted to so sorry for taking so long on getting back on this.

As this is a check in on how people feel we would like to know mainly if you would like the accepted answer pinning to the top again or leaving as is: unpinned. Why, or why not, should the accepted answer be pinned?

---

You can see the list of affected questions using this SEDE query. Since many users are specifically concerned about this change for story identification questions, you can see the list of affected story ID questions using this SEDE query and may choose to vote for the accepted answer in order to ensure it is at the top whether or not the pinned option is chosen.

---

Outcome

The voting as of 09/12/2022 is 26 in favour of going back to pinning the accepted answer and -8 in favour of leaving it as is, unpinned. In light of this I have added the status-review to this post for the CM team to look into repinning the accepted answer for us.

Answer 1

Thanks for your patience in getting this change enacted here.

I've now changed this site setting back for the main Science Fiction & Fantasy Stack Exchange site, so that the accepted answer will once again be pinned to the top of the answers list for a question. This community's consensus seemed pretty firmly in favor of making this change, so it was pretty straightforward.

(Accepted answers were already pinned on this Meta site; that has not changed.)

Let me know if you have any questions!

Answer 2

Let's go back to pinning accepted answers.

(Usual disclaimer that this is my personal opinion, mod hat off, and I do not represent the view of the whole mod team here.)

Anecdotally, in every case I've seen on this site over the last year where the accepted answer is not the highest-voted answer, it's been a bad thing to not have the accepted answer pinned to the top.

• Let's get the elephant in the room out of the way: story-ID questions. We have a whole site policy on question closure in our biggest tag which is based on acceptance. That policy hasn't changed, but it gets a little more confusing to apply when the accepted answer isn't at the top. You click through to what you think is a dupe target, the top answer is one story, but no, wait, that wasn't correct and the accepted answer is a completely different story.

• Another common occurrence on SFF is when new canon is released that makes a previously correct answer outdated. Example. If the poster of the old answer is no longer active, the only way to get a correct answer is for someone to post a new one. And right now, the only way to get that correct answer to the top is to wait however many years necessary for it to get enough upvotes to pass the old, perhaps 100+ scoring, top answer. Back when accepted answers were pinned, the new correct answer could reach the top much faster, if the question poster was still active. (That's not always the case, of course - askers can go inactive just like answerers can - but at least sometimes there's a quick fix from an OP, rather than waiting for dozens of votes to tip the balance.)

A single user choosing the answer they want is not a reliable metric for correctness, true. But neither is the lottery of HNQ upvotes. Either with or without pinning accepted answers, we'll sometimes have wrong answers at the top. That's unavoidable. It is an imperfect universe. But:

• If a bad/wrong answer is accepted, the active community will normally react to that mistake by downvoting it, which provides a visible indication, on the answer itself, that it's not a good answer. Even if it's pinned to the top, the negative score sends a useful message to readers. A misguided OP only has a single vote, which is easy to counteract by just a few other voters among the active community members.
• If accepted answers aren't pinned to the top, and a wrong answer gets a lot of upvotes (usually from HNQ drive-by upvoters), it's very hard to counteract that positive score with the relatively few votes of active curators. Without pinning accepted answers, we can have wrong answers at the top with no indication that they're wrong (except comments). The only indication would be when you scroll down and find another answer accepted.

We also need to accept (pun unintended) the limitations of the system here. As far as I know, SE has no plans in the foreseeable future to make accepted-answer-pinning adjustable per tag (e.g. pin them to the top for story-identification and not for anything else), nor to create a different kind of pinning to the top that doesn't depend on the OP alone. Nij mentions the possibility of some kind of pinning that's orthogonal to OP preference, which could be a great idea, but in practice, we have only two choices - rank answers by score, or pin the OP's choice to the top - and other may-never-be possibilities are a distraction.

Answer 3

Don't go back to pinning accepted answers.

The good reasons for removing the pinned behaviour still suffice to justify the current position.

• The purpose of ranking answers is to make them easier to find. Rank by score is intended to make the best answer as determined by the community be the most visible; rank by changes is intended to make recent changes more obvious for voting and commenting to be updated as well. In neither case does the opinion of a single user help this, much less because they happened to be author of the original text of the question.

• Acceptances are inconsistently used, including in their frequency. Acceptance can be applied for any reason, or for none, and there is no formal systematic way of changing it. Also importantly, there is nothing that makes them choose, even when they have explicitly noted that the answer is most helpful to them. This should be enough to remove any positional emphasis from acceptance.

• User confusion over the meaning of acceptance stems from both these things, as well as their own expectations. The checkmark does not indicate accuracy reliability, consensus, or even meta-quality. Emphasising an answer in this way has caused confusion across the network, including on SFF. (Yes, users have been confused by the lack of pinning as well, bit this only because they were following a habit created by the use of a bad system) (and this is a result of rare coincidence, to the extent that it is remarkable when it does occur, rather than so regular that it can be read into discussion by default)

• In the rare cases where pinning a specific answer to the top (regardless of score or timing or other properties) could be beneficial, this is an orthogonal consideration to what the original question author likes most. Tying the two actions together means both must carry the downsides of the other, without adding any further upsides. Even if pinning is needed and brought back, it should not be connected to acceptance at all.