17

There is a on going Meta SE post about unpinning the accepted answers on a post from the top of the list of answers. In fact, Stack Oveflow has already unpinned the accepted answers.

For Stack Overflow, there are good reasons behind the change. As a developer, I'll often see older questions on Stack Overflow that have answers that are outdated due to program languages changing overtime, thus at the time the question was originally answered and said answer was accepted, it may no longer be valid/the best practice in the present time. At the same time, someone may have posted an updated and valid answer that has received more votes that would be buried down below the outdated accepted answer simply because it has the green check mark.


But does this make sense on our site? You decide!

SE has two scenarios laid out, but only one of the following will be done (quoted from the linked Meta):

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.

Thus, do we (Arqade) want to keep the accepted answers pinned to the top, or do we want to unpin them?

Note: SE plans to collect feedback before the end of September 19th.

10
  • 2
    I'd like to first request some additional information before I make a judgment, namely a statistic: How many questions with an accepted answer have one or more OTHER answers with a higher score than the accepted answer?
    – Nzall
    Sep 16 at 17:53
  • @Nzall Here's something that might work. Took me a minute to figure out to query the data Sep 16 at 18:01
  • @Nzall Here is a percentage I came up with, I reused part of the query Wondercricket posted. Take this with a grain a salt, but I think I did it correct: data.stackexchange.com/gaming/query/1462149
    – Timmy Jim Mod
    Sep 16 at 18:44
  • Is there a SEDE query like Wondercricket's above, but sorts according to either question score or popularity (in terms of views)? This is so we can check how this will affect the top-voted or most popular questions. Sep 17 at 3:45
  • HNQ, drive-by votes, etc. Answers outscoring the accepted ones are not always what they seem to be.
    – dly
    Sep 17 at 11:17
  • 1
    @galacticninja I believe this is what you want. You can switch between the results of the "sorted by score" and "sorted by views" by clicking on the "[1/2]" part in the "Results" tab.
    – Timmy Jim Mod
    Sep 17 at 18:59
  • If it's not too much trouble, then would it be possible to have a separate tab in the answers area for "previously accepted answers" or something like that?
    – Cyanite17
    Sep 17 at 19:19
  • @PvParkour that would have to be something the SE team would implement. There are probably similar suggestions already made on the linked SE Meta posts though I haven't looked.
    – Timmy Jim Mod
    Sep 17 at 20:18
  • I forked Timmy Jim's query to also include the relative age of the accepted answer vs the top-voted answer. A positive number indicates the accepted answer is newer by that many days.
    – Schism
    Sep 18 at 20:00
  • 2
    After reviewing the results of @TimmyJim's SEDE query above, I believe that most of the top-voted and most popular questions would benefit if the accepted answers are unpinned. Top-voted answers tend to be more accurate, have broader solutions, or are more up-to-date, compared to accepted answers with lesser votes. While I noticed that there are cases where accepted answers are more up-to-date, I've also noticed that most question-askers tend to not change their accepted answers in light of new information probably because most of them are not active on the site anymore. Sep 19 at 2:43
15

One thing to keep in mind is that in a lot of cases, there is also the possibility that an accepted answer is actually the NEWER answer, which takes into account new information or applies to a more recent client version.

I modified Wondercricket's query from the comments to account for dates, and in roughly half of the roughly 3000 queries that have the highest answer having more votes than the accepted answer, the accepted answer is the more recent answer that most likely contains additional information or corrects outdated information.

So my vote is AGAINST unpinning accepted answers, because in our context, our users tend to reassign their accepted answers to account for outdated information quite often, and after checking a bunch of the questions in Wondercricket's query (not all 3000, I do NOT have the time for that), I've found that often times, it's simply the newest answer that is most accurate, regardless of whether it's accepted or not.

Speculation: IN THEORY, it might be better to show the highest voted answer by default, UNLESS the accepted answer is significantly more recent than the highest answer. However, I doubt Stack Overflow will go that way.

4
  • I agree. Only if the highest voted answer is more recent than the accepting event the accepted answer should be unpinned. If this isn't possible - leave it.
    – dly
    Sep 17 at 6:49
  • 3
    Hm, that's a good point. Not sure if this is possible, but what if the asker's activity factored in? If a user had been inactive for a certain amount of time, their accepted answers would be unpinned, or something like that.
    – SaintWacko
    Sep 20 at 23:41
  • 1
    Can you show us the modified SEDE query? The one that shows which cases have the accepted answer being newer than the highest-voted one? All I see is this: data.stackexchange.com/gaming/query/1462149 which does not show that. Oct 4 at 0:58
  • @user1271772 -- the sign of the AcceptedRelativeAgeInDays column of my fork indicates this. A positive number indicates the accepted answer is newer than the top-rated answer.
    – Schism
    Oct 4 at 3:23
10

I think accepted answers should be unpinned. While we may not see this issue as much as Stack Overflow, I have certainly seen cases where the accepted answer is no longer the best answer, or even is outright wrong, but is still the first answer anyone sees when they visit the question.

While accepting an answer is important, votes are a better indicator of how useful the answer is to the most people, not just the person who asked the question.

2
  • 1
    ^ Agreed. Often users might be deterred from looking at other answers because of seeing the green check mark Sep 16 at 18:03
  • 1
    The outdating argument is good. For example ≈90% of Minecraft commands answers that were written before 1.13 are no longer valid. On the popular questions there are often ≥5 outdated answers before you get to a valid one. Unpinning the accepted answer could at least sometimes reduce that by 1 and eventually the correct answer could rise to the top. Sep 18 at 2:30
5

With games-as-a-service being the industry trend I would vote to unpin accepted answers from the top.

Constant changes to games will see to it that even solid answers from the past can be quickly outdated.

3

At the time of making this post:

Summary of votes on the answers on this post (excluding answers that don't distinctly state a side):

Post Keep Pin/Unpin? Total Votes Positive Negative
Nzall's Keep Pin 16 15 1
SaintWacko's Unpin 12 9 3
childe's Unpin 5 4 1
Robotnik's Unpin 5 3 2
Schism's Keep Pin 1 0 1

Net Votes (Total Positive for all related Posts - Total Negative for all related Posts):

Votes To Keep Pin Votes to Unpin
13 10

While using the votes as a way of gauging the community consensus isn't the most accurate way to go about this (people could vote on multiple posts), it does provide some insight.

Since there doesn't seem to be an extreme towards either proposal, it seems that leaving the accepted answer pinned is what the majority of the community wants. Should we want to change this in the future, we should be able to by making a request out to SE Community Managers.

Thank you all for participating!

3
  • The above table currently states that Schism's answer is "Unpin", but it actually states "I vote to continue pinning the accepted answer to the top.", i.e., "Keep pin." Also, at the time of writing this comment, the votes have changed. I currently count 14 votes for keep pin vs. 13 votes for Unpin. Oct 5 at 3:11
  • @galacticninja I adjusted the votes based off of Schism's stance that I originally read incorrectly. As for the current vote tallies, I think this answer should remain as is since its an "as of" answer.
    – Timmy Jim Mod
    Oct 5 at 3:32
  • Yeah, it seems that the status quo will remain. I was just trying to mention that it's closer to 50/50 rather than a significant majority for "Keep pin." Maybe that should be indicated for Arqade in the related post in Meta SE? Oct 5 at 3:46
1

Something not really taken into account is that the highest scoring answer is sometimes not the best/correct answer either. It could be well written but objectively wrong, it could be outdated, or higher scoring by virtue of being posted earlier.

For cases where the OP is still active, this will remove their ability to 'pin' the most up to date answer by way of moving the checkmark to the new 'correct' answer.

Having said all that, my vote is still yes, it should be turned off by default. In the general case, a highly-voted answer is more likely to be useful to more people, and we have other avenues (comments, flags, post notices, deletion) to deal with highly-voted, incorrect or outdated answers.

Additionally, (and for any SE employee reading this): personally I would like to see "Always show the accepted answer above other answers" as an option that can be enabled at the profile-level. This would give users the ability to control it themselves, in much the same way we can choose to dim or hide ignored tags.

1
  • I also thought it would be nice to have a user setting to still show accepted answers first.
    – Timmy Jim Mod
    Sep 17 at 8:32
0

I vote to continue pinning the accepted answer to the top.

Several answers describe the case where the question asker is inactive and does not change the accepted answer. While not ideal, I don't think this is too much of an issue: the top-voted answer, if more relevant, will be the second answer in the list.

However, in the case where the question asker is active, and updates the accepted answer, unpinning would result in the new "best" answer being arbitrarily low on the list of answers by score. Allowing the asker to pin the newest answer allows it to be promoted immediately.

In other words, I think the worst case when we unpin the accepted answer is much worse than the worst case when we leave it pinned.

1
  • Open to hearing the reasoning behind disagreeing here -- this is meta; let's discuss!
    – Schism
    Oct 4 at 3:21
-1

Maybe have it on a tag-by-tag basis with it on by default?

For example, if someone tags a question with minecraft-java-edition || minecraft-bedrock-edition, it stays on because, as SaintWacko mentioned, the game updates frequently and has had & will do various changes.

Meanwhile, something like final-fantasy-15 is not being worked on anymore, so it's pretty probable that the accepted answer to a question about the game will stay correct in the future. Not that this couldn't cause problems, mind you, specially when it comes to technical questions.

1
  • 3
    This would have to be something the Stack Exchange team would have to implement, and probably too specific of a feature. I suppose they could implement a way for mods to set this option on a per tag basis, but I don't see that happening.
    – Timmy Jim Mod
    Sep 19 at 18:58

You must log in to answer this question.

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