7

It seems that Stack Overflow is getting a change and is no longer going to pin the accepted answer at the top and will now default to sorting by post score alone. In the post announcing this they posted a question asking how this should apply to other sites. Below are the choices being given and I was wondering how others thought we should tackle this.

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

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

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

Here are some stats about what questions would be impacted on our site. This comes from a query I got from another meta site question.

We have 8 questions with an accepted answer having a negative score and another answer with a higher score.

https://data.stackexchange.com/boardgames/query/1457490/questions-where-negatively-scored-accepted-answers-is-outscored#resultSets.

We have 220 questions with an answer that has a score at least 3 higher than the accepted answer.

https://data.stackexchange.com/boardgames/query/1457505/questions-where-accepted-answers-dont-have-the-highest-score-by-at-least-3-votes

We have 469 questions with an answer that has a higher score than the accepted answer.

https://data.stackexchange.com/boardgames/query/1457440/questions-where-accepted-answers-dont-have-the-highest-score

13

I would strongly support making this change. The elevation of the accepted answer is primarily valuable in situations where the asker is in a unique position to evaluate the answer differently than any other voter. In particular, when questions are about the asker describing a problem that they face and answers are about solving that problem, only the asker can know whether their original problem has been solved.

This very rarely applies to our site. Most of the questions I see are about understanding rules of games, and in those cases other users with more knowledge about the game are better able to evaluate the correctness of answers.

In addition, the linked meta post links to a Stack Overflow Meta post describing an experiment that shows that even Stack Overflow itself, a site that is more focused on problem solving, not elevating accepted answers produces an improved experience. That's what prompted them to make this change in the first place.

0
0

On Stack Exchange the asker can actually test out the answer they were given and see if it solves their problem, here however those who ask questions often have no where to go to see if the answer was right or works, aside from other users on the site. We're also dealing with games where the rules can change over time, like this question where the accepted answer was correct at the time, but is now wrong in some ways. While an asker can change the accepted answer (and the answer can be edited when it's out of date), this doesn't always happen.

I support this change for these and many other reasons, across the entire stack exchange platform as a whole, but even more for our corner of it. A wrong answer shouldn't be permanently highlighted because the user who accepted it no longer exists or has simply abandoned their account - and allowing newer correct answers to rise past the old accepted answer provides a mechanism for that.

I do think that an accepted answer should still be involved in the ordering process - but as a tie-breaker, in essence equal to ½ of an upvote for the order of answers before the random sorting takes over. And there should be exceptions, particularly for meta sites where SE staff answers actually set or explain policy, but that's a case for a separate tool for forcing specific important answers to the top before either accepted or vote based ordering gets involved at all.

2
  • I would like to point out that just because a user can test an answer and it works for them doesn't mean that it is the correct answer. I have seen plenty of answers on Stack Overflow that while they might appear to work at a glance they can cause future issues down the road or introduce various security issues. Examples of this include sql injection, things stopping working due to code combability or just plan bad performance due to doing things in an inefficient manner. Could also be something of the fix only works for some cases but not all.
    – Joe W
    Sep 19 at 14:31
  • @JoeW True, there's always a possibility of "Solved the current problem, but introduce several more." And with SO that can be because of bad answers or incomplete information. All developers know this happens. But that's kind of tangential to the point I was making in reference to this site, where, since you don't have the rigidity of maths or coding, answers can be objectively wrong and still "work" and questions often get asked built on a false premise, like those about where the free parking money in monopoly comes from built on the false premise that money exists.
    – Andrew
    Sep 19 at 20:15
-2

I'm against this in general. Sometimes a better answer is simply written later, which leads to it having fewer upvotes. Here's an example.

OP should get to decide what answer solves their problem. If the answer is bad or wrong, downvote it; the blurring out should make it clear to others that they should probably look at a different answer.

3
  • 1
    It's true that there are examples where this would lead to a less desirable ordering, but I have also seen examples of the exact opposite, where a lower-quality answer with a lower score is elevated by the asker's choice. The existence of such examples doesn't meaningfully inform us about the overall value of making or not making this change.
    – murgatroid99 Mod
    Sep 20 at 20:14
  • The issue with your logic here is that when an answer is written late enough to have significantly less attention than the existing one, it usually also doesn't get revisited by OP either, so no change in accepted answer status most of the time - so that better answer is lower down the list under either system.
    – Andrew
    Sep 21 at 14:42
  • @Andrew OP is more likely to see their question has a new answer. They receive a notification after all.
    – Allure
    Sep 21 at 16:03

You must log in to answer this question.

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