Recently I ran some stats on the SO dump to search for question with answers marked correct that have less than -1 votes.

The list was illuminating: What interesting stats can I obtain from the Stack Overflow data-dump?

It is slowly getting rectified, is there a better way of getting this stuff fixed, should there be an official section on SO for this?


First of all, if the original poster of the question felt that this answer fixed his problem. He has the right to accept it (be it wrong or not).

Secondly there are cases where the original poster is just going in the wrong direction with his question and the community votes some answer that rectifies his trouble in a way he is not looking for (and then voting for it immensely) But he can still accept any answer he wants.

The thing is that an accepted answer is necessarily not the correct one. It's just the one the OP wanted.

So I think we shouldn't do anything about this. This is a system to help others, the OP can accept a solution but we as a community can show our support for the answer we think is right and it will be the 2nd one after the accepted one.

  • 1
    +1 I agree .. the system addresses the issue already. – tomjedrz Aug 18 '09 at 17:50

I think it would be a good idea if we could have an OP accepted answer, and a Community accepted answer. There's nothing saying the two couldn't be the same, but I've definitely seen cases where there is a question with a clearly correct answer, but the OP marks an answer that is wrong.

For instance, if someone asks, "How are statements ended in C#?" and someone answers "with a ?" that would obviously be wrong, but nothing is preventing the OP from marking that as accepted. On the other hand, if the Community had the opportunity to accept an answer as well, then there could be a second accepted answer that will most likely be correct. Maybe there could be a rep requirement or something, I don't know, but I thought it was worth bringing up.

  • I think this idea has merit. It could work similar to close on questions and require 5(?) votes from people with more than a certain rep level. Is this more or less responsible than closing questions? – ChrisF Jul 8 '09 at 13:09
  • @ChrisF Yeah that's what I was thinking too. I don't think you would close a question even if the accepted answer is wrong, because the question is valid regardless (presumably). I do think that having a Community accepted answer would help a lot of people, though, for the questions that have an accepted answer that the community thinks is not ideal, even if there's no real "wrong" or "right" answer, but it works very well for clearly "wrong" answers. – Joseph Jul 8 '09 at 13:20
  • That would be the highest-voted answer, which is always right after the accepted answer. – David Thornley Jul 8 '09 at 13:58
  • @David In a lot of cases I would hope that would be true, but there are indeed cases where the highest voted is in fact, NOT the correct answer. Also, having a community accepted answer would help solidfy the point that the highest voted answer is viewed (by the community) as the correct one, except when it's not. – Joseph Jul 8 '09 at 14:19
  • 2
    We have this already; the highest voted answer is the community-accepted answer. – tomjedrz Aug 18 '09 at 17:48
  • @Joseph .. can you post a few examples of where the accepted answer and the highest voted answer are both wrong? – tomjedrz Aug 18 '09 at 17:49
  • @Joseph: How can we possibly handle situations where the highest voted answer is incorrect? That would require having super-answerers who we trust to pick off incorrect answers, which is an extra layer of complication. Moreover, how would we pick the community accepted answer except by votes? If I see an incorrect answer high on the list, I downvote and comment. Is there a better way to do this? – David Thornley Aug 18 '09 at 18:15

How about adding something to the recent activity page, that says something along the lines of 'You have a question with an accepted answer of less than 0 points.'

  • 2
    That would work for regular users, but perhaps not for long standing questions/answers that have been effectively abandoned. – ChrisF Jun 28 '09 at 16:28

I think the voting system is working as designed in these cases. Anyone who visits a page where the accepted answer has a large negative score and the next answer has a large positive score can draw their own conclusions. It has been stated before that in many cases the OP is exactly the wrong person to accept the best answer, or they wouldn't have needed to ask the question in the first place. This is why the "community accepted" answer, the one with the most upvotes, appears directly below the OP's accepted answer.


Upvote the better answer, or write your own..

One benefit of the later is it will appear in the users "Replies" box, which might prompt them to look at the answers again

  • +1 - use the system as it stands to deal with the issue! – tomjedrz Aug 18 '09 at 17:47

You could reset the question (not accepted anymore), if it hits a threshold (e.g. -5 votes). And you have to clarify what happens with the already earned rep for accepting this answer.


I've got one of those, an accepted answer with a -12 ranking. The thing is, it is a correct answer. At first, I answered it very in a very flip way with an obviously dangerous suggestion, but over time I improved it until it's actually what I feel a deserving answer, but people continue to pile on and vote it down in spite of the improvements.

So if you ask me what to do about "obviously incorrect" answers, I say that a negative score doesn't necessarily mean it's "obviously incorrect".

  • That question really belongs on serverfault now .. and I do think the iptables approach is much safer – waffles Jul 8 '09 at 22:06
  • @JP has a really good example of this, he fixed it up, I upvoted him, but he is still stuck in -2 stackoverflow.com/questions/746548/… – waffles Jul 8 '09 at 22:08

I don't think that this issue is as important as others who have commented/answered, and I don't think any changes are necessary, as the situation resolves itself the vast majority of the time.

1- The original poster (OP) has solved the problem, or at least thinks it is solved.

2- The community has indicated what it thinks is the best answer through the voting. Quite likely the community has also (in comments) indicated problems with the accepted answer.

3- The Googler (or Binger!) who lands on the question page will see the accepted answer, the community-voted best answer, and quite possibly a number of other answers, and so presumably has the tools to solve their problem.

4- It is dangerous (and arrogant) to not accept the possibility that the community could be wrong and the OP and writer of the accepted answer correct.

I do not think the community should be able to "unaccept" an accepted answer through downvoting. I suppose a moderator could be given "unaccept" authority, but that concerns me too because moderation on these forums is at times heavy-handed.

  • Really? stackoverflow.com/questions/168427 – waffles Aug 18 '09 at 21:44
  • How does the binger know that the green answer is not correct, or that those numbers on the left side are relevant? – waffles Aug 18 '09 at 21:45
  • @Sam Saffron: I think it is OK for the system to assume that the binger will have the brains and curiosity to read the whole thing. – tomjedrz Aug 21 '09 at 1:55

Can we at least not keep the accepted answer at the top if it has been outvoted by another answer? At least if the accepted answer has a negative score?

  • Disagree. Shading and noting the accepted answer despite the votes is a good feature. Votes aside, the accepted answer has worked at least once, for the OP. – tomjedrz Aug 18 '09 at 17:47
  • @tomjedrz: But there are some questions that answers don't "work" for -- take this one for example: stackoverflow.com/questions/514492/what-is-windows-powershell The accepted answer is ridiculously inaccurate and the OP is now totally misinformed and may not even know it. – arathorn Aug 18 '09 at 18:11
  • @arathorn: A poster who accepts the first answer that sounds good, doesn't check back, and ends up down a rabbit hole he gets what he deserves. – tomjedrz Aug 21 '09 at 1:53

You must log in to answer this question.

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