On politics.SE it happens annoyingly enough that someone asks a question which seems ok (i.e neutral and objective enough), but then they accept a hugely downvoted answer, e.g -20 relative score. That gives such an answer undue prominence. Yes, it's pretty obvious to a minimally-informed user that the accepted answer is (very) contested... but still I think it's an abuse that could be easily curtailed. Namely:

  • Make answers with a big enough negative score unacceptable, i.e. remove the checkbox from them, automatically. I'm not sure what the best threshold would be, but I guess -3 (upvote-downvote difference) is a reasonable cutpoint.

Yes, there are some alternatives to this, namely deleting answers, but there's little agreement on politics.SE to delete bad answers that aren't outright abusive. Given the more subjective nature of the topics, you can probably guess why. (There have been some delete/undelete wars [cycles], complaints on meta etc. Some have featured exactly this kind of very-negatively-scored accepted-answers.)

  • 2
    Here's one case where it would be an issue: meta.stackoverflow.com/questions/363148/… Sep 27, 2019 at 22:17
  • 6
    The better approach, that has been batted around here before, that neither gives undue prominence nor robs anyone of their agency is not to take acceptance into account when sorting answers. Just like self-answer acceptance. Sort the answers strictly by score. Don’t let OP overrule the community.
    – Dan Bron
    Sep 27, 2019 at 22:17
  • 1
    It works fine enough as is. There will always be edge cases regardless of which strategy we use. I don't see how changing the ordering in this way improves anything at all. Just as people sometimes accept bad answers, people also upvote bad answers.
    – Kevin B
    Sep 27, 2019 at 22:24
  • @SonictheAnonymousHedgehog: I'm not saying it should be done for meta, where voting has fairly a different meaning. Sep 27, 2019 at 22:25
  • 1
    @DanBron: that's not ideal. Sometimes a less-scored answer is added later with much better info but the page is no longer on HNQ etc. I just to prevent the most obvious abuses with very negatively scored answers on top. The accept button provides some latitude for the OP who is [presumably] still around, while most of the flock is gone. What you propose would work ok if there were no HNQ. Sep 27, 2019 at 22:26
  • 2
    See also meta.stackexchange.com/questions/264045/… which proposed to make accepted answers on -3 or below non-special Sep 27, 2019 at 23:34
  • 1
    Of course this applies to more than negative score answers. An answer that received 20 upvotes and 15 downvotes has a score of +5, but it is contentious, and, in terms of this question, I'd say it is in the same category as a negative score answer. Sep 28, 2019 at 1:18

2 Answers 2


Personally I'd welcome a change in the sort order instead. When an OP accepts their own answer it doesn't rise to the top. The same logic could be re-used for negatively voted answers. (Or all the time, for that matter.)


Accepting the answer is entirely up to the asker - they can accept whichever answer solved their problem or answered their question best.

It may not be a popular answer, and it may even be entirely wrong, but that is what the Accept button is for - for them.

Scoring is the community rating of the post.

Personally I think the two things are very separate and should be kept that way. Sometimes I want to accept an answer with negative scores, as it is the right answer for me.

  • 5
    Yes but the (very) subjective user choice of ticking the box affects the ordering of the answers for everyone. Actually for self-answers, that doesn't happen, i.e the order is only the community's choice. Why should it be different for fringe answers [i.e. very substantially downvoted] provided by someone else? After all, the user account system is rather easy to game. On politics.SE there are some serial sockpuppets. Sep 27, 2019 at 22:21
  • 2
    Sometimes the community is wrong (yes, I know this is me being devil's advocate) but to be honest it doesn't matter. You can see the score, so if you want to go with the highest scoring, it's very obvious
    – Rory Alsop
    Sep 27, 2019 at 22:23
  • 1
    Next time a conspiracy theory is the accepted answer because it "worked for" the OP, I'll have someone to blame. (I'm kidding.) The rejection of this (and as a far I can see all other proposals to change/reform the accepted-answer dynamics) leaves the "constitutional hardball" of using the delete button more often as the only real alternative. Sep 28, 2019 at 9:17
  • 2
    I suppose SE sites where the accept button can be abused heavily should try to codify more of their procedures/criteria, like Skeptics did. There mods won't bat an eye at deleting highly upvoted (never mind accepted) answers for breaking the site's [highly localized] formal rules. Sep 28, 2019 at 9:32
  • That's broadly a good idea @Fizz, however I think mods do that anyway if something is dangerously wrong (eg Electricity, Mechanics or Security sites for example)
    – Rory Alsop
    Sep 28, 2019 at 11:52
  • This is the bulldog clip Q&A, where the correct answer is obvious and provable yet the OP wants to choose the incorrect answer. You should have seen all the comments before it was moderated into its current state. The correct answer has the most votes, but the OP likes a different answer; now some have buyer's remorse but edits to weasel out have been rolled.
    – Rob
    Sep 28, 2019 at 13:01

You must log in to answer this question.

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