9

Today we changed the way we sort answers on Stack Overflow. We no longer pin the accepted answer (with the green checkmark) to the top of the list of answers. By default, we now sort strictly by votes (descending order by highest score), and the accepted answer's order in the list is based on its score.

. . . .

Would you like to have the accepted answer unpinned on your site? StackOverflow.com

There is a longer discussion at the reference, but I've quoted the significant part above. Should the first answer in the list be the accepted answer (pinned), or should it be the answer with the highest votes (unpinned)?

11
  • 2
    I have a definite opinion on this, as a user, but I'm going to try to find time to type up a "pros" answer and a "cons" answer. Lay all the cards I can see on the table and then see what everyone else thinks.
    – T.E.D. Mod
    Sep 17 at 14:53
  • Don't get it. What has SE research to say on this? Any insights? UX:SE or Psycho:SE? This is user interface & cognitive science stuff, to ascertain 'best'? But IMO, isn't the best option always & already: let the common non-reg user have on option to sort for this or that? And only really relevant if here are many As? (As the checkmark has always been '1 man, 1 supervote', ie: neither 'democratic wisdom of masses', nor 'proven expert choice'? Therefore: quite limited 'meaning' or 'value' anyway, except for QP himself (and likeminded readers)? Sep 18 at 17:03
  • 1
    Should we unpin accepted answers from the top of the list ? - Certainly ! Provided they're not mine, of course... :-)
    – Lucian
    Sep 18 at 20:08
  • 1
    I think that a good case can be made for both pinning and unpinning but, on balance, I'm not in favour of unpinning. Quickly posted first answers often have an advantage and accumulate a lot of votes while a later, better researched (and sourced) answer gets largely ignored simply because some people read no further the first answer (especially users who clicked on the hotlist). OPs (in my opinion) do usually accept the best answer to their question, which is a big help to quality, late answers. Sep 19 at 9:21
  • 4
    If you simply have the highest voted answer on top then there's the risk that erroneous "common knowledge" answers (where visitors vote for the answer that matches their wrongly-held belief) will rise to the top, especially if the question hits the HNQ list. For questions with a lot of answers, there's then the potential that the accepted answer doesn't even appear on the first page (although that would be rare on History SE).
    – Steve Bird
    Sep 19 at 13:46
  • 1
    @LarsBosteen In case of a QP having left the building, your example gets reversed: accept tick stays forever and then a vastly better late A will suffer even more than just from time-skewed voting accumulation, since it will never be accepted, it can't get. Sep 20 at 9:52
  • 1
    @LаngLаngС True, but that's a relatively small proportion of cases. Further, with Qs that were on the hotlist before, a better new answer has no chance of getting enough votes to the top (and that's the problem with Hotlist criteria so heavily weighted in favour new questions - it's just plain dumb to assume that new questions are always more worthy of a chance on the Hotlist than an old question with a good new answer. So much for SE's claim on best answers going top). Anyway, as I already said in my previous comment, there are good arguments to be made for both pinning and unpinning. Sep 20 at 12:03
  • 1
    @LarsBosteen How many those are, IDK? That's exactly why 'I don't get it': for the effects of (un)pinning we lack data, and args can be made in favour of both, while my solution would be to focus on actual user/usage behaviour data, realising that this only affects posts with lots of As, when the simplest solution is giving users needing As an option to sort (& tell Google snippets that 'the model' has problems). And you're also quite right: Q-age & HNQ and the detrimental design and its effects are the much bigger mammoths, stranded in that ocean ;) Sep 20 at 12:12
  • 1
    Let's just stick to the way it's been for the last 13+ years! Keep the accepted answer pinned :) Oct 4 at 0:41
  • 1
    @justCal - (wiki) Answers added.
    – T.E.D. Mod
    Oct 7 at 1:46
  • Did they consider the potentially more useful alternative of allowing only one answer to float above the accepted answer? I.e. the accepted answer will always be in first or second position. This allows a clearly better answer to be at the top, but also ensures that the accepted answer is well positioned to handle the cases where the popular answer is actually not a better answer. Oct 8 at 1:13
2

One basic dynamic I'd like to point out here, that will ..er.. underpin everything below, is that an answer being the top answer under a question is a huge advantage (assuming it doesn't suck). People read answers by scrolling down from the question. The further down the page you get, the less people are now reading. Answers get less chances for upvotes if less users are reading them. So yes, it is a big deal which way things are done.

Arguments in favor of Unpinning accepted answers

  • (The obvious)Removes the annoyance of inferior accepted answers being prioritized over better answers.
  • Makes asking push questions less rewarding/tempting.
  • Lowers the incentive to answer with what the questioner appears to want to see, rather than what the answerer believes the truth to be.
  • Its a particularly good fit for this site. On a technical site, users really do have specific problems they are coming to get solved, and it makes much more sense to differentiate great answers that didn't actually solve the person's problem. History is more of a fuzzy science, so a case could be made that we here can usually identify what correctly answers a user's question better than they can themselves.
3
  • If you use an advanced search,( is:answer isaccepted:yes score:0) you will find 46 accepted answers score=0. By my tally, 9 better answers would move up. 1 maybe 2 better answers would move down. The rest unchanged. I support this option.
    – justCal
    Oct 7 at 2:15
  • @justCal - Well, my issue here is that higher scoring answers aren't (IMHO) necessarily better ones. Often they are, perhaps even usually, but not always. Currently, the checkmark is the only ... er ... check we have on inferior but popular answers.
    – T.E.D. Mod
    Oct 7 at 3:19
  • One of the features this would fix is that one-shot user, that accepts the first answer, then never looks back. Good answers arriving later should be on top. Neither option is perfect, but to me unpinning the accepted answer is a good tool to discourage abuse, including low quality answers being quickly posted and accepted, push questions, and some vote-irregularity manipulation. Good answers will get their votes regardless. Its bad answers that need dealt with.
    – justCal
    Oct 7 at 3:45
2

One basic dynamic I'd like to point out here, that will ..er.. underpin everything below, is that an answer being the top answer under a question is a huge advantage (assuming it doesn't suck). People read answers by scrolling down from the question. The further down the page you get, the less people are now reading. Answers get less chances for upvotes if less users are reading them. So yes, it is a big deal which way things are done.

Arguments in favor of Pinning accepted answers

  • (The obvious)Removes the annoyance of misguided but popular answers being prioritized over superior accepted ones.
  • Provides much more incentive to be responsive to the question author, rather than just "play the crowd"
  • Provides more incentive to write questions, since the author is far more powerful than a normal user.
  • Our future voting dynamic will be consistent with our past one, which will make it easier to interpret the voting results on old questions.
1
  • @justCal - I believe it would resort the stack, which is the problem. Answers that outvoted accepted answers despite being placed under them will be difficult to tell apart from answers that merely outvoted accepted answers after the change happened (unless you remember when the change happened). That's important information (at least to me) lost.
    – T.E.D. Mod
    Oct 7 at 3:15

You must log in to answer this question.

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