41

I think that the new accept rate feature should count only questions with upvoted answers towards the accept rate percentage. Currently answers with zero upvotes (and with even negative score, I think) are included.

Currently questions without upvoted answers are considered "unanswered" and thus shown in the "Unanswered" tab.

Also I don't think it is fair that the asker is "punished" if answers are no good, that is, do not get upvoted and are not worth accepting.

Otherwise the accept rate percentage -feature seems nice.

Improve this question
2
  • 8
    The problem is that any time a number is associated with a user, the user feels he must make that number as big as possible. Perhaps we need a high-rep role model with a low accept rate to put people at ease. I see that litb is at 50%.
    – Anton Geraschenko
    Commented Aug 23, 2009 at 16:49
  • 1
    I want to have a 100% acceptance rate, but I can't because my question didn't get answered? Huh? How is that right?
    – Nick Hodges
    Commented May 10, 2011 at 18:28

4 Answers 4

Reset to default
8

I am actually more concerned about closed questions, which I should have removed from the calculation, and probably will sometime today. Closed questions are not included in the calculation any more.

I agree with Kobi, as the question owner, this should be motivation to get people interested:

  • edit your question to refine it and provide status updates, and implicitly bump it
  • offer a bounty

Now let's put on our "game theory" hats. How can one improve his or her accept rate percentage, by any means necessary?

  • accept an answer, obviously -- a net positive
  • delete the question (only works if no significant upvoting and the question has few answers, otherwise not possible) -- probably a net positive
  • convert question to wiki -- neutral, as this also prevents you from ever earning any more reputation on that question.
  • answer your own question, wait 2 days, and accept -- neutral to negative

Bear in mind that any accept rate of 70% or higher is considered "excellent", so you don't have to sweat every single one. There are occasions where accepting isn't practical.

Improve this answer
11
  • 19
    70% is your rule of thumb, but you have no idea how it will work out in practice.
    – tvanfosson
    Commented Aug 23, 2009 at 13:07
  • Regarding conversion to wiki: Does this only affect future votes? I thought I would also lose any rep gathered by that question so far...
    – Oliver Giesen
    Commented Aug 23, 2009 at 15:40
  • 1
    @Oliver, yes, conversion to wiki only affects future votes.
    – Jonik
    Commented Aug 23, 2009 at 16:11
  • 43
    If you want to enforce 70% to be "excellent" (nobody will see it this way intuitively), display a label like "Excellent" instead of the percentage (70%). Basically, put the percentages in the FAQ and just gloss over the details with a tagline of "[Excellent|Good|Fair|Poor] acceptance rate".
    – Andrew Coleson
    Commented Aug 23, 2009 at 22:24
  • 6
    Having said that, I still agree with others that your game theory is going to backfire into discouraging users from answering questions from people with low accept rates, and additionally make people with validly-low accept rates feel bad, which is a terrible way to keep people engaged with the system. If I come to SO only to ask difficult questions, get an automatic negative connotation on my questions, and people are (however "wrongly") discouraged from answering, that sounds like a completely negative cycle.
    – Andrew Coleson
    Commented Aug 23, 2009 at 22:26
  • 1
    I personally don't think the accept rate will discourage answers if the question is good. It's a useful metric to have. Like rep info, it tells you something about a user. If the question is poor and the accept rate is low, it might encourage people to not respond. Isn't that a good reason for implementing this?
    – nagul
    Commented Aug 24, 2009 at 7:54
  • Bug? I have an accept rate of 10 out of 11 on SO, yet the only unanswered question that I have, is one without any answers. stackoverflow.com/questions/1108699/…
    – Jan Aagaard
    Commented Oct 7, 2009 at 20:23
  • @Jan Aagaard: Are youkidding? stackoverflow.com/questions/323367/… - stackoverflow.com/questions/1172127/…
    – Ladybug Killer
    Commented Oct 27, 2009 at 21:19
  • 2
    @Jeff Atwood If your view is that people should edit their question or offer a bounty in an attempt to get a good answer, why don't you count questions with no answers? I don't see a difference between a question with no answers and one with one wrong answer. In both cases, the questioner should update his question in an effort to elicit a good answer.
    – David Kanarek
    Commented Jan 25, 2010 at 0:27
  • 3
    If 70% is excellent, why isn't it highlighted in green?
    – RandomEngy
    Commented Mar 28, 2010 at 1:43
  • 2
    @AndrewColeson, re "[Excellent|Good|Fair|Poor] acceptance rate": if there's poor, opposite of good, and there's excellent, then the later should have an opposite, too: perhaps atrocious?
    – msh210
    Commented May 19, 2011 at 20:03
6

This was probably thought out. The asker should work actively to get the correct answer - edit and bump it up, or offer a bounty. If the question has no answer it can be made wiki. If the issue was resolved, he/she can add the solution and accept it.

Improve this answer
3
  • 2
    Absolutely. If you had a problem, eventually you did something about the problem. Go back and document it and accept your own answer. When searching for technical solutions, I often find people who asked the question months ago, but got no answer. Certainly they figured out some solution. What was it?! This motivates people to go back and answer that.
    – Jeremy Stein
    Commented Sep 3, 2009 at 18:22
  • 5
    That's not always the case; sometimes the answer is you just do without whatever it was you wanted - i.e. drop the feature.
    – L. Cornelius Dol
    Commented Oct 27, 2009 at 17:38
  • 3
    ...or you put a project on the back burner for a while and never resolve the problem.
    – devuxer
    Commented Mar 15, 2010 at 5:27
3

I like Anton's comment on the question:

The problem is that any time a number is associated with a user, the user feels he must make that number as big as possible. Perhaps we need a high-rep role model with a low accept rate to put people at ease. I see that litb is at 50%.

Perhaps acceptance rate should only be shown for users with lower than a particular rep threshold, since it's mainly a tool for distinguishing question pumps from team players.

That said, I like the OP's suggestion too - acceptance rate should only take into account questions with upvoted answers. Questions with no satisfactory answers should not compel the poster to accept one of them. Edit: wow, I can't believe I said that. Unanswered question -> unanswerable question == potentially bad question. Either work harder to make the question answerable, or if the OP found a workaround or some sort of solution himself, document it as an answer and accept that answer.

However, one of my questions on SO is sitting with no answers accepted, and it doesn't bother me at all. I see my acceptance rate as less significant than my answer:question ratio (112:4) or my overall reputation (~3500).

PS. I'm going to repeat a comment by Andrew Coleson from elsewhere in this question, as I think it's excellent and deserves to be highlighted:

If you want to enforce 70% to be "excellent" (nobody will see it this way intuitively), display a label like "Excellent" instead of the percentage (70%). Basically, put the percentages in the FAQ and just gloss over the details with a tagline of "[Excellent|Good|Fair|Poor] acceptance rate".

Improve this answer
1
  • I think your original position wasn't unreasonable. Unanswerable questions shouldn't really have an answer on them to begin with, so ideally they should fall into the zero answers bracket. Whether zero answer questions should count towards the accept rate is a separate debate.
    – user200500
    Commented Jan 1, 2013 at 9:20
1

I wholeheartedly support this proposed change. Currently, there is a perverse incentive structure going on. If an answerer sees a 100% accept rate, they'll basically know that they can make any effort at answering whatsoever and if no one else chimes in, they'll get an accepted answer.

The question asker has been able to say "I accept any answer, no matter how bad or unhelpful it is."

Improve this answer
2
  • 1
    well, if there are users who are accepting nonsense repeatedly, flag their answers and the mods will look at the pattern of behavior. Days or months of 'bad' behavior can be undone in a single click..
    – Jeff Atwood
    Commented Mar 28, 2010 at 2:11
  • It might not be so blatant. It's quite easy to give a lazy answer that seems well-intentioned. Miss a basic detail of the question or link to a solution to a vaguely similar problem that doesn't apply. But the root problem is that you're encouraging the wrong accept percentage. 100% is listed in green and bold. 75% is in dull gray. But which person is actually more helpful to the community? I would argue the one who only accepts real answers to the question that was asked, not the one who clicks "accept" on every question, regardless of merit of the answer.
    – RandomEngy
    Commented Mar 29, 2010 at 22:42

You must log in to answer this question.

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