Possible Duplicate:
Would it be useful so show people’s “hit rate” - i.e. percentage of answers accepted?

To compliment the question acceptance rate, I suggest adding an answer acceptance rate. This would be the number of answers a user has accepted over the total number of answers they've provided. A high answer acceptance rate is indicative that the user's answers are usually detailed, thorough, and provide the asker with what they need on a regular basis.

Also, as less popular topics garner fewer views (and votes), this could also be a way to encourage users to participate in these topics, as they typically will also get fewer answers, leading to a greater chance of the answer being accepted, raising their answer acceptance rate.

There is also a related question here, of which this question is largely a duplicate of, but I am proposing my feature request as a companion to the question acceptance rate, displayed and calculated roughly the same way (exclude CW, etc). I also favor adding both this and the question acceptance rate to the profile as well as being displayed below the username beside that user's answers.


6 Answers 6


I think simply "no".

The "accept rate" for question is only on the user's side. The percentage of accepted answers means nothing, as it relies only on what the person who asked a question does, not the person who answered. Many times questions remain not marked as answered, even though there are highly voted answers.

And the fact an answer is not accepted doesn't mean it's a bad one, just that it wasn't the one accepted by the question author. So in general, such percentage would show nothing.

Edit : Eventually a ratio based on reputation per answer/question posted would make more sense, but it would still be highly biased.

Edit 2 : Thinking more about it, in fact, there is another issue. The "accepted" answer doesn't mean this answer is the best in quality (as finally the proposed percentage would suggest), but simply that this answer was actually working for the person who asked. I often see great quality answers, detailed, precise, which are not accepted, for example simply because the person who asked the question forgot to mention a fact which makes this answer invalid, compared to a one-liner on the same question, which would actually give the correct solution.

I agree that there is no reason to accept an answer which doesn't solve the problem asked. But on the same way, I can't see such "accepted answers"/"total answers" ratio be an indicator of quality. It simply means that this user managed to give answers that people chose to accept.

My main concern, in fact, is that in the end, such ratio would increase the effect from the question ratio. The question one would already push people to not lose time answering someone "not likely to accept". But such "answer ratio" would push people to not answer too much unless they are sure it will be accepted. so finally, it will become that once someone will post an answer which looks like "the one", people would be less likely to add an answer (which could add details, or another worthy solution), since it would only decrease their ratio....

(I'm not saying everyone would react like that, far from it, but it's on the same idea as the "question ratio" concerns.)

Edit 3 (answer to comment) : Why would you answer a question if you don't think that the OP would like your answer?

There is more than "the accepted answer" and "bad answers". There can be several good answers to a question, but only one will be accepted. Typically, if someone is asking for the best way/program to do a task, the fact that the question's author accept one answer doesn't mean all other answers are worthless.

More than this fact, on questions where I see that the correct answer has been posted (or even already accepted), I add an alternative solution if I have one, for the sake of exchanging knowledge.

In the end, such ratio would influence people to simply leave a question as soon as an answer has been accepted. The goal of these sites is not only to provide a single answer to a particular person, but also to provide solutions to someone else finding the question by some Internet search. And maybe this someone would prefer an alternative solution, than the one "accepted" by the original author.

  • 1
    Your concerns are valid, but with the introduction of the question acceptance rate I suspect that users will be more likely to answer questions where the asker is likely to accept one, reducing the possibility of a great-but-not-accepted answer. And we've still got the reputation system to tell users what other people think of their posts. Commented Aug 23, 2009 at 21:59
  • 1
    The "accepted" answer doesn't mean this answer is the best in quality (as finally the proposed percentage would suggest), but simply that this answer was actually working for the person who asked. There's something to be said for continually providing answers which the OPs find helpful. Such an answer ratio would push people to not answer too much unless they are sure it will be accepted I think this is a good thing to reduce noise. Why would you answer a question if you don't think that the OP would like your answer? Commented Aug 23, 2009 at 22:39
  • +1: added also a thread about the populist badge that resumes your thinking. Commented Aug 24, 2009 at 15:51
  • 2
    +1 I agree with Gnoupi. There are too many brilliant, detailed answers on SO that weren't accepted, sometimes simply even because the asker didn't understand them. Also, often answers that point out flaws in the OP's basic architecture go, even though perfectly correct and valuable information, unaccepted. I think often rightly, because they were not an answer to the question. Still, those answers are important for SO's quality as a knowledge base, and it must not become a problem to write them.
    – Pekka
    Commented Feb 3, 2010 at 18:57

I've been a proponent of some sort of ranking that took the number of accepted vs. the number of answers total. I think it could be a very good indicator of a user's "success" as a member of the sites. Someone who has a high number of accepted answers could be seen as someone who is giving high quality answers.

  • 1
    not really - lots of answers may just not be seen
    – warren
    Commented Oct 11, 2009 at 4:57

No. It goes against the spirit of the Populist badge.

As Gnoupi said, there would be virtually no incentive to add a good answer to an already answer-accepted question. It's just the same than bounty questions, there is much less interest on them once the bounty is awarded1.

1. For bounty it is by-design. In the general case it would be counter-productive IMHO.


I don't think percent is a good metric for this.

If you have a high number of answers, you would have a low acceptance rate.

Not to mention that this would be an easy one to game, just delete all of your answers that aren't the accepted answer. Instant 100% acceptance rate.

Total number of accepted answers would be a better metric.

Just looking at Jon Skeets answers he has about 50 accepted answers, on Meta, out of about 250

That's about 20%

My educated guess, is that less than half of his answers on Stack Overflow have become the accepted answer. ( Look at his lowest scoring answers )


I think I'd rather see a badge awarded for a certain number of answers having been accepted. Say 25 for silver, or 100 for gold?


Points/answer is a better measure of quality than answerer's acceptance rate and it's already weighted to favor accepted answers.

It should be called Quality Rating rather than "acceptance rate".

Accepted answers are judged by the OP, but up-votes include the opinion of anyone interested.

Points/answer is in keeping with the Populist badge.

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