Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 158 Stack Exchange communities.

What is meta?
Here's how it works:
  1. Any Stack Exchange user can ask a question
  2. The community provides support, votes on ideas, and reports bugs
  3. Your voice helps shape the way Stack Exchange operates

Completely rewritten. More specific, more complete. Give it a chance.

I have rewritten the question in hope of salvaging respect with my peers.

One commenter below helped me find an analogy that I think will clarify my intentions, and make a more specific proposal. I'll re-tag the question accordingly.


The intention of this discussion is strictly to consider how reputation is handled, in particular when an existing accepted answer is unaccepted... nothing more.

If there is confusion yet, or parallel details I may have overlooked, please ask questions in the comments section and I'll edit this Question to address them.

As It is Today:

When the OP has already accepted an answer to his/her Question, and at some point another user comes along and provides a new answer. The OP finds the new answer more suitable, so he/she unaccepts the current answer and accepts it instead. The new answer's user is awarded +15 reputation, and the original answer's user -15.

The Concerns

If the purpose of earned user reputation on Stack Exchange is to encourage users to participate in an online community and to convey to their peers the expertise in their answers, then is it appropriate to revoke the reward earned for their accepted answer when a new answer is chosen?

  • Should unaccepting the original answer cause reputation loss?
  • Should a Question be worth a finite 15 reputation, or should accepted answers be considered independent?

The Analogy

As there can only be one accepted answer to a question, the scenario of a business contract is fitting.

A company (OP) defines an offered contract (Q) with a certain criteria, and awards the agreed contract (AA = Accepted Answer) and its compensation (Reputation) to the bidding company best suited to their needs.

Company A and Company B then come along and bid, and the contract is awarded to Company A. They are awarded the contract until it's term is up or the criteria has changed. Later, Company C comes along and the offering company is impressed.

Using this analogy, SO currently rewards Company C with the new contract and compensation, and revokes the contract from Company A. While this may seem fair at first, the issue is SO also revokes their compensation.

In the real world, the contract is honored until the criteria changes or the term of the contract expires. Meaning when the offering company awards Company C the new/altered contract and takes it back from Company A, but they do not ask nor can they expect their money back... services were rendered... the answer service its purpose.


My suggestion is that Stack Exchange limit questions to a single accepted answer, but that previously accepted answers honor the compensation offered... to treat them as separate contracts if you will.

As this would mean a question could produce an indefinite amount of reputation, I propose that some kind of terms be put it in place (much like the terms on a contract) to prevent exploitation. Ideas:

  • Time Period - if the accepted answer stands for X Days or Months, the user retains his/her compensation, else it is revoked.

  • Criteria Changed - if the question is edited the existing answer retains its value, as it was right for the original question.

  • OP Decision - let the OP decide/provide reason why the answer was unaccepted, and that reason determine whether or not the user keeps his reputation.

I suggest #1 as it seems the most unbiased, subject to the least exploitation, easiest to implement.

share|improve this question
Are you saying the OP shouldn't be allowed to change their accept after a certain amount of time, or just that the rep shouldn't be lost? – jadarnel27 Jan 30 '12 at 21:28
Just that the original correct answerer not be penalized. The OP should always be able to accept another more correct answer, and that answerer be properly rewarded. – one.beat.consumer Jan 30 '12 at 21:56
"Second person is rewarded" Given that the later answer was a self-answer the second person wasn't actually rewarded. However it will help anyone who has a similar problem in the future in that they'll see the "better" solution – Some Helpful Commenter Jan 30 '12 at 22:10
@one.beat.consumer Then what you seem to be suggesting is exploitable to generate extra reputation, by switching the 'accepted' answer from one person to another. – Andrew Barber Jan 30 '12 at 22:22
@AndrewBarber I posted an answer using your thoughts. If you would like to post your own, I'll mark it properly. For friendly debate, one thing to think of is that a Question already provides an indefinite amount of reputation simply because of upvoting from users. Is the possible exploitation of 15 points such a serious matter? And would it not be easy to monitor if it were implemented such as a moderator flag so it were logged? Just thoughts. – one.beat.consumer Jan 30 '12 at 22:40
@one.beat.consumer I haven't an unaccept in a couple months and I get one today. I never should have commented on this. :P – Some Helpful Commenter Jan 30 '12 at 23:27
@ConradFrix Yeah, as you can see it got me flamed pretty hard. Some cutthroat folks around. At least you read it an made an intelligent comment. ;) – one.beat.consumer Jan 31 '12 at 0:30
"Please be considerate and not downvote simply because you feel some noob is whining about losing 15 points. It is an honest and serious question." Once again, note that votes are different on meta. Downvotes mostly indicate disagreement. We mostly think this is a lousy idea. – Anthony Pegram Jan 31 '12 at 4:13
If you consider the response to this question "flaming" then I strongly recommend you don't visit the rest of the internet. The harshest thing anyone seems to have said is "Don't worry about it" and "you're wrong". Nobody is picking on you, we just think this is a terrible complaint about a non-issue. Perhaps a thicker skin will solve both issues. – JNK Jan 31 '12 at 13:11
@AnthonyPegram Thank you for explaining. I rewrote the post entirely in hopes of refocusing it on topic and minimizing peoples concern for the OP and New Answerer. Let me know if you still think it is lousy. – one.beat.consumer Feb 1 '12 at 0:52
This is still an increasingly complicated "solution" to a problem that nobody else thinks is a problem but you. – JNK Feb 1 '12 at 0:59

The accepted answer is the one that the OP considers the most useful. That might change over time, either because the world changes or the OP gets more time to try out the answers.

A new answer can always be better that the existing ones. That's life!

share|improve this answer

You are not being penalized, but neither are you being rewarded for having an out of date answer.

While it might seem unfair to you, the OP got a better answer, and he decided to reward someone else instead of you.

It is not like your answer is more accurate now, so I don't see why you should get the +15.

share|improve this answer
Try to understand what I said. It is definitely fair for the OP to reward a better answer, and for that answer to be rewarded. But removing reputation from someone who did provide a correct answer before, is by definition penalization - they are two separate transactions, one should not affect the other is all i'm saying. – one.beat.consumer Jan 30 '12 at 22:20
Taking away an extra is not a penalty. It is the lack of you getting a reward. – soandos Jan 30 '12 at 23:42
Think of it like this... an athlete wins a race and gets his medal (reputation in this analogy). next year he races again and someone else wins - they get their medal, but he does not lose his. To take away his medal because someone won the race later is penalty... another decent analogy would be the American judicial system. – one.beat.consumer Jan 31 '12 at 0:36
@one.beat.consumer - your analogy doesn't work - they are different races. A better analogy is that several companies compete for a contract. Company A is awarded the contract, but a few months later the criteria change. Company B is a better fit for the new parameters so they are awarded the contract instead. – JNK Jan 31 '12 at 13:21
You're starting to understand. Yes, it is more like a contract exchange. I am saying it should be treated more like races when it comes to the reputation. Continuing your analogy, when Company A fits the bill and serves the length of the contract (a correct answer for some given time) and next year Company B is better fit for the renewal, do we take Company A's compensation back? No. Company B continues on with pride, but we don't say "hey, now that our criteria has changed, can we get our money back for what you've done for us" – one.beat.consumer Jan 31 '12 at 17:26
It's still the same contract though, is the issue. It's not like a new question was asked, it's the same question, and this kind of thing happens ALL THE TIME in military contracting. Mcdonnel Douglas or whoever starts ramping up production of a plane because of a contract but it gets taken away and given to Boeing instead. They don't pay both companies, even if McDD sank money into it already. – JNK Jan 31 '12 at 18:46
At this point we are splitting hairs. I used your analogy to rewrite the post, which I am grateful for. The subtlety is in the length of time the answer is correct. My suggestion then was to make a rule that says "if the answer stands for X amount of time, it remains rewarded" as if it were a contract already served. Thanks for all your input. I appreciate you following it to the end. – one.beat.consumer Feb 1 '12 at 0:48

As the others stated, you aren't being penalized.

Another way to look at this is that disallowing the OP from un-accepting an old answer and accepting a newer, better answer is a disincentive to update old questions!

If a new solution for something comes out, we really really want to have that in an old question if it applies - those old questions have higher google-fu than new ones, and likely get views still.

If you don't allow an OP to accept that fancy new answer, why would anyone go to the effort to post a new answer to an old question? They won't!

You don't have to monitor all your answers, but nothing is stopping you from looking for old questions that could be updated and posting your own answers. You will make the site better and likely garner some rep in the process.

share|improve this answer
I NEVER suggested the OP not change the answer, or that a New Answer not be rewarded. Not once. – one.beat.consumer Jan 30 '12 at 22:21
@one.beat.consumer - you implied it. The alternative is allowing OP to accept multiple answers which seems like a bad idea too. You probably should make a concrete suggestion or stop complaining that people misunderstand you when you are unclear. – JNK Jan 31 '12 at 2:42

Note: Thanks to Andrew Barber (see his comments on the Question), and @JNK'sMetAccount.

Altering the award system in this way would treat accepted answers as separate transactions, or "contracts" to use the terminology from the Question.

Using a "time accepted" measurement for answers could be used to ensure quickly replaced answers were revoked (reputation and all), and answers that held true for a significant period of time remained rewarded (keep reputation). This measurement would also help to minimize possible exploitation.

share|improve this answer

You must log in to answer this question.

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