What do you do when an accepted answer is changed from correct to incorrect?

Question

I'm more than 99% sure the wrong answer was accepted here: Rotation angle of a giant lily when a child crawls on its rim

(Of course, I'm subject to significant bias saying this, since I submitted a different answer.)

Edit: In the particular question I linked, the wrong answer's been changed, so I'm happy about that.

Ordinarily, if a wrong answer is accepted and a better answer exists, the better answer will still get a lot of up votes so that a user browsing the site can see the community's opinion. In this case, I think the scenario is different.

The incorrect answer was originally a correct answer to a simplified version of the problem. That correct answer got a lot of upvotes. Then the author generalized the answer incorrectly, resulting in an incorrect answer with a lot of upvotes.

So now, the incorrect answer has a lot of upvotes, and is marked accepted, even though I believe strongly that the community as a whole would not support it as the correct answer. Is there some course of action I can take to call the community's general attention to this situation?

Edit: I'm not completely sure whether this is an "honest" meta question, or whether I'm actually just fishing for people to support my answer. If this question gets a couple of downvotes, I'll take the hint and delete it. On the other hand, behind my emotional desire to be validated, I really do want the correct answer to be accepted in general and think there is some actual validity to this meta question.

Answer 1

If it is edited into an (not necessarily) incorrect answer, but a conflicting with the earlier, or critically stiuation changing details are added after the answer beeing accepted, a moderator should revert it to the earlier version and the user should be told to post instead a new answer, since these changes constitute the definition of a NEW answer.

Answer 2

I would say:

1. Comment -- with good argument, this should make answerer correct the answer
2. Downvote to enforce reading the comment
3. If answerer is not responding (for instance lost interest in SE), edit the answer

About the mutating question problem, this should be resolved by adding new question, not new answers. In SE philosophy, a question/accepted answer pair should later serve as instant-information page -- one Googles the question, reads the answer and has his problem solved; a clutter of answers to different stages of question evolution effectively masks any useful information. So, if you see this happening, ask the OP to make a new question.

Answer 3

Good question. In my opinion, this shows the fundamental weakness of the StackExchange based approach. Everything is static and any change (either in question or answer) is not reflected at all and leads to inconsistencies. So assuming SE people won't change the engine underlying voting, one can only suggest that things not be changed unless absolutely necessary. It's fine to make the question/answer more precise but to change them completely is just bad.

Actually, this is an issue just for questions because changing them makes all the given answers invalid. But if someone wants to change the answer I suggest they just post a new one. This also suggest a solution to your problem: revert the accepted answer to the old, correct one (albeit simplified) and let ftiaronsem post a new answer with his new solution.

Answer 4

The usual method in teaching is that the teacher asks the students questions. The students provide answers (guesses), and the teacher grades them. This works because the teacher is assumed to have more knowledge about the subject than the students.

Stack Exchange tries to turn this on its head. The "student" asks a question. The "teacher" provides an answer. Then the other students (i.e. readers) grade them. This kind of teaching method is suitable for subjects far simpler than physics. Physics is about learning how to make calculations. Stack Exchange is mostly about the transmission of data.

It shouldn't be too surprising that the result is kind of rough sometimes. You're letting the inmates run the asylum. I think it was Winston Churchill who said that something to the effect that democracy was a horrible system of government but seemed to work better than the alternatives.

But really, why should it be such a horrible thing that a wrong answer, or even a complete crank answer, gets voted up and accepted by the questioner? It's not like someone is going to be executed as a result of the decisions made here. A student who relies on the internet for answers to their homework shouldn't be shocked when (a) they occasionally get wrong answers and (b) they usually don't learn anything.

If you want to learn physics, reading answers is not going to help you. The people who are learning physics here are the ones answering the more difficult (for them!) questions, not the people reading other people's answers.

I think things will work better if we all understand that the primary purpose of this site, just like the evening news, is entertainment, not truth. That said, I've had some good use out of this system and will likely continue to use it. I don't expect it to converge to the "truth" and won't lose any sleep over its failure to do so.