It seems to me like we have developed a problem with a particular type of answer: the answer that ignores the puzzle completely and thinks of a "creative" workaround.
What I'm talking about are two examples in particular, which I'll call "The oracle was lying!" and "Just stab them!" answers. Both of these answers ultimately completely ignore the puzzle and create extraneous answers that are ultimately completely unhelpful toward developing a solution to the obvious puzzle. These two types of answers I personally characterize as:
"The Oracle was lying!": The question presents a set of conditions. The answer assumes the question didn't specify all the relevant conditions, and creates new ones to fulfill the reqiuirements of the puzzle. In essence, it declares something completely new to be true, then answers the question using it. "The Oracle says..." => "The Oracle lied!"
This answer is a serious issue because ultimately it's not answering the question at all. It's just declaring things that aren't necessarily true (and frequently there's no reason to believe they are true), then completely avoiding the puzzle using them.
Just stab them instead! This type of answer also totally ignores the puzzle portion of the puzzle, and instead MacGyvers a solution using the parts of the puzzle.
"How do I get the sword on the train?" => "Kill the train operator with your sword."
"Which door does the Prince knock on?" => "The Prince waits until she comes out in the morning, then knocks on the door."
As above, this type of answer doesn't really answer the puzzle. Sure, it works as a solution, but it doesn't answer the question the OP asked at all.
(Side note: these answers are frequently characterized by one-line or one-sentence responses. Not always, but usually.)
It's worth mentioning that we have actually discussed this once before, to the conclusion that we shouldn't be judging whether these answers are really answers, and should downvote them for being low-quality instead.
The reason I bring this up again is that I'm not sure our old decision is sufficient anymore (though I'm but one voice among many), and here's, in brief, why:
- I believe we can now create objective criteria which we can use to categorize action on answers
- I think these criteria cover a smaller, more actionable subset of the problematic answers from our first discussion
- I believe the prevalence of our questions on the Hot Questions list, and simply that our site is growing, is exacerbating this problem
- I firmly believe that, at some point, while there is onus on the asker to foolproof their question, this only goes so far, and it's not really the OP's job to predict all the poor answers that may come in - it may not be reasonably possible, either.
Still, the previous discussion is worth reading over. If we create criteria, they must be as specific as possible, and must be as objective as possible.
So my questions to the community are:
- Is this analysis sufficient/complete enough to express this issue?
- Is it a correct analysis in the first place?
- What do we do about these answers?
- Are they even answers to what the OP was really asking? (Similar to answers against the spirit of the question)
To address possibilities for (3), I can think of the following possibilities, but these are by no means our only options:
- Community flags as not an answer; answers are converted to comments
- Community flags as not an answer; answers are just deleted
- Community generally agrees these answers are of low quality, and downvotes
- Moderators and users with privileges aggressively protect hot questions
If we opt for the deletion route, we need to figure out what, specifically, makes them not answers. We as moderators aren't here to judge the accuracy of an answer, and this is important to keep in mind. Acceptability criteria can't involve determining if the answer is actually a solution to the problem.
Do you all think this is a problem? If so, what should be done about it?