Examples: Which way is the bus going?, https://puzzling.stackexchange.com/questions/569/getting-the-coin-out-of-the-bottle, https://puzzling.stackexchange.com/questions/568/getting-the-100-dollar-bill-under-the-inverted-pyramid, https://puzzling.stackexchange.com/questions/172/how-to-escape-a-blender

On the "coin out of a bottle" puzzle, @WendiKidd posted this comment:

This question appears to be off-topic because it is one of those questions that can have many different answers (as proved by Kevin's answer of cutting out the bottom of the bottle). We've discussed this on questions in the past and on meta; if you can make up just about any answer to the question, it's not an in-scope puzzle.

I kind of agree -- the question is dangerously close to being a "poll"-type question, but these criteria seem to disqualify every on the site.

We need to decide what makes brainteaser-type problems where you are placed in a hypothetical situation on-topic. So, how should we determine which of these we want, and which of these we should close?

  • $\begingroup$ I think that the answer by Kevin to the bottle question violates the spirit of that kind of puzzle and therefor isn't really an answer, and so shouldn't count against the question. $\endgroup$ Jun 4 '14 at 14:21
  • 2
    $\begingroup$ @Donald Actually... I think the point we're making is that (in our opinion) answers like Kevin's are exactly what the puzzle is looking for. "Come up with any possible way to escape this conundrum!" And that's the problem. (Though of course you're free to disagree.) $\endgroup$
    – WendiKidd
    Jun 4 '14 at 16:03
  • 1
    $\begingroup$ @WendiKidd When you cut the bottom off of a bottle, it isn't a bottle any more, and therefore, in a very real sense, you have violated the conditions of the puzzle. I agree with your answer, but I think don't think that it applies when all but one of the answers involve "cheating". $\endgroup$ Jun 4 '14 at 16:22
  • $\begingroup$ Oh to have these be the speculative questions we are being asked again. $\endgroup$
    – kaine
    Nov 18 '14 at 21:10

This is me paraphrasing what a bunch of other people have already said in past discussions, but I feel this basically comes down to this:

If a puzzle has one clear, demonstrably correct answer, it is on-topic. If the OP has to return and say "oh, that's the answer I wanted! Technically that other one's right too, but that's not what I meant" then we're on the wrong track. If you can imagine infinite possibilities for how you could escape the blender/get the coin out of the bottle/make the best of whatever situation, you're definitely on the wrong track.

This does not mean that questions can't have multiple answers. They can! There can be multiple ways to arrive at a solution, as we have often seen on our questions so far. And if there are some puzzles that, say, have 2 clearly defined, explainable solutions... That's one of those "call them as you see them" situations. We'll have to judge these as we go along. But in general, I think there's a big difference between "This has one solution" or "This has two solutions" and "Let's imagine ALL THE SOLUTIONS!" And I think it's pretty clear when reading a question which ones are which. (And that the latter are off-topic.)

  • $\begingroup$ For completeness' sake/full disclosure: I cast a close vote on all four of the questions you've listed as examples before I saw this post. (Hey, at least I'm consistent!) $\endgroup$
    – WendiKidd
    Jun 4 '14 at 0:54
  • $\begingroup$ If someone could come up with 2 non-trivial alternative solutions that have not already been listed for 2 of the examples (except for the bus) I will reconsider my view on this but I currently do not agree with your opinion of these questions having an unreasonable number of realistic possible solutions. $\endgroup$
    – kaine
    Jun 5 '14 at 14:59

I think there is a difference between what makes a good puzzle and what makes a good question on the site. To be a good puzzle of this sort, the test is that if you are told the answer you respond "Of course, how didn't I see that?" I would accept "Which way is the bus going?" under this criterion (even though I had a different answer). For this site, the challenge is that people don't remember the puzzle exactly as asked, so there may not be a clear answer. Often, the puzzle will be a classic, maybe misstated, but identifiable. In that case I would cite a classic answer and ask OP if that is what s/he meant. Otherwise, I would make whatever assumptions are needed to make a good puzzle, state those in the answer, and give an answer. This is very fuzzy, but as the site depends on the interest of the participants if the puzzle is posed badly enough, nobody will be interested.

If I had to respond to the title of this thread, I would say we can't make clear ones. It is hard. A challenging one is the Kruskal origami question. I would accept it, but would understand the opposite view.


For good discussion I will include a counter viewpoint though I also voted to close some of the examples as they went against our stated policies. I don't nessesarily agree with these policies.

The way this site is supposed to work (for a particular situation) is as follows: someone finds or creates a puzzle for which he either does not know the answer or wishes to determine if others can find a superior solution. He submits the question to the site and recieves several answers. Based on the quality and cleverness of the answers (and several other factors) user vote on the best answer. Regardless of the votes, OP choses which answer helped him best in resolving the issue that brought him to the site.

If I found the inverted pyramid and did not know the answer, I cannot really improve the problem beyond the way it was stated in the question. A perfect answer would be the following:

"This puzzle plays on your assumptions that you want to retain the $100 bill but explicitly states that you are to remove it. The classic example is therefore to burn the bill. Several other solutions are convivable such as merely removing the bill slowly and relying of the force of a sharp pyramid to cut the bill."

It is worth noting that the only other answer given to the inverted pyramid question (cutting it) is likely infeasible so any solutions besides burning are debatable.

If the bottle/coin/cork question had said "you can't cut the bottle" rather than "you can't break the neck" I would not know any other answers than my accepted one!

The best example of this is the tiny man in a blender question. This was a famous interview question at google until it apparently was in that internship movie (didn't know that until reading your answers). I reallly would like to know what other answers are possible to that question that Google would have attempted. Only (really) two answers were provided to that question and there is not a number that should be consider problematic for this site.

Currently we average 1.6 answers per question while Area 51 claims that 2.5 is a good number.

I still agree that multiple answers to a puzzle is sometimes a bad thing. If someone says "I'm playing hangman and he wrote _ _ _ _ and I've not guessed yet; what is the answer?" that is a terrible question. If, however, someone asks a puzzle which is designed to make you cut Gorian knot, that kind of hits at the heart of the nature of puzzles as opposed to just arithetic or math.

The questions used as examples don't really have many non-trivial answers and can be answered reasonably easily. These should not be closed merely because they could have several answers as the Stackexchange system is designed to handle multiple answers with varying degrees of usefulness. This would cripple the usefulness of the site and prevent some interesting exchanges.

I don't enjoy they types of questions that are examples here but that does not mean they should be closed or banned from this site!

  • $\begingroup$ Without comment on the rest of your post, it's worth noting that the Area 51 guidelines are just that: guidelines. It's up to us, ultimately, to analyze the health of the site. I don't think that particular point is an issue right now. $\endgroup$
    – user20
    Jun 5 '14 at 15:37

You must log in to answer this question.

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