Should Elchanan Mossel's dice puzzle be undeleted?

I'm fairly surprised to see that Elchanan Mossel's dice puzzle was deleted. The question seems to be attributed -- this is Elchanan Mossel's puzzle, and that is mentioned in the title itself. This makes the plagiarism reason given a bit mind-boggling to me.

But ffao, the poster did not mention Gil Kalai, the author of the website the poster likely saw this puzzle at!

Suppose a friend of yours comes to you and tells you a puzzle he read in a magazine. Do you need to credit this friend whenever you tell this puzzle? I'd say surely not, since it's not his original creation to demand attribution for, rather he should credit whoever actually made the puzzle. I believe most people should agree with this; if you do not believe me, this is what Gareth has said earlier on this subject:

Sorry, that may not have been clear. I mean: If Way2 shows you a puzzle and you want to post it here, then you need to credit not Way2 but whoever actually made the puzzle, and if Way2 doesn't show you that information then you need to find it out some other way. – Gareth McCaughan♦ Jun 21 at 11:03

In this case, Gil Kalai did not originate the puzzle, Elchanan Mossel did, and he is the one that deserves the credit for his work, credit he got in the most visible place in the entire question.

Regardless of what the poster of this puzzle may or may not have done while posting other puzzles, I believe this one to be properly attributed. This question also has very high-quality answers, and keeping them deleted feels like a disservice to this website. Should we undelete it?

• Archived page for low-rep users. – Apep Dec 1 '17 at 17:20
• Did discussion in chat inspire this? :P – Quintec Dec 1 '17 at 17:21

I think the following things are reasonably clear:

1. It's a very nice puzzle and some nice answers were posted: it would be a shame to lose those.

2. At present it is basically copied from another source, without giving it due credit (though it does, indeed, credit the original source of the puzzle in its title).

3. The person who originally posted the question was not willing to attribute it properly.

The following thing is less clear but I think definitely also true:

1. It would not be appropriate for the person who originally posted the question and was unwilling to attribute it properly to receive a pile of lovely tasty Meaningless Internet Points for it.

Accordingly, here is what I propose to do.

1. Make a new question which (1) links to Gil Kalai's blog and (2) isn't so close to being a copy-and-paste from there. (Note: it's a very short and simple question and there's only limited scope for wording it differently...)

2. Make it community-wiki -- there's no more reason for me to gain Meaningless Internet Points from it than for the OP to.

3. Merge the existing question into it, which will attach its answers to the new question.

Any objections?

[EDITED to add:] This proposal seems to have had general approval, so I have done the above. I have undeleted every answer to the original question that had a positive score. One answer that I haven't undeleted gave a wrong result and complained that the question was ill-defined; the version I've now made is worded differently and I have no reason to think the author of that answer would have said the same about it, so I'm leaving it deleted.

Here is the new question.

• As somebody who didn't see the post being discussed, I'm a bit confused about the problems with attribution. If source A tells me about B's puzzle, isn't it B that needs the credit? A seems rather incidental. Is the problem that the post used A's telling of the puzzle while only crediting B, or something like that? – David Richerby Dec 2 '17 at 15:04
• Is it relevant that the OP of the post being discussed is someone with a longer history of plagiarism, so it's harder to assume they were acting in good faith and just didn't know the rules? I mean, normally I'd think this deletion route was pretty harsh, especially if it was a new user. – Rand al'Thor Dec 2 '17 at 15:55
• The post did indeed copy A's way of describing the puzzle and not give credit or a link or anything. I would prefer not to get too far into specifics, but I agree with Rand that a new user, or another user to whom we had reason to get the benefit of the doubt, both should and would have been treated more leniently. – Gareth McCaughan Dec 2 '17 at 16:59
• I really like your solution of merging the old answers onto the new CW question. It might be fair to deprive the OP of those MIPs, but I was feeling bad for humn/lauir and the other answerers. (Though this may play hilarious havoc with the HNQ list algorithm, with all the answers being older than the question!) – Rand al'Thor Dec 3 '17 at 11:24

The question in question uses almost the exact same phrasing as created and published by Gil Kalai:

You throw a die until you get 6. What is the expected number of throws (including the throw giving 6) conditioned on the event that all throws gave even numbers?

The phrasing in the question for comparison:

You throw a (single) die until you get $6$. What is the expected number of throws (including the throw giving $6$) conditioned on the event that all throws gave even numbers?

I can't find the original proposition by Elchanan Mossel, so I will assume for the purpose of this question that it is different than above.

As is stated in the referencing guide, using the exact phrasing is still against site policy:

Do not copy the complete text of external sources; instead, use their words and ideas to support your own. And always give proper credit to the author and site where you found the text, including a direct link to it.

Because the phrasing of a question is a material difference for non-trivial sentences, reuse of that exact phrasing without proper attribution is still against site policy. With respect to the policy, the answer to "Should we undelete it?" is not as it is now, but...

The meta post concerning how plagiarism should be dealt with suggests editing the post to fix the issue for a single occurrence or flagging for moderator deletion for repeated occurrences.

I would propose that in cases where repeated occurrences occur, individual posts that are deemed to have significant merit can be edited by users who would want them to be kept. The user can then flag them for undeletion. (I'm assuming these operations can be done in this order by users with sufficient reputation.)

In this particular case, I see two options available:
• the post should be edited to be written in a way that conveys Elchanan Mossel's original puzzle without relying on Gil Kalai's phrasing; or
• the post should be edited to include attribution to Gil Kalai near the quote that is used from his site (this would also involve moving the (single) to its own sentence outside the blockquote).

• Your whole answer seems to be predicated on the assumption that Mossel's original formulation was substantially different from Kalai's telling. Is that really likely? The question is a fairly simple mathematical thing and there are only so many ways of phrasing it. In particular, the sentence containing the question is a very natural phrasing and I imagine that the great majority of mathematicians would choose almost those exact words to ask that question. – David Richerby Dec 2 '17 at 15:09