Is there a case for treating the contest math tag a bit like homework?

Question

There has been some debate about live contest questions and how to handle them - and likewise about homework, which is a side issue to the question I want to ask, which is about the use of the "contest math" tag - which is perhaps underused anyway.

Would it work to have a policy by which "contest math" questions should generally be answered by hints rather than full answers - at least initially?

There would seem to me to be two possible advantages:

1. One purpose of contest questions is to challenge participants to stretch themselves, and to dig around in their mathematical knowledge for clues how to proceed. Giving answers short-circuits that process and cuts out the imaginative aspect of problem-solving. Since it will be rare for the actual solution to a contest problem to be significant scaffolding for future learning - it is the process of solving which is most in question - help with the process seems indicated, rather than "help" by giving the "answer".

2. Holding off providing a complete answer would help to preserve the integrity of "live" contests, and enable appropriate action to be taken in cases where action was felt to be necessary. I know issues should be flagged, but answers can come up very quickly.

Unlike "homework" the "contest math" tag can fairly be added by anyone who recognises a contest problem. It might be helpful if brief contest details were provided in a comment (if not in the original question) - since this also would indicate the level of mathematical sophistication appropriate to the question.

Answer 1

I disagree with this proposition.

Problems from ongoing contests should receive neither solutions nor hints until the deadline.

Problems from past contests should not be treated differently from any other question (that is, it depends on the context and the judgment of the answerer whether a hint or a full answer is appropriate).

I think that we should generally encourage people to give more information on sources and motivation.

Since it will be rare for the actual solution to a contest problem to be significant scaffolding for future learning

As someone who learned about contest problem solving mostly by spending time on failing to solve a problem and then seeing a full solution, I strongly disagree with this.

Answer 2

I agree that problems from old contests where the deadline has passed should not be treated differently from other problems. However, I wanted to make one caveat. I think if a question was asked while the contest is live and someone notices that they should still flag for a moderator to add the "This was asked in a live contest but the deadline is passed now" moderator message. This is useful both as a warning to people that they might get caught, and for locating past attempts at cheating when doing an investigation.

Answer 3

I fully agree with Phira's answer, but I'd like to be a bit more explicit yet.

As I see it, math contest questions (whether already tagged as contest-math or not) fall into two categories:

1. Problems from past contests: I see no particular reason for treating these in any way differently from any other questions on this site. Some users may want to provide full answers, some may prefer to give just hints, and I expect this would, in any case, frequently depend on details of the specific question.

2. Questions from ongoing contests: These should not be answered at all. Period. Not even with just hints, and certainly not in full.

   Instead, the procedure I'd recommend to follow, if you spot a question taken from an ongoing contest (based in part on this proposal for handling such questions by David Savitt, one of the organizers for Canada/USA Mathcamp), is:

   1. Flag the post for ♦ moderator attention, noting the contest it is taken from, and ask for it to be closed or locked (but not deleted, unless explicitly asked to do so by the contest organizers) until the contest is over.

      It may be a good idea to ask the moderators to also contact the contest organizers, even if you've already done or intend to do so yourself, since it can be a bit tricky for the organizers to contact the mods if they don't already have an MSE account and the 15 rep needed for flagging posts.

   2. Leave a comment stating which contest the question is taken from, when it ends and what you've done and/or intend to do about it, and ask other users not to answer it before the contest ends (just in case it takes the mods a while to respond to the flag).

      (In fact, if the question is very new and hasn't been answered yet, the first thing I might try would be posting a pseudo-answer saying that it's from an ongoing the contest. This is so that anyone else currently in the process of writing an answer would hopefully see and click the "a new answer has been posted" notification bar — which is a lot more prominent than the corresponding notification for new comments — before posting their own answer.)

   3. If possible, contact the contest organizers. Let them know what's happened, how we usually deal with such situations^* and what they can do to help.

   4. If you have the rep to cast a close vote, that probably won't hurt, either. Just possibly, it might even allow the question to be closed by the community before the mods get around to handling it.

   Some parts of these suggestions are intended to minimize the chance that anyone will post an answer if the question doesn't yet have any. Contest questions which already have answers are, in some ways, much more problematic, but I don't see any reason not to follow the same general procedure in that situation as well. Certainly one should not try to delete or otherwise hide the existing answers, unless and until the contest organizers ask for it — it's very likely, thanks to our efficient notification system, that the person who asked the question will already have seen them anyway.

^{*) I realize that the details of that are still not as firmly established as they ideally should be, but I would hope there's at least some level of agreement on the general principles by now. If nothing else, in the absence of a formal policy to point to, one could always link to recent meta threads on the subject and let them form their own impression.}