Suspecting cheating at competitions

Question

Once in a while, a new user posts an immaculately-written challenge with test cases formatted like

1 ≤ m ≤ 10000 1 ≤ n ≤ 100

SAMPLE INPUT 10 9 8

SAMPLE OUTPUT 1

Maybe they say they're especially interested in Java code that runs quickly.

These make me suspect an attempt to cheat at a competition on an external site, which is disallowed. Sometimes it's blatant, but other times I'm unsure. Googling for snippets doesn't find anything and I don't really know what programming contest sites to look at.

Accusing them of cheating seems rude, especially in ambiguous cases -- maybe they just like how a site presents challenges and mimicked their style? And if they deny copying and claim it's original, should I believe them? Is a sense that something's off enough to justify closing and/or flagging?

Answer 1

We should not consider every problem in this format in a running competition.

These competitions generally runs for only a few hours. After they end, they are just exercises. They may still maintain a scoreboard somewhere, but it is not much different from PPCG upvotes and reps.

Some websites don't like people posting code in their own forums or something. But I assume that's just for not making it too easy. I doubt any of them will ban discussing the algorithms, or interfering with other websites, as long as it is not in a running competition. (If they really take it too seriously, it's really their bad.)

If they are really running competitions, it's unlikely they will get it satisfying our rules and receive an answer in their acceptable languages in time. Some websites even have mechanisms for detecting cheating (mostly for similar submissions in their own sites, and has no effect on copying the algorithm, though). So I don't think it is too much of a problem if we didn't notice some of them, yet.

It could also be a homework problem. But any problem could be a homework problem.

It could only annoy us for not satisfying the rules and unwilling to change. I don't know whether there is more we can / should do for this. One reason I could think of is copyright.

But if we don't find anything feasible, we could usually still just close them and/or asking for changes for the normal reasons. We could ask for its source and tell them copying the algorithm problems without modification is the reason it probably won't satisfy the rules, and sometimes uninteresting, but don't get distracted from the real problem, namely not satisfying the rules itself.

Answer 2

Be more aggressive about rewriting questions if they're a long way from typical PPCG conventions

If a question is defined in terms of test cases (with the exception of test-battery, where the number of test cases is typically very large and there's clearly no other way to specify the behaviour), it specifies a single language when there's no good reason, it has excessively strict I/O requirements, or the like, the challenge is only going to be improvable by changing it to work more like challenges normally do on PPCG. So we'd want to edit it as it is.

We often try to stay near to the OP's intent for challenges even when it makes no sense. However, if the challenge is a long way away from our conventions, doing that is probably actively a bad thing; if the details of the challenge change in such a way that they no longer answer the original question, there's at least no scope for cheating on a competition with them (and even if there is no competition elsewhere, we'd still want the best possible competition here). (For example, most other competition sites I'm aware of have a fairly strict runtime limit on their code, the challenge being "solve this so that it runs on all test cases in under 1 second each" or the like. Here, a pure code-golf is normally but not always more interesting; note that the original victory condition is not fastest-code either because it normally cares about the code being as fast as possible to write and only sufficiently fast to run.) It may well just be that the OP has competed elsewhere and is emulating the style of other sites rather than copying questions; in this case, fixing the question to follow our style is probably only going to help.

Note a special case: if the challenge is plagiarised, it probably needs to be deleted on copyright grounds. However, if the original can't be found, that probably isn't the case.