8

I just got a review audit where the "late answer" was a duplicate of the accepted answer, but flagging it on that basis failed the review. That seems like a bug to me.

Question here: Is it better to call a function that doesn't have an effect at that point, IF it improves code clarity?

I found that by clicking the question to view the entire Q&A, and noticed that the copy was from a generic "user12345" account (probably with 1 rep as well) and the original from a real-looking account. Viewing the whole Q&A is often useful when reviewing, especially late answers (which are often repeats of earlier points, disagreements with those, or random addendums better submitted as edits). I looked at the duplicate and thought "weird, rep harvesting maybe?" so I clicked "needs moderator attention" and wrote "copy and paste of accepted answer". FAIL!

I think that this review type can't ever work, because to avoid the problem I found you'd need to make up whole new "late answers" which by definition the review bot can't know anything useful about. Either that or to pass you'd just need to take any action at all, including upvoting the original of the duplicated answer. Which I didn't do, BTW, because I had no idea or opinion about the question or answer.

(this is not a duplicate of "What are review tests", it predates that and focuses on a particular defect within the process. See the comments below).

  • I can't seem to locate the review you are talking about. Link? It sounds like maybe the user on the review was changed/incognito to trick the reviewer, when the actual review you were reviewing was the accepted answer, just showed a different user during the review to trick you. Or did you see two seperate answers by two differnet users on the same actual question page? – CRABOLO Sep 16 '15 at 1:12
  • I didn't realise the exact question was at all relevant so I didn't include it, but I've added it. The fake answer was from one of the generic "user12345" usernames, the original was a filled-out username (edited that into the question) – Móż Sep 16 '15 at 1:58
  • 2
    Possible duplicate of What are review tests (audits) and how do they work? – CaldeiraG Aug 29 '19 at 7:28
  • 1
    @CaldeiraG This was asked before that section was added to the FAQ, and I referenced it when adding it back then. – Sonic the Anonymous Hedgehog Aug 29 '19 at 7:50
  • @SonictheAnonymousHedgehog alright, didn't noticed it, there was another person with the same doubt on the comments. – CaldeiraG Aug 29 '19 at 8:04
9

The post in the audit was the accepted answer. You flagged that answer itself — not a duplicate answer, since in all likelihood no such duplicate was ever posted — and, of course, if it had been a live review flagging a good accepted answer like that would have been quite wrong.

How can this be? Well, audits will deliberately fake post details so you can't tell if a post has been accepted or upvoted, or who wrote it, in order to avoid making audits absolutely trivial: you have to look at the contents. But this means that naive attempts to suss out duplicate answers will get you in trouble when simply not looking for duplicates may not.

But all is not lost: if you pay attention to where you land on the page when you click the link to the right of the answer in review, you can always tell. The link given will always scroll you to that particular post; if it leaves you at the top of the page, you know it's been deleted (and is therefore a known-bad audit), and if it takes you to an upvoted answer unexpectedly, you know that that's the real deal and that the review is a known-good audit.

P.S. Looking for duplicate answers is very helpful, so please do keep doing that, just make sure to look at all the information available so you don't run into audits.

  • 1
    Also, a shameless plug: my SOUP user script includes a feature to show the other answers directly on the review page. It will not show the answer being reviewed twice, even if it's an audit with faked details, so it would've also saved you from this confusion. – Ilmari Karonen Sep 16 '15 at 12:15
  • Thanks for the answer. The SOUP script sounds useful, if I'm going to put in the effort to game the review system automatiing that as much as possible is sensible. But at this stage I'm just not going to waste time reviewing where the random robot goes. – Móż Sep 16 '15 at 21:49
  • @Ӎσᶎ It's really not gaming the system to integrate robust dupe-checking on the same page. So I'd recommend that. – Nathan Tuggy Sep 16 '15 at 22:23
  • The general idea of "tool to work around features designed to prevent fake reviews" is gaming the system. The fact that as currently designed the audit system deliberately fails honest reviewers is why I'm calling this a bug. And having to have a tool, or a list of known problems, and pay attention to something other than reviewing, is why I'm not doing it any more. The "game" of "can I beat the bot introduced to make reviewing harder" is not one I want to play. Unless they introduce a "beat the bot" badge for passing 5 or 10 audits. – Móż Sep 16 '15 at 23:41
  • @Ӎσᶎ: That doesn't make any sense. The design of the queue excludes any possibility of checking for dupes within it; therefore, dupe-checking must, logically, start from the provided link which opens with, as it turns out, just enough context to avoid the problem if you know exactly what you're doing. The audit behavior is certainly quite problematic, and I don't consider it at all praiseworthy, but to say that good reviewers will fail this by design is quite wrong. It is only those who are making a good-but-not-quite-good-enough attempt that will fail. – Nathan Tuggy Sep 16 '15 at 23:45
  • Let us continue this discussion in chat. – Nathan Tuggy Sep 17 '15 at 0:58
  • 3
    'The post in the audit was the accepted answer. You flagged that answer, and, of course, if it had been a live review that would have been quite wrong'. I disagree with this, as I followed the exact same reasoning of the OP, four years later. If the reviewer saw someone trying to re-enter an response to a question as his own, asking for moderator attention would be the correct review step to be taken – caxcaxcoatl Aug 29 '19 at 2:38
  • @caxcaxcoatl: How does the edit look? – Nathan Tuggy Aug 29 '19 at 7:09
  • @Nathan Tuggy, yes I think I can agree with that, now. I still do not agree with how the system behaves in this case, but since it has no concrete consequences I don't see the point in complaining. – caxcaxcoatl Aug 29 '19 at 17:52

You must log in to answer this question.

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