137

On the one hand, I see very few posts in the review queue when I visit.

On the other hand, I see a lot of pretty awful first posts sitting on the front page.

This leads me to suspect that robo-reviewers are blowing through the queue.

Given how empty the queue is, could you please increase the number of people who have to check in before a post leaves the queue?

Of course, some honey-pot solution to detect and disable the auto-clickers would be even better.

2
  • 8
    On Stack Overflow only, I presume? The other sites don't have near as much volume for this to be a real issue.
    – ale
    Jul 11, 2013 at 20:00
  • 7
    I'd say that new questions who don't get a revisit from the person who asked them should expire. I think that it is expected of users to respond to comments and answers to their own question and if not, I see no real use for the question since it will sit for eternity without an expected answer. If not expire, at least allow some users to accept for them.
    – djangofan
    Jul 11, 2013 at 22:44

4 Answers 4

98
+150

I'm pretty sure it was three reviewers when these queues were originally created. I never really understood why it was lowered to just one. Sure, that cut down on the robo-reviewers casting three upvotes on everything because there wasn't a No Action Needed button, and review audits didn't exist at the time either. Now that we have all this stuff, can we bump it back up?

At this point it seems like we've cut out the abuse of users destroying our site with the upvotes and stuff,now they're destroying it with the No Action Needed button. Really, it only takes one good reviewer out of the bunch to make something happen with that post. A single user reviewing it (flagging, editing, whatever) should still dismiss it from the queue. Once it's flagged or edited, there's really no reason for additional people to be looking at. Presumably that person took care of the problem. However, I think it should take the full three No Action Needed reviews to dismiss it legitimately as not requiring any action.

3
  • 14
    Yes, I've been running into instances of spam that slipped through review due to a few people spamming "No Action Needed" who either weren't yet banned by the audits or who found ways to game them. I then had to go back and check everything else they'd reviewed to handle the other garbage they let through. If we made this require at least two reviewers, the odds of this stuff surviving would be significantly reduced. Jul 11, 2013 at 19:54
  • 4
    I think people just don't understand the meaning of "No Action Needed" so majority of those cases are innocent mistakes. I'm almost sure that if notified, those people will become better reviewers, and we should also make it more clear what is the real meaning of clicking that button while reviewing. Maybe showing confirmation dialog box for the first X reviews. Jul 11, 2013 at 22:28
  • Has this been changed? I'm still seeing NaA posts from 1-rep users pretty consistently.
    – Jack
    Dec 20, 2013 at 2:12
57

Yes!

More reviewers!

I think that currently we only need one reviewer per post. On SO, at least, this number can be raised significantly. Up to 3 or even 5. Since the FP queue is usually empty, we won't have any problems with more reviews needed per post.

14

Additionally, I think the system should track when a user presses "No Action Needed" and when someone else takes action. An automatic review ban should be placed if a user consistently presses "No Action Needed" when action was needed.

This would cut back on robo-reviewing significantly. The review ban could last as little as two days; just enough to get the point across. It just needs to provide a link to their review history so that they can meta-review their reviews.

I've started a more comprehensive Meta question here.

3
  • 3
    Possibly flag vs no action needed. A comment action could be anything from "this needs minor attention" to "everything you've done is wrong Jul 11, 2013 at 21:47
  • @RichardTingle: Well, even then. If a flag gets declined, then the action they took was invalid and thus it could be argued that the No Action Needed button was the right course of action for that post... -- There would be so many lines that would have to be drawn for this type of automated suspension that I don't really see it working out well.
    – animuson StaffMod
    Jul 11, 2013 at 22:04
  • @animuson I'm about to make a meta question concerning this - I'm curious what your thoughts are when it's posted.
    – user206222
    Jul 11, 2013 at 22:05
3

I would suggest to give some guidelines to the reviewers. And force them to mark as checked some conditions before being able to skip or mark it as no action needed. This conditions could be

  • Format: There is code with the proper format?
  • Tags: The question is relevant for the tags marked?
  • Grammar: It is easy to read and correct (e.g:no i, i'm )?
  • Inside scope of the help centre: If not please leave a comment and vote the corresponding flag.

This way it will not stop much the speed of reviewing but it will help to make sure it is not just clicking no action needed per se and he did some effort

You must log in to answer this question.

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