12

Please provide 10k+ users more close votes.

Why?

  • Putting questions on-hold in a timely fashion is critical in providing useful feedback to new users. It's rare for salvageable questions to be put on-hold without a constructive comment having been left. Quickly putting questions on-hold keeps the question from attracting answers that may be invalidated by the suggested edits.

  • Anecdotal evidence shows that we're seeing a number of low-quality questions coming from users with question bans on SO. I have seen a number of comments where the OP indicates they can't ask questions anymore on SO.

  • We want to maintain site quality. Programmers has one of the highest answered rates (98%) for the Technology sites. But at an average of 43 questions per day, it's easy for us to fall behind.

  • It's rare for our close review queue to be completely empty. While we don't have the challenges of the SO close queue, I think that for smaller sites having a backlog in the close review queue is not a good thing. A backlog indicates that new users aren't getting the feedback they need in a timely fashion.

How many?
For the sake of consistency, I would suggest an additional 26 (total 50) so it matches what SO allows.

From the help page on close privileges

You may cast up to 24 close votes per day (50 on Stack Overflow).


  • Option A:
    Give all 10k users additional ### close votes

  • Option B:
    Give 10k users ###/2 close votes
    Give 20k users ###/2 close votes

  • Option C: Tie votes into close queue activity (ie. +## at 250 reviews; +## at 1k reviews) along with a tie into reputation levels.


Why not open this for everyone?
Once you've earned 10k on Programmers, you should have a good feel for the site dynamics and be active in self-moderation. That's why we allow vote-to-delete at 10k+. Adding more close votes recognizes the commitment that's been made and allows those users to continue shaping the site's culture.

Won't this lead to more questions showing up as on-hold?
Perhaps it will, but that only indicates we have been letting those new users down. Questions are put on-hold because they don't meet the site's criteria for high quality Q&A. With established users frequently running out of close votes on a daily basis, new questions that should be put on-hold are slipping through the cracks. Putting a poor question on-hold encourages the OP to revise the question; get the real answer they needed; and provide meaningful Q&A for the site.

  • 2
    Perhaps if a user is banned from asking questions on SO, it should also apply here. – GrandmasterB Mar 3 '14 at 20:56
  • 1
    @GrandmasterB - there is some merit to that. Many of those who are q-banned on SO end up becoming q-banned here too. They generally fire off 3 very low quality questions which are then closed and down-voted appropriately. – user53019 Mar 3 '14 at 21:16
  • I bet that "43/day" doesn't account for deleted questions, and that real number is more like 50/day, with 5-10 sure targets for closure (and subsequent deletion) – gnat Mar 4 '14 at 8:26
6

I'm not totally opposed to looking into upping the limits here, but I want to run some data first. However, there's two things at play here that we need to deal with separately, while at the same time not forgetting that they're symptoms of a whole.

Let's go by each problem, on its own.

Q-blocked users trying their luck here

This is a growing problem that illustrates just how imperfect the question block system really is. While it does the job well enough, people end up deleting and re-creating accounts, going to sites where their questions are at best tangentially related to the topic, or hooking up with voting rings to get their blocked accounts unblocked for a short time.

For now, this needs flagging, not just close votes. We need to make the system better and it's something on my list to do, it shouldn't be such a manual process to police in these cases. However, if you see someone that is very likely unable to ask questions on one site bringing the same poor quality here - flag it. Your mods can and will get me or other members of the community team involved in slowing these people down over the network.

More close votes would help, but I'd much rather fix the underlying flaws.

An increase in poor quality questions

Traffic is growing here. I was just looking at stats to determine if the rise is coming from folks with more engineering related searches, or the old 'best book to learn books' kind of thing. It's a bit of a split, but looks like it's landing firmly in topics we actually want, which is good. That means the right kinds of users are finding the site.

It also means that we're getting more clueless folks tripping over us in search results. While this has been becoming increasingly evident to folks that use and review the site daily, it's .. now .. being evidenced a bit more by stats and data. Programmers is far from the size of 'Bedford Falls', it's a city, and it's growing. We want to make sure you're equipped to deal with bigger city problems as they present, but I'd like to observe this for just a little while longer before proposing the changes to the team.

I'm marking this as deferred, for now, but it's not in any way forgotten - It'll take a day or two to refine some queries, at which point I'm going to start watching it.

  • please make sure that your data includes deleted questions (official "43 questions / day" most likely doesn't take these into account) – gnat Mar 4 '14 at 8:28
  • 3
    Thanks for the quick update! Some additional thoughts: I am leery of flagging q-banned users from SO. Except for ChrisF, none of the Progs mods can see if a user is q-banned on SO. The only way we find out is when a migrate fails; but we frequently don't get to that point because the question isn't worth migrating. Robert Harvey would have our heads on a platter for it. :-) Many (all?) of them are help vampires, so a stake through the heart via close votes may be most appropriate. Flagging seems like it would increase the amount of manual effort and generate a lot of false positives. ... – user53019 Mar 4 '14 at 11:54
  • @GlenH7 That's .. part of the weirdness of post blocks I'd like to address. However, the PSE mods can always find a community manager or SO moderator lingering around somewhere that can check. For now, as wonky and manual as it sounds (and is), that's the best course when it comes to users trying to sneak questions into other sites after being blocked. The problem with showing it to all mods is, it's calculated on the fly, and not exactly cheap. I'm thinking it over now, hoping for that flash of brilliance .. any minute noooowwwww .... (crickets) – Tim Post Mar 4 '14 at 11:57
  • 1
    @gnat Yeah, I am. I think there's a saved query I can build from, but I also think it's missing things. Hopefully I'll have something stitched together by the weekend. – Tim Post Mar 4 '14 at 11:58
  • 2
    ... That said, many of the regulars here are very happy with where the site is headed. This request is an attempt to get and stay in front of the maintenance that's required to keep that quality up there. Truth be told, we've had a few successful "conversions" of otherwise low quality salvaged to higher quality due to quickly being put on-hold. The OP wanted an answer, so they edited and the Q was reopened - everyone won. But we see the positive signs too and want to keep the momentum. – user53019 Mar 4 '14 at 11:58
  • And you're always welcome to swing by The Whiteboard to bounce query ideas off of the regulars in there. There's a few I can think of who would be ecstatic at the chance to help guide an efficient query out of all of that... :-D Both MichaelT and JimmyHoffa have already come up with some fairly decent manipulations of what's available in just DE.SE. – user53019 Mar 4 '14 at 12:00
  • 1
    @TimPost one thing I've stumbled across for around noon (CST) is that a number of the people who are more active in closing (and migrating questions) is that, well, we're out of close votes (if you've got 3 of the most common closers out, its often hard to muster timely votes) to migrate a question (rather than waiting another 6-7 hours for people to refresh votes). One of my pet peeves is the "belongs on" comments that lead to multiple postings. – user40980 Mar 4 '14 at 17:12
  • 2
    @TimPost gotta say... its painful recently with being out of close votes for the day typically at about 10am or so... and then watching questions like Do you like my idea? come through and seeing the '0 votes remaining' for close votes. Sure, the mods are picking up the spill - but they shouldn't have to. And then if I happen to click on /review, I see another dozen items in the queue that I can't vote on either (and that means when votes refresh I'll be out of them soon again). Have you thought any more about this? – user40980 Aug 7 '14 at 20:58
  • 3
    @TimPost It is now September. We're getting bombarded with homework. Things hurt. Have you given any more thought on this? – user40980 Sep 2 '14 at 22:32
  • 2
    @MichaelT I would say, either this or Refund close votes for questions deleted on the same day would be really helpful – gnat Sep 3 '14 at 18:57
  • 1
    @TimPost 6 months later; a reply would be greatly appreciated. Please also reply to the link gnat posted in the comment above mine. Thanks in advance :) – durron597 Apr 1 '15 at 15:28
  • Tim, regarding your advice on flagging, I would appreciate if you take a look at particular case asked about here: Is it okay to flag for moderator intervention attempts to circumvent question block at other site? – gnat Apr 28 '15 at 13:05
  • 2
    @TimPost Its been another couple months and September is right around the corner (with all the student homework that brings). Have you thought about this any more? – user40980 Aug 14 '15 at 20:58
  • @MichaelT: Just for kicks, I looked at questions closed by month. Turns out September is near the middle and the busy/bad months are May, July and August. – Jon Ericson Aug 15 '15 at 7:32
  • @JonEricson you are likely missing out on the "we've been doing a very concerted effort at cleaning up old questions through July." Furthermore, September is a state of mind in students that can begin in August and last through October (and happens again in January and when the term finishes in May). Data.SE also misses out on deleted questions - which we are one of the more active sites for feeding the roomba and 10k votes - especially when it comes to "do my homework for me" questions. – user40980 Aug 15 '15 at 12:44
4

Request for stats to take a glance at (yea, this might be another question, but its directly pertaining to this one and maybe getting a glimpse into the "running some data first").

  • Questions closed per hour on the site
    • Three or more community close votes
    • Two or fewer community close votes (mods step in)
  • Number of outstanding review tasks
  • For the people who have cast more than 25 (5/work day) close votes over the course of the week
    • Sum of remaining close votes available to this pool of people
  • For people who have cast more than 100 (20/work day) close votes over the course of the week
    • Sum of remaining close votes available to this pool of people

The graph should be for 24 hours for work days (a separate one for weekends wouldn't be bad, but thats not when things happen much or are too interesting for draining the close vote pool).

The information that I'm trying to glean from this is about how fast are people using up their close votes in various ways (trying to handle the queue) and when is it necessary for mods to start picking up the poor question closes that come in (because the close vote pool is drained). It would also give a hint as to how much increasing the close count for people would help, or if it would be more useful to have additional moderators who can rapidly cast close votes (though noting that this has the side effect of some decrying the 'categorical mod closings').

To this end, it might be helpful and/or interesting for this information to be 'live' while a community manager capable of identifying bad questions (and closing them) on P.SE is able to look at the activity over the activity for a week to help with either guidance as to help prevent these questions from getting asked in the first place, or looking at ways to encourage people who are less active in community moderation to partake in it too. (being able to look at a question and say, "why isn't that getting closed? Oh, of the people that cast 20 votes/day, only 2 of them have any votes left")

4

According to the data, of 464 users who can vote to close, 35 have hit the limit at least once:

days 
---- 
583  
488  
176  
68
31    
27   
21  
13 
12   
10   
10
...  

(35 row(s) returned)

For reference, 19 of those users have at least 10k reputation as of today.

Just two users have run out of close votes 1,071 times compared to 553 for the field. On Stack Overflow, 3787 separate users have hit at least 24 close votes in a day. Looking the other sites, there just aren't any that have such a small number of users who 24 close votes in a day so frequently:

Users  Days Site
-----  ---- ----
 3787 65564 Stack Overflow
  111  1391 Mathematics
  108  2836 Ubuntu
   59   999 Server Fault
   35  1624 Programmers
   33   313 English Language and Usage
   26    98 Meta Stack Exchange
   22   279 WordPress
   19   187 Magento
   16    43 Android Enthusiasts
   15    17 Gaming
   13    87 TeX - LaTeX

Raising the limit on Programmers would reduce the odds these users hit the limit, but I don't know if that would really help get bad questions closed faster.

Finally, if we were going to increase the limit, it doesn't make a whole lot of sense to just raise it for 10k users. From a very practical standpoint, changing it for everyone is a site setting, but changing it for users over 10k would require a code change. For another, the setting has already been raised to 50 for 3k+ users on SO, Math, Super User, Ubuntu, and Server Fault without any particular problems. I'd suggest removing that portion of the feature request that needlessly confuses matters.

  • 1
    stats for 28 questions a day don't seem to account for deleted questions, right? If that's right, consider re-checking your reasoning, and in particular whether given request is just a simpler way to address what was asked for in Refund close votes for questions deleted on the same day. If you run with dev access, you can check my close votes tab to roughly estimate for quick deletions – gnat Aug 15 '15 at 7:53
  • 1
    @gnat: True enough. Counting deleted questions, the site gets ~46 questions a day. But if you similarly count deleted questions on other sites, Programmers still falls short of Android Enthusiasts and all the sites above it. Deleted questions also don't explain why such a small number of users on this site have had trouble running out of close votes. – Jon Ericson Aug 15 '15 at 8:12
  • 46-28=18, that's 18 delete- (and hence, close-) worthy questions a day. Account for these and you will get 6 remaining votes for questions that aren't close- but not delete-worthy. Not much of a headroom – gnat Aug 15 '15 at 8:17
  • ...worth keeping in mind that Programmers isn't a big site. And is likely to remain such for quite a while. My understanding is, it is now officially okay for site to be considered healthy while remaining small: "Small, consistently active sites are great!" (typo in prior comment, should be "6 remaining votes for questions that are close- but not delete-worthy" - feel free to edit if you want to) – gnat Aug 15 '15 at 8:24
  • @gnat: Maybe don't bother closing ancient questions then? But more importantly, why do you have to personally vote to close every closable question that comes along? Where is the rest of the community? (So you understand, by "small", we are talking on the order of single digit questions a day. Programmers is large; it's just not massive.) – Jon Ericson Aug 15 '15 at 8:30
  • 1
    19 upvotes, 51 answers, featured status and 1500+ views on STCI Phase 2 suggest that site community rather strongly supports getting rid of broken windows. As for me personally, don't you worry, personally I am having no problems with current CV limit. What bothers me is that other folks affected by this issue feel frustrated and as a result have harder time contributing content to the site – gnat Aug 15 '15 at 8:36
  • alternative worth considering is to empower holders of silver/bronze tag badges with some extra CVs. Hi rep folks tend to have such badges in popular tags, which, in turn, tend to get more close-worthy questions. Extra votes in these tags would make for some headroom for concerned folks. (As of now, Programmers have 345 bronze tag badges and 13 silver tag badges) – gnat Aug 15 '15 at 8:57
  • 2
    I can honestly say that part of the reason I never hit my CV limit anymore is because the limit is so low there's little point going out of my way to clean up broken windows. Back when I did that, I often found I had no votes left to deal with the brand new terrible questions that kept appearing (which were often very similar to those broken windows). Remember that PSE is unique in having a massive scope change that makes almost every question asked a couple years ago automatically a broken window. – Ixrec Aug 15 '15 at 9:30
  • fresh example to ponder on - 5 minutes ago someone (not me) has spent their CV on a close- and delete-worthy question: 10K link – gnat Aug 15 '15 at 12:17
  • ...and one more note regarding stats here, would be helpful to see how many folks over here hit not 24 but 22-23 votes a a day, to account for those who force themselves to skip close-worthy questions so that they have at least 1-2 votes just in case if something terribly awful pops up. This comment above suggests that such practice may be widespread – gnat Aug 15 '15 at 13:41
  • 2
    @JonEricson Closing old questions is something we do because we consistently get "but this old question is still open" justifications for why new questions should remain open. This leads to a poor user experience to the asker in giving them expectations that such material is acceptable when it is not. Closing new questions promptly is important for not having a poor user experience for people answering the crap questions (that they don't know it is). Getting +20 rep on one day and then losing it two days later when the post is deleted. – user40980 Aug 15 '15 at 14:47
  • 1
    Jon - if the request's aspect of "give it only to 10k+" users is causing headaches for implementing this, then please, please feel free to ignore that part. It's there only because prior conversations implied that requesting more close votes for all eligible voters would be met with even stronger resistance than asking for just 10k+'ers. If it's trivial to flip Programmers over to a 50 votes per voter site, then please ignore the prattle about 10k+'ers and flip the switch. – user53019 Aug 15 '15 at 22:14
  • You have a valid point regarding overall site volume, although it's a bit dampened by the deleted questions. But the point of this isn't to be able to vote on "every question that gets closed." The point is that there's still a fair amount of cleanup left from the NPR days, and regular site volume has a lot of stuff that needs to be closed. We do have a strong base of close voters, and that base continues to grow with time. But we'd appreciate more tools (votes) at our disposal to help deal with the increase in low quality / off-topic questions. – user53019 Aug 15 '15 at 22:17
  • And for the record, I up voted because I appreciate your running the numbers! :-) – user53019 Aug 15 '15 at 22:17
  • 1
    @gnat: That's fine. – Jon Ericson Aug 17 '15 at 21:20
3

Because the single anecdote proves the rule, right?
I'm adding this answer to show how failing to close bad questions quickly impacts new visitors to the site and gives them a bad impression.

This gem of a question "https://softwareengineering.stackexchange.com/q/233104/53019" triggered this answer.

And it occurred on an afternoon when many of the regular reviewers were out of close votes for the day as hinted at in this chat comment

Now, there's nothing horrifically wrong about the question that's asked. But it is off-topic and it's not a good fit for the site. That type of question occurs often enough that we have a meta post explaining why it was closed.

The problem is that the person answering the bad question didn't realize they were walking into a trap. Their answer was down voted and they reacted negatively to comments that were left trying to clarify what happened.

First part of comment chain

 

Second part of comment chain

And the user is new to the site, having joined only 6 days ago.

user profile snippet


While it appears everything worked out well in this case since iCobot wasn't too rattled by the interaction, I think this occurs more often than we realize. If so, then we are losing otherwise valuable contributors because of a bad interaction from seeing bad questions that weren't closed off quickly.

Hopefully that puts a real example behind the site quality reasons justifying this request.


And for additional perspective, here's a screenshot of the 10k tools from today looking at questions needing review / close votes.

first half of 10k tools

 

second half of 10k tools


As further proof of the confusion caused by not getting questions closed quickly, the comment thread has continued into the next day.

comments, round two, part 1

 

comments, round two, part 2

You must log in to answer this question.