36
$\begingroup$

It seems like there are routinely low effort, sloppy PSQ questions from new users that immediately after being posted get upvotes, sometimes multiple.

Is this a case of sock puppet accounts? What are the mechanisms that exist to stop this from happening if it is sock puppet accounts? Do people get banned for this?

$\endgroup$
  • 23
    $\begingroup$ Sometimes these are just "real life friends". Sometimes these are really just class mates. Sometimes these are socks. Sometimes these are boots. In either case, it's very annoying. $\endgroup$ – Asaf Karagila Jun 14 '18 at 18:12
  • 1
    $\begingroup$ This is a really interesting one in terms of MSE norms - sometimes there have been badly posed no effort questions, which nevertheless have liked (but not upvoted) for one reason or another. So it seems consensus is no to upvote a badly posed question, even if it is a nice question as far as one is concerned. I recently answered one (which hadn't upvoted) where one of the comments suggested maybe answerers were upvoting to make the question sound better (!) Wow - cannot believe people would do that for that reason. But anyway that comment has definitely made me circumspect on upvoting, since $\endgroup$ – Mehness Jun 14 '18 at 23:41
  • 8
    $\begingroup$ There is one person who creates multiple accounts, and uses them specifically (after answering a few questions to earn sufficient rep) to serially upvote the poorest of poor posts, as well as to answer fully closed questions (in comments). This person is known by math.se mods, who, from what I understand, have resigned themselves to playing "whack-a-mole" - deleting one account of this person, after another, more than a dozen, and counting. But the mods, despite playing an on-going game of "whack-an-account", have done nothing further. This person then goes on to create more accounts... $\endgroup$ – Namaste Jun 14 '18 at 23:56
  • 2
    $\begingroup$ and the mods know this, but only account when other users devote our time to flagging fishy behavior. And then, only after one of this person's accounts has transgressed site policies/ only then, a mod will delete that one account. I am convinced more can be done, by engaging SE staff, but all I know is the mods on this site seem to have decided it's not worth their time to look much more deeply into stopping this person who regularly comes back to upvote the poorest of questions, and answer questions which were rightfully closed, in the comments. $\endgroup$ – Namaste Jun 15 '18 at 0:01
  • 2
    $\begingroup$ On Meta Stack Exchange there exists sock-puppets tag. Maybe if somebody has time to have a look at some of the posts there, they could be able to answer your question in detail. This question seems relevant: How can you detect if users have created sock puppet accounts? This one also mentions briefly what is expected from mods: How should sockpuppets be handled on Stack Exchange? $\endgroup$ – Martin Sleziak Jun 15 '18 at 5:28
  • 2
    $\begingroup$ I see this as one more example of a nonlinear scalability barrier. As user numbers grow, all sorts of odd behavior patterns can be observed. Those disproportionally affect the people engaged in the vetting/sorting/moderating business. Stemming twice the number of posts with twice the number of voting people does not really work because each person gets to see almost all of the new bad stuff. That's psychologically overwhelming. Calls for technical assistance are typically answered with "It's insignificant, don't let that affect you." Alas, there are limits to the applicability of such advice. $\endgroup$ – ccorn Jun 15 '18 at 5:44
  • 3
    $\begingroup$ Non-serious proposal: Buttons "blue pill" and "red pill". The blue pill hides the review queue functions and shows you only high-score stuff that is at least one year old. The "red pill" sends you back to the unpolished reality. After lots of review tasks, the system automatically activates the blue pill. Oh, and moderators should be rotated frequently, for their own good. $\endgroup$ – ccorn Jun 15 '18 at 6:09
  • 38
    $\begingroup$ "moderators should be rotated frequently, for their own good." Won't they get dizzy? $\endgroup$ – Gerry Myerson Jun 16 '18 at 1:32
  • 3
    $\begingroup$ Well, at least we should not let the system grind them to dust. If moderation at scale requires being non-sentient, that's a problem which needs a solution. $\endgroup$ – ccorn Jun 16 '18 at 6:04
  • 6
    $\begingroup$ @amWhy Concerning the user you mention, there's simply nothing other than deleting the accounts when we find them (and we find a number ourselves, but flags from users help finding them faster) that can be done. If somebody is determined to create an account, they can. They can't keep it if the leopard's spots show, but … shrug $\endgroup$ – Daniel Fischer Jun 17 '18 at 21:55
  • 2
    $\begingroup$ Some examples would be nice. $\endgroup$ – user 170039 Jun 18 '18 at 3:41
  • 7
    $\begingroup$ @Gerry: That depends on the rotation. If you rotate the moderators by rotating their entire frame of reference, they should be fine. $\endgroup$ – Asaf Karagila Jun 18 '18 at 11:28
  • 28
    $\begingroup$ A much bigger problem is when reasonable questions from new users immediately get downvotes. $\endgroup$ – Phira Jun 19 '18 at 16:56
  • 11
    $\begingroup$ My impression is that questions receive too little votes in general. Which leads me to the unorthodox idea of assigning much more weight to both up- and downvotes on questions. This might shift the existing market-like mechanisms more toward the question side. Have to think about that. $\endgroup$ – ccorn Jun 19 '18 at 17:39
  • 2
    $\begingroup$ Actually, @Phira, you're wrong. You just can't see the waste that's fallen on the way side due to failure to respond to reasonable questions about context. Far more poor questions get blessed, and quickly answered (because low ability answerers) and upvoted than do reasonable questions that might challenge our leaders in the pack for capping each day, week, month, are able, or have the time, for answering. $\endgroup$ – Namaste Jun 28 '18 at 23:25
10
$\begingroup$

Voters on questions can be roughly taxonomized into the following groups:

  • Math.SE regulars who are considering answering the question and find it interesting/high-effort.
  • The OP, their friends, their classmates, and random Googlers who need the answer to the question and would find it useful (for their exam, homework, research, etc) but don't care about the question quality or site norms.
  • Tourists from "hot network questions" or social media.

There is a scale separation of roughly an order of magnitude between the number of voters in each group. A detailed research-level question on PDEs will get 0-1 votes. A question from this week's Calculus AB worksheet will get 10-20 votes, and a random Batman graph will get 400-500 votes. It's just the nature of the math.SE beast.

$\endgroup$
  • 4
    $\begingroup$ This doesn't answer the question of the immediateness of it. Why is it that in less than a few seconds some posts have upvotes, not enough time to even read the question to tell if it a problem from this week's calculus worksheet? $\endgroup$ – user223391 Jun 24 '18 at 15:30
  • 5
    $\begingroup$ One reason might be that one need in general less than 3 seconds to fully understand those questions? @ZacharySelk $\endgroup$ – user99914 Jun 24 '18 at 17:28
  • 1
    $\begingroup$ This is a rather superficial non-answer. Yes, this happens, but this doesn't address the trolls who hang around for the thrill of "upsetting the order" by upvoting and answering crap, while other, better questions, go ignored. $\endgroup$ – Namaste Jun 28 '18 at 23:05
  • 1
    $\begingroup$ @amWhy I'm rather skeptical any such trolls exist. There are far easier, and flashier, ways to disrupt the site than to secretly upvote borderline questions... $\endgroup$ – user7530 Jun 28 '18 at 23:17
  • 1
    $\begingroup$ Perhaps you've got more to share, @user7530. $\endgroup$ – Namaste Jun 28 '18 at 23:19
  • $\begingroup$ @amWhy I hope you get paid something from MSE to legitimize your passion in deleting non graduate level questions $\endgroup$ – AmateurMathPirate Jul 1 '18 at 6:28
4
$\begingroup$

It is generally impossible to perfectly police a site the size of the stack exchange, so site is moderated automatically by users(and some dedicated moderators) via the review system.

https://math.stackexchange.com/review

Invariably this means some amount of error,since there is some fuzzy logic the idea being that over time more people will downvote then there are sock puppets. All questions are designated as important when they are new, but this status very quickly degrades if they are not "popular". This means a question should show up, be down voted, and then drift into oblivion without causing too much disruption to the site's usability.

Since most users are either searching a specific question, and not arriving on the site to just see "what's hot", there is a fairly small pool of questions which both answer the question and are popular, these will float to the top of a search.

Users which repeatedly receive bad reputation are banned https://math.stackexchange.com/help/answer-bans

This means generally speaking it is not necessary to manually weed out sock puppets, since the accounts they promote are automagically found to be disreputable by the community.

See also

On Meta Stack Exchange there exists sock-puppets tag. Maybe if somebody has time to have a look at some of the posts there, they could be able to answer your question in detail. This question seems relevant: How can you detect if users have created sock puppet accounts? This one also mentions briefly what is expected from mods: How should sockpuppets be handled on Stack Exchange? – Martin Sleziak

$\endgroup$
  • 4
    $\begingroup$ Statistical significance can be surprisingly irrelevant. Consider facilitation of access to unlicensed copyrighted works. This is probably not significant here, but I expect that there are provisions in place for dealing with such. Or consider highly disturbing content. Statistically insignificant, but once spotted, it is upsetting enough to warrant action. Also note that long-term sorting effects are due to users like those asking here; no credit due to the machine there. At least I would be positively impressed to see some supporting machinery for recurring tasks in place. $\endgroup$ – ccorn Jun 19 '18 at 15:20
  • $\begingroup$ Posts with significant amounts of both upvote and downvote activity could be automatically flagged as warranting moderator concern. Most aren't controversial enough to be polarizing, so the likely source would be sock puppets upvoting while users downvote. $\endgroup$ – Cristobol Polychronopolis Jun 28 '18 at 14:22
  • $\begingroup$ @amWhy My limited contribution to this particular subdomain of the stack exchange neither precludes my ability to understand how the system works by using stackoverflow, nor removes the possibility that I might peruse this exchange upvoting questions and answers, though not frequently needing to ask new ones. I find your comment in poor taste. $\endgroup$ – awiebe Jul 1 '18 at 8:27

You must log in to answer this question.