48

I'm a high-rep user on Stack Overflow. I shouldn't care overmuch about downvotes directed my way. In general, I don't, except: I've asked only two questions; against hundreds of answers. A single user can downvote all my questions without triggering (as far as we know) the vote fraud algorithms. Questions are featured much more prominently than answers in a user's profile, amplifying the apparent effect of those downvotes.

I bring this up because I've noticed a pattern. Periodically, batches of my posts (including my questions, which are quite old at this point) will get downvoted at the same moment. This has happened maybe three or four times. I tend to vote-to-close about a dozen or so questions a day, and posters will sometimes make clear their displeasure at these closures. I suspect, but can't prove, that these users are going around voting down all those folks that voted to close their questions. If they limit themselves to one or two votes per closer, no algorithms (again, theoretically) will get triggered.

I feel a bit like a whiner bringing this up, but:

  • I'm proud of my contributions to SO, and don't hesitate to point people to them, including from my resume. I don't want graffiti on my house, and I don't want garbage in my profile.
  • I wonder what will happen when I have another question to ask - will the accumulated downvotes prevent me from asking?

I've written to the team before about this, and Jeff was kind enough to roll back a previous batch of downvotes on my questions. That's not a good way to deal with this, though. As it stands, I'm reluctant to vote to close bad questions.

I propose extending the vote fraud detection script to include this scenario. Roll back downvotes if the voter is voting down more than N posts by users who recently (successfully) voted to close one of their questions. Bonus points if the downvoted posts in question are quite old (and thus no longer visible).

Again, I can't say that this is a real problem, but I suspect it is.

EDIT: The question downvotes I received yesterday, and several others besides, appear to have been scrubbed from my questions.

22
  • 6
    As much as I feel it's whining too, I see the merit. But I'm downvoting because a) I feel that you're not fighting the cause but the symptoms, and b) you can't stop other users from being pissy.
    – jcolebrand
    Commented Jun 29, 2011 at 2:36
  • @jcolebrand: Sometimes fighting the symptoms is a good start. But I understand where you're coming from. Commented Jun 29, 2011 at 2:46
  • 12
    @jcolebrand: If the cause ("users being pissy") is unfixable, then what's wrong with fighting the symptom? Especially since the symptom is the bad part -- I don't really care if I have some cold virus in my body, as long as I don't get a headache and stuffy nose.
    – jscs
    Commented Jun 29, 2011 at 2:50
  • 4
    This happens, nothing you can do about it. Better they downvote a question than a perfectly good answer. Commented Jun 29, 2011 at 2:56
  • 2
    @HansPassant they are downvoting both questions and answers.
    – jcolebrand
    Commented Jun 29, 2011 at 2:58
  • 1
    @Hans, not in my experience. Commented Jun 29, 2011 at 3:00
  • 1
    @HansPassant that's just it. If they've figured out the rules for downvote revenge reversal, then they will stay under that threshold (which we should not discuss on purpose)
    – jcolebrand
    Commented Jun 29, 2011 at 3:02
  • 2
    @jcolebrand: I knew that metaphor was going to cause trouble. What about a chronic disease, though (the unchangeable "users being pissy")? Mightn't you like some painkillers to allow you to go about your day? Or do you think that it is a terminal illness for the site?
    – jscs
    Commented Jun 29, 2011 at 3:02
  • 4
    @KolyaMiller while that would be true for some users, not all of us have tremendous amounts of questions. A) things get answered before they become our questions, B) we're pretty good about figuring stuff out on our own, C) you're probably right ;)
    – jcolebrand
    Commented Jun 29, 2011 at 3:13
  • 1
    Maybe Micheal is a special case. Downvotes on questions are free and he has few enough questions not to trigger fraud detection. If he had ten questions they would get rolled back. Face it, revenge downvoters for closed questions are probably not clever enough to make just two downvotes. Commented Jun 29, 2011 at 3:20
  • 4
    @RickSladkey you don't have a careers.SO page do you? You don't consider this site to be a good representation of your meta skills as a programmer do you? Or the network as a meta representation of your skills as a human? What a shame. I can think of a lot of reasons why Michael wouldn't want his questions downvoted out of spite. (same as getting his name spelled right)
    – jcolebrand
    Commented Jun 29, 2011 at 4:09
  • 1
    @jcolebrand: Whoa, whoa. I didn't downvote the question or say Michael doesn't deserve those votes. Of course he does. I was just speculating on how the idiots got away with it in his case. Commented Jun 29, 2011 at 4:17
  • 4
    Sorry, I jumped on this Maybe Micheal is a special case. ... he's not a special case, I get those downvotes too, I just wanted to play devil's advocate here to get convinced otherwise. I've had a really good answer get downvoted to oblivion because it was about the word "breasts" on EL&U.
    – jcolebrand
    Commented Jun 29, 2011 at 4:19
  • 1
    @Michael: Sorry about the spelling, I cannot correct it now. I meant that you in particular are vulnerable in the current fraud detection system. A solution might be to reinstate the cost of downvoting on old questions. Commented Jun 29, 2011 at 4:24
  • 1
    twitter.com/#!/jonskeet/status/86097267035095040 ... interesting
    – jcolebrand
    Commented Jun 29, 2011 at 16:14

5 Answers 5

3

Are you sure the revenge voting doesn't already account for this?

Your edit indicates that the votes were indeed scrubbed and a simple implementation of the revenge voting detection would simply check for rapid successive vote activity directed at a users' posts (questions or answers), why it would only check one type or the other wouldn't make much sense since the penalty for being downvoted on both is the same (for the poster, not always for the voter now).

1
  • I suspect this is a new thing - I had a bunch of older downvotes, and they all appear to have been removed today. I'd understand if the team wishes not to comment on the exact changes. Commented Jun 29, 2011 at 15:46
13

Looking at your stats. You have 21k rep and a total 4 downvotes on your two questions. Looking back to May 12th I see a total of 3 down votes and 600 positive rep due to other activities.

This problem seems very small and not worth making changes for additional revenge vote detection. I recommend against such changes.

4

I think this is valid. I'm in the same situation - 3 questions vs 345 answers. My questions get DVs occasionally when I vote to close (which I do a lot), or I post a comment to another high rep user who answers an easy question instead of voting to close (like this).

This exchange this morning got me a downvote on each of my questions pretty much immediately.

Edit:

Kon appears to have deleted his answer, but the gist was I told him I DV'd his answer since he was answering a softball question for easy rep instead of voting to close and reduce the noise.

1

I disagree with @jzd. He's picking on numbers and not addressing the real issue. I think that the Stack Overflow community is beautiful, and it's a shame that there are people that act this way. Members of the community, such as Michael Petrotta, are working hard helping other people with their problems/questions, so when people use their power to downvote as a revenge - it makes a community contributor leave with a "bad taste" and that's not the way it should be!

This is something that should be considered. I would implement one of the three:

  1. A fraud system like Michael suggested
  2. Exposing the names of the downvoters next to each question/answer
  3. Hiding the names of the members who flagged a question for closing.
3
  • 2
    #2 is not an option. Downvotes are completely anonymous by design. #1 is a great idea, #3 is possible.
    – Cody Gray
    Commented Feb 19, 2012 at 8:13
  • @CodyGray - one thing that I never say is "not possible" - as an engineer you're superman - there's nothing you can't do! ;)
    – Nir Alfasi
    Commented Feb 20, 2012 at 2:38
  • 2
    Yes, it wasn't really a claim of technical impossibility. Anyone who's been around Meta long enough has seen enough crazy stuff to know that would be an absurd claim indeed!
    – Cody Gray
    Commented Feb 21, 2012 at 18:28
-4

In Slashdot the way this is prevented is you cannot go to a user's profile view his posts and down mod them. Instead, if you want to mod comments, you're given a limited number of comments which are randomly drawn from all comments and are in threads you did not participate in.

A first step to a fix is limiting the number of down mods a person is given in a time period.

The second step is stopping users from down answers which compete with their own by preventing them from voting in threads they are a participant in. Perhaps this is already done here?

Yes, this has the draw back of not permitting the most knowledgeable people- those who know enough to answer the question in the first place - to also perform the useful function of evaluating other posts. But this is only a slight drawback however, for the following reasons.

  1. they have already contributed maximally by posting their answer to the question

  2. crowdsourcing the up/down vote is a great application of the knowledge of crowds and the single vote of the participant who is excluded from voting will not be really missed.

  3. If someone is moved by an issue so strongly they want to down mod other participants' comments, the system forces them into a choice–down mod your enemy or supply your own answer. This is a good thing since strongly motivated people are more likely to want to argue their side rather than just down mod, although they will do both if given the chance.

The third step is to remove the ability to down mod more than a few times the same user's answers across all questions within a certain time frame. Also, you cannot up mod the same user's answers across questions, although the number of permitted up mods should be greater than the number of permitted down mods.

This prevents people gaming the system either way, by acting out against an enemy or by creating alliances between users.

For example, you can down mod userX's answer to a question say ONCE a week. You can up mod userX's answers say three times a week. We don't want to discourage appreciation and positive ranking too strongly, only when the number of positive votes permitted could lead to a quid pro quo arrangement between users.

Of course the actual number of times and time frame both need to be experimented with to get them right.

So suppose you see userX's answers all the time because you innocently share the same interests. This is just exactly the type of situation in which gaming the system is most likely to take place; you may either come to like userX or come to loathe userX.

Either way, we don't want alliances and we don't want revenge taking to be easy options. If you are in a forum with the same userX over and over the best way to participate is contribute an answer. If you like userX's answers then you have plenty of up mods to offer, but not an unlimited amount such as could be used to bribe someone.

Similarly, you don't have enough down mods in your quiver to hurt someone you don't like.

8
  • 7
    We already implement some of these things. The rest don't seem like good ideas. Some people post consistently bad content. I'd like to retain my ability to vote that down. Votes are not against people here, they're against the content or usefulness of questions/answers. Yes, some people misunderstand that, but that's hardly a failing of the system, and redesigning the system to interfere with that goal would be counterproductive. This site is intentionally very different from Reddit/Slashdot/etc.
    – Cody Gray
    Commented Feb 19, 2012 at 8:12
  • Can you be more specific as to the implementation of features you're talking about? Of course votes are not supposed to be against people here, but is that a problem? If it's not then let's just say that- there's no problem here.
    – user179159
    Commented Feb 19, 2012 at 15:05
  • The idea that one person needs to continually down mod someone else b/c their posts are continually bad is fallacious in my view because it implies that crowd sourcing does not work, i.e. it implies that without your vote, the person would not get enough down mods. But the crowd will down mod them , that's the whole hypothesis behind sites like StackExchange and Slashdot.. it's the wisdom of crowds
    – user179159
    Commented Feb 19, 2012 at 15:10
  • 5
    Not really, no. Stack Exchange and Slashdot don't share any wisdom. Search Meta for more information about the tactics implemented here, it's too extensive a discussion for a comment. And this is precisely how "crowd sourcing" works (if you want to call it that)—members of the crowd vote down posts with incorrect or misleading information, sending it to the bottom of the list, and vote up posts with useful or correct information, sending it to the top of the list. Placing arbitrary limitations on that (like you can only vote for a user twice a day) gets in the way of the model.
    – Cody Gray
    Commented Feb 20, 2012 at 1:42
  • 2
    We have scripts that run to detect abuse, but beyond a pattern of abuse, I don't think vote counts should be limited.
    – Cody Gray
    Commented Feb 20, 2012 at 1:42
  • Cody Gray said: Placing arbitrary limitations on that (like you can only vote for a user twice a day) gets in the way of the model. – user179159 replies: This statement of yours seems to me to be in need of some supporting evidence outside of just opinion. I understand you think it's true, but of course that doesn't make it true. What we do know for certain is that without these proposed limitations, or something like them, then we're inadvertently facilitating the real problem the OP was complaining about.
    – user179159
    Commented Feb 21, 2012 at 15:37
  • You took the conclusion of a relatively long commment out of context, and then claimed I provided no support for that claim. No, not in the conclusion. You have to read the rest of the comment. Note particularly that A) it's difficult to provide as extensive a justification as I otherwise could in a comment, and B) this is all opinion-based, the best thing I can do is give you a well-reasoned opinion. Before getting angry at people on Meta who assume you know how the site works and its philosophy, I recommend spending a bit more time browsing the other discussions and answers.
    – Cody Gray
    Commented Feb 21, 2012 at 18:28
  • 1
    Ahh I was not aware of myself getting angry. I re-read my comment to see if perhaps I came across wrong inadvertently. I don't see that and neither do people I've asked to review my comment. The OP's problem is real in his instance, that is, I believe him and I don't see how it's being addressed.
    – user179159
    Commented May 29, 2012 at 11:37

You must log in to answer this question.

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