Possible Duplicate:
Should the weight of downvotes be increased?

I think the current voting system has a problem as it causes users to act dishonestly and complicates their decision making process. I will try to explain this in more detail below.

I get annoyed that when I cast a down-vote on a post that I think is bad someone comes and up-votes the post just to make the total zero again, where as if I didn't down-vote the post would remain with zero votes. This means that in these situations between the option of down-voting and not down-voting I have an incentive not to down-vote. This means that I have to guess if someone else is going to counter-vote my down-vote which complicates my decision making process.

The effect of the current design of voting weights is that if someone thinks that a post really has a problem but suspects that another person might disagree he or she has an incentive to act dishonestly and not cast a down-vote:

  • without the down-vote: the post has 0 total votes, the OP receives 0 reputation.

  • if the down-voter acts honestly and down-votes (question case) and someone else counter-votes: the question has 0 total votes, the OP receives +3 reputation.

  • if the down-voter acts honestly and down-votes (answer case) and someone else counter-votes: the answer has 0 total votes, the OP receives +8 reputation, the down-voter receives -1 reputation.

The reputations are small so it might not cause an major noticeable problem. (However I suspect that a few users have got a considerable part of their reputation out of such counter-votes which is annoying.)

It is a bad design of voting mechanism, it causes users to act dishonestly and complicates a user's decision making process (because users have to guess what others think and if there is any other user who would counter-vote). A good voting mechanism should be transparent and encourage users to cast votes honestly based on their own evaluation of the question independent of other user's opinion and votes.

Increasing the weight of a down-vote to -5 or -10 would not completely solve the problem as down-votes still have a negative utility on casters.

It might be a good idea to discuss this with someone working in Algorithmic Game Theory / Mechanism Design.

  • "t is a bad design of mechanism for voting, it causes users to act dishonestly and complicates a user's decision making process (because users have to think if there is any other user who would counter-vote)." [citation needed] – Cat Plus Plus Jan 22 '13 at 1:24
  • 4
    Also, why would your downvotes be more important than my upvotes? Maybe I think you're wrong about the downvoting. – Cat Plus Plus Jan 22 '13 at 1:25
  • 1
    @CatPlusPlus, no one says anyone's vote is more important than other people's votes. Regarding your first comment, What do you exactly want a reference for? I can try to give one if you are more specific. – Kaveh Jan 22 '13 at 1:29
  • 7
    If the other user has said that they upvoted only to counter your downvote, then this is wrong and I hate it too (though because it's democratic voting, we also have no right to tell them stop!). Otherwise, maybe you were just wrong about the post. – Lightness Races in Orbit Jan 22 '13 at 1:29
  • 4
    @LightnessRacesinOrbit, I don't mind other people disagreeing with my vote. What I think is a problem is that my vote causes the reverse of what my intention, i.e. my down-vote causes the OP to receive reputation. I will edit the post to make it clear the main point I am trying to make. – Kaveh Jan 22 '13 at 1:31
  • 1
    @Kaveh: How do you know that the other voter wouldn't have upvoted the post even without your downvote? – David Robinson Jan 22 '13 at 1:32
  • 1
    @Kaveh: I agree in part. I disagree that the OP gaining rep is a problem; I don't care about their rep. I care that the question/answer then has a score that's not representative of what people have decided in accumulation that its quality is. I also can't find a suggestion in your question to fix this that I can support; sorry :( – Lightness Races in Orbit Jan 22 '13 at 1:33
  • 1
    @David, what I have done is the following: I have waited a reasonable time before down-voting a post I think is problematic to make sure someone is not up-voting the post for itself, then have down-voted and seen that the question got up-voted in a short period of time afterwards. I haven't done complete calculations to see the probability of such events however my feeling is that this happens way more often than I think would be coincidence. – Kaveh Jan 22 '13 at 1:35
  • 7
    My upvote here was no pity upvote ;-) Indeed, nowadays I hardly ever downvote zero score posts before first leaving a comment without downvoting (urging the author to change something), as I also feel (though I might be wrong) that too often my downvotes were just undone for no good reason—even when I did comment. – Arjan Jan 22 '13 at 1:37
  • 3
    not a problem – Doorknob Jan 22 '13 at 1:41
  • 1
    Unless, @Doorknob, many people don't downvote to circumvent it, as stated in the question (and is my personal experience). – Arjan Jan 22 '13 at 1:42
  • 1
    @Arjan I... really haven't done that/haven't heard of anyone not downvoting a bad post just because of this. – Doorknob Jan 22 '13 at 1:43
  • 3
    @Doorknob: Yes you have. For example, Arjan just a few moments ago. – Lightness Races in Orbit Jan 22 '13 at 1:43
  • 1
    Thinking of it, my upvote was a pity upvote. I don't know (yet?) if I'd like changing the weight of receiving a downvote. And I guess I'd never have upvoted the above if its score was not -6 at the time. – Arjan Jan 22 '13 at 1:56
  • 3
    Other proposals in (the answers of) Towards Preventing “Pity Up Votes” – Arjan Jan 22 '13 at 2:18

The problem here is your motivation for downvoting. When you downvote you're thinking:

I want this person to lose reputation as punishment for this bad post

But I'm thinking:

I want other people to know this isn't right

On my scoreboard, up and downvotes are worth exactly the same. Now you can still argue that the pity upvotes are wrongly bringing the post back up to zero when it belongs at -1. But did you know that some people think that once there's one downvote, everyone piles on and adds more even if they're undeserved? I think that the number of people do either of those things is pretty small and that they roughly balance each other.

So I don't even see a problem. The current system works for me, because I don't care what rep other people have. For those who feel differently, let's compare these two ways to get rep:

  • ask a crappy question, get a downvote (-2) receive a pity upvote (+5) for a net of +3. Keep that rep for a while but risk losing it if the question is deleted
  • suggest a crappy edit, get it accepted, 2 points, can't lose 'em

Guess what the smart kids are doing these days?

  • +1 You have a point. But I still think this is a bad thing that users get reputation this way. In an ideal SE we wouldn't have reputations I guess, but in the real SE we have. It sends a bad message when a user asks a bad question and gets positive reputation out of it. You may not care about such cases, you have a different utility function which doesn't give you incentive to act dishonestly, but for those of us that care about these cases and don't like causing an OP of bad question getting a positive reputation because of our down-vote this is not the case. – Kaveh Jan 22 '13 at 2:25
  • OK, but then I'm saying "please join me in stopping the horrible edits that get approved" - your motivation can be the 2 points the bad editors get. It's a much bigger problem than cancelled downvotes or never-issued downvotes. – Kate Gregory Jan 22 '13 at 2:39
  • I am for that also. :) – Kaveh Jan 22 '13 at 2:41
  • "I want other people to know this isn't right" — True for me too. But that's also invalidated by pity upvotes. (If any. But +1 for "some people think that once there's one downvote, everyone piles on and adds more even if they're undeserved?") – Arjan Jan 22 '13 at 9:00

From what I can tell, the factor you're raising here that differs from existing questions is that you believe the problem is widespread, predictable and a practical problem (as opposed to simply a major bloody annoyance for people like me who just can't help but care deeply for whatever reason).

If so, I believe the canonical answer, which follows a statistical analysis by @JeffAtwood, is:

Based on this data, "sympathy" or "pity" upvoting is not an actual problem -- if it was, I would expect to see much higher rates of an upvote following a downvote.

That doesn't take into account the claims of people such as @Arjan:

nowadays I hardly ever downvote zero score posts before first leaving a comment without downvoting (urging the author to change something), as I also feel (though I might be wrong) that too often my downvotes were just undone for no good reason—even when I did comment.

That's because stats can't tell you how people would have voted in an alternate timeline. I'm not sure I can believe that this behaviour is vastly widespread, though.

  • As for Jeff's statistics: what if many people, like me, often don't downvote zero scored posts as they think it might cause pity upvotes? That cannot be measured the way Jeff tried, I guess. (But I am no statistician.) – Arjan Jan 22 '13 at 1:43
  • @Arjan: Added response to that factor. – Lightness Races in Orbit Jan 22 '13 at 1:44
  • Thank you. That is interesting. However I am not saying this is a problem with major effect (as I wrote in the question). I think the problem it causes is it complicates the decision making for casting a down-vote and more generally is what is called "untruthful mechanism" in algorithm game theory and mechanism design literature. The annoyance comes partly out of this complicated decision making process, you get annoyed IMHO because the result is different than what you expected when you cast your down-votes. – Kaveh Jan 22 '13 at 1:44
  • 1
    @Kaveh: Yes I agree it's wrong and bad and shouldn't happen. But not only do you not propose a solution that can work, but you also don't demonstrate that it's a practical problem (something I've oft tried, and failed at just as hard). Until someone succeeds there, I can't see that there's anything to be done here. – Lightness Races in Orbit Jan 22 '13 at 1:45
  • 1
    @Arjan: Are you really arguing that while sympathy upvoting isn't a problem, people irrationally responding to (nonexistent) sympathy upvoting is a problem? And therefore, that we should take measures to discourage sympathy upvoting? – David Robinson Jan 22 '13 at 1:46
  • Well, @David, I feel boldly stating "sympathy upvoting isn't a problem" is too easy a way to try prove yourself right, while in my comment I actually do explain why I doubt those statistics. – Arjan Jan 22 '13 at 1:50
  • @LightnessRacesinOrbit, I admit it doesn't seem to cause a major problem over the network, however I think the weights for votes and more generally the voting mechanism can be designed in a better way. – Kaveh Jan 22 '13 at 1:50
  • @Kaveh: I'm sure that that is true. – Lightness Races in Orbit Jan 22 '13 at 1:51
  • 1
    The Arjan's behavior is the kind of dishonest behavior that a bad mechanism design creates for users: if they vote based on their own opinion then the result can be reverse of what they intended to do. They have to think about other people's opinion about the question to decide what they should do: cast a down-vote or not. A good voting mechanism would not require a user to think about these. It is probably the case that it is not causing a major problem, however I believe this is a practicable problem that can be fixed. – Kaveh Jan 22 '13 at 1:58
  • I'm sure "the Arjan" agrees too – Lightness Races in Orbit Jan 22 '13 at 2:00
  • 1
    @Arjan: You're missing my point. Even if we grant your premise that this behavior exists, and that that's the reason no evidence of sympathy upvoting exists, than the problem would not be sympathy upvoting- it would be a reaction to sympathy upvoting. – David Robinson Jan 22 '13 at 2:02
  • 2
    @DavidRobinson: I think we can accept in pretty good faith that problems caused by the problem are inherently included in this discussion. – Lightness Races in Orbit Jan 22 '13 at 2:08
  • @LightnessRacesinOrbit: Those aren't problems caused by the "problem" if there is no problem to begin with. Those are the only problem. And in that case, the best answer is just for Arjan not to do that. – David Robinson Jan 22 '13 at 2:16

The root cause of the issue is that the voting system is horribly skewed in favour of upvotes. A much simpler solution would be for upvotes on answers to cost reputation.

Let's put our money where our mouth is...

  • Peer Pressure is missing from my collection... – yannis Jan 22 '13 at 1:45
  • I'll do you one better: an upvote should cost the caster the amount which is bestowed on the recipient. Especially on new beta sites. In seriousness, though, this isn't as completely cracked as it seems at first blush. Ideally, voters should be thinking about upvotes just as carefully as they do about downvotes. Maybe they should even cost more than downvotes do. – jscs Jan 22 '13 at 4:14
  • 3
    an advice from fellow downvoter: to get desired downvotes, you'd have to either cut the answer to something like "voting system is horribly skewed" or suggest something sufficiently trollish for answers upvote cost, like +2 or +3 rep – gnat Jan 22 '13 at 7:18
  • 1
    No thank you. Why should I pay for the privilege of upvoting? – user102937 Jan 24 '13 at 18:59
  • 2
    @RobertHarvey Why should I pay for the privilege of downvoting? – yannis Jan 24 '13 at 22:44
  • Why should downvotes only cost the victim 2 rep? Clearly, upvoting and downvoting are not mirror images of each other. – user102937 Jan 24 '13 at 22:45
  • @yannis, that's badge 36 not the correct badge 38 id – new QOpenGLWidget Feb 2 '19 at 20:21

This was much much "worse" when upvotes on questions were worth +10.

You're missing the fact that a "game" has to give out points for participation. When someone counter-balances your downvote with an upvote, they're saying "it's bad, but it's not that bad." While I agree truly bad questions should stay negative, in this case someone disagreed with you that it was a bad enough question to stay negative, so he +1'd it. So the system says "hey, tahnks for playing +3" to the OP, and you get -1 for being such a spoiled sport ;)

  • :) but then why don't we just give the points right away when someone asks a question but only after someone down-votes and someone counters it? Don't those who post decent by no vote questions deserve reputation for their participation? This gives users incentive to post bad (but not too bad) questions in place of posting OK questions. – Kaveh Jan 22 '13 at 2:07
  • Good point, there is no reward for throwing a question out there that has a completely neutral response. The only solution I can see is increase downvote weight – bobobobo Jan 22 '13 at 2:37
  • Still much worse on StackApps! – new QOpenGLWidget Feb 2 '19 at 20:23

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