Possible Duplicate:
Should we reduce rep bonus for upvotes on posts with a negative score?

The issue of "Pity-Upvotes" has received a bit of attention, most notibly here: Is there an actual "pity" or "sympathy" upvote problem?. It seems that I encounter this on a pretty regular basis... truly terrible questions (one sentence, "do my homework", etc) or answers (obviously just a comment, wrong answers, etc) which receive three or four down-votes and suddenly an up-vote appears which I can only presume is a pity-vote ("It wasn't soooo terrible...maybe I'll just boost them back to -3").

I don't necessarily have an issue with the thought behind this, and sometimes merely mediocre posting ends up a -6 or more, but because of the skewed weight of up and down-votes the targeted post will often end up in the black despite being a pretty poor effort.

A proposal for remedying this issue: weight the reputation value of up-votes as only +2 (on par with down-vote) when the aggregate score is less than -2. This preserves the value of up-votes when there is a legitimate difference of opinion on the quality of the question/answer (a good posting will sometimes have a less than zero score, but almost never more than -2) and gives users the ability to vote "this question/answer isn't as terrible as the score would suggest" without rewarding lousy posts. Does this seem like a reasonable proposition?

  • 7
    Bad posts will get voted down just fine, I don't see a wide-spread 'problem' here. Aug 3, 2012 at 22:45
  • 2
    Curious how many pity upvotes this question gets... Aug 3, 2012 at 22:47
  • 1
    @MartijnPieters I'd be curious about seeing the numbers here. Anecdotally, I feel like I see this very frequently. Jeff Atwood published some stats that suggested it was an issue, but not a big one. But he only looked at the pattern of -1,-1,+1... My experience is that it starts to emerge only after -4 or -5. Specifically, I think it would be useful to see what percent of questions or answers has more than four or five downvotes and then receives an upvote or two.
    – Ben D
    Aug 3, 2012 at 22:52
  • 2
    @BenD - I have upvoted one or two horrible questions because I didn't think a brand new 1-rep user needed to have -10 on their first question ("Welcome to StackOverflow, we hate you and your question" - is what that essentially screams to them). Sure, it's pity, but -9 will still get the point across that it is a bad question. I have never seen nor would upvote an extremely negative post up to (or pass) 0. So I don't think it throws off the entire system..
    – Ason
    Aug 3, 2012 at 23:06
  • 2
    @Ason - I agree on the warm welcome for newbies... in fact, that's what brought this question up for me. I often want to help undo the terrible reception that new SO users receive, but I don't necessarily want to give them 5/10 rep. However, it's a very interesting point - pity votes are often given to new users who just don't know how to ask questions yet... it may be worth erring on the side of giving too many points to them to keep them interested in participating...
    – Ben D
    Aug 3, 2012 at 23:14
  • 2
    @BenD - Like you said, newbies are often chased away by the downvote band-wagon, and I feel as though +5 rep won't crush the working of SO's rep. However, I can agree with you that pity-voting around the 0 marker isn't beneficial and would likely confuse those seeking an answer. I cannot think of an elegant solution to that though...
    – Ason
    Aug 3, 2012 at 23:39
  • Notice no pity upvotes here =P
    – casperOne
    Aug 4, 2012 at 14:23
  • 2
    @casperOne I see three upvotes, all of them may be pity upvotes. Aug 4, 2012 at 14:37
  • Here, I'll cast a pity downvote for the downvoters who were countered by the pity upvoters :)
    – Benjol
    Aug 7, 2012 at 9:12
  • This question is NOT a duplicate of meta.stackexchange.com/questions/1390/… - That question suggests a universal parity of up and down votes. This question is more nuanced: it suggests a parity ONLY WHEN THE VOTE TOTAL IS NEGATIVE. Many folks in the comments, answers, and close-votes seem to have missed this fact.
    – Ben D
    Apr 2, 2013 at 5:51

3 Answers 3


A proposal for remedying this issue: weight the reputation value of up-votes as only +2 (on par with down-vote) when the aggregate score is less than -2.


There is a point to having upvotes give more rep than downvotes. For example, consider a bad question that was downvoted to -4. The OP comes by and corrects it a few hours later; now it's a decent question.

Why should my upvote for this now decent question give less rep? Just because 4 other people judged a worse version of the question?

Personally, the kind of "sympathy" upvotes you're talking about aren't hazardous to the site. I'd say the problem pity upvotes come from marginal questions/answers that aren't bad enough to need a downvoting, yet somehow attract a single upvote because it's a 1-rep new user and they just want to give them some rep.

  • The point about modified questions is well-taken, though I'm not convinced that this is often the case but it's a strong argument against the proposal.
    – Ben D
    Aug 3, 2012 at 22:57
  • 3
    No. The problem of voters not returning to change their votes on posts after they are edited is not solved by the way votes are weighted. Rep is only one of many things affected by the post score. Other things, such as the fact that posts are removed from the front page of the site after its score is less than -4 or the automatic ban for consistently low scoring question/answers are not affected by this at all. To solve your problem, something else entirely should be done.
    – Yi Jiang
    Aug 4, 2012 at 0:47

Take this with a grain of salt, as recently I'm been feeling particularly disillusioned with Stack Overflow, but I've come to believe the problem is much more fundamental than relatively benign "pity votes".

There are a huge number of low-rep users with voting privileges that are posting and upvoting bad content. I'm sure it's not malicious, they simply don't know the difference between good content and bad content. They don't understand that they're posting crap, and when they receive three crap answers to their crap question they continue to post in a similar vein without improving. They're content to upvote pretty much everything, even garbage.

I see a similar problem with the edit queue. Users just accept everything, further devaluing rep. I'm continually shocked when I encounter users with 3k or 4k+ or more rep, users who are approving the most trivial broken edits and presumably upvoting similarly bad content. Maybe (probably) it's me and the tags I frequent (PHP, JavaScript, CSS...) but I feel like half the content flowing through these tags should be closed as too localized, NARC or duplicates, but almost none of it is.

The up/downvote weighting throws things further out of whack. The work of one bad upvoter requires five good voters to counter, or the people posting bad content continue to gain rep for it. There has to be five times the number of sensible, good users who can recognize good content for every single user who upvotes bad content or rep continues to get awarded and devalued for posting bad content. This is the real problem with "pity upvotes"; it only takes one to make posting bad content worth-while.

Worst of all, I think the number of bad users who upvote bad content now dwarfs the number of good users and I really can't see how the site can combat this - there just aren't enough people downvoting bad content to compensate.

That got really ranty, so... yeah. Repicide.

TLDR: Pity votes aren't the problem, bad voting in general is the problem.

  • See stackoverflow.com/questions/11804556/… and its two upvotes for an example of the posts I'm railing against. This question is crap. Awful English, it's hideous to read, it's too localized in its current form, and it would require editing the entire thing to make it worth while as a work of reference to future users. Yet, it's been upvoted twice, presumably by people who don't understand that Stack Overflow is supposed to be a work of reference. How to solve this?
    – user229044
    Aug 3, 2012 at 23:50
  • 2
    I agree with you very much in general, but that specific question comes out on the okay side for me. It's stating a clear problem; maybe it's a bit chatty in explaining the various scenarios, but still. It's fairly localized, true, but if every OP put this much effort in their question...
    – Pekka
    Aug 3, 2012 at 23:52
  • I guess it's not the worst example I could have chosen, it was just the top question on the front page when I checked. Its particularly bad English rubs me the wrong way, but I try to overlook that as it's a particularly fixable problem via the edit button.
    – user229044
    Aug 3, 2012 at 23:56
  • 1
    Yeah. But in general I agree. I ignore most incoming questions these days, there is so much repetitive crap I won't even bother to downvote or closevote any more.
    – Pekka
    Aug 4, 2012 at 0:09
  • @Pekka - I think that the repetitive crap is a way bigger issue... it does seem like users will often post on SO before even googling, much less actually searching SO for similar examples. This further degrades the site because correct answers to stupid questions dominate a lot of users' rep. I find it very frustrating that so many of my highest-rep answers are to questions that shouldn't have been asked in the first place, but they are obviously correct so they get upvoted. Anyway, off topic, but I agree that it's a much bigger problem.
    – Ben D
    Aug 4, 2012 at 0:39
  • 2
    I think my problem basically boils down to Stack Overflow being more or less a crowd-source debugging system these days. People post their blobs of awful code, with no idea of how to walk through and debug it themselves, and ask "what is wrong?" The bulk of questions I see are "Here is some code, it is not working, how do I fix it" and that is the definition of too localized. Yet those questions get upvotes and answers and people are encouraged to keep coming back and debugging their software through Stack Overflow instead of learning to do it themselves.
    – user229044
    Aug 4, 2012 at 2:59
  • @meagar Yeah. case in point from a minute ago... I fear it may be an unfixable problem, but we will see
    – Pekka
    Aug 4, 2012 at 8:24
  • @meagar: Stack Overflow is a big place, so what you see depends on where you look. The PHP tag may be overrun with such questions, but I find the tags I frequent have plenty of reasonable questions. Aug 4, 2012 at 9:00
  • @Nicol yeah, you have a point, but it's a general phenomenon IMO. Those tags that are (rightly or not) the hottest names in the industry - for ease of use, perceived career prospects, or whatever, will tend to be flooded with bad questions.
    – Pekka
    Aug 4, 2012 at 9:11

Giving the upvote and the downvote the same weight in reputation effect is a bad idea.

The stack exchange sites are post oriented, not user oriented. Thats one of the reasons why downvotes are equal weight for the posts, but also why the effect for users is less for downvotes. We encourage most contributions. But some have a slight effect on your reputation.

  • 1
    I understand this, but I'm not calling for equal weighting. I'm calling for equal weighting only when there is a near-consensus that the question/answer was bad. My solution would, I think, actually make for a more post-oriented outcome, as it would encourage people to upvote questions which they legitimately feel are being under-rated without fear that they'll reward mediocrity.
    – Ben D
    Aug 3, 2012 at 22:55

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