What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

Probably everybody who is very active on SO has experienced cases of random, serial upvoting as the expression of thanks from a (usually newbie) user. The pattern is usually 5 or more upvotes on your latest answers.

The upvotes will be detected within 24 hours, and removed. (Or so I assume.) No harm done except maybe a minor irritation for people after one of the "rep cap" badges because they are shown a higher reputation than they actually have.

Still, I was thinking whether, in an attempt to channel the positive attitude into more productive areas, it would be worth showing a tooltip on the third upvote for the same person within time frame x:

Please use one upvote, a comment, or some quality content of your own, to say thank you

Large numbers of upvotes for the same user may become invalid as part of our spam detection mechanisms.

Does this make sense? Maybe it is too localized and rare, but it has happened to me three or four times in the last two weeks alone. You decide.

Clarification: I am not advocating putting up any kind of barrier on upvotes. The barrier is already there. I'm suggesting informing people of the barrier and the futility of their ten or so upvotes for a single person because those votes are likely to be revoked.

Update: @devinb makes the good point that "saying thanks" is not the only possible reason for serial upvoting, and the message wouldn't make sense then. The point still stands: Your votes aren't going to count, and when it comes to up voting (which is always some sort of positive behaviour, however misguided if done indiscriminately) I think it's fair to inform people of this.

share|improve this question
4  
(-1) for the reasons here –  devinb May 12 '10 at 9:43
    
Although, I am very much in favour of being civil and polite, or "nicely discouraging" of negative behaviour. I like that :D –  devinb May 12 '10 at 9:44
    
"Probably everybody who is active on SO"? Now exactly what are you saying about my activity level, hmmmm? –  Pops May 12 '10 at 14:21

6 Answers 6

up vote 21 down vote accepted

This is basically training people to a) avoid upvoting or b) avoid detection by the vote fraud algorithms.

What good could come of it?

share|improve this answer
    
Hmm. Re a), true; if you do not want to make serial upvoters aware of the fact that certain patterns are detected, then fair enough. I was just thinking of making them aware because serial upvotes are rarely fraudulent and more often a misguided expression of gratitude. But as I said, fair enough. Re b), though, it stands to reason that whoever wants to seriously commit vote fraud has read your blog entry on the issue and knows exactly what to watch out for. –  Pëkka May 12 '10 at 10:30
3  
Roughly 10% of the questions on Meta are about the score. –  dbasnett May 12 '10 at 13:48
4  
I as a new user on one of the sister SE sites (Gaming), was helped by a user with a very exhaustive answer. I upvoted his answer. I browsed his user profile and saw a bunch of related questions(Skyrim related) that I had/or would have had. They were the accepted answers. So I voted all the helpful ones. But the system reversed all of my votes. Today same thing happened with a different user on ME3 tag on Gaming. Even here on SO, browsing Jon Skeet's profile or Eric Lippert's profile I do the same. Should I have not cast my vote if I thought it was helpful? –  One-One Mar 13 '12 at 16:20
    
I had no idea about the serial vote reversal. I searched all over Gaming meta, searched all over SO meta. What is serial vote reversal? At this point I know if I vote someone a lot in a short span of time, my votes will not count. BTW this has nothing to do with the ME3 missions going on over at Gaming. BION. –  One-One Mar 13 '12 at 17:24

The barrier against serial upvoting is destructive. I just hit it yesterday.

Example: http://stackoverflow.com/users/1539315/mattstep?tab=reputation. He's got a pretty low reputation, probably because he's only been around a year.

But he's the guy who writes the code that these answers address. Yes, he's the #1 world expert. So I went through and upvoted a bunch of his answers, because, well, just maybe he's the world's most qualified person to actually answer them.

But nooooo, those upvotes might be fraud. Sigh.

share|improve this answer
5  
You should never ever, ever vote according to the user who made the contribution. Vote according to content. Did you actually read and understand all of the user's posts or was it just blindly upvoting? –  Lix Jul 22 '13 at 17:43
    
You're saying that when someone who wrote the code says something about that code, it's not useful? That's wrong. And of course you can go though 10-20 posts and upvote them after reading them. It's not even hard to do that in a reasonable way. We're talking a paragraph or two, not an academic paper. –  James Moore Jul 22 '13 at 18:36
1  
That is exactly what I'm saying. What happens if the author of a certain piece of code makes a mistake? I mean, I know they're programmers, but sometimes, we make mistakes. This is what the voting system is all about. You can't vote on content without verifying what the content actually contains. It sounds to me like what you're really trying to do is show your appreciation to the programmer. While that's totally cool, upvoting the user is not the way to go about it. If you happen to meet the user in a chat room - go right ahead and shower him with compliments :), but blind votes don't help. –  Lix Jul 22 '13 at 19:47
    
It's conceivable that the "blind" algorithms who monitor the voting system (and trigger reversals) will make mistakes too. They would see a suspicious voting pattern when in fact the voter has read the contents of each post and it genuinely contains quality content. –  Lix Jul 22 '13 at 19:50
    
Take a look at this post I answered a little while ago: How to know if I am a serial up voter? –  Lix Jul 22 '13 at 19:51
    
No, I think we disagree on what an upvote is. I 100% think that anything written by the code's author that seems reasonable gets an upvote. I'd be happy to hear that there was a bot dedicated to doing that. It's not happy-face for the author, it's a "what this person has to say is almost by definition important". –  James Moore Jul 22 '13 at 20:01

Any "you shouldn't do that too much" will end up being a barrier to a good behaviour.

In this case, if you put limits on a max of upvotes, it will be associated with the idea that upvoting is not good, and limit upvoting in the end.

I think that the impact of "serial upvoting" is relatively limited, compared to the risk of making people upvote less.

share|improve this answer
2  
it's not me who is suggesting putting barriers on anything. The barrier is already in place in the form of silent deletion of serial upvotes. I'd like people to become aware of the barrier. If it were up to me, there would be no such barrier in the first place (or one that is manually sorted out by moderators). –  Pëkka May 12 '10 at 9:55
    
@Questions - in this case, if it's only about notifying about something which already exists, I guess it's more or less ok. However, even if it exists already, the fact of pointing it with a warning might act as a "barrier", for users. when it's not visible, you don't have this effect. –  Gnoupi May 12 '10 at 10:02

I never experienced this issue myself, but I can imagine it happening. Your proposal is reasonable, well worded and easy to understand even for new users, so I suggest implementing it.

I would make a small adjustment to the message: "You have upvoted the same person many times [...]"

  • Avoid to display the actual number required
  • Make clear that it is the same person whose posts were upvoted
share|improve this answer
1  
Good point about avoiding actual numbers. –  Pëkka May 12 '10 at 10:03

Isn't it possible that someone started reading through a user's answers and thought - that's a good answer, hey that one's really interesting, gosh, I didn't know that, and voted accordingly? This seems like perfectly reasonable and correct (for SO) behaviour to me.

share|improve this answer
    
The question is not about the behavior, it is about warning users about the behavior. –  mafutrct May 12 '10 at 9:49
3  
@Neil yes, absolutely, but no matter what your motivation, your votes are going to be removed by the system anyway, notification or not. The notification only makes them aware of the futility of their votes. –  Pëkka May 12 '10 at 9:53
    
@Questions I'm not aware that "serial upvotes" are actually removed in the same way that downvotes are. –  nb69307 May 12 '10 at 9:56
    
    
@Questions The blog says "suspicious", not "serial". If I answer a couple of questions in quick succession and they get upvoted by the same person, I'd be a bit ticked off if those votes were removed, and AFAIK, they won't be. –  nb69307 May 12 '10 at 10:03
    
@Neil I had a "serial upvote" this morning: 7 upvotes in 30 second steps. I'll keep an eye on them and see what happens. (They clearly qualify as "serial upvoting" to me, though, different from the scenario you describe.) –  Pëkka May 12 '10 at 10:06
    
I've been a serial upvoting victim and the votes were removed the next day (although the vote history still shows up incorrectly in my Recent Activity - i.e. the bad votes were removed, but the remaining good votes that day still think that they were made after hitting the rep cap). –  Ether May 12 '10 at 14:56
2  
As far as I know, serial upvotes are not removed simply because they came in rapid succession -- it's only if a significant percentage of the user's votes were used on the same person that the detector is triggered. Then again, no one other than the devs (and possibly the moderators) knows for sure. –  mmyers May 12 '10 at 15:16
1  
@Neil @mmyers just as an update, the seven serial upvotes I talked about (May 12 00:30 - 00:59) are still in place, so they don't seem to be removed as strictly as serial downvotes. –  Pëkka May 14 '10 at 21:35

This one is a difficult problem, because it isn't necessarily an expression of thanks. If you look at Marc Gravell's profile page, and look at ANY of his answers, chances are you are going to want to upvote it. Which means, if I land on his profile page and then only click the question titles I think are interesting, I'm going to end up "serially upvoting" him.

For that reason, the message you describe is ridiculous. We are assuming intention and purpose behind a 'pattern' we have discovered. Another valid situation is that if Jon Skeet answers the first three C# questions in the "recent questions" list, and I look at all of them, chances are, I'm going to want to up vote his answers, because they'll be solid. And then a message pops up that says I shouldn't be upvoting anymore, even though I may not realize that I've upvoted the same person three times.

Further, if the last part of the message (the functional part) is only displayed whenever you have crossed the threshold, then you have indicated to any spammers what the threshold is.

I feel that any way this is implemented would lead to many mis-fires, confusion, and ultimately very little benefit.

share|improve this answer
    
@devinb good point about how it's not necessarily an indiscriminate gesture of thanks. However, don't forget that if you upvote more than two contributions by Marc right now, all your votes are going to be automatically removed within 24 hours, so your expression of respect for their answer is futile. I'd like them to become aware of the futility and express their respect/thanks/whatever otherwise. –  Pëkka May 12 '10 at 9:50
    
@Questions Well, we don't want to in any way discourage upvoting. We also don't want to encourage strategic voting. The people who are unaware of this rule can just continue voting as they did before. Every now and again someone might lose two or three upvotes because of the rule. BUT, if you tell them that their votes are being lost you are saying to everyone that they shouldn't "just vote" but rather they should be monitoring who they vote and when. I say " to everyone " because the fact is that the more involved users are already aware of this, but it is because they chose to... –  devinb May 12 '10 at 10:05
    
investigate the inner workings of voting and such. In programming terms, your API should expose functionality, and hide implementation details. This is a similar situation. The people who want to know the implementation of the voting fraud mechanisms have many ways to discover it, but we should let the rest just continue on their way happily as before. –  devinb May 12 '10 at 10:07
    
@devinb valid points. Still, I would find it fair to convey the information that if you do a lot of votes for a single user in a short time span, the votes may be lost - doesn't have to be in a tooltip but somehow. But maybe it's an issue so rare it isn't really relevant - it could be. –  Pëkka May 12 '10 at 10:07
1  
@devinb: I think you are criticizing the serial upvote detection (validly), not Pekka's proposal, which is still sound. Whatever the actual "invalid" behaviour is, the system should warn the user when he is approaching it. –  Ether May 12 '10 at 14:58
2  
I actually have no problem with serial upvote detection. But I believe that it is something which should run under the covers. We want the barrier to entry to the site to be very low, and every pop-up is an additional piece of information which they need to keep in their heads. Essentially, the thought of a user is "If I didn't need to know this, then why did they tell me?". Because we don't want them to change their behaviour, we should not be popping up a message that implies that we want them to change. –  devinb May 12 '10 at 15:43

You must log in to answer this question.

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