29

In this question I asked for a practical reason to lock votes after 5 minutes and Bill the Lizard and Arjan were very nice to explain the mystery.

Well, vote locking seems to deal with strategic downvotes fine, yet vote locking isn't a useful feature in itself. If it was the code, I would call it 'ugly hack', in a sense that it deals with a problem by creating another problem. It deals with few 'reputation optimizers' by annoying the whole community.

There's probably no need to illustrate my previous point: I've provided an example in the linked post, and there're many more complaints about it in the meta. Sure, in 99% cases you have no urge to change your vote, yet the remaining 1% is very annoying.

So, how can we adjust vote locking to reduce collateral damage? I see two options right off the bat

  1. Simple tweak. Lock downvotes only for the users who have answers in this question.

  2. Radical. Allow no downvotes in the question where user has posted answer himself. As Kyle Cronin puts it, This has some drawbacks, but can you really trust someone who posted their own reply to be objective about judging the other peoples' replies?.

Any ideas?

12
  • 6
    Disallowing downvotes might make people vote before they post? (The simple tweak seems nice though!)
    – Arjan
    Jan 19, 2011 at 15:28
  • 2
    @Arjan Of course, there're some technical details to be worked out :) ("Godzilla has eaten 3 of your downvotes because you've posted an answer here.") Yet, the idea seems viable too. Jan 19, 2011 at 15:34
  • 4
    @arjan A simple workaround for that problem would be canceling down-votes when you post an answer. Jan 19, 2011 at 15:34
  • 5
    #2 is in my opinion obviously bad; #1 seems fine since it's just a weaker version of what we have now Jan 19, 2011 at 15:35
  • 23
    Considering downvotes on answers can often indicate that an answer is wrong, I'm inclined to think that such a downvote would be more trustworthy if the voter happened to know that it was wrong by, say, providing the actual correct answer.
    – Grace Note StaffMod
    Jan 19, 2011 at 15:35
  • An advantage of the second is that it at the same time prevents poor sportsmen from simply down-voting competing answers without canceling the down-vote. Jan 19, 2011 at 15:36
  • Note that strategic downvoting might also apply to questions. (At least, I noticed an odd downvote revocation once, and figured that downvoting a question might make it less likely to be answered by others? Maybe that is not an issue though.)
    – Arjan
    Jan 19, 2011 at 15:38
  • 1
    @Grace People don't usually downvote to indicate merely wrong answer (well, most of them). They comment and sometimes post correct answer. The amount of downvotes made by average user is disproportionately low to the number of wrong answers on SO :) Jan 19, 2011 at 15:59
  • 1
    This is possibly a duplicate of meta.stackexchange.com/q/22771/130885
    – endolith
    May 18, 2012 at 0:36
  • 1
    Damn. I'd previously upvoted this question, but I changed my mind and no longer agree with it. So frustrating that I can't change my vote!
    – Ben Lee
    Jan 17, 2013 at 22:14
  • @GraceNote ..."if the voter happened to know that it was wrong by, say, providing the actual correct answer" or leave a comment telling what he belive is wrong so somebody can check...
    – SPArcheon
    Jan 22, 2013 at 13:37
  • 2
    @GraceNote: It takes a lot more effort to write a correct answer than to identify that an existing answer is incorrect. Doing both is great but, sometimes, I only have time for the former. There's still value in that action. Feb 14, 2013 at 13:52

8 Answers 8

21

Simple tweak. Lock downvotes only for the users who have answers in this question.

For the sake of voting, I put both original suggestions into separate answers (CW, so no rep for me).

3
  • (This might need an enhancement of the puppet detection mechanisms, but surely the bright folks at SE can figure that out.)
    – Arjan
    Jan 19, 2011 at 15:47
  • How in the world does locking votes help anything? People keep repeating this as if it is a solution, but it's not. You can still downvote everyone else's answers and push your own to the top of the heap. The only thing it prevents is undoing the damage afterward.
    – endolith
    May 18, 2012 at 0:35
  • And what if you want both to give a good answer AND point out how the other answers are wrong? If you answer a question you're probably the most qualified to point out mistakes, so you should certainly be allowed to downvote.
    – Calimo
    Jun 17, 2013 at 13:57
13

What if you eliminated vote locking, but charged a reputation penalty for removing a downvote past the lock period? So your -1 reputation would still be locked in, but you could still reverse the downvote later.

10

I initially upvoted an answer, only to realize a few minutes later that it actually didn't answer the OP's question.

So I answered it myself, and wanted NOT TO DOWNVOTE, but to remove my initial upvote.

Because of this provision, I can't.

If this feature intends to prevent tactical downvotes, then please let me retract my upvote.

6

One can still do strategic downvoting.

One loses one vote per answer, so if there are another 5 answers, he will have lost five points by downvoting them. Yet, if one gets just one extra upvote from this maneuver, he still won.

As far as reputation is concerned, no-one can see who voted for whom, so there is no threat of revenge.

Edit- Looks like someone figured out a way to see who downvoted him.

On the other hand

This site has a "vote-cheating mechanism" to prevent "sock pupetting", "revenge downvoting", "upvoting all of someone's answers as thanks for answering his question" and other such sundry activities. One can just add to this list "downvoting all answers where he has an answer". After enough such votes, they (the downvotes) will be undone (like sockpuppet points).

5

Show canceled votes in some moderator tool, and have them take disciplinary action if a user exhibits a strange pattern. (i.e. cancels votes on competing answers very often)

Perhaps displaying the number of down-votes, canceled down-votes, down-votes on competing answers and canceled down-votes on competing answers makes it easy to see such behavior. And as a bonus show users who simply down-vote competing answers.

9
  • IIRC, there's just no concept of 'disciplinary action' against a user on SO. Nobody has been expelled from the site yet. Jan 19, 2011 at 15:36
  • @Nikita Users can be contacted by email of moderators to suggest behavior changes, and ultimately suspended for a period of time if they do not reform. We don't expel, no, but there's no upper limit to how long a suspension can be.
    – Grace Note StaffMod
    Jan 19, 2011 at 15:37
  • Errr, surely people are being expelled, @Nikita!
    – Arjan
    Jan 19, 2011 at 15:41
  • Related (and downvoted): Allow moderators to view (serial) up/down-voting?
    – Arjan
    Jan 19, 2011 at 15:43
  • @Arjan That's a user merging, if you fully read Jeff's answer. That doesn't seem like expulsion at all.
    – Grace Note StaffMod
    Jan 19, 2011 at 15:46
  • @Arjan There're some blog posts dealing with the question of voting fraud and Jeff states that nobody will be banned or penalized. I didn't know about suspension, though. Jan 19, 2011 at 15:56
  • I finally get it, @Nikita (and Grace): removing sock puppets does not expel the original user, indeed.
    – Arjan
    Jan 19, 2011 at 15:58
  • Since exploiters do it to gain reputation, a reputation penalty just like with posts flagged as offensive could be one example. Jan 19, 2011 at 16:02
  • Actually, CodeInChaos, most exploiters lose most of their reputation because the merging process will invalidate all of the cross voting.
    – Grace Note StaffMod
    Jan 19, 2011 at 16:04
1

During the initial answering period after a question has been asked, the current number of votes on each answer should be hidden, and the answers should be presented in random order. This has two benefits:

  1. Tactical downvoting will no longer exist, since downvoting everyone else will have no visible effect.
  2. Currently, there's a bit of bandwagon effect where people pile on votes to the answers at the top and ignore the answers at the bottom, even if the answers at the bottom are better. With the answers in random order and no vote counts shown, each one will get (on average) equal visibility and equal consideration from voters, and they'll have to actually read them and judge them on their own merits before voting, making the vote count more objective.
2
  • 2
    The "initial answering period after a question has been asked" lasts virtually indefinitely. When would we start showing vote counts again?
    – Cody Gray
    May 18, 2012 at 0:50
  • @CodyGray: Depends on how much attention the question gets, how many users the site has, how many upvotes the answers receive, etc.
    – endolith
    Oct 11, 2013 at 21:02
1

Simple tweak. Lock downvotes only for the users who have answers in this question.

As long as the lock is removed on an edit. To prevent strategic downvoters from editing to fix, you disallow them (them=people who have downvoted and answered on the same post) from editing. Point is, all downvotes should be reversible on edits.

Though I still see problems here, I've often downvoted+edited posts, disallowing edits on a smaller site can be dertimental.

Aside from this, it may be pretty easy to create a sock and use it to strategic downvote. We may need to improve the vote tools to prevent that.

Radical. Allow no downvotes in the question where user has posted answer himself.

Bad. BAD. VERY BAD. Especially for smaller sites. On smaller sites, the number of users who fall under can downvote&&is online is extremely small. So answers can go by unnoticed for a while; every downvote counts. Again, I've had plenty of cases where I've downvoted "rival" answers on Chemistry, IMO legitimately (following the same vote metric as other posts).

Alternative

Why not just add "profusion of strategic downvote-like patterns" to the mod tools? This may accomplish the job in a non-blocking way.

1
  • Good alternative suggestion.
    – AndrewC
    Oct 25, 2012 at 8:17
-5

Radical solution. Allow no downvotes in the question where user has posted answer himself. As Kyle Cronin puts it, This has some drawbacks, but can you really trust someone who posted their own reply to be objective about judging the other peoples' replies?.

edit
The original discussion has some good points for and against this approach.

For the sake of voting, I put both original suggestions into separate answers (CW, so no rep for me).

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