17

This Meta question: Give an incentive for finding duplicate questions deals with various ways of rewarding the finding of duplicates.

This concrete suggestion: Reward finding duplicate questions - +10, +2, -5 was declined as too extreme - from what I gathered, it is out of the question to the SE team to even think about subtracting reputation from "bad" closevoters. (A notion which I personally think may be misguided. But well.)

So, a more modest suggestion.

  • The first user to come up with a duplicate suggestion is rewarded with 5 reputation points if the question ends up getting closed with their suggestion.

  • The user gets to keep those reputation points in any case, even if the question is later reopened.

  • Subsequent users voting for the closing are not rewarded.

  • To prevent abuse, a duplicate suggestion must be voted on at least 4 times to make the closing final. If there are three different duplicate suggestions, one of them must garner at least four votes. Yes, that would increase the number of closevotes needed in some cases - but you can make the argument it should be that way under the current system, too.

  • To prevent abuse by people mass suggesting fraudulent duplicates, there would have to be limits - for example, voting rights could be restricted if you have a significant number of duplicate votes that no one agreed with, and a moderator flag raised eventually with the option of suspending users with egregious records.

31
  • 5
    Personally I think 5 is too much, it's as high as upvote. Would be happy with +2, same as a good edit. Mar 19, 2014 at 14:17
  • 12
    This still has the exact same problem as the duplicate of this proposal you linked. People go around proposing tons of questions as duplicates of posts that aren't really duplicates in the hope that some of them get confirmed, giving them rep. As much as I dislike duplicates, especially duplicates that aren't closed accordingly, closing questions as duplicates that aren't actually providing answers is far worse.
    – Servy
    Mar 19, 2014 at 14:17
  • 1
    @Servy yes, there would have to be checks and balances for that. It's misuse that could be addressed like extremely bad reviewers should be addressed.
    – Pekka
    Mar 19, 2014 at 14:21
  • 2
    The only way to balance this would be for bad duplicate suggestions to give negative rep but that would require a way to say that a suggested duplicate was actively bad rather than just not the best one Mar 19, 2014 at 14:21
  • 5
    Tell that to the suggested edit system. It hasn't been able to really deal with that problem throughout its entire history.
    – Servy
    Mar 19, 2014 at 14:21
  • @Servy that's because that system is broken - with no good way to report bad reviewers to mods, and no clear guidelines on whether mods can/should do anything to stop bad reviewers. (I added a paragraph on how to deal with that.)
    – Pekka
    Mar 19, 2014 at 14:22
  • 2
    @Stijn Not really. If the rep gain is just 2 or 5, answering duplicates is still likely to generate more rep unless the culture of the site actually changes such that those answers just don't get upvotes, unlike now. This proposal wouldn't curb that much at all, because it just gives so much less rep than answering.
    – Servy
    Mar 19, 2014 at 14:23
  • 3
    @Pëkka Well, we've had years to try to fix it. If we've been unsuccessful there, what makes you think we'd be successful here? If we can't get the tools put in place of that system, what makes you think we'll get the needed tools here? I'd much rather see suggested edits fixed, if that's feasible, than see a bunch of dev time spent on this.
    – Servy
    Mar 19, 2014 at 14:24
  • 1
    @Pëkka Duplicates have been somewhat of an issue for a while, and while, yes, it would be nice to have something to deal with them, I most certainly would not say that they need an urgent remedy. After all, you've just posted a duplicate of a 4 year old question (oh the irony) that also feels it's an urgent issue, yet we've survived for the past 4 years. Saying that a proposal is of a lower priority than other issues is a reason to not implement it. Development resources are finite, there are tons of good features that aren't "good enough" to warrant actually getting developed.
    – Servy
    Mar 19, 2014 at 14:30
  • 3
    @Pëkka Your point that answering easy dupes is true enough, and this feature request wouldn't address it at all. People would still be incentivized to answer dups even with this feature, because they are so much more likely to gain more by answering. It's not enough of an incentive to close, and of course the greater the close incentive, the greater the abuse of closing. It's a lose/lose.
    – Servy
    Mar 19, 2014 at 14:32
  • 7
    I personally don't feel closing as duplicates is ever worthy of reputation. A badge for it, maybe. But reputation, hell no.
    – animuson StaffMod
    Mar 19, 2014 at 14:33
  • 10
    @animuson I couldn't disagree more. Finding a good duplicate is hard work. Sometimes harder than answering. Why should Googling a cheap answer to an endlessly duplicated question net you more rep than finding a good existing original? If we continue on the current course, SO's question base will grow into the dozens of millions - but then the place will be a support forum, not the collection of canonical Q&A with consistent great quality that was originally envisioned. Do we really want that?
    – Pekka
    Mar 19, 2014 at 14:34
  • 7
    But by awarding reputation, you're telling users to trust them more. They may not have a flipping clue what the question is actually about, but because they matched a few keywords in a search and said "this looks like the answer" they should get reputation and the trust that comes with it? That sounds incredibly broken.
    – animuson StaffMod
    Mar 19, 2014 at 14:38
  • 3
    @Pëkka But it does mean that the fix is worse that the problem itself. Nobody's saying that there isn't an issue here, just that your solution causes more problems than doing nothing.
    – Servy
    Mar 19, 2014 at 14:41
  • 5
    I'm with animuson and Servy on this, I see the potential for abuse being too high to offset the benefits. Too often do I see people in the close vote queue blindly going along with the first duplicate close vote, without bothering to check if it's correct. Looking through the moderator flag queue, there are 30 questions currently flagged for being incorrectly closed as duplicates of something they did not duplicate. We already have real and growing problems with vote fraud, and introducing reputation points into the closing mechanism will I believe cause even more issues. Mar 19, 2014 at 15:44

2 Answers 2

2

Please no!

Please lets not clog up the re-open queues by incentivising closure

I see way too many false positive dupes for this to feel a "safe" thing to do. The current mechanism for closing as dupe requiring only 5 or 3 votes mean that too few people sometimes have the ability to close something as dupe. Not to mention the "gold badge hammer".

I've seen too many cases of dupe flagging where the title might look similar but the body asks a wholly different question.

I frequently see people erroneously flagging dupe just because an answer might be the same, despite the question being wholly different 1.

I've even seen a case of a gold badge dupe closure happening where the "duplicate" was for a different programming language.

In short dupes are too easy right now, and require too little agreement for them to be a good way to record rep. People are too lax when checking other's dupe suggestion. We shouldn't incentivise this behaviour.

how is this different to other rep

Rep is achieved through other users giving it. A user's action does not directly earn themselves rep, it's third party recognition of that contribution that earns the rep.

On the other hand the suggestion in this question is a sort of "self-service" rep: the user's own action directly earns the rep.

This means the rep is not earned through quality but quantity. Offering rep for high quantity with no link to quality is simply going to drive down quality. It will drive users to make poor choices on flagging dupes just because they get something for it.

Under what circumstances might this work?

Rep demonstrates other's agreement, or applause of your contribution.

If (and only if) the dupe mechanism were wholly re-engineered so that users could flag a dupe in the form of an answer, and others could then up-vote "this is a dupe" answer, then (and only then) I'd feel comfortable with it having a rep pay-off.

But adding an additional incentive that can be scored too easily through sloppy reading? No!


1 I'm NOT talking about the etch-a-sketch FAQ. I'm saying the question "what's 17 + 25" is not a duplicate of "what's the answer to the ultimate question of life the universe an everything"

1
-1

This part sells it for me:

To prevent abuse, a duplicate suggestion must be voted on at least 4 times to make the closing final. If there are three different duplicate suggestions, one of them must garner at least four votes. Yes, that would increase the number of closevotes needed in some cases - but you can make the argument it should be that way under the current system, too.

Like the other people objecting, I find the bigger problem is when we close things as duplicates when they aren't. In my opinion there should be no "gold badge hammer". I have seen gold badge hammer applications that were clearly incorrect, too.

Rewarding dupe finds while simultaneously raising the bar on dupe closes may be the answer we need.

One of the great powers of Stack Exchange is its built-up history. We don't want someone to search for a question and have to sort through 10 different posts, all identical. Finding those duplicates does take some effort, though, so some small reward seems appropriate. I could quibble over 5 points vs some other value but broadly I like the idea.

It's kind of a conundrum for me personally because I often feel VTC can go off the rails and dissuade people from using the site for some really frivolous, arbitrary, inconsistent reasons, but at the same time, there is true value in finding real duplicates. Requiring 5 votes all with the same link to help ensure it actually is a duplicate seems like a good idea anyway.

You must log in to answer this question.

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