Skip to main content
Meta Stack Exchange

You are not logged in. Your edit will be placed in a queue until it is peer reviewed.

We welcome edits that make the post easier to understand and more valuable for readers. Because community members review edits, please try to make the post substantially better than how you found it, for example, by fixing grammar or adding additional resources and hyperlinks.

8
  • 1
    You make an interesting comment that gamification is eventually braking down any voting system. However, I believe that gamification is not the worst reason for the system to break down. It is the growth of the site, with more and more questions and answers that makes that the presence of the newer posts is very much diluted. This is an effect independent from gamification. Therefore I believe that your suggestion may work in making sure that questions and answers receive equal attention. Commented Nov 19, 2019 at 20:45
  • @SextusEmpiricus Yes, these are certainly very valid considerations. It might be generally true that the larger a field grows, the worse the sociological issues become, so yes, the growth is a huge factor. (One may also think that the simplest solution could be to remove the incentive of gaming the system. But I am afraid that this will also decrease the incentive to participate quite a bit.) Commented Nov 19, 2019 at 21:16
  • 1
    Counterexample: the Advogato trust metric is a quality indicator that is provably ungamable except through social engineering. It's fairly different from SE rep because it only gives a true/false answer instead of a quantitative one. Commented Nov 19, 2019 at 21:36
  • @rockwalrus-stopharmingMonica I do not know anything about that metrics but as you say it does not seem to allow one to "rank" posts relative to each other. Wikipedia also mentions that "Despite the trust metric, posting privileges to the front page of Advogato have been gained by controversial individuals, leading some to claim Advogato's trust metric solution is faulty.", but as I said I don't know anything about this metric, let alone have the ability verify or falsify this statement. Commented Nov 19, 2019 at 21:41
  • 1
    @Schrödinger'scat The proof divides the community into three parts: good users, bad users, and confused users. Good users only trust good and confused users, while the other two groups can trust anyone. (The assumption behind the names is that bad users abuse however they want to while confused users just have misjudged the character of some of the bad users, but the proof doesn't assume that behavior.) The proof shows that the number of bad users that end up trusted can never be more than the number of confused users times a tunable constant. Commented Nov 19, 2019 at 21:52
  • @rockwalrus-stopharmingMonica As I said, I have no idea about this metric. But I fail to see how any metric can be immune against a large number of users "conspiring" by deciding to vote favorably on each others posts regardless of the content. Commented Nov 19, 2019 at 23:34
  • @Schrödinger'scat The algorithm starts by trusting a few hard-coded users, and then trust flows through them at a limited rate to the users they trust, and through those users at a limited rate, etc., with each user consuming one unit of trust until there is no more trust to flow. It gets a little more complicated to deal with users who trust each other, but those details aren't important. Users that aren't connected to the trust source never get any trust, while the number of users that can receive trust through a confused user is limited by the amount of trust that is allowed to flow through Commented Nov 20, 2019 at 1:21
  • 2
    @rockwalrus-stopharmingMonica Replace "few hard-coded users" by "few competent and fair referees" and we get rather close. ;-) That is, if one could make sure that there is a group of neutral and competent referees, a lot of problems could be solved. If it were just not for the if ... ;-) Commented Nov 20, 2019 at 1:24