Vote Quota as a function of Reputation
I have no formal education on voting systems, so if someone can enlighten me with more info I'd appreciate it. But it occurs to me that being a member in good standing -- in any social context, not just social media -- gives more weight to your opinions and thus, your votes.
So if Piefed has a daily vote quota (political implications aside), is it possible for it to assign that quota on a per-user basis, as a function of user reputation? So the higher your reputation, the more daily votes you have to give (no limit on the number you can receive). And when I say function, I mean mathematical function, not necessarily 1:1, but maybe linear, logarithmic, etc.
And more than it being technically possible, is it even a good idea? What negative side effects exist in such a system? It seems to me that this would encourage new users to participate rather than lurk, in order to increase their reputation and afford them the ability to vote on more posts/comments. But maybe incentivizing this is a terrible idea, which I'd love to hear about.
14 replies
It's easily doable but probably a bad idea:
An incentive to karma-farm.
Receiving downvotes would feel worse and some people have a really hard time with them already.
I noticed an earlier thread where reaction emojis were implemented, which I think is a nice addition and a lot more granular than simply +1 or -1. BBS forums have been running on that model forever. Contextually, the emojis have different meanings too, say 🤔 for example. A well reasoned post could receive a lot of these from people who admire the thought put into it even if they don't agree. It could also mean someone has said something so wildly out of pocket that it could be used to expressed incredulous concern, without being outright hostile. Naturally 👍️ and 👎️ still exist as proxies for voting, not much you can do about that I think. Though it would be interesting to see the sums of each reaction to a user's comments/posts on their profile, ordered from greatest to least.
I feel like if you want to really combat karma farming, you have to make it totally irrelevant so that nobody would want to bother with it, possibly even remove it altogether. But then how else do you build up reputation? Tough question, I don't know a good answer.
There is a irl human phrase "if you want to make friends, you need to first be a friend." The same is true of online interactions, so I don't think we need to try to avoid karma bc it implicitly exists in everyday existence - it is even a good thing?
Why else would we go to "doctors" or "lawyers" or "police" rather than "some guy" down the street? Perhaps all we can do is to ensure that standards exist that truly distinguish those that deserve the accolades vs. people who merely pose long enough to earn them and then switch to betray the system when it becomes possible to do so.
Read up on game theory if you are interested in more stuff like this. Karma is a natural and inevitable consequence of the system, and even arises as an emergent property from nonliving computer programs with a mere few (hundred?) lines of code.
This was how Reddit worked, and it was exploited in bad ways. You would get people posting porn and other stuff, plus communities designed to pump Karma for this purpose. Then you had people pushing for high Karma and then selling their account to advertisers, who used the high Karma to place ads where everyone could see them.
On the flip side, you had people ganging up and Karma dumping people they didn't like. It was kinda chaotic if you were just a regular person who wanted to share opinions and get help with stuff. If you wanted to exploit the system, it wasn't that hard to do. You almost couldn't win by being honest or genuine; you'd say one thing out of the way and a group would target you and systemically keep you down by vote bombing you. So each comment could only count 100 votes either way; after that they stopped affecting you, but it was per comment, so they'd bomb all your comments via your history. They'd just go down the page. Maybe even had a script that would downvote everything, hit next, and repeat. You'd be deep in the red for helping people, because in another, unrelated conversation, you said you liked Ending 2 more than Ending 3 in Mass Effect 3 or something equally asinine. (I'm picking on that game because after a trilogy of choices that matter, you were literally given three doors with no explanation given as to which did what, and told to pick an ending, so it didn't matter which one you picked, you were picking at random unless you had a guide telling you which did what.)
I haven't used reddit in so long I don't think I've ever known about this system. Though the random porn-on-main posters now make more sense.
Here's a question: is there anything preventing Lemmy/Piefed/mbin from being subject to the same problems? Even without the technical incentive to karma farm, there's still a social incentive to be seen as trustworthy or influential. We just naturally crave that kind of thing even if it has no tangible benefit; number go up, number make happy. So what do our current systems do to combat karma farming, account selling, vote brigading, and other misbehavior?
Well, there's a lot of shitty behaviour that people brought with them from Reddit. I made a comment that wasn't completely in favour of generative AI and some guy banned me from like six AI subs. I'm neither for nor against in general, but I do have some feelings both ways — but that's not the point.
and nothing of value was lost
I find the idea of merging these two really unpopular features together quite amusing, ngl.
As for my opinion on it, one's "reputation" or "karma" should never limit one's ability to interact with any platform because such a limitation gives an incentive for karma-farming - thus decreasing the quality of content on such platform for every participant. Just leave "reputation" as a stat that only admins can see to help them with their moderation duties.
I feel I should point out there's quite a political dimension to this. I mean feasibility and karma-farming aside... There's several schools of thought on how society/power should work. This suggestion sounds to me like you're going for ideas from Timocracy or Meritocracy. Which is fine, we've been discussing these ideas since Ancient Greece... But I'm fairly sure the average audience here is more subscribed to Egalitarianism. It's an ideal, so we probably don't really have it implemented anywhere in the real world. But the basic doctrine is: all humans are equal in fundamental worth or moral status. As such, they should be accorded equal rights and treatment under the law. I think this transcends country governments... We're also shaping small online societies here. And part of our "law" is platform design and what kinds of algorithms we decide to implement. So the biggest question in my eyes is: do we strive for meritocracy? Or egalitarianism? With the next question, what downsides do we get? Because they all come with a lot of side-effects.
Except the vote quota goes against the spirit of Egalitarianism, and instead twists back around more towards Meritocracy. Consider some quick thought experiments:
The busiest people - like doctors or nurses - are afforded the least consideration by this type of quota. A monthly quota would be more equitable... at least for these scenarios, but we could surely keep going. e.g. someone with a terminal disease could attempt to work a year's worth of voting crammed into a single week while in a hospital - is it more equitable to deny such a scenario or allow it?
Or am I misunderstanding something here? Those who can "work" - or dare I say "game"? - the system, not necessarily intentionally btw as the above thought experiment hypothetical scenarios show, are afforded the highest merit in terms of being able to vote. In other words, the "merit" there lies in the regularity of voting, more so than the total amount. To state explicitly, someone who attempts the same amount of votes daily as someone else who spreads them out more evenly in weekly or monthly batches, receives higher privileges. As in, higher meaning "not equal", based on a talent performed by the user.
Also, by definition all Lemmy users have nearly infinitely more merit than users on PieFed instances that have not disabled this opt-out feature - as Lemmy users can vote all day long whereas the latter are restricted by the quota. (In practice though this is limited by physical constraints such as time and even amount of content that exists to be voted on).
The only thing that truly seems equal to me is no quota. All users then have the identical opportunity to vote. Once the restrictions come into play though, it sets up reactions to and against them whereby those opportunities are never distributed fully equally, given how different people experience life and their time here on the Threadiverse. i.e. it seems an attempt at least at "equity" but... without knowing the pre-existing imbalances, how is this "fair and balanced" approach justified in countering those (which are what again, exactly)?
And perhaps that's what you meant by the "audience here", as in Lemmy is Egalitarian in this regard. But I wanted to point out how radical the departure of PieFed is from that: now, any community located on a PieFed instance that implements this quota is imposing a merit-based criteria whereby those who play along most closely with the criteria are afforded significantly higher (in some cases) consideration than those who try to use nonstandard timing scenarios such as those briefly outlined above.
All of that said, everything under discussion above pales in comparison to the fact that users whose contributions have been shadow-censored are not told about this. Lemmy and PieFed both have a modlog for instance, but afaik this rejection does not get placed into it, unlike removal of a comment or post. No acknowledgement will be sent to the originating instance (at least atm?), no ability is provided iirc for users of other platforms (Lemmy, PieFed, Mbin, nodeBB, Friendica, Mastodon, etc.) to see in advance whether their contributions are likely to be rejected, afaik no acknowledgements are offered in the community sidebars like "this community is housed on a PieFed instance that implements a vote quota, thus not all your votes will be accepted if you have already made [vote quota amount - which btw is set at what value, and would this displayed value be kept current, i.e. what happens when the admins change it?], prior to the daily reset [which likewise is when? and in whose timezone?]". The result is that some (by default it would be all actually, as this is an opt-out feature rather than opt-in) PieFed instances are manipulating votes that no longer match the counts on the originating instances? Or at least that is my fear - tbh I am fuzzy on the actual implementation details. Or probably that's too far, bc the instance where the community is housed at is by definition the source of truth, but it is definitely a complicating factor, and not transparent at all atm.
Finally, this system is ridiculously easy to circumvent, only requiring additional alts that so many of us across the Threadiverse already use for other reasons (like modding Lemmy communities or testing how features work differently across different instances). It seems akin to asking people to write down a name - any name alongside their vote. Therefore... people will do that, using whatever name they want to, and go about their day. I have the feeling that you know better than I how well any system can be expected to work that has no actual ability behind it provided for enforcement. And even worse still, wouldn't the concept of perverse incentive seem to apply here, where people are now encouraged to spread their votes across multiple alts rather than use a single one? I've seen several people claim that they reached the daily quota in merely one hour. Conversely, so very many accounts may as well be inactive, e.g. have 3 comments spread over the course of 3 full years with zero posts or votes. So setting the vote quota according to comparison to an "average" or even standard "median" value seems of little utility to me.
And since despite all my words I still have yet to answer your question: I think I would probably prefer some sort of meritocracy, even of the type that OP describes. Though I realize that I am naive in this so I also don't trumpet that thought loudly. Then again, I also don't understand the push-back that many people tend to offer: what exactly is wrong with wanting to incentive behaviors that we consider good while not extending those privileges to others until they earn them? At least at some very low, extremely minimal level - not differentiating between 10k vs. 100 but differentiating between -10 vs. +10. e.g. when I made a PieFed.social account I had to wait a week or two prior to being able to send any DMs. But... isn't that a good restriction, to help protect against spammers? I was glad to see it in fact! I am not "entitled" to be able to do anything at all on someone else's machine, and even though I had someone I wanted to send a message to all that time (and did so the moment that I could), those restrictions are in place for a reason and do not seem overly onerous to me. (Or would this rather be called egalitarian since the same rule is applied to everyone equally?)
Yeah this is the kind of analysis I was looking for. Thanks for this insight.
And thank you for the thanks! 👍
Thanks for your detailed comment. I'm afraid I have to study the code first, before I can address most of your points. Like whether incoming votes are concerned or it doesn't affect Lemmy users.
Is there a better way to do it than a fixed point in time? I mean we have lots of algorithms and maths available to do things in more clever ways. It probably can't go through all the votes all the time, though. As most of the strain on a server like mine is due to vote federation and I wouldn't put more expensive database lookups there...
Idk. I'm probably annoying everyone because I keep reiterating my perspective underneath some post every other week... But I think the voting system here is broken anyway. I liked more traditional internet forums. I want to read someone's Home Assistant stories, hobby projects and blog posts. However on the Lemmyverse, that gets like 5 or 10 votes. While reposting the news or taking some meme picture from Reddit and reposting it here gets you 300. So if that's the "merit" with the general audience... I don't think I want it. (I mean I get it, politics and the outrage of the day are popular topics... It's just that I don't like it in the amount it creeps into my feed. And any popularity vote gives me more of it...)
I'm wholly on board with fostering positive behaviour/atmosphere. I guess the issue is always to figure out how to measure merit, or desired contributions. There's lots of bad ways to do it. And with how little our voting system reflects what constitutes discussions I myself like... I'm not sure if any amount of duct-tape on top is going to fix it. It's probably not making it considerably worse, either... Question remains, what can we do about it? I mean what "succesful", commercial platforms do is get rid of the downvote button, keep the upvote one but it doesn't really do what people think it does... But I'm a bit split on doing it that way.
Sure we can do anything we want. But what should be done? And on that note, are we absolutely certain that we want a vote quota at all?
I asked Rimu and he said it had nothing whatsoever to do with network strain. If it did though, then this would be an exceedingly poor "solution" to that? Like why not instead turn off voting altogether if you don't want that type of "information", or at least deliver content in batches - do you really need every vote the very second it comes out, or could a batch be sent every ten minutes, so six times per hour each than sixty? People using 3rd-party apps fat-finger swipe erroneous votes all the time anyway, so perhaps have time to upvote, downvote, cancel all votes, then upvote an item of content again, and only the final result needs to be sent rather than each item along the way.
Anyway, nobody wants to tell you what to do with your own instance, though there is a very big difference between making your own decision and having the vote quota feature be opt-out rather than opt-in. At worst, Lemmy instances might end up defederating your instant, along with all PieFed instances (or just those that enable this feature?) - bc while anyone can do whatever they want, so too can others in response. But even in that extreme hypothetical, you would still be able to receive all content from all other instances, due to the one-way nature of the defederation process.
And if you did want to restrict votes, why discard those of the most highly active people across the entire Threadiverse? Especially those who upvote the most highly? Why not do something, pretty much anything, else besides that? Like ban the lowest-reputation people who downvote but never upvote or post or comment, only ever downvote. I would start by asking the questions like: what type of information do you want to receive on your instance, vs. what type of noise would you rather do without? I would be extremely surprised if after thinking about it, you came to the conclusion that you wanted to ban all the happiest upvoters trying so hard to make people feel welcomed in this dry barren wasteland with barely any niche content that so desperately needs to be nourished if this Threadiverse experiment is going to survive. We seem to agree on that point then: this place needs MORE of such content, hence more upvotes. However, we now have a throttle that does the exact opposite - and even if you choose to opt-out on your particular instance, many communities have already been established on instances such as PieFed.social where this behavior is turned on. This will have a potentially dramatic cooling effect on other such communities migrating there, and I've already seen a rapid drop-off in people choosing to post to those communities, after PugJesus announced his departure, thereby bringing further attention to this issue.
If you find details by looking into the code, please do let me know - I do not want to spread misinformation and will correct what I said above if you find that any of it is wrong, plus moreover not spread it further.
And I would be against removing the downvote functionality. I easily upvote 1000-fold more than I downvote (I would check real quick on lemvotes, but PieFed.social has disabled that - oh wait, actually I was fairly accurate in that guess as you can see here), and I find a utility in downvoting poor-quality content independently of a mod outright removing it for violation of the rules. The fact that others misuse it is something that I would love to see addressed, but not by taking away a feature that does have a legitimate usage, when used properly.
Extremely unfortunately, this shadow-censoring of people's votes without disclosure is going to drive people away from using PieFed, and back into other platforms such as Lemmy, the new PyLova fork, or even Reddit. There is little to nothing that can be done about that at this point.