Who decides what you see on the fediverse? A look at voting patterns
-
@rimu Vote?
This is about Threadiverse - part of Fediverse that is link aggregating, not microblog
Like https://piefed.zip for example
-
Out of the 37,000 people who voted for posts or comments in the last month, the 10 most prolific voters (0.02% of us) cast as many votes as the bottom 59%. Here’s how that looks, visually:
As you can see, a lot of people didn’t cast many votes. Someone cast 23k votes, with a group of 13 each casting at least 10k votes.
“But of course most people aren’t really engaged, most of those 37k people are just NPCs who don’t really matter” you say, “Rimu you’re just including them to make it seem worse than it is”, you might say. Ok, cool, let’s pretend the bottom 85% of us don’t matter and just look at the top 5000 voters. Here’s how the distribution looks among them:
Still super unbalanced. Let’s analyze this a bit.
Among those 5000, the top 147 (2.94%) cast as many votes as all the others (4853 people) combined. Among those 5000, the average number of votes cast in a month is 1142. Among the top 147, the average number of votes cast in a month is 6868.
How do you feel about a tiny group having this much influence over what news you receive?
I am not concerned that some people vote a lot.
I sort posts and comments by New. Given this, I expect that how others vote doesn’t much affect what news I receive - maybe not at all, though I haven’t looked at the algorithms to confirm.
I sometimes look at the votes. It is mildly interesting to see what attracts a lot of votes, either up or down. But I am more interested in the posts and comments than the votes.
I sometimes vote: up votes for what seems sincere, interesting, informative, evidence based or clearly expressing an opinion; down votes for the more extremely contrary, but mostly up votes.
It is important to see a diversity of content. Anything that is popular enough to be influential is interesting, even if it is utterly false or nonsense, merely because it is influential. Opinions and perspectives different from my own are helpful if well expressed. It is even good to see how extreme people can be, sometimes: how hurt, upset, deceitful, ignorant, unreasonable, deluded, disrespectful, abusive or manipulative they can be. It makes it difficult to decide what is worth my time, let alone someone else’s, and it often depends on my mood and how much time I have for a bit of a wander and a ponder. So, overall, it is far too complex for a simple net up/down vote to be very helpful.
So, I don’t think voting reliably distinguishes what is worth more of my time and attention from what isn’t. I suppose it is disrespectful of me to think that how many people vote is not very useful information. I sometimes wish I could filter or weight the votes according to how much I respect the opinions of the voters, but most voters are so anonymous to me that I couldn’t judge the value of their opinions even if there was some mechanism to rank or weight them.
I think what I see is more dependent on what people post than how others vote on it.
-
That would just cause those 10 people to hate me. I don’t want to anger the people who determine what gets seen and what gets buried. That demonstrates the whole problem with having 10 people controlling everything - it’s a power imbalance.
Anyway you would not recognize most of them. The people who vote a ton are not the same people who comment and post a lot.
I don’t want to anger the people who determine what gets seen and what gets buried. That demonstrates the whole problem with having 10 people controlling everything - it’s a power imbalance.
I’m a week late to this discussion but, Piefed doesn’t have algorithmic sorting like Reddit, so does it matter? After reading most comments in this thread, it seems like most people (myself included) sort their main feeds by Subscribed/New. Occasionally I will also sort by Popular/New or Popular/Hot just to see what’s ticking in communities I don’t subscribe to.
Respectfully, I don’t think this is a great idea, Rimu. The 240 number seems arbitrary and personally, I do not feel as if I am being manipulated, or influenced on this platform by someone who happens to upvote a bunch.
-
Out of the 37,000 people who voted for posts or comments in the last month, the 10 most prolific voters (0.02% of us) cast as many votes as the bottom 59%. Here’s how that looks, visually:
As you can see, a lot of people didn’t cast many votes. Someone cast 23k votes, with a group of 13 each casting at least 10k votes.
“But of course most people aren’t really engaged, most of those 37k people are just NPCs who don’t really matter” you say, “Rimu you’re just including them to make it seem worse than it is”, you might say. Ok, cool, let’s pretend the bottom 85% of us don’t matter and just look at the top 5000 voters. Here’s how the distribution looks among them:
Still super unbalanced. Let’s analyze this a bit.
Among those 5000, the top 147 (2.94%) cast as many votes as all the others (4853 people) combined. Among those 5000, the average number of votes cast in a month is 1142. Among the top 147, the average number of votes cast in a month is 6868.
How do you feel about a tiny group having this much influence over what news you receive?
It’s been a couple weeks and I wanted to add another perspective about this.
I noticed a user downvoting posts, https://lemvotes.org/user/FarStare@lemmy.dbzer0.com if it helps to have specifics. In the more than one full year of existing, they have never once contributed via making a post, never once commenting even, just doling out up- and down-votes in roughly equal proportion, and even those nearly exclusively via posts but almost never comments.
I am not concerned about receiving votes from PugJesus - either up or down, as PugJesus has earned the right to speak on behalf of the community. But who is this other person who is influencing how others see my content? Are they male or female? Do they have a degree of some kind? Do they primarily visit meme communities, as the vote pattern might suggest, or other communities as well?
So how do I feel about people who use the Threadiverse more than others? My answer is more complex and nuanced than simply yay or nay. I might prefer the opposite of the implemented feature even - the more someone posts (including comments?), the more upvotes they are allowed to make? Have they not earned such?
Also, if votes were exposed, each individual could decide for themselves how they wanted to make use of voting - perhaps someone would ban PugJesus, or conversely write special code (a 3rd party app / API recipient?) to highlight posts that PugJesus upvoted (isn’t that how “boosting” works? at least re-tweeting does).
It’s really difficult to guess what someone wants, as each person wants different things. Which is why ultimately in an effort to make things more fair & equitable, the solution can end up making it less so.
I like the follow-up suggestion to leave this new option OFF by default, so that an instance admin would have to specifically opt-in, perhaps as part of an overall suite of features limiting things like posts and comments as well. Although really, I think other approaches might work better - e.g. why would someone want to limit themselves to “some” of PugJesus’ content, wouldn’t they rather simply have it all or else ban him from the instance and therefore have none? I am not seeing the benefit realistically of having “some”, especially when the delimiter is not based on community or overall content but simply what time of day (based on an arbitrary threshold for when they starts and ends to restart the next period) the votes are cast?
