Debate thread: About downvoting (modnote:no, not being removed again)

For downvoting.  
Against the mods still planning on removing it, just through subtle changes over time to avoid backlash.
 
“Atrophy the monster one limb at a time. Don’t try and fight it head-on immediatel.”

@byte[]  
That can’t be done now, not after five years under the expectation they’re private.

@The Smiling Pony  
But doesn’t the same go even more strongly for removing them? It’s a smaller deal to make them public than it is to remove them. So I’m not arguing the point, but if you accept that point, then it applies even more strongly to downvotes as a whole…

@Cirrus Light  
2 claps  
👍

@Cirrus Light  
Expectation of a feature existing is severely different from the expectation of privacy.

@Cirrus Light  
I believe the difference is: One is removing a function that’s been abused heavily. And the other [making who downvotes visible] Open up users to being attacked for what they disliked.
 
There’s already users who ask to have their Favorites cleared because they don’t want to be associated with certain content. If downvotes were visible, people who enjoyed the content/artists would know exactly who to dislike themselves.

@The Smiling Pony  
@Zeb  
But if downvotes are public, then users still have the choice to downvote or not. They can maintain their anonyminity by not downvoting… Which is kinda part of the point. The hope that this would deter them from mass downvoting or downvoting things that don’t deserve it.
 
And if you want anonyminity then make an account to do your downvoting on. Though I guess that conflicts with the “x downvotes per time” idea. Sorta. Mostly. Depends.

The idea of public downvotes seems like a great way to give people fodder to make auto-block lists, ala the SJW shared twitter blacklists.
 
A great way to divide the community further.

@Cirrus Light  
Mmmm…It’s sort of like the demonstrably flawed logic that non-anonymous uses won’t troll or be as obnoxious as users tied to known username. For some, it might. Others may not be discouraged in the least, and others might even troll more. I can see something similar happening with public downvotes. It honestly would get pretty ugly and probably cause new, different headaches instead of reducing them.

@Nonymous  
@ArrJaySketch
 
Maybe, maybe not. We’re working with a lot of uncertainties. The things you’ve mentioned aren’t the sort of thing that’s being incentivized, just stuff that might happen, and probably will just through sheer volume of users (ie, if 1 a thousand do something, and you got 10 thousand…)  
 
I still think it’s a less drastic measure than removing downvotes altogether, though.
 
I guess at some level it makes me think of choosing a meeting time when I first started a university brony club. We got a meeting time most people could make - but kept changing it to try to meet everyone’s needs. In the end, the changing time just meant less people knew when it was and made it. So we just took the time the most people could make, and nailed it in, and it’s been at that time for like 5 years now. So any solution chosen is going to cause some issues or displease some people.
 
So it’s really just about pleasing the most people or causing the least issues - or really some balance of the two.
 
The thing most people have been in favor of is keeping downvotes and hiding, it seems to me, anecdotally, though until someone does some survey anecdote is all we’ve got - and even then, there could be a strong selection bias. It’s clear some threads had more downvote supporters than others. Perhaps sort the 100 most active users and survey them? I dunno.
 
I would also like to see neural networks and Wilson scoring given a shot, though, but I’m also a believer in technology and science (after all, this site wouldn’t exist without them) :P

@Irony  

 
Yet another problem with downvotes; despite the name they’re not “voting.” They’re leaving an incoherent comment. Or a “rating” at best.
 
@Cirrus Light  

 
@byte[]  

 
But how do you maintain the anonymity of the people who already downvoted something before it was public? Do we erase all the existing downvotes and start fresh? Create a New Public Downvote and retire the ability to add the olde tyme anonymous ones? (And who codes the ability to handle two different kinds of downvotes into the site?) It just doesn’t work. (And I say this as someone who is perfectly happy to have my own downvote history go public, personally. If I didn’t like something, I explained why in the comments, so my negative opinions have never been anonymous anyway.)
 
If downvotes were turned public, it would be useful only if it would allow us to also downvote users. If someone downvotes all sorts of things they should be filtering, then everyone else can downvote the person. And if the person gets enough downvotes, they’re banned from upvoting/downvoting for a short time. Then a long time. Then forever. Not a site ban; they can still post and comment. But since we know their downvoting behavior is worse than worthless, actually damaging to the goal of getting meaningful image scores, they just can’t do that. We’d be left with a rating system run by people who care to do it right, the scores of images would be critically meaningful opinions by people who like that specific sort of image, rating it against its actual peers, and everything would work better.

@Ferrotter  
Creating two different kinds of votes is easy. I could code something with that with just my rudimentary first-year programming.
 
And a mob-sourced system sounds… like a bad idea, to put it politely. The issue here is people getting downvoted unfairly, but now you want to get the unfair downvoting mob more power to effect the people they downvote?

Downvoting a person would be even worse than anonymously downvoting art.
 
Imagine the scripts that could be made.  
Someone hurt your feefees? Add them to our blacklist and watch them get hundreds of downvotes automatically from everyone else subscribed to our Meanie List.

@Nonymous  
I’ve actually seen that in action.
 
Was a reputation indicator where you could +1 or -1 someone on their account. Popular user has a run in with someone watch their friends dogpile jump the opposition.

@Nonymous  
You meme a lot “feefees”, but you know damn well any feature and “list” like that would be used by 4chan et al when they get triggered by whatever “liberal degeneracy” sets them off just as much. I’d much appreciate it if instead acting like a channer funnyman or whatever you at least tried to converse in a reasonable way about a topic that’s honestly been a large headache to deal with.

@byte[]  

[underline added] Okay, these kinds of questions don’t just go away…
 
So what are you calling the “parameters”? You’ve got μ_p̂ defined in terms of p̂, q̂, n and z.
 
Z is something that depends on the distribution of the data set, right? So are we assuming a standard bell-curve distribution?
 
And n is something we pick when we take samples - if we were, saying, conducting a study or survey. But in the case of the site, n varies from image to image.
 
So, given the metadata of an image, we can get p̂, q̂ and n. Z we do not know but perhaps are assuming a standard bell-curve distribution, or drawing the data from elsewhere? It’d have to be elsewhere, because I think if we just used a bell-curve distribution then it’d just tell us how far we are from a 50/50 score. So I’m not sure where you’re getting the z-values from. Perhaps either paramaterizing the bell curve from other images or drawing a curve from other data or what…
 
But even with those, we’re just getting μ_p̂ = [some number] ± [another number], which is, two possible solutions for “as we take an infinite number of samples, the percentage that fall in this range will equal this”, right?
 
Probably way off the mark, here, but I’m guessing because you can’t get right if you don’t know where you’re wrong, so you can show me where I’m wrong to correct my line of reasoning/understanding.
 
Btw, cute avatar.

@The Smiling Pony  
I’m sorry you don’t approve of how I act, but I can tell you straight up I’ve never seen people on 4chan gang up on others the way I described. Demonize them all you want but people that have actually been there know how things are. “We’re not your personal army.”

@Nonymous  
Ah, I see; the threads directing people to that Applejack image, or to the various black humanization images a dozen times before that, including the goading of creating and uploading edits of them, weren’t attempts to gang up on others. Right, my bad, sorry.

@Nonymous  
Well, people say that… I got a friend who’s a regular there. He’s mentioned it’s really just a matter of learning to put things the right way, or do things the right way to get them to be your personal army, and while I’m don’t really get the art of mob-rousing, I see how it is one and some of its basics.
 
And even while one person says “we’re not your army”, others will go join the lynch mob. I’ve seen it once when I got involved in a thing - the whole name-war over “Pacific Glow”. I think most people on 4chan actually preferred “Pacific Glow” over “Molly,” or posted near a 50/50 split on their site, but here on Derpi a ton of BGP’s with no other site activity came out of nowhere to support “Molly”, and I don’t think many, if any, came to support “Pacific Glow”. Because “lel drug reference”. Then they got really mad and salty when they didn’t get the pony tag named after Ecstasy/Crystal MDMA.
 
Looks like a similar thing happened with people downvoting the AJ trans image, and I don’t doubt there are lots of other examples of this kind of thing.

How about showing (in each user’s profile) which tags most often appear in their downvoted images? A user who downvotes 1000 images tagged Flash Sentry images would have “flash sentry” on that list. To avoid the list from getting too long/noisy (these kinds of automated statistics would collect unrelated/collateral tags), it could only display the top 10-20 tags. It does not directly reveal which images they’ve downvoted, but it shows what kinds of images they usually do downvote.
 
You just need an associative array for holding tags and their counts. Tally up all the tags in a user’s downvoted images and then use your favorite “find the N largest values in an array” algorithm to generate the list.
 
That list could be then used to… erm… “suggest” which tags they should filter.
 
^{Actually, I’d like to see this kind of tag list made from my upvotes and faves. I should go do some data mining…}

@The Smiling Pony  
There’s a difference between people voicing their distaste for something, like the content of an image or the work of the moderators, and a group of people colluding to silence/discredit/harass people that say things they don’t like.

I’m not going to check all 8 pages but has there been any response to just abolishing the entire vote system altogether? Seems like it would solve this mess as well as some of the pissing matches I see with an image’s score such as what the top voted image should be (Yes that’s still a thing for some people).

@Corrderio  
There’s been some talk of adding a hide feature on top of downvoting. And some other stuff I think.

@Nonymous  

 

@Corrderio  
I think it’s safe to assume that would have the same effect that removing downvotes had, we’d see more honest comments telling people why they don’t like the image, which would be more mean messages to delete, which would be more work.
 
Now removing upvotes, THERE’S a great idea. That would encourage more positive comments, obviously.