@Background Pony #6996
@Cirrus Light
Restatement(Sorry for that. It might still be hard to read due to my florid prose from whoever knows where.)
Here are my thoughts on this issue:
I don’t think going back to the old system, i.e. giving up, is a preferable solution because then the problems of the downvoters commiting voter fraud by creating multiple accounts and refusing to give constructive criticism would be left unresolved. Arguing that downvotes are beneficial is as superficial and naïve as arguing that upvotes are beneficial because then the artists would not know what are the pros and cons of their art as they are simply told by the votes that this amount of people like it and this amount of people don’t like it — furthermore, when the voters do comment, they still don’t give the basis of praise or criticism in most cases — and are left wondering what is right or, more occasionally, what is wrong. Concerning which, I am not saying that such vague votes result in every artist not improving as there have been at least 2 cases that I know of in which the artist decided to judge his/her own art and improved, but whether their critique of their own art is the same as the critique of the downvoters is unknown. However, these exceptions are very rare and renders the main argument for going back to the old system negligible.
Also, simply adding the Hide function is equivalent to this solution because voter fraud and silence would continue, but the consequences of adding a Hide function are uncertain as I have never encountered an image board that had such a thing. Although, Hide’s consequences could probably be approximately predicted by considering the results of having filters, but according to The Frowning Pony in the Downvoting is being replaced by Hiding thread, people have tended to not use filters and instead downvote whatever they would’ve used a filter for, which could probably be explained by supposing that they either didn’t know or refused to take responsibility for their browsing (something the DWM movement refused to do whenever kids encountered porn), in which case, perhaps the Hiding function would not be utilized much, by virtue of which, renders it pointless.
Making downvotes public or applying the voting system upon comments would have negative consequences because, in the former, people will go out on a witch hunt for downvoters, of which I am certain will happen because in the replies section of the tweets by For Exposure, an account that tweets statements by people who want free artwork or free work to humiliate them by exposing them — ironically enough — , but they don’t give out names because of what I’m about to tell you, there were significant amounts of people wanting to know the names and even addresses of the persons quoted to “punish them”, so publicizing downvotes is a terrible solution, and, in the latter, people’s inclination to rage and shitpost would be triggered which I’ve seen often enough in the comments sections of FiMFiction to not read them anymore, so that is also another catastrophic solution.
One solution that seems to be likely to work is to add to the Up- and Downvoting functions a system that demands reasons for liking or disliking, but, since the moderators don’t want to moderate any further than the comments sections, I’m not sure what could be done to force people to give constructive criticism other than a least lower bound for the number of characters typed.
I’m left undecided on this matter.
Further thoughts on this issue:
It appears to me that a solution, for specifically voter fraud, will require statistics to solve in which perhaps indexes for Upvotes and Downvotes will need to be developed from a graph, where the x-axis is Upvote/Downvote density and the y-axis is time since the picture was uploaded. Although, concerning how to remove false votes and what situation would the graph be most applicable, whether it should take data from the accounts or the pictures, I’m not sure.
The placement of where the owners of each account on Derpibooru probably have to be accounted for, but in a way such that one does not know their exact position, so perhaps knowing the city and state of the country they’re in would be enough? But again, I’m not sure as I’m not a statician and having only began teaching myself statistics, so I’m left thinking about possibly utilizing the techniques in the following papers, if any of y’all are interested, Redistricting and the Will of the People by Jonathan C. Mattingly and Christy Vaughn and Redistricting: Drawing the Line by Sachet Bangia, Christy Vaughn Graves, Gregory Herschlag, Han Sung Kang, Justin Luo, Jonathan C. Mattingly, and Robert Ravier, which are papers that have investigated the voting system of the USA, so they might shed some light on the voter fraud problem. I haven’t quite completely understood the papers myself (to say it again, I’m not statician). Regardless, perhaps some progess could be made going from them.
I think a probability measure and distribution might have to be used if one was to consider accounts that have certain tags and filters on or off.
This thought has a lot more uncertainty, considering that this unconstructed, partial solution enforces less voter fraud rather than influence people such that they wouldn’t do so.