A solution to the downvoting/flagging problems on Steemit
from the art of EscorpioTR
The platform has adapted a simple solution to addressing spam, plagiarism and abuse.
What used to be called a downvote has become a flag, and the community has struggled with the consequences of this solution, both when it was a downvote and after the action was renamed to flagging, ever since it was implemented. Many saw it as a detriment to the community and rightly so, but they failed to address the problems that downvoting/flagging addressed: the lack of Admins/Mods and the subsequent need of the community to police itself.
The issues surrounding the community with downvoting have been discussed at length in numerous topics. Currently it seems that every other day someone is affected negatively in their user experience with the problems of downvoting, and are voicing their concerns with this problem. The issues range from whales downvoting to remove rewards, as exemplified by the numerous grievances on that topic over abit and smooth's experiment, to flag wars ensuing between users from disagreements and squabbles.
The current system has it's inherent flaws. The fact that downvoting was renamed to flagging is one. This confusion is part of the fix from months ago addressing the previous issues of downvoting and it's consequences of affecting reputation and visibility while acting as a downvote and removing rewards.
Adding onto the confusion was the change in the list of reasons why flagging is acceptable. It shows that the rebranding of downvotes to flags was a bandaid solution and the need to express the obvious intended consequences of downvoting to remove rewards, so expanding the list of reasons for flagging to include disagreement on rewards didn't help to clarify more than it added to the confusion: is it a flag, or a downvote?
The other problem is that gang stalking and griefing cannot be countered by the community. If the community is to police itself we are powerless against these problems. To my knowledge that has never been addressed with solutions but it needs to, as people are being chased away by combined efforts of bad actors, and it doesn't have to be that way.
Flagging in internet terms denotes marking content as abusive, either outright abuse such as threats, personal attacks and other cruel, malicious acts, or spamming, trolling, and plagiarizing. Flagging is used in order to curb and minimize visibility or outright prevent such content/acts. On this platform this is one of the three(or two in case of no upvotes) simultaneous aspects that flagging is used for. The other aspect that is inline with flagging is reducing reputation, or the numeric of the good standing through contribution to the community. One cannot have reputation without contributions, and without worthwhile contributions that the community appreciates, one cannot have high reputation. The aspect that is not really in line with flagging is reducing payout, but as something abusive, spam or plagiarizing will hardly have any reward, this isn't that much of an issue, it's built into the simplicity of the system. Flagging is equated and has been equated to a downvote. Many will argue that it is a simple solution, but this solution of downvotes to police the community is marginally effective at best against large SP holders who can use their vests in a detrimental way to the community or against gang stalking activities, while it still can be used to grief people and all these actions chase users away from the platform.
Flagging's simplicity is that it acts just as the opposite of an upvote. While an upvote increases reputation, visibility and payout, a downvote or flag decreases reputation, visibility and payout. So what is the problem? The big problem, that many have acknowledged, is that this is a determent to community building and engagement. I think we can all agree that downvoting does little to those ends, and community policing should increase engagement and overall retention and not be incentivized as a way to engage in flag wars or griefing. While it's simplicity is to be applauded the consequences of this system are that a few bad actors can gang stalk, engage in senseless flag wars, and there is little the community can do against large SP holders who are considered bad actors and detrimental to the community, and there is no incentive for those bad apples to correct their behavior.
What's one solution?
De-incentivize downvoting. By increasing the amount of Voting Power spent on downvoting by ten times, or maybe less, large SP holders such as whales will consider alternatives to downvoting and it won't be as appealing to use a downvote. This isn't the end to the solution, there could also be a curb on how a downvote is weighed in respects to an upvote, maybe 33.33% less than an upvote in terms of vests, that will minimize the effect and in turn abusive downvoting.
This opens up the problem of spam. Spam cannot be countered by a downvoting system stunted like I suggested, nor can abusive content be effectively policed, as a bad actor could not be effectively countered under such a system. To solve that, one solution is to implement a whole new system designed for flagging and in turn counterflagging.
This system of flagging will become unlocked at a certain reputation level. One way to determine that level is to average the reputation of the community for everyone with a certain amount of posts/comments. So we will average the reputation of everyone with over 100 comments and that will be the designated level one must achieve to unlock such a system. This is partially to stop sibyl attacks. Also, a more veteran member will undoubtedly be better suited in engaging at effective policing, primarily because the group they belong to wants to see the community benefit and grow.
The function of a downvote under the new system will remain to affect rewards. Other than that, it will be the flagging system that negatively affects reputation and visibility.
(some terms: f=flag, cf=counter flag)
For effective flagging it would need to affect visibility with just one flag(1f:0cf=content hidden, negative mark against authors reputation) per content, and all users that can access the system will have the same weight for flags, regardless of reputation or vests. The reason for flagging should be divided in three categories: spam, abuse, and plagiarizing. Spam denotes posting things unrelated to the comment or post. Abuse would include trolling, threats, personal attacks, and such. Plagiarizing would be the last category, self descriptive. As a person engages in any of these categories, granted they are weighed independent of each other, the weight of a flag on their reputation will increase with each successive flag. Each flag decreasing a certain amount of reputation based on a curve that tracks/adds up previous flags. The time window for flags to be irreversible should be relatively short, enough so that a person clicking accidentally will have plenty of time to undo their flag, but not to much so that this doesn't become a game. By having a small window for reversals people will be cautious of their flags. They will be cautious because once their flag is locked in people will have a chance to counterflag, especially if it is abusive or perceived as abusive.
CounterFlagging
The author which got flagged won't have that option to counter flag so as not to nullify the flag themselves. I suggest a ratio of 2 counterflags to 1 flag for the person flagging to be countered and their reputation to be affected, which will lead to the forth category for flagging(this case flagging a flag/counterflagging) that will be weighed on previous actions just as the other three, in order to decrease one's reputation for repeated offenses. I suggest that ratio because one counterflag to one flag will reveal the content (1f:1cf=content visible/no negative mark against reputation), restore the reputation and nullify that flag, making countering abusive flags relatively benign.
Group Flagging/Piggyback Rule
The problem that needs to be addressed at this point is the group flag. To combat that and not allow for a group effort to drive someone's reputation into the dirt, each content will be affected by only one initial flag. So if there are multiple flags they will be counted as one flag, each excessive flag will be purely for gestural reasons. To further this, counterflagging can still happen successively even in lieu of numerous people flagging one content, but only the first flagger will be punished as they are the ones with any weight behind their flag. Each excessive flag will be null in the system, besides the first one, they will simply be gestural. So if someone counterflags it will reveal the content, regardless of numerous people having piled on the flag side. This might not seem logical or fair even, but it is. Flagging was effective with just one person. Everything after that initial flag was excessive, granted people should still have the option of showing support there's no logic to piling on, other than for gestural reasons. Even though someone might counterflag and reveal the content, even though it got numerous flags, this doesn't mean much if the content is perceived as abusive or worthy of a flag, it will take just another person to hide the content and affect the author's reputation. Equally even if there are two successive counterflags and the first flagger's reputation is affected, if it is abusive counterflagging, it only takes another person to stall the efforts of the counterflaggers, and then two more counterflags would be needed to affect the two flaggers reputation, the first flagger being one, and the one that flagged the content after the two successive counterflags punishing the initial flagger being the other who's reputation could be affected with two more counterflags (2f:4cf). In this situation with a piling on of flags after the initial flag, two counter flags and a follow up flag, the ratio of flags to counterflags would be 1:1, making the content visible(2f:2cf). For content to be hidden it will take a ratio of >1f:1cf. The nullifying of excessive flags is a shared property with counterflags. Lets continue with this scenario and recapping the timeline:
1f+nth gestural f=hidden content, authors reputation negatively affected- 1f:0cf is >1f:1cf ratio for nullifying a flag.
1f+nth gestural f+ 2cf=revealed content, initial flaggers reputation negatively affected- 1f:2cf is = 2cf:1f ratio for affecting a flaggers reputation
1f+nth gestural f+2cf+1f=revealed content, nobody's reputation affected- 2f:2cf is = 1f:1cf ratio for nullifying a flag
If there's more flags and therefore the flag to counterflag ratio is > 1f:1cf then the content will be hidden and the author's reputation will be affected negatively, being 3f:2cf. For the counterflagging to affect the flaggers reputation now, there would need to be 4 more people counterflagging (3f:6cf). It will take 1 more counterflag (3f:3cf) to reveal the content, but one more flag (4f:3cf) will hide it. Each effective flag will make it that much harder to counter flag, but not impossible, since each account has one voice over the content. To make this harder for attacks to happen, the time window for flagging and counterflagging should be a set time from the creation of the content itself, so that people can petition others to either nullify flags, to effect counterflags or simply to reveal content and not affect anyone.
What this suggestion aims at doing is stopping gangs from driving people's reputation into the dirt, giving people who appear to have responsibility the tools to affect visibility and reputation of bad actors regardless of vests, as abusive behavior should be countered by people and not money in an effort to incentivize good behavior and community growth in turn, while addressing pertinent or perceived issues by implementing a new system.
This new system is not simple but I believe it's complexity is necessary. The problem of the community policing itself is complex, and this new system was aimed at considering that complexity. This can be implemented from a programing perspective, even though such an overhaul has no precedent on this platform it doesn't mean I want to set such a precedent only wish to address the numerous issues and confusion that downvoting/flagging is riddled with.
Any feedback is appreciated, and the input of the developers would be greatly welcomed.
I think this idea has some potential. I really only see a hint at one flaw at the moment and you may have covered it and I simply missed it.
What is to stop a person from creating multiple accounts, even a dozen, and doing some simple posting with them for a week, and then having an army of accounts they can use to counterflag/flag.
Yes this would be another form of Sybil attack. Yet this is also why the "whales not being able to vote" experiment was a short term solution at best. If you draw a line in the sand then the exploit will usually find out how to spread themselves out across more accounts rather than one and with their accounts spread they effectively have the same power as before the "whales cannot vote" occurrence. So, what I am seeing here is that they likely could also find ways to game this system.
The reputation system didn't initially exist. We had a big problem with spam in late July early August. People tried to combat it, but what the spam bot makers did is create an army of accounts that were small bots that all up voted each others posts. So the people trying to write bots to counter spam and the people trying to counter spam were stopped by that sybil attack.
They instituted the Reputation system ONLY as a means to deal with that. A person with a higher reputation down voting something would reduce the reputation of the target, and it could not be countered by a lower reputation. Up votes similarly could increase reputation though it tended to only work for up votes from people of higher reputation. This made it possible to effectively stop this spam in its tracks.
It was very bad for a moment before that. You could see a new post every couple of minutes and they were almost always one of these bots. You had to wade through that mess to try to find legit posts.
So it worked well for that.
It also made it so people can have a high reputation due to votes from people while not having high steem power. So they can theoretically counter a high steem power person being able to hide someone's post. This is really the only place in the entire platform where someone can impact such a person without having massive power themselves.
I was only providing that for historical information in case it might help you in some way.
The suggestion I made was to tie the flagging system to a high reputation of 55-60. If they make a bunch of accounts for example and upvote their post for one week, that won't give them that high of a reputation, but let say that they do. Once they engage in illegitimate flagging they would have to counter the whole community because if they flag stuff and get counter flagged their reputation will start dropping, and with each successful counterflag they would lose more reputation than the last flag, and I think exponential is the way to go. Also of course tying it into categories, because that way people have a way to legitimize the flags, and by only counting in those categories for example someone engaging in spam won't suffer exponentially more damaging flag were they to write a comment or post seen as abusive and flagged as that, giving them a chance to change their ways and not simply run the off.
The problem is still spam. Spam under the current system is only dealt with on the UI side, there is no way to stop someone with even one account to post massive posts and load up the system, theoretically this could happen in a matter of hours, gigabytes upon gigabytes. To counter that we need to have a threshold that people need to stay above and not sink to oblivion, and I am talking about oblivion where they have no way to create content and post it on the blockchain, any kind of content even memo's to wallets, so when someone gets nuked they have been neutralized and the only way they could get back would be to attempt to revive their accounts through proxies.
The same for curation system, we cannot stop sockpuppet accounts from hiding content and demoralizing people or attacking people's payouts. If we set a threshold for 30-40 reputation to unlock the ability to upvote and downvote content they will effectively stop these account as they would need to create content and get voted to unlock it, and like in the case where people try to sibyl attack through flagging, in this case lets say they try to attack through counterflagging and attempting to damage legitimate flaggers' reputations they could be policed by a small group and their accounts brought below the threshold to flag/counterflag.
So back to the curation system, if they were to artificially boost their accounts over that threshold (obviously fairly easy and not a major obstacle) and they were engaged in malicious downvoting, even with the stunted 10x more draining voting power DV, people would go to their content and flag it. For them to be marginally successful they would need to wait 30 days and get past the limitation of flagging something, but that could be eliminated outright and a time limit would start not at the creation of content but at the first effective flag, and they would have 30 days to petition the community to reverse that action.
I will write an update and include these changes soon, awfully busy at the moment but thank you for the feedback.
The guns are on the table, the only way to address a person with guns is to run faster than a bullet, catch some bullets with ones teeth or dodge them, seeing that nobody can pull those three off the sensible option is get some guns and fire back. We have to deal justice with mallets that aren't equal for everyone, instead one person holds a mallet that trumps 90% of people's combined mallets and can smash at will because the system is built as such.
Perhaps your the Neo of The Steematrix - the man with the power to stop bullets?!
No, I consider myself more like Agent Smith, the casual agent that breaks free of the matrix and leads the revolution, while neo is an anomaly I am the anomaly.
I am the black cat. I am the black cat. Deja vu? ;)
You are more like Tank :D so you have to die really soon :P
I want to be Sati - the little Indian girl from the Underground station :)
LOL :D
I wonder how many words how much I have added to the platform, this post alone is about 2k, and I think I have written some replies that are over that.
No matter, this will continue to affect the community so it will always be an issue that is worth opening the discussion about. So far my suggestions seem viable, effective. They have put up a wall so far, it's hard to dismiss that because of the numerous views that it has and the lack of discussion about it. If there is something wrong, nobody has been able to point it out so far, what is clear is that my suggestion should have focused more on the downvoting issue.
That's not hard to implement: revoke voting all together for people reaching below a certain reputation, so anybody with less than 35 or 40 reputation has no say over rewards, done. What's the drawbacks? One possibility is the new users that have no option to vote will think it not fair. A solution would be to inform them that in order to have a say over rewards or curation and they need to gain influence through creating content or commenting to be granted that privilege.
It's not that hard to implement like you say - just like its not difficult to get that polar bear back to its real home. A striking image.
Difficult=time and effort, considering that both time and effort can be mitigated among those more capable and with better equipment/tools and working together it will be mostly effortless.
If more people would join this conversation, instead of sitting on the sidelines, and give their suggestions or point out the problems with it both time and effort will be mitigated among many minds with many ideas and perspectives, and not simply me alone.
P.S. All hail the @baah!!!
https://steemit.com/steemit/@mindhunter/now-i-ve-reclaimed-my-no-1-human-steemit-poster-crown-it-s-time-to-press-the-follow-button-just-once
A very interesting solution to a very nagging problem! Let us hope Steemit Inc. is in listening mode here.
Upvoted / Resteemed / Added to Auto-vote ... AND FOLLOWED ... I am a man of my word ... you are now my single 'Following' after you finally published content. CONGRATULATIONS @baah!! :)
@baah you are a very special person because @mindhunter does not follow ANYONE...just one....lol
I did promise him if he produced one excellent original content article on Steemit I would FOLLOW him :) @mindhunter is a man of his word :) Old skool.
Good for you!😀
This is the first time in a loooooong time where I've seen #comments outweigh #upvotes on a post - and the last offender to do that was ... ME!!
Yup, and I think I understand why that is, the loaded guns are on the table for everyone to consider, and it's turning people away from considering them as useful. The traffic is another evidence of this, the ones that are turned off by this see it as creating more problems but in reality this could help change everyone's behavior and make people a lot more considerate.
And evidence of that is Switzerland
http://world.time.com/2012/12/20/the-swiss-difference-a-gun-culture-that-works/
Can a Steemit Singapore mutate into Switzerland??
https://steemit.com/philosophy/@mindhunter/singapore-like-steemit-leads-the-good-life-under-a-benevolent-dictator-but-how-long-can-the-american-dream-last
Never seen people carrying weapons in public, except for police, in Switzerland. I doubt that “nobody bats an eye at the sight of a civilian riding a bus, bike or motorcycle to the shooting range, with a rifle slung across the shoulder” is true.
Like in other countries you must have a license; to obtain it you must pass a theorical and practical exam and also and primarily you must prove that you need to carry the weapon “to protect yourself or other people or objects/properties from a specific danger”. See for example here.
There's not any exam, from the link.
There's no need to prove you have to protect yourself, are you even listening to what you're saying? Exactly what our how will that look? I need to protect myself from threat a?
Switzerland had and continues to have a flourishing gun culture.
The given link is official. I was referring to the sentence “nobody bats an eye at the sight of a civilian riding a bus, bike or motorcycle to the shooting range, with a rifle slung across the shoulder”: people would bat their eyes, and civilians can't carry weapons, unless what I've already written:
Buying a weapon to keep it in your house? Sure, you're right, you can, and it's what you're talking about (“acquisition”). But when I've written carry, I mean in public, along the streets, at the cinema, on public transportation. Talking of the topic, there's no sense in using carry to refer to the fact that you can walk inside your house carrying your gun.
The next time you go to Switzerland, please take a pic of these people walking the street or riding busses with rifles.
Note: my experience is on french-speaking Switzerland. I can believe that other cantons could have different laws, though we should check first if it is a federal law that single cantons can override. From the given link, it seems its nationwide.
Oh, and I forgot to point at your griefer attitude:
It's not me saying it: I've translated a part of a text you find in the link, where you can find a summary of what suisse law says.
I have, the text is very succinct and never mentions an exam or justifying your acquisitions of weapons with such a outrageous "i must protect myself and my property from threat a". You can point out the paragraph that describes that, but I don't think you can as I read the entire document and it mentions no such thing. They have open carry, meaning that you can carry anywhere, unless the business has a rule against it of course.
So to recap: no exam is needed to acquire weapons, you can carry weapons in public/open carry without exam, just the permit, nobody has to justify carrying weapons.
So to recap:
But you've read the entire document, so this is invented from here
I still like my old Amazon review star system idea:
https://steemit.com/steemit/@mindhunter/why-not-use-the-amazon-review-star-system-to-cast-our-vote-instead-of-up-down-votes-to-allow-a-fairer-voting-system
As dwinblood answered that question of why I agree with him.
For example, making a system for creating content brings about the problems of spam, plagiarizing and abuse in other forms, so to deal with that there needs to be a way to either hide such content or outright remove it. In creating a system that offers people ability to vote over rewards for content there needs to be a way to deal with those gaming the system for their own ends. That is why with just a upvote only system it can only work if there is no rewards incentivizing others to game for their own ends, and even then it can be abused, marginally but still abused by sibyl accounts.
For some it's hard to deal with the reality that only guns can protect us from people with guns, and no amount of bulletproofness will suffice to deincentivize a would be attacker, only a gun will. I realize that what my system does is basically put a bunch of loaded guns on the table, which turns off most unwilling to look at the gun as a tool that is just as useful as it can be abused.
This problem seems as intractable as trying to convert a DASH disciple to Steemit! I don't know which is tricker??
I don't want to attempt any kind of conversion, but clearly DASH doesn't deal with content and it's inherent problems.
I've just this second put down my spoon to your egg analogy. The war is over. Let my philosophy slowly sink into her ... or not!! Thanks for all your support to this old <3 :)
As do the other 99.9% of cryptos as well. That is why their disciples are so vociferous! ... there is just a coin ... and PrivacySend ... and a credit card ... and a HUGE Darkcoin pre-mine!
P.P.S. A NEXT POST IDEA FOR @baah - could you do a lyrical analysis of my favourite REM song 'Ignoreland'?
I know that this is vitriol. No solution, spleen-venting,
But I feel better having screamed. Don't you?
http://www.azlyrics.com/lyrics/rem/ignoreland.html
Check out this story about why we yell in anger
http://www.awakin.org/read/view.php?tid=958
Wise words @baah ;)
Indeed, it's a very powerful message egging us to recognize why we do the things we do.
If you listen to Stipe's distorted voice on that 'Ignoreland' track you'll hear the lyric 'Bah FucYou Yeah!' :)
I have fans everywhere, lol I must be famous :P
Keep producing quality content and you'll always be famous on here @baah :)) You might even get a presentation slot at #Steemfest2 - so be careful for what you wish for :)
@baah, this seems like a really viable path forward, if anyone is listening.. and I sincerely hope they do.
You're exactly right about flagging having a negative name-- in broad Internet terms it has to do with abuse/cheating not the value of content.
To extend what you're saying here... if we could only get to a point where the flagging/downvoting power rests in the hands of only trustworthy ACTIVELY CURATING community members, that would help a lot. Maybe it could even work in a similar way to how we currently vote for witnesses... so we could also cast votes (AND withdraw them!) for a slate of "Community Curation Managers" and that power rests only with them.
As for the protests of "stake holders," it seems everyone has (conveniently) forgotten what an "investment" actually is. When you buy stock in a company, all you're buying is the opportunity to benefit from gains in the company's stock value... you are not buying some "magical rights" to be paid dividends or income... and in most conventional investment projects you're subject to "1 share, 1 vote" limits. Steemit is not a company that "hired" anyone to do a "job;" and nobody was promised a "salary."
That could work but the downside which I mentioned with any committee is that the individuals cannot handle the ever growing community. The ones on the committee could be rewarded for flags and such to incentivize their efforts but this can also lead to abuse, as can any position of power.
I think a viable way is to allow the community to do the policing, the only problem is that our current tools are ineffective at dealing with collusive power, griefing and gang stalking like I mentioned. There is no reason why downvotes shouldn't drain the voting power by 10x more than an equal upvote, as they aren't to be used extensively and are negative actions overall.
Some the Hammer helps indeed!
I think we should have Fun with Flags instead of the war - need people to engage and comment their flag so we can see which great community we have.
https://steemit.com/positivity/@uwelang/no-to-flag-wars-but-yes-to-fun-with-flags-positivity-is-needed-it-seems