Simple Redesign on Downvote for More Positivity
This post is a suggestion to address side-effects of downvotes.
Upside and Downside
Downvote(a.k.a. flagging) is very important tool to prevent reward abuse. Without this, there is more likely to have airy posts with self-votes, which never help the growth of Steem. Also it is essential for adjusting hyped reward. It is strong disagreement on other upvoter's opinion and this up-down dynamic can make the system healthier if appropriately used.
However, downvote also can be used for trolling other's legit contents. The problem is more serious when a troll has lots of voting power. Only-downvote period also exacerbate this since no upvotes are allowed to opposite trolling flags.
It is double-edged sword, as like many other stuffs. If we empower the right to downvote, we can reduce reduce abuse more but also opens more leeway to trolls. How can we balance it?
Principles against Reward Abuse
There are two important aspects of reward abuse. First, the goal of reward is earn some free money based on the abuser's stake. But if the rewards are relatively less than his/her stake (e.g. an abuser owns 1% Steem Power but only get 0.1% of daily reward), there is less incentive to do. Secondly, reward abuse is not one-time thing. It is repetitive, for days or weeks, tens of posts and comments per day.
The reward abuse is obviously bad for Steem. However, if we attempt to remove it perfectly, it would cost significantly high. Rather, based on two aspects mentioned above, we can apply soft-penalty on the abuse with the following principle.
- A reward abuser must not earn a proportion of rewards more than the proportion of stakes one owns.
- Rewards should converge to zero as a reward abuse continues.
In addition, we have the most important thing. Fights against the abuse is not individual's job, but the community's responsibility.
- Penalizing reward abuse should be addressed by collaborative actions
Simple Idea: Asymmetric Voting towards Positivity
In short, a desirable way is to partially reduce the abuser's reward in the beginning, and increase the penalty as the behaviors keep going. At the same time, the policy should not give much power to trolls. Here's a simple suggestion for that purpose: Asymmetric voting.
- Downvotes consume 5x (subject to change) voting power than upvote
For example: A troll with 2,000 MVESTS (1 million STEEM) only reduce 200 MVESTS vote. But the downvoter has to use a voting power as the same as 2,000 MVESTS upvote.
Why asymmetric? The reason is that trolling tends to be an individual action while preventing abuse is more likely to be a collaborative move. 5x requires five times more voting power to have the same magnitude of the current downvotes. Abusers may get some rewards first time. But as they repeat the abusing, more people will join to downvote and they will get fewer and fewer rewards. Meanwhile, big whale troll can be nullified by a 1/5 size dolphin. This asymmetric vote weights positivity.
Conclusion
With linear reward system, our reward distribution will be fairer. However, threats from rich trolls remain and they can easily spread negativeness in Steem community. Asymmetric voting towards positivity can weaken troll's influence while still can discourage abuses with collaborative action.
This is a good idea and an interesting proposition for bridging two conflicting ideas:
We can debate the 5x factor specifically, but that down votes should "cost" more than an up vote in terms of recoverable vote power is wise.
I've seen several ideas about this and I will generally side with ideas which are simple and a clear, direct aspect of the system.
For example, I disagree with the suggestion by @kyle.anderson for this reason. He suggested:
Not to single it out for any other reason but an example, but his idea requires much more additional code and collusive action. Your proposal allows for the same system of independent action which is collaborative.
The difference between collusion and collaboration are that the former is planned, organized and arbitrated. The later is not necessarily so, and can be the independent alignment of views. It's this kind of stuff that allows for people of disparate views to come together to achieve something together without planning it, and a wonderful feature of the current system which should be preserved. I don't think we want to "bake in" collusion when so much effort has been made to reduce and disincentivize the negative effects of it so far.
As a related side note, the "flag" terminology needs to be removed from the steem front end finally. It only adds to the confusion. Keep the message imparting the seriousness of it, but remove the terminology.
Yeah I mean this is an interesting perspective but I really don't think it is needed. You make good points about the issues with the downvote only period. I really don't think these downvotes are as troublesome as you might think.
There is definitely a problem with some users abusing their stake to downvote against users. A better solution lies around forming a better resolution system where stake can be pooled for the purpose of manually restoring payouts to posts that have been abused. This is basically a charitable service for the network so it would need people to stand behind the idea for the sake of the steem blockchain as a whole.
If the platform gets popular I think we can expect an increase in the amount of abuse. If it devalues someones post just because some hypothetical dark whale and following cronies start flag a persons account to the ground, then I dont see people hanging around to get their hard made posts and oncome devalued. Not where money is involved.
I do like this and while the community is small it could be feasible, but I dont know how well even that would work. I dont see a case by case basis working if steemit blows up. How many people are even on steemit? Im not sure how large this platform has gotten.
This is all hypothetical ofc, steemit might remain a smallish community relative to the youtubes and reddits out there. I just think if reddit earned money by upvotes on posts, we woulndt even have a reddit in its current form.
When the platform gets more popular there is going to need to be more stakeholder involvement in maintaining an appropriate content payout profile. Period. This is what will make or break steem.
When steem gets bigger the manual adjustment is really going to be supplemented by many more bots voting. Curation bots will become smarter and stakeholders can use them to address problems like flag abuse directly. Manual review is going to be a part for a long time IMO.
Curation bots? I never considered that. How would they function? Search out patterns of abuse and allow decisions to be made without human intervention?
I have faith in the community but I worry what will happen when organizations get involved. I guess you are correct that right now its not a worry.
Curation bots are around now doing just that. They will only improve in scope and efficiency.
I disagree with you. I can only see a "resolution system" as unnecessary bureaucracy, and wonder how do you imagine this would really work? It sounds complexifying, not to mention that it would require more new user onboarding.
Like I said, this will probably have to be 'charitable' until its effect on steem price as a whole is tangible. Bots and users that ensure proper payouts are going to be a cornerstone for steem progress.
We can already do that. I was disagreeing with your "a better resolution system" idea in so far as you described it.
Well the resolution system would be something community directed, or at least stake directed. This would be a system where users who are being wrongly flagged or abused can get a human appeal with the prospects of a group with lots of steem power to help the user out - a user who had previously been essentially removed from the view and conversation.
An appeals process is really antithetical to the current philosophy, which is that of incentives only. I'm in favour of retaining this.
Appeals are centralized by nature, as they have to be to some authority. I've debated ideas around that in the past, even proposed some, but ultimately concluded that they are too complex and require a layer which I do not which to see - management.
Users can still act towards a common in the system as it is stands (they do, the whale experiment is one for example), there is just no formal mechanism for this. I think we should not introduce one.
You don't want formality, that is fine. An appeals process would be entirely a layer above steem. It is entirely based on incentives - incentives stakeholders have to keep the network free from abuse. Absolutely aligned with the current philosophy.
There shal be a formalization of 'the whale experiment' and things of the like. There must be if the platform grows.
If you really mean a layer above steem, i.e. an app on steem, I'm all for that, and in fact have suggested it.
If you're going to define "incentives" so broadly, it's true what you say. Obviously I meant direct reward based incentives, not more general health-of-the-system ideas, which I agree are relevant. Not to be baked in though on a systems level.
The formalization of the whale experiment are really direct changes. We have the current proposal (which is looking like it will become a reality) of the linear rewards curve as a response to this.
This costing of downvote voting power opens the door for more spam and abuse to happen as the drainage will surely happen before the abuse stops. A stunted flagging system is only one part of fixing the downvoting abuse.
@LukeStokes mentioned how downvoting falls under negative bias and there is no reason why not to reflect that in the code.
https://steemit.com/flagging/@lukestokes/hey-steemit-let-s-talk-about-flagging-again
I have suggested exactly what you are saying a couple months ago, it's in the thread I linked above, a link to my article.I think I need to polish up another post about it in the future.
So you're in support but it's not enough? That's good. Step by step. In fact, clear and small steps should be the way of any change. It was one of the issue getting KF 17/18 through, waaaay too much in there.
A downvoting system that drains more voting power than an equal upvote won't solve abusive downvoting as much as it will enable abuse and make countering it SIMPLY for visibility that much harder.
Another problem is that yes, a downvote needs to be weighed in respect to it's negative bias impact on users, so that griefing and abusive downvotes are not motivated with the current equal worth system. One negative action can easily negate numerous positive actions, and as such minimizing the effect of negative actions and keeping it at a ratio that balances these effects.
The reason I am for a changing downvoting is because this scenario has repeated a few times now:
Bad whale actor ensues flagwar.
There's no way to counter a bad whale. The problem is that there are two different things, flagging is what a downvote is, yet flagging isn't just the renaming of downvote because the reason it was renamed was to be perceived more severely than simply a Downvote, but that only led to the confusion which amended the list of reasons one should flag with "disagreement over rewards". Flagging is all about abusive behavior and how to deal with that. Flagging therefore should get it's own system which doesn't differentiate between actors vests and their behavior, if a whale wants to spam, or plagiarize, or troll, or spew hate and spread malice, they should be dealt with by the community, but specifically the individuals, and not be untouchable because of their bulletproof vests, there's not anywhere written in stone that a downvote is equal to an upvote, or that without that there's no incentive to hold SP, it's not written anywhere that holding SP is dependent on a equal weight voting system or one without a dedicated flag. It's not written anywhere that policing the community should not be incentivized and those that endeavor to write the wrongs of the bad whales should waste voting power and fight a losing battle, because in the end the negative impact couldn't be mitigated ever by an equal weight system.
A simpler way is not always the answer, sometimes we need to observe the complexity of the problem and simplify it's solution sufficiently instead of making small steps/simple solutions, chasing our tails..
We have a system of reputation that seemingly does nothing more than effect visibility above and below a threshold. We have no way to deal with spam, even at 40 megabites a day, spam still can happen and it needs witness support to exclude an account from spamming. So far spamming is not a problem, it can become a problem as a person can register numerous accounts and overload the system with data by simply pasting images in a comment box! Correct me if I'm wrong, but once you upload an image there's no deleting it, even if you discard/cancel the comment or post you uploaded, the link seems to disappear into redundancy but that's not to say it doesn't work, because apart from not being able to browse all these different images directly, the data is still there and accessible with the link but the limits that are in place to stop spam can still be abused and I think would be abused no matter what they were in frequency.
Besides spam there's the problem of dropping visibility and popularity for these bad apples leading to other exploitative behavior like creating a spam post and filling it with enough comments that it doesn't peek anyone's interest, and under the cloak of a nuked account, they can vote on that account's comments, creating a progressively deteriorating situation in the worst case they can spam numerous comments, upvote them and wasting people's voting power, time and effort as well,trying to counter them, and they can still extract wealth at the same time.
I think that the flagging problems and our inability to police effectively will continue to persist, I don't see a simple solution creating much good in such a complex problem, but you never know. I do think that that by creating a meaningful reputation system which goes up with upvotes but only goes down with flags while downvotes still weigh on the payout yet drain 10x the voting power (shaking fist at proposed/upcoming stoopid change over 10x increase in voting power used for each vote!) of an equal downvote is a viable solution. If we give all the users the same tools at flagging, not one more powerful than another, but simply lock these tools behind high reputation, vests won't have the ultimate say on the platform, and they shouldn't as this is SOCIAL MEDIUM first, and crypto currency second. They can also then tackle spam, first by effecting the reputation of the author much more readily but also because the limit for creating content doesn't exist now, but it's equally not written in stone that everyone should be able to produce content on this social medium, yeah I know, call me Hitler but reaching Zero reputation should lock the person's account from posting, and a reputation under 30-40 should equally be locked from curating, or show me the stone! If we lock curation we can do away with alts and bots flagging without any content, or downvoting without any skin in the game. Yeah, I'm in support but it's not enough.
Your position doesn't appear to be coherent. You say it will actually enable more abuse
But then conclude that you're in support of it
Which is it?
The original intention of Steem is not actually against abuse as such you may be surprised to hear (I certainly was). From the whitepaper:
I suppose what this says is that people will try whatever can get away with, and that's fine as the coin should still be active, and thus potentially valuable. The problem is that this doesn't help us much with the social network.
I understand what you're saying about flagging being actual flagging, not downvoting. But this is simply not the case as it stands. If you advocate making flagging into a separate thing (I sounds like you do) I plain disagree. For me, behaviour is in the eyes of the beholder. We should support differences of opinion with tools but we already have that. I think improvement is all that's needed, not another system.
Let's see what happens with linear voting and then discuss the effect of so-called "bad" whale using their stake. It might be enough, especially coupled with what @clayop is suggesting here.
On rep, yes it should mean something on the blockchain. As has been pointed out plenty, this is a rather large oversight so far. But going as far as your "Hilter" solution, I'd oppose that.
This should definitely be a way to grow the Steemit platform and strenghening the will of the majority!
Lately I saw 4 posts with 400 Steem reward downvoted to zero by 1 downvote... Which is absurd...
deserve upvote and resteem...
Seems for those of us at the low end, we don't even have the option to downvote...
No option to flag you mean? I haven't heard of this before, are you sure?
I like the direction on this. Make it more costly to downvote.
But why can't downvotes just be an admin function without monetary penalty? Enough downvotes and your post is hidden and your rep goes down, but if the post earned rewards you should be able to keep them. imo at least.
thanks for sharing, i like your post
Interesting post, thanks for sharing.
nice