My take on Self-voting, Vote-buying and Reward Pool Rape
tl;dr: Linear flags were a mistake.
And nobody asked for them in the first place.
Background
On the 14th of June, 2017, this release hit Github:
https://github.com/steemit/steem/releases/tag/v0.19.0
It is version 0.19.0 of STEEM, title: Equality.
It introduced linear rewards.
Before, rewards were distributed by the infamous n^2 curve.
Simplified, this meant that a vote was worth less on a post, that nobody had voted on before, and worth more on an already popular post.
Whereas with linear rewards, a vote with a certain amount of STEEM Power behind it, will have the same impact regardless of which post it is for.
Together with changing the 'reward-curve', the update also changed a single vote's impact;
A 100% vote will be 4 times more powerful once HF19 is completed.
On the 25th of June, enough witnesses had upgraded to v 0.19.0 and the scheduled 'hardfork 19' happened.
Before, the community had complained about the unfairness of reward-distribution on STEEM. To get high rewards, you needed a substantial amount of r_shares ( powerful votes ).
It either took seriously good content or serious collusion to get really high rewards on a single post.
With HF 19, all those issues with collusion instantly disappeared.
Now, to abuse the rewards system, you do not need a substantial amount of STEEM Power anymore.
You can simply vote yourself.
As HF 19 ( Equality ) hit the blockchain, the effects of the changed 'reward-curve' and the increased vote-impact resulted in a ridiculous reward-drain.
Some payouts increased up to 10x, while the top-payouts did not shrink as much as one would have expected.
Strangely, nobody had predicted this effect ...
The effect on the community was disastrous.
Most users realized they did not need a network of readers anymore, they could just vote for their own posts and comments only and maximize their own rewards.
More Background
The original STEEM Whitepaper had a mechanism to prevent this sort of 'vote abuse'.
Interestingly enough, the updated Whitepaper kept the same passage.
With linear rewards, it just does not make as much sense:
The Story of the Crab Bucket
A man was walking along the beach and saw another man fishing in the surf with a
bait bucket beside him. As he drew closer, he saw that the bait bucket had no lid and
had live crabs inside.
"Why don't you cover your bait bucket so the crabs won't escape?", he said.
"You don't understand.", the man replied, "If there is one crab in the bucket it would
surely crawl out very quickly. However, when there are many crabs in the bucket, if
one tries to crawl up the side, the others grab hold of it and pull it back down so that
it will share the same fate as the rest of them."
So it is with people. If one tries to do something different, get better grades, improve
herself, escape her environment, or dream big dreams, other people will try to drag
her back down to share their fate.
Source: https://steem.io/SteemWhitePaper.pdf
It then goes on about how with enough 'crabs' in the 'bucket' the more 'abuse' would be negated and that it never is a perfect system and so on.
What this all fails to mention:
With linear rewards, it takes as much power to fight abuse ( pull another crab back down ) as it takes to just vote yourself ( just keep climbing ).
Before, ( with n^2 ) a vote had a much larger impact on popular posts, for either upvoting or downvoting.
Oversimplified example:
An account with the voting power to reward a single post -which nobody had voted for, yet- with 1$, would increase the payout of a post worth 100$ by 10 $.
It could also downvote the 100$ post by 10$.
The numbers are made up - it is just an illustration
Non-linear flags
I would like to propose a new hardfork: Quality
Non-linear reward distribution should create a self-regulating mechanism, just as intended in the original Whitepaper;
Only content that a large group of stakeholders agrees on could generate large payouts.
A vote on already popular (exposed) content would have more impact.
In theory*, this would create a curation process, where good content would automatically rise to the top.
*This only works with half-way responsible shareholders, or to be more clear, if @ned stops giving away stakes to $%@§heads.
I wrote 'non-linear' because there was some discussion around other 'curves' - sigmoid for example.
I think there never was anything wrong with n^2.
I want a vote to have the most impact on the most exposed content.
The last 6 months have shown that hf 19 was not a good idea.
The quality of the trending page is even worse than before, self-voting is rampant ( and very lucrative ) and vote-buying is 'an interesting phenomenon'.
I understand, STEEM is now becoming 'The blockchain for Smart Media Tokens (SMTs) and decentralized applications'.
I understand, Steemit is mostly a proof of concept and that STEEM is designed to host all sorts of applications.
I still think that Steemit should at least make an effort to ensure good content, to get STEEM into the hands of talented people of all sorts.
Thank you for reading,
Please also consider voting for my witness @felixxx here:
https://steemit.com/~witnesses
I also tend to think linear rewards were a mistake. I would love people to open up on this topic that's why I've resteemed your post.
I'm not saying linear reward were definitely a mistake but I do think so. I've been thinking about it daily pretty much since it was announced.
We must get the Steem economics right. They are the second most important thing after the code. Both can lead to either Steem prosperity or systemic failure.
https://steemit.com/steem/@dantheman/evil-whales
Before, when the reward curve was n^2, Steemit wasn't nearly as popular as it is now. You needed to be upvoted by someone with 1,000,000+ Steem Power to get any decent rewards. But now, someone with just 20,000 Steem Power can upvote you and get you a few bucks.
Today, now that the platform is so much more popular (and the number of posts is many times greater), there just aren't enough people with 1,000,000 Steem Power to receive a vote from one of those users. There are 10 new blog posts each minute and just a handful of whales.
I believe if we returned to the old reward curve, almost every user would become dejected after they realized how difficult it'd be to earn any meaningful amount of Steem. Today, it's tough enough for new users to become successful, and with an n^2 curve it'd be many times more difficult for new users.
I'd like to see some more ideas thrown around regarding downvoting and making it more popular. There's a lot of highly-intelligent people on this platform, and I'm sure we'll come up with a workable solution.
It's not about Steem's popularity. Not everyone can get paid on Steem. That's a mathematical fact.
Only the Steem concentration is changing. That's not the issue here. Reddit and wikipedia are 2 of the most popular website yet they don't reward any of their users.
Dan explained some of the reasons why non linear is needed in a comment under his latest post and I commented about it.
https://steemit.com/eos/@dan/proof-of-good-governance#@teamsteem/re-dan-re-mikepm74-re-dan-proof-of-good-governance-20180101t023815634z
Followed your link(s) and read up. Lots of theory relating to an idealized perfectly functioning Steemit, but a bit short on objective and achievable goals.
@shenanigator presents us a convincing case about negative long term consequences for non-linear voting and your reply gives me the impression that you oppose non-linear simply because it can't make an imperfect world perfect. On the surface that might sound like defeatism, but, when really thinking about it, it sounds even worse to me - it sounds like you're saying that since it's impossible for everyone to get paid, it would be better to pay even more to those who do get paid!
If we go to that other extreme, is there any justification for further increasing already large reward amounts? Can we justify in a democratic setting the idea of giving even more to those who make the most, and less to those who make the least?
If anything, the only thing justifiable in my view would be non-linear in the opposite direction: the fewer votes received, the more relative power behind them. In theory, voting power should equal, regardless of who has it or how much has been receiven, but if we're going to try to ethically justify a modification to that ideal equality, wouldn't it have to be progressive, and not regressive? We are living in the 21st century, are we not?
I'm really befuddled here.
Certainly you aren't arguing in favor of increasing the relative power of each vote as the number of votes per post increases, are you?
Is wikipedia is a good comparison? Are we looking to become something that resembles wikipedia? Or something more like Facebook, with circles of collaborators that grow, intersect, multiply and that have open membership? The later would be a Steemit where small groups could actually distribute meaningful rewards among themselves, as well as participate in more well known authors' threads, for example.
I'm not sure I'm explaining myself very well since my main point is the same but I'm getting no feedback on it.
Don't you think it's best to aim for the most widely distributed reward structure possible? Even if it's not possible for everyone to get paid?
Or, perhaps, precisely even because everyone can't?
I don't pretend to know the absolute truth on this matter and I respect the opinion of everyone. I'd love to be convinced either way whether it be for or against.
Steem aim at is making the most useful information at the top of Steem. The most useful information is specific to each individual. We don't have the same perspective.
Those who have the most to lose, the largest investors, must have larger votes.
Someone cannot claim the labor of others nor can they claim the investments of others.
Those who advocate for non linear reward aren't looking to destroy their own investment which would happen if we don't get the economics right.
That's the issue here. If it cost more to police Steem than to abuse Steem then in the long run the abusers will own most of the Steem.
This is what Dan is arguing in his latest post. He's arguing that it cost more to police than to abuse Steem thus the abusers are at an advantage.
If he's right then it needs to be corrected.
I've shared some other thoughts in a reply below to @shenanigator.
We're on exactly the same page.
Is this an absolute target, as in on the main page? Or is it a relative and distributed target for each tagged topic?
It seems to me that they already have an appropriately weighted advantage: they have voting power correspondent to their investment - everyone has that same "equal" percentage that we've all agreed is fair.
The task would be to present rational arguments in favor of giving the richer an unfair advantage by making their relative percentage stronger. I think the progressive counterargument would eventually win out, but first we must debate the ideas.
As for the rest of what you said, I have problems following what's being said. I'm not even sure if policing is necessary, but that's a whole other argument that I think is premature at this point.*
Before moving forward I think we clearly need to decide if we're in favor of making the equally distributed percentage based voting power of richer Steemians stronger relative to poorer Steemians - which is what I understand you to say when you say "the largest investors, must have larger votes". Correct me if I've misunderstood.
*See my most recent post for an introduction to why policing is a throwback to centralized authority and not necessary in a positively aligned environment - Steemit would be grass roots at its best if allowed https://steemit.com/steemit/@cryptographic/enlightened-self-interest-and-steemit
Just a random thought here. The old system made it beneficial for people to just upvote the same old people all the time. This doesn't mean they are creating good content. Just that people always upvote them.
I think that the issue is we are upvoting individual, and not content.
So many people are auto upvoted, they could post a PoS, like a literal picture of a PoS and get over $100.
This isn't helping Steem. And I think that people with bot armies, aka, the person I have been posting about lately, would get extreme rewards just because they have 1000s of accounts.
I've shared some more thoughts on the matter in my reply below.
I don't think people are coming to Steemit because it's such a great website (like Reddit and Wiki). I think they're coming here because of the revolutionary concept of getting paid for online contributions.
Not everyone can earn a lot, sure. However, everyone who contributes to the success of Steem can earn a little. If we make it nearly impossible for people to earn a little by going back to the old reward curve, I believe that will negatively impact Steem's popularity.
Furthermore, if we return to n^2, that will have a negative effect on the value of Steem Power. Today, if the average user goes out and buys 1,000SP to increase their voting power, that 1,000SP has significance. Under an n^2 curve, an additional 1,000SP would have no meaningful effect on the value of your vote (It might go from $0.04 to $0.05). How many people are going to go out to purchase SP, when they can't see it adding a worthwhile amount to their vote? I don't know the exact number, but I'm certain it's fewer than today.
I don't pretend to know the absolute truth on this matter and I respect the opinion of everyone. I'd love to be convinced either way whether it be for or against.
I agree with your first point that people don't come here because the Steem content is great and that's an issue.
I've been looking at the youtube trending page everyday for years now and someday I don't even click on any of the video because they're all bullshit video created to make money without any real substance. That's the second most popular website where all the supposedly best content creators are and yet all the population of the world is sometime not enough to produce quality content.
We can't expect Steem to create super duper content right now even on its trending page because there's so few content creators compare to Youtube or reddit but it will come. A lot of it is a question of numbers.
As for the main point, if everyone vote for themselves, Steem become pure POS minus the reward that goes to the witnesses.
The minnows would make cents as this is the value of their votes but that's not even the issue. When new Steem are created, if no new buy orders are created on the order books, newly created Steem dilute the price of Steem and doesn't increase it.
If there is 1 buy order at 5$ for 10 Steem those who receive the newly created Steem can now sell at that price too but in the end not every Steem can be sold at that price, in fact as I said, the price of all Steem has been diluted if no new orders are created thus the reason why it's important for us to get the Steem economics right.
So I do feel the n2 curve disincentives self-voting and if it doesn't then it's easier for larger shareholders to police the abusers.
Right now it's profitable for anyone to abuse the system while under a n2 curve it's only really profitable for larger shareholders and these will be easier to spot because there is infinitely less of them and they will also be easier to police them by the even larger shareholders.
One more thing, Steem popularity shouldn't be part of the equation here. Steem should aim to become addictive whether or not there's a money incentive. Other website have achieve this and there's no reason it can't be achieved.
Let's not forget, the inflation is going down overtime, meaning there will less Steem for a growing number of Steemians.
Problem is that n^2 leads to centralization and exclusivity. In my opinion, that's not the right way to encourage universality. I think we need to be inclusive and broaden the reward distribution as much as possible. I also argue in favor of the complete elimination of downvoting for the same reasons. My thesis is that if we force ourselves to move forward in the positive, leaving the holding back of others to others, and keep an eye on the widest contributor base possible, with time even the abusers will come round (or go somewhere else 😉).
I also favour the elimination of down- voting. My opinion is just as valid as yours so why should you be able to down vote it? I'm not saying every opinion is right, but every steemian has the right to be heard. Down voting is just a form of censorship. If quality content were actually rewarded then there would be no need for down votes as bad content would never be visible. The issue is with the up voting system, not the down voting system.
I agree with your sentiment but we need to combat spam and plagiarism. I've expressed my opinion here in detail: https://steemit.com/steemit/@vimukthi/a-philosophical-and-economic-outlook-at-steemit-sbd-and-flagging-and-suggestions-for-a-better-future-based-on-positive
Short version is just make flags consume 3% VP instead of 2%. You could even make it 4% and that would really help to make the flagging wars die faster. Positive reinforcement works better than negative reinforcement.
I downvoted a post which took my content off of reddit and was profiting off of it (and they literally shared the post as is - it's not like it was a review or lets play etc - i.e. something transformative). They literally took my content as is and started profiting off of it. I sorted it out with the admins in the end, but downvoting those who steal from others is definitely a legitimate use of downvoting as they shouldn't be doing that shit.
That's an interesting case you raise @birchmark. I hadn't given that scenario any thought. You are absolutely right that others shouldn't profit off your hard work. I'm not sure what the solution is here. I've decided to leave it to those more experienced with Steemit than me to work it out and instead focus on my blog :)
There probably isn't an easy solution to these situations unfortunately - I've also been on YT for about 1 3/4 years and the amount of issues and drama and changing policies over that time has been ridiculous.
I agree with you that when money is involved though people shouldn't be downvoting just because they disagree. I don't think that should be policed or anything, but I believe it is bad etiquette. Even on reddit downvotes are primarily meant for content that doesn't contribute to discussions or is inflammatory / needlessly nasty, not things you merely don't agree with, so people shouldn't really be downvoting comments they simply don't agree with on here either if it is simply part of a discussion and everyone is being civil.
I don't see how this necessarily follows at all. Downvoting is a necessary evil to combat posts that are overvalued due to inorganic voting as one example (a greater evil in my humble opinion). How else would we deal with this?
Those kind will just keep getting more powerful and powerful and bury the potential gold mine of minnow posts in their mediocrity whilst getting paid in doing so.
This, my friend, is something that is contrary to universality or wider appeal to talented users that may not be so affluent to use the "tools" that the not so talented yet affluent use to make a bucks. We are allowing this platform to become a Ponzi scheme of self-voted and bid buying. Flagging may be are only hope of correction but I think we collectively need to realize what is happening and what the implications will be if we do not nip it in the bud now.
Vote me.
https://steemit.com/@mahikaler
You guys are IMHO completely wrong about the problem here! :)
The biggest issue is not self-vote but it's whales "sharing" votes between each other. And no matter which linear or non-linear reward calculation there is, they will still give votes to each other.
If you think i am wrong do a test. I know that @teamsteem you are influenced enough, that if you do a post with nothing in it or maybe just a line of text. You will get more than 200$ reward. Some of the voters wont probably even look at your post before the vote?
And don't get me wrong ... I actually don't have an issue with that either, I always defend such votes, with they must have proven themselves in the past, but i don't agree with talking about linear votes after HF19 as being the main issue. HF19 provided many, many of minnows including me, to get at least a small share for the work they are doing. And also it gave the possibility to the whales to "share" their power, because 1 small vote (1-5%) from a whale to a minnow means a lot to them/us. Before HF19 even if you would get a small vote from a whale, it wouldn't matter? Because you would need at least a few of them? Maybe my logic isn't correct, i signed up on steemit just few weeks before HF19. But i see a lot of whales cooperating with different curation groups, sharing their votes and power, so the "lucky whale votes" actually do happen.
I do believe that linear voting is OK and would rather see some limitations. I've red before about the high reward cap per account per week or maybe even limit on how much a person can vote to another person? And the last option ... maybe there just isn't anything wrong with the current system? We are still getting more and more accounts, more and more posts, and hell even rewards are with increasing crypto space more and more appealing?! So is there anything wrong?
Or maybe I'm just an idealist that believes most of the community will not vote only for themselves and that good content will prevail?
Anyway I wish you all the best in 2018 and sorry for the long comment xD
Hear hear here @felixxx. According your proposals on this post. How these changes are going to suposedly be able to combat The Elitist Club of Wealthy Circle Jerks so that they are not the only privileged ones in claiming their Primae Noctis - Droit de Cuissage - Derecho de Pernada to rape first all the maidens of steemit's rewards pool?
"The Elitist Club of Wealthy Whale Circle Jerks.." It wouldn't be bad if the content up on the trending pages are actually good but pictures of concrete walls with ridiculous analogies fetching $250+ is just bad. Buying votes and spending your own money is the only way to even get to the 3rd tier or get some kind of exposure.
That's right @adonisabril ¿Doesn't it?
As for 'Buying votes' you may well want to read my latest post about the subject. }:)
Cheers!!
I read your post, @por500bolos. It was even more unexpected than the Spanish Inquisition.
Hahaha, glad to know you went thru unexpected depthts to find and read my post @penston.
However, as for The Spanish Inquisition goes. I guess that like a heretic, I should burn you at the bonfire or something. Since you did not interact with the post in any way!!
You did not vote it, you did not drop a comment on it nor you resteem it. ¿How would I know how unexpected it was for you?
Run Galileo Galilei run. Next time you won't escape. LoL }:)
I agree on this.
Or as stated in the OP
Linear rewards helped the bigger userbase. They leveled the field, and everyone has the same possibilities for earnings now, by abuse or legit means.
@teamsteem
Always on point as I cannot agree more to your suggestions. Little wonder I borrowed your steem dance video in my thank you post of yesterday where I mentioned you. Hope you ok with it. Keep it coming.
Vote me .
https://steemit.com/@mahikaler
Thanks for bringing up this topic. You are absolutely right self voting is bad for this platform. Actually I never understood why this option was available in the first place. It is an invitation for abuse. You don't have to feel bad by using that offer. This really needs to be changed. I came across two accounts from the same person which vote each other on posts and comments. Their articles are complete linguistic made up nonsense. But content will be financially honored.
I agree. I believe in this platform but people trying to game the system may lead to its downfall. Hopefully, creativity and posting original content will be what pulls Steemit forward rather than the reverse.
Not much to add here...
One would consider a near billion marketcap enterprise would have a better usable UI than an 80ies style IRC chat.
There is still so much unused potential, I sometimes have the feeling I go mad, seeing it this neglected...
I think you know I like the idea of a sigmoid reward curve as it is getting more flat in the end than n^2 which avoids extreme rewards, but anyway: you are completely right that the linear reward curve makes self-voting just too attractive and should be changed to anything non-linear again!
Even if it is not the exact topic of your article I should add that also the restriction to four fully rewarded posts per day to prevent (self-upvoted) spam posts was very reasonable in my opinion. Who is really writing more than four good posts per day? At least I could never achieve that ... (and yes, yes, I know: if I can't that doesn't mean nobody can ... but I just doubt that many can do that. :-) )
As @timcliff emphasized correctly in one of his recent comments most well intended measures can be circumvented with some efforts but I still think at least it shouldn't be made that easy to exploit the system ...
Sigmoid looks great on a first glance, but at which point should the slope start to decline ? ...
That huge problem aside, it would counter the idea of getting more influence over the rewards, the more popular a post gets.
The 4 post a day limit never applied to comments, so it is not even a question how to circumvent such a limitation for abuse.
Yes, this had to be discussed. I think one could reach a consensus here ...
One could actually also let the sigmoid curve become linear (after x has exceeded a certain value)! You see I am far from having an elaborated solution, but am contributing so many highly interesting thoughts, haha. :-) I am just an ordinary Steemit-user giving the implementation into the hands of professionals who are getting paid for doing so. :-))) (Read that with a twinkle in your eye ...:)
At a certain point, yes. However one could select kind of a 'stretched' sigmoid function, so that one could receive 'quite some' upvotes before the flattening takes place.
One could set a limit for the number of fully rewarded comments as well. In addition one could limit the maximum percentage of voting power which can be used to upvote comments. What about a restriction to (let's say) 15 % of the own voting power for comments?
I fear all those complicated solutions will get in the way of SMTs and other possible applications of STEEM.
Keep it simple.
This is an interesting debate and clearly not an easy one. Might a combination of the techniques maybe work:
Half (or any other percentage) of the reward pool is distributed the way it is done today with linear votes.
The other half is distributed using the n^2 technique.
I personally would also like to see a reward which is not based on SP but per head, but I understand that this might be dangerous due to sybill attacks. However, maybe the option to show posts with the highest number of votes might help as well, as this would show you the posts that post people have liked.
Its the old quantity vs quality, steemit needs both. Not only good content but decent quantity of users. The larger proportion will like me not be that good at blogging and or quickly run out of good contents. So self voting gives thoses like myself incentive via self voting to keep coming back. Also to keep trying, you never know might get a luck whale vote. But in the meantime i can self vote in order to keep a bit of SP and a bit of steem coming in, in return I vote content and use the platform.
Not going to happen.
With linear rewards, whales are better off voting for themselves, just like everybody else.
(Or sell their vote/Rent out their SP: Even more profit)
Until they realize the error of their ways.
Many whales have heftily invested in STEEM (in more ways than one), and once they understand the ramifications of their "feast today, famine tomorrow" dilemma, I'll bet a good number change their tune . . . of those who are actual abusers.
I think one of the biggest issues is that if 90% of the whales change their tune, then the 10% can feast for years while the 90% are upholding the platform. It's the Prisoner's dilemma.
This epiphany you hope these whales will have should have happened like yesterday.
hence lucky, might catch one in a generous mood who throws some crumbs to the mob =).
Very insightful observation ;+)!
Peace.
Yeah , you're right man about the selfvoting, but there is one thing i must think and it is that when i work on a post and i put lot of effort into it , i think it deserves to pe upvoded by myself , it is a way to appreciate myself for the work i do .
Hear, hear!
Keep on steeming
Yeah, but people are naturally biased towards themself. That's one reason I don't vote on myself. That and I don't want to perpetuate the notion that my voice is more important than someone with less stake than me. By affirming bid bots and self voting, the majority has affirmed the opposite proposition. I've heard it said, "Well, I have a higher stake so my comment deserves to be higher up there." Bullshit
Ant this is anoyher sensible topic that needs more discussion. I appreciate that though
I agree with all the cons that came with the HF19, but im not convinced taht going back to the non-linear voting would help good content to get exposure...one of the pros of the lienar voting is that unknown good content can get a somewhat good payout due to only handful of voters
Unknown good content ?
*selfvote to prove hf19 quality curation incentives.
you didnt prove anything at all...you just confirmed that there are cons...you didnt address the matter "good content can get rewwarded if only one wealthy investor spots it" at all:D
I read it all. I do not agree. I think hf19 gives a fair chance for new-comers and beginners. Its hard to get in, and before it was even harder. Quality will prevail, but in order for the whole platform to be popular enough you need to give new players a fair chance.
XD
Edit: selfvote to prove hf19 quality curation incentives.
I agree that limiting self votes is desired. But so many workarounds makes it a hard problem to solve. Sometimes its better not to limit, since only the week suffer.
yeah you tell em @felixxx
nice post sir. we really have to improve content quality evaluation !
i really appreciate your effort for steemit !
and it seems to pay off now.
happy new year. enjoy the "unexpected" price of steem over "5 $" ... ;)
greetings
It wont change anything at all. Whales will still write some crappy texts about nothing and they will stell get shitloads of cash while newbies will gain nothing at all.
It is just like the situation in a world where 1% has 99% of wealth