Understanding Steem's Economic Flaw, Its Effects on the Network, and How to Fix It.

in #steem6 years ago (edited)

pexels-photo-1133504.png

I'm sure most users on this network have seen the same "vote-farming" and flagging dramas on Steem over and over again for the past year. It's not hard to see that a lot of users are doing this "vote-farming" thing, including the largest stakeholders on this platform. It just means the following under Steem's current economic model:-

  • Self-voting
  • Vote-selling
  • Vote-exchanging

Please keep in mind that all of these are economically equivalent. For example, I can sell my votes by delegating to a bidbot, or I can also achieve the same effect by using all my votes on my posts only, or give someone $X amount of votes to receive the same amount back. All of these actions are the same: vote-farming. Vote-farming at 100% just means using 100% of our votes only on ourselves. This is indeed the best move under the current economic model, which is fundamentally flawed, and a lot of users including whales are actually participating in it.

Anyone not participating in this activity will just be losing out BIG TIME over the months and years. Like I've said in my previous post, someone with the same amount of Steem Power (SP) as me in the beginning of this year is now already ahead by 50,000 SP just selling their votes, while my votes mainly went to curation while taking my time to post. So I'm not holding back as much anymore, and I'm playing the vote-farming game on my alt-account @etherpunk to keep up until Steemit Inc and the top witnesses decide that some of us are right about our diagnosis and proposal to fix the misaligned economy. An economic system is flawed when it relies on the altruism and sacrifice of some users.

The path of profit maximization is again, vote-farming at 100%. This means maximum profit with the least effort. If you think there's a risk of people flagging, sure it may happen. But over time, you'll come to understand that Steem's current economic incentives just makes flagging arbitrary, futile, and ultimately meaningless. Nobody will even bother flagging anything else other than outright fraud, theft, or just plain retaliation. You can put up 10 random one-sentence original posts a day and just vote yourself up all the time while even buying some extra votes for the ROI too. It's very likely that nobody will stop you these days. Even if someone does, they'd most likely be unable to sustain their flagging on your account over a long period of time, unless you're just pure evil. Bidbots also generally do not have high standards. How else are they going to remain competitive?

It's now a move that I actually encourage everybody to do. Vote-farming at 100%. Instead of selling votes, you can also avoid the middleman and just vote on your post all the time for the same effect with more profits. Now you might think: alright, so everybody just minds their business and do their vote-farming at 100%. Under this scenario, what's the actual point of the ~8% annual inflation on Steem which just goes back to the pockets of the voters? This is why I'm saying that anti-abuse initiatives are contradictory under the current economic model. We're fighting against the tide of incentivized spam and abuse. My as well just set Steem at 1% annual inflation without reward pool just to keep the network running, all while people post whatever they want as usual without voting.

The point is: Steem's economic incentives should be aligned to encourage allocation of inflation to contributions that benefit the platform. While it's impossible to design anything that encourages all SP holders to allocate 100% of the inflation to other contributors other than themselves, we can go for about 50%.

This is the proposal. The best way out of this situation is likely to:-

  • Increase curation rewards (50% is a good start)
  • Move away from pure linear and into something like n^1.3
  • Subsidize downvotes (cheaper downvotes, separated pool)

Doing so will make curation more likely as the economic incentives from doing so will have a better chance to outperform "vote-farming". This "hack" is really about giving people half of every votes given out under a curation game. Some of us have been saying this and repeating the same arguments for well over half a year, but not many seemed convinced at all. I hardly expect anyone who have only put in less than 20 minutes of thought is seeing the problem and solution clearly. But that's generally all the time most people have these days. Some of us have put in dozens if not hundreds of hours just refining our thoughts about this one single problem.

Some words about bidbots: the proposal above will simultaneously make the pricing for votes much less predictable than in a linear economy. It won’t kill the vote bidding business but just introduce some extra risk (which there’s essentially zero at the moment) which would end up being more like promotion with less ROI, so it wouldn’t serve as a bypass to the act of holding SP for more earning power. Yes, there's a leak here that needs to be patched up. More profitable curation services should also emerge after the proposed changes.

The proposal has non-trivial costs, but as @trafalgar says it, nothing is as costly as a completely broken content quality indifferent economic system for social media outlets. Remember there's nothing wrong with profit maximizing behavior, one should expect it. The problem is when we don't align it with the behavior we want that's best for the entire system. To further repeat the point, I'll copy and paste @trafalgar's words here as I think he generally explains the situation much better than I do:-

We currently have a system where stakeholders are incentivized to engage in content indifferent voting behavior (vote selling/self voting) to maximize returns. This completely defeats the protocol’s ability to foster engaging platforms that revolve around content discovery.

We need to move away from an economic system that makes it too costly to vote on work that stakeholders actually deem valuable if we want this place to survive.

  1. Upping curation rewards is an obvious answer, but by itself, it is easily circumvented. Bid bots can just offer to share most of the curation rewards. Self voters can spam hundreds of posts and just upvote the ones that were least voted on by others after 15 minutes to avoid others 'stealing' their rewards.

  2. Non linear should be seriously considered. Having reward amounts/SP that differ per post depending on its popularity is the core component of moving away from content indifferent voting behavior. Yes it has a cost, but the n^2 we were use to was far too steep. At n^1.3 (or any crude approximation of this), we should get most of the benefits at the lowest cost.

  3. Downvotes incentives should also be encouraged. Of course greater downvote incentives has downsides, but in combination with 1 and 2 above it should allow for curation rewards to out compete mindless vote selling/self voting.

Content indifferent behavior is the bane of this system. No social media platform can survive if it's unable to at least sort exposure and rewards by some subjective standard of quality.

Note that the 3 items in our proposal have been considered very carefully to cover all the angles so we can get out of the situation posed by Steem's current pure linear economy with low curation rewards and costly downvotes. Just implementing one item is likely not enough. We need 3 of them. But of course, we only want to achieve an intended effect, so any better implementation suggestions are welcome.

Also, suggestions like passive staking for investor-types so they don't influence the content voting and ranking, etc are flawed. The same problems will still happen since we'd still be under a pure linear economy with low curation rewards and costly downvotes. We'd just be reducing the reward pool here to be reserved for passive staking. The shrunken reward pool will then remain the be used in the same way, which is vote-farming, instead of content and curation.

It's my wish that more community members around here take notice about this problem and solution. If you're convinced, please make some noise about it. I've also written this about this a few times earlier this year. Some of us even wrote about it more than a year ago. The arguments and proposal remain to be in the same vein.

Please save Steem from all the unnecessary suffering and restore our sanity. We just need more users to stop running around headless, stop getting involved in senseless witchhunts under a flawed economic system, and instead seriously consider our specific problem statement and solution here. Some of us are really surprised that nobody at the top is onto this matter like an eagle over the past year.

It's a mistake to pin Steem's problems on culture and greed. Altruism and selfishness should be one and the same, which is what our proposal seeks to achieve in providing voters ample returns if they play the curation game instead of mindless vote-farming. The correct move then is in designing economic incentives to align profit maximization with activities that are good for the platform instead. It's as simple as that!

As always, thanks for reading.

Note: I want to make it clear there's nothing wrong with self-voting or selling your votes. It just has the potential to hurt the platform when there are no checks and balances in a flawed economy.


You can vote on your favorite witnesses here: https://steemit.com/~witnesses. I'm also running as one. If you trust in my choices for other witnesses, you can also set me as your voting proxy on that page.

Sort:  
Loading...

First of all I want to say that I am very glad for the work being done by @kevinwong and @trafalgar to enable participation on the current Steem economics, while also pushing solutions forward. I very well agree with most of the points brought up here. The increase in curation rewards and downvotes incentives are particularly significant, as we tried several attempts to overcome those in our product development. The lack of proper curation incentives drove us to tricky implementations which we had to tweak continuously in the way we reward our curators/moderators. The lack of downvotes incentives made us think of complete different ways to sort our contents, compared to what the blockchain can offer. Built-in solutions, like the ones aligned here, are ideal and could lower the hassle significantly.

I am collecting inputs, like this one, trying to have an overview of the hottest community requirements. Utopian supports the maintenance of Steem forks (see @reggaemuffin ' s one) which enable proposals and testing of disruptive changes. It makes sense to bring such inputs in the right places (github issues) and incentivise the development, via community participation.

Thanks a lot for taking this up @elear. Much appreciated. It would seem like a great time to be setting up extraneous protocols to get important discussions going on with good info, and tying it along with development / testnet

The lack of proper curation incentives drove us to tricky implementations which we had to tweak continuously in the way we reward our curators/moderators.

Can you elaborate on this? Do you mean to say that if curation rewards were higher, you would spent less on incentives for your curators?

The lack of downvotes incentives made us think of complete different ways to sort our contents, compared to what the blockchain can offer.

As you clearly agree with the points in the post, could you give more details on your comment? I feel that your words are too broad and don't know what to take from them.

There is an entirely different way to supplement the voting pool while implementing restrictions for self voting and vote selling. If the voting pool is reduced, there would little to no incentive to use the platform. After all it is slow, clunky and has a very small population with respect to other social platforms. Better rewards with incentives to do good or restrictions for bad actors could be implemented. But where would we get the extra revenue to do this?

Here is the idea:
Start a centralized exchange to complete with the likes of Binance, UpBit, Coinbase, Poloniex etc. yet have STEEM and SBD as the common trading pairs. If Steemit Inc were to create and run it, they can generate massive revenue to supplement their Software Engineers. Excess revenue can go toward supplementing the voting pool or towards rewarding good actors and good witnesses. There is no problem getting support and usership as All major exchanges can’t resync their nodes in a timely manor. People would use the new centralized exchange as they can’t use Binance or Poloniex. By the way both Binance and Poloniex STEEM wallets have down for a month. This isn’t the first time for Binance either. Their Steem wallets have been down for a full month before. Poloniex Steem wallets have been down for 6 months in recent past.

But get this, an exchange with all the major cryptocurrencies traded with STEEM and SBD can be created by anyone with the right technical knowledge. It doesn’t require Steemit Inc to do it. Whales could even pool their resources to fund startup costs They won’t need to self vote as the exchange will be so much more profitable.

I don't disagree with any points made here, I think its been obvious to anyone paying attention for way over a year.
Though I would have one major statement to make and one question.
There is no way you will keep profit seeking gamers away from working the system. It takes to long to implement hard forks and is to easy for a gamer to come up with work arounds. The problem is now how Steem works but how PEOPLE work.
This is never going to be fixed by trying to create rules that MAKE people do what we want... If anything it will only make them leave. Then the price of steem will crash and everyone will still be complaining as the rewards will be even worse than they were with people uprooting them selfs.

So the REAL question here is of which everyone actually needs to be asking. How do we provide a quality return (which means, reliable, easy and high enough %) on the investment whales make?

This is the answer. If we can find a way that is better for the over all eco system yet still provides a reliable, consistent and profitable return then its all taken care of.

Almost no investor is going to invest in a platform not to make money. Also most whales do NOT have the time nor desire to curate. Plus higher curation rewards wouldn't change anything anyways. Just use curation bots. Easy peasy.

Glad everyone is thinking about this, but simple economics shows that its the ones with money who hold the influence. So have to find a way to GIVE them INCENTIVE to move in the direction the community wants.

SteemOn!

Hey thanks for dropping by @quinneaker, I think it just takes a lot of repetition and refinement in arguments to get the points through. There's a bit of momentum this time around as ned has considered it in the comment section of his recent post.

Check out @trafalgar's response above. He's answered your concerns very well!

Yes, all progress takes vast amounts of collective discussion, meditation, pondering and even arguing. Thats why this is great. Though with that being said my statements/questions still seem to be the key in that process. I read a LOT of comments including many from @trafalgar but did not see one that answered clearly and specifically my questions/statements. I can look again as its a critical issue.
Best regards and SteemOn!

indeed,most of big stakeholder of sp are self voting on their own post,but we can't blame them if they run and act like that on the steem blockchain.

Because they are investing and spend their own money to raise their SP to do some propits of it's own post.

But still we are glad that some community giving a free curations specially to minnows and new comers with good article and writing.

like @curie , @acidyo specially you and to other steemians who support those good author..

Happy to know it also that you finally run your own thread and being one of my new witnesses @kevinwong

If you don't mind can i know you witnesses too! because i have special witness that i supported if you my 2 witness count's on you i'll take your proxy if not i give you a vote alone.. :))

The past 2 weeks have been brutal for minnows. Our post payout are decreasing by the hour and by the end of 7 days it's probably gone down at least 15%.

Due to this factor, many have chosen to leave and not waste anymore time on STEEM. Out of 15 friends I've gotten on to STEEM, only 4 remain and only 2 of the 4 are active everyday posting contents and engaging with other people.

So I think it is a huge problem from the way the way things are now and I don't think STEEM will achieve it's potential we all hope for. Even when SMT starts, I don't think it'll make much of a difference. More stuff on a broken system doesn't fix anything.

Flagging content is all we can do and as minnows, nobody actually wants to flag because of the fear that we'll get flag too. And whale flags are obviously way more powerful than minnows. I personally think flagging can solve this problem as only "content with effort" (like this post) will be valued. However seems like minnows doesn't see this as their problem. They see it as "the whales will sort it out" problem. I think everyone has to take ownership of this and be more responsible on this whole issue or else we're doomed.

Anyone not participating in this activity will just be losing out BIG TIME over the months and years. Like I've said in my previous post, someone with the same amount of Steem Power (SP) as me in the beginning of this year is now already ahead by 50,000 SP just selling their votes, while my votes mainly went to curation while taking my time to post. So I'm not holding back as much anymore, and I'm playing the vote-farming game on my alt-account @etherpunk to keep up until Steemit Inc and the top witnesses decide that some of us are right about our diagnosis and proposal to fix the misaligned economy. An economic system is flawed when it relies on the altruism and sacrifice of some users.

From since I've known you which is January, you have not done what most have done which is self-voting and that's like 10 months. You've waited long enough for something to happen and have been vocal about this issue. I'm with you in this even though I'm directly affected by vote farming.

I'm just praying this changes happens sooner rather than later. Bless ya man!

Yes, the flagging is not effective over the long-term mostly because of linear. When everyone can reach for the nearest switch for max profit on nothing to do with content, there'll be too many people doing it. If you're not joining, you're losing out over time. If you're flagging, you're losing out over time too.

Edit: Flagging should be contingent though. It has some social cost attached to it. That's why it's better to have an aligned economy and just a small amount of free downvotes.

Yeah that way of flagging would make much more sense. If flagging is costing me my upvotes, why not just upvote someone else or myself? If this method of flagging was implemented, I think it would solve a lot of problems with this issue of rewards and upvotes.

You summed it up perfectly Kevin. It's why I have delegated a lot of my SP to a bidbot. I regret not doing it a long time ago, as I was just wasting the earnings potential of my SP. You're also right that this is a big issue and generally symptomatic of the broken economics that have plagued steem, for a long time. I am not certain what the correct equation is for effect of stake on the reward pool, but it's very clear it's not linear. (Though this was clear when linear was implemented from the original mechanism).

Hopefully you can use your influence to get more people onboard with creating a fix. It doesn't even need to be the optimal fix (as we may not know what that is until further experimentation), but at least some iterative progression would be nice.

@elear of Utopian is helping coordinate this matter, hopefully getting the ball rolling etc. The other side effect of the current economy is that nobody cares. Every second you care, the more you lose out lol. I hope the top witnesses are compensated well enough to do so. I wonder if I should make another move to only support witnesses that prioritise what some of us think is the most urgent matter that doesn’t require immense R&D to fix.

Every second you care, the more you lose out lol.

Sums up my entire experience with this platform.

@elear of Utopian is helping coordinate this matter

Good to hear that Elear is also on board.

I wonder if I should make another move to only support witnesses that prioritize what some of us think is the most urgent matter that doesn’t require immense R&D to fix.

Only you can tell you what to do with your stake. :)

Every second you care, the more you lose out lol.

Sums up my entire experience with this platform.

From a human standpoint, it's always worth caring. Whether this results in money or recognition, however, is another matter. In it, everyone is himself the best psychoanalyst, how he deals with a situation that depresses him or does not bring him enough profit. The question was always "How much is good enough?" and will always be the same. But one has to face it and not on the surface. To do this, one has to go deep and deeper into what is called self-knowledge.

To sprinkle in a little bit of philosophy here.

;-)

It's a solid proposal and worth a big vote. I don't necessarily agree with all parts of it, but I like your contribution to the discussion.

First, I think passive investors should be paid a return to keep them from mucking with the content through either posting or voting. If the rewards must go through the system, then higher curation rewards certainly would be preferable to expecting people to post junk for reward votes. But it still perverts things somewhat. I'd support it if there's no realistic option to simply pay passive investors some return, but I'd much rather pay them to stay on the sidelines and not feel like they need to post or vote on junk.

As for the downvotes, I think it could be abused unless the rep score accounts for voting behavior somehow. I'm not against your suggestion, but I'd have to think about it more, and I suspect it would need another component to be successful.

As for superlinear rewards, I just don't see it. I still think that would be a mistake because we've been there and seen that things work better with linear. Superlinear is a faster route to more concentration of wealth and power, which is the wrong direction. I think believing that superlinear would work also assumes that larger SP holders would vote for the best content, which is not necessarily the case, though I do appreciate that you are trying to find the right economic balance with curation rewards also. I'm probably one of the people who would benefit from superlinear, but I don't know that it's in the platform's best interest.

Would it be possible to fork some testnet where various ideas like these could be tested?

I've found a place for passive investors that doesn't involve messing with the natural flow of content production or consumption. I explain in detail within this post:
https://steemit.com/steemit/@nonameslefttouse/turning-rage-into-the-soothing-steem-now-blowing-out-my-butt-as-i-mince-my-words-for-you-the-reader-of-this
the common struggle someone like me faces as a content producer here after all these years and I offer a solution that could easily fit in with what @kevinwong talks about here. The post leads to other posts with more information. The comment sections are jammed with even more. People are scratching their heads wondering how to bring new money into the platform and I'm clearly stating the fact this is the entertainment industry and a functional promotional system using ads is one way to bring this new money in. I've explained how, but I need more people to see it.

Yes Utopian is working to setup a definitive place to study the proposal and read opinions about it. @reggaemuffin seems to be handling a testnet so it’s likely we can test these things.

First, I think passive investors should be paid a return to keep them from mucking with the content through either posting or voting. If the rewards must go through the system, then higher curation rewards certainly would be preferable to expecting people to post junk for reward votes. But it still perverts things somewhat. I'd support it if there's no realistic option to simply pay passive investors some return, but I'd much rather pay them to stay on the sidelines and not feel like they need to post or vote on junk.

The same problem is going to persist unfortunately, as we’re only effectively shrinking the rewards pool for the passive staking allocation, while the economic incentives for the rest of the users remain the same.

As for superlinear rewards, I just don't see it. I still think that would be a mistake because we've been there and seen that things work better with linear. Superlinear is a faster route to more concentration of wealth and power, which is the wrong direction. I think believing that superlinear would work also assumes that larger SP holders would vote for the best content, which is not necessarily the case, though I do appreciate that you are trying to find the right economic balance with curation rewards also. I'm probably one of the people who would benefit from superlinear, but I don't know that it's in the platform's best interest.

Previously we were at superlinear (n^2). Today we are at linear (n). The proposal suggests somewhere in between (n^1.3) which should be more balanced while making it less predictable for people to put a price on votes, providing variance/contrast, and abuses would only be possible by piling up votes for a more substantial payout so its more manageable for the community.

As for the downvotes, I think it could be abused unless the rep score accounts for voting behavior somehow. I'm not against your suggestion, but I'd have to think about it more, and I suspect it would need another component to be successful.

Cheaper downvotes is more for contingency as most wouldn’t practice it anyway. It just makes it easier to take down abuses or overvalued posts from any gaming. All said, there’s some social cost to cheaper downvotes, but I think the benefits outweigh the downsides of this proposal.

Superlinear is a faster route to more concentration of wealth and power, which is the wrong direction.

I would prefer a curve which started as n^2 / exponential (thus flat), and then later changed into linear which would work against self-voting as well as excessive rewards.

@clayop had a similar idea.

I appreciate your initiative bringing back some movement in the rewards discussion. A higher curation reward share and free flags both sound good to me.

Although I share @whatsup's concern that moving back to a superlinear curve will not incentivize voting on better content but on known successful authors, I also feel like the overall situation can't get worse than it is now. After the last big changes to the reward system we said we'll have to test and change more, so let's do that!

I'm also not convinced that paying stakers an automated return wouldn't improve the content rating - as long as they get the same they'd get with vote farming. Why should anyone put in the work if they can make the same or more without. What I can see though is a risk that the reward pool would shrink to a small fraction of what it is today. Definitely a lot bigger change than your proposal, so I'll support that for now.

Hi thanks for dropping by @pharesim!

Although I share @whatsup's concern that moving back to a superlinear curve will not incentivize voting on better content but on known successful authors, I also feel like the overall situation can't get worse than it is now.

Curation services will emerge out of the proposal and we'll have the best of both worlds (something like n^1.3, which is between n and n^2), just need sometime for it to play out until it reaches the new economic equilibrium. Cheaper downvotes are a contingency and coupled with n^1.3, we'd have a manageable playing field for community self-reg.

I'm also not convinced that paying stakers an automated return wouldn't improve the content rating - as long as they get the same they'd get with vote farming. Why should anyone put in the work if they can make the same or more without.

With the proposed changes, passive staking could work (or else we just have a shrunken rewards pool under the same econs), although stakeholders will most likely delegate to curation services which should more or else bring about the same returns as bidbots.

Well, probably everything has been said by now. Flags aren't free, they have a social cost. And a big issue in the reward economics is that we force people to participate in something they don't care about. Those will always go the easiest route to make their profit, which under superlinear means automation towards successful authors. With n^1.x less than in the old days, but probably still noticeable.
The reward economics would drastically change when staking is more profitable than vote farming, the assumption that it'd stay the same doesn't make sense to me. There'd be no more bid bots and self voting, as those looking to maximize profits wouldn't participate in that econ any more.
But that's running in circles with assumptions on both sides, no way to tell with certainty without testing in the live environment.
Anything that makes bidbots less profitable is good, and that's something your suggestion will definitely achieve!

I wrote a post 2 years ago and the conclusion is a linear reward distribution mechanism depends on downvoters to succeed if possible. IMHO efforts should be made towards finding and incentivizing good downvoters. What I have done and am still doing is in this direction, see @adm and the old downvoting event. However, encouraging downvoting is hard and can easily be abused if designed poorly; it's also controversial and comes with costs since most people don't understand it, E.G. due to downvoting my witness has been voted down from top 20 long before. I had another post analyzed curation rewards distribution mechanism and perhaps helpful for you. Good luck.

what's yoyow doing in terms of voting economics?

The yoyow blockchain doesn't have voting right now. I don't know whether nor when it will be there. 3rd-party platforms running on yoyow e.g. biask.org have off-chain voting but I don't know the details.

What do you think about a 10% separate downvote pool @abit? Linear does give everyone the ability to immediately gain the maximum returns per SP exercised being entirely content indifferent, so it does seem like an unmanageable field for downvoting activities. And thanks for the links! Will check them out!

Sorry, I don't know whether a downvote poll will work, or how it will work. Who can use the energy in the pool? How to prevent it from being abused?

Downvoting is a really bad idea, it destroys social bonds.

Just saying. "Punishing" bad content should be done on the social level like in the tight-knit human communities of earlier ages when people "behaved" because they didn't want to become pariahs in their communities. The threat of "not getting the community upvotes" anymore (because of bad behaviour) is much more effective and less damaging than outright downvoting

Coin Marketplace

STEEM 0.23
TRX 0.22
JST 0.036
BTC 98531.65
ETH 3359.79
USDT 1.00
SBD 3.16