You are viewing a single comment's thread from:
RE: Improving the Economics of Steem: A Community Proposal
I do not support all three of these things together. While I might be able to support higher curation rewards, I do not see this as the most important element to work on at this time and I do not want to see SteemIt, Inc's dev team pulled off of SMTs and other work that ads appeal to those outside of our community.
Curation is a large account game, it means very little to most of our users.
We had everything in this mix before and it didn't matter. minus a downvote pool. The big accounts just followed Authors they thought would be successful and upvoted them for rewards. There wasn't any curation in it the first go around either.
I think fixing the economics has priority over SMTs at this point. If the economics don't work then who is going to be using Steem for the SMTs to matter?
The issue is that unless we have communities to test these economic theories, no one really knows if what they are doing to 'fix' the economics is going to work.
We've had a community to test this for over a year now (steem has existed longer but I'm talking of current situation with bidbots), and as you can see, there's really not one to speak of.
Community. We need that ies or it doesn't count. When we have hundreds of communities (or fine maybe just dozens) each one can do something different and PROVE their theories. Stabbing in the dark is how we got here, and some people think we are worse off than 2017 in this very comment chain.
Should we go linear? How about super linear? Parabolic? I don't know, but I know that if there are all three at the same time we can test them
We've already tested the current rules and they clearly don't work as well as we need. So we're not stabbing at the dark here at all. But it all depends on how long the SMT's take time to get out, we can't really wait for year or two when we have competitor coming out that's backed by big money in few months only, that's specifically learned from the issues Steem is facing...
We have dropped 30 places on Coinmarketcap in the last two years. We are worse off now than then.
Not only would that require for us to wait for Smts and communities, it'll also require some SMTs to reach a level of recognition, legitimacy and market confidence that would dwarf the underlying Steem token for voting behavior surround that SMT to be indicative of its viability in terms of economic incentives.
This is basically impossible in any reasonable span of time.
Just because we have to speculate over the the economic equilibrium of a change doesn't mean we don't have any intelligent things to say about it.
100% hyperinflation is a bad idea, because in 30 years, the currency increases by 1,000,000 times
n^2 is a bad idea, because someone's who has 1000x your stake will have a vote 1,000,000 your weight
linear and low curation is a bad idea because it leads to content indifferent profit maximization voting behavior (ie self voting and vote selling)
I don't need to test these to tell you they're completely off. I can also give you numbers around around curation rewards, rewards curve, separate downvote pool size, curation curve that are at least not completely off and likely conducive to getting the behavior we want at an acceptable cost.
It's hard to do worse than what we have so what do we have to lose? The status quo doesn't merit caution. Why fear changing it?
That's not factually correct. Curation rewards fundamentally encourage content agnosticism and financial manipulation of voting to maximize rewards extraction IN ADDITION to self voting and vote selling. You can verify this by considering whether any minnow ever gamed curation for rewards, without doing either of the others. Someone might have, but it's so small a financial return for small stakeholders they didn't do it for long.
Curation rewards are only significant incentive to game for financial rewards. They do not encourage anyone to curate content qualitatively, and since author rewards are highly gamable, are essentially merely additional incentive to game author rewards. They don't even matter to folks without substantial stake, and no one with substantial stake is able to curate both for content quality (actual curation) and maximize financial extraction. Folks intent on ROI ignore quality for gamability.
All curation rewards are counterproductive to curative purposes. Even if 99% of rewards were curation rewards, this would not change. Conflating curation with extractive mechanisms replaces content quality with financial return as incentive to vote, no matter what the curation rewards rate.
Boy I guess this is the part I am very unsure of:
I actually think it would be very easy to do worse. That doesnt meanim against everything, actually I am testing 60/40 curation on weedcash.network right now; its an incentivized test net.
The problem with testing our economic equilibrium accurately is that it requires people to actually care about the currency. That means the stake has to command a certain level of real world legitimacy and confidence for the voting behavior to be indicative of anything.
This is why I'm against testing it out with SMTs and other currencies that have yet to be really established.
Smart traders paper trade their systems first, and then find that reality behaves differently. This is why large quant funds do back-testing first and then forward-testing with small allocations before scaling to meaningful size.
I go back and forth on this a lot. I think test nets are important, and SMTs will give an opportunity to test a lot of things. On the other hand, the most important impacts will only occur if the change happens universally across Steem instead of being opt-in.
Long-term, I don't think it all matters that much. SP holding will be about RC's instead of vote value, and most of the value distribution will be SMT's within communities. We do have to actually survive to the long-term.
While I responded negatively to your other comment about the tax, I think you are correct about the overall impact (and need) for these economic changes.
I'm also skeptical that we can survive through to the long term without a reasonably functional economic system in place for our base token. As a pure bandwidth token, many other projects have beaten us to it, and can do what we do faster and better.
I think test nets are better used for technical tests of function than economic behavior. We're interested in the latter which can only happen if people really really cared and are playing with real money. Maybe I'm being insufficiently imaginative with SMTs that can be pegged to certain other currencies etc, but I don't see people caring about them for a long time for voting behavior surrounding them to be a useful indication of anything.
I'm all for SMTs, and I think even if these economic changes are implemented alongside them they'll take no more than a weeks worth of extra dev time.
This.
What would appeal the most to investors is a functioning content discovery and rewards social media platform where there's actually an incentive for people to take part and engage in rather than one that forces stakeholders to defecate all over the front page if they wish to retain their stake.
Imagine how good SMTs or Communites would have to be in order to turn this place around if our economic incentives continue to force people to spam, self vote and sell votes. Realistically, what are the chances of Steem based SMTs taking off to the point where they're not just successful themselves but can carry the failure of the entire ecosystem when Steem is spirally down in CMC charts and Steemit is dropping Alexa ranks because there truly isn't any reason to be on a platform whose front page is a dumpster. SMTs and Communities would need to be impossibly, unfathomably good.
Now imagine how good a more reasonable economic system would only need to be to fix this. It just has to make an intelligent attempt at aligning better rewards with behavior we want, such as people to actually vote based on their subject opinion of a contents appeal. The answer is economic reform just has to be sensible. This is by far the most important and most cost effective change we can make.
It'll only be a small exaggeration to say that @Vandeberg could probably bash out a pretty sound economic system in an afternoon (maybe a week). Yet this would be the one change that would totally turn this place around. Not only that, a functioning content discovery and rewards system would greatly magnify the value of all the other initiatives. SMTs, Communities, Marketing. They won't get us far if our core value proposition is the one thing we're failing the hardest at.
For the record, no engineering efforts are being diverted. This is just the start of an important discussion
"Imagine how good SMTs or Communites would have to be in order to turn this place around if our economic incentives continue to force people to spam, self vote and sell votes. Realistically, what are the chances of Steem based SMTs taking off to the point where they're not just successful themselves but can carry the failure of the entire ecosystem when Steem is spirally down in CMC charts and Steemit is dropping Alexa ranks because there truly isn't any reason to be on a platform whose front page is a dumpster. SMTs and Communities would need to be impossibly, unfathomably good."
Those words are gold. I actually starting to question whether Steem has to compete with other social networks. They do know blockchain exist and will do their best to stay in the trend. The worst part is that people don't really like being on all of those platforms. They fight for those that deliver the best features and unite people together to share unique and amazing content. Steemit was those things in the beginning, then it faded somehow but why? I'm still trying to figure it out. It seems like the level of hype is correlated with the number of quality posts which is weird and shouldn't be like that.
I believe Steemit as a website and company is missing very critical point and that is strategic marketing. I've never seen Steemit doing that and that could be a problem. People outside of crypto are not aware it exists unfortunately and that's not cool.
Those things are pretty obvious but were not addressed yet so decided to share. Thanks for starting this conversation. I know you create lots of cool posts so I know how you feel to produce great content and simply having 0 feedback because people are not on this platform. There are lots of people playing games built on Steem and that is an advantage. So maybe that what Steem in particular has to be focusing on? Your thoughts?
In fact it is quality posts that market Steem. Good content attracts eyeballs, and that attracts those eyeballs to the platform, where those that like it can invest in Steem. Curation and the rewards it delivers is intended to encourage creating good content, and is thus the strategic marketing mechanism for Steem.
More mechanisms besides blog posts are being created to market Steem, but like DLive and now Drugwars, they aren't loyal to the platform. Few mechanisms have proved to be as powerful at creating value as social media, and all that is necessary to grasp that fact is a look at the growth of the FAANGS in the last decade.
Thank you @trafalgar. Im a steem newbie of just 16 months and I have earned every steem I have, rather than buying in. I curate manually for both @ecotrain & @freedomtribe. This EIP excites & encourages me - thank you! Im a single mom living in Chiang Mai, Thailand & MAKING SURE to get a flight to BKK for SF4. Hope to thank you in person for all you do.
Is reasonable development that imbues the investment vehicle with increased value, producing capital gains.
Content isn't their focus. ROI is. Presently ROI is able to be extracted immediately from rewards, and cash is king. That needs to end, and ROI enabled from funding development that improves and imbues value in the investment vehicle, Steem, needs to replace it.
To listen to the audio version of this article click on the play image.
Brought to you by @tts. If you find it useful please consider upvoting this reply.
Anyone that weighs in on this post without actually digesting @vandeberg's latest post might be making a huge mistake.
This is the whole argument. I'm curious to what you really think of it. And no, not if Traf and his likes will profit or not.
I agree with this comment! Read Vanderburg's post also
That quote is factually incorrect. Code is infinitely mutable, and all financial incentive to corrupt curation can be eliminated. While some people will continue to act irrationally against their financial interests, currently it's irrational to not degrade curation for financial gain. Eliminating, rather than merely tweaking that incentive, can be done, and I propose it below in reply to the OP.
Traf doesn't want the vast majority of active stake taking part in mindless self voting or vote selling just to get staking returns, because that's seen as the norm now.
Traf can't afford/is unwilling to not do this to defend his own stake, but would much prefer a system where everyone is incentivized to act honestly with respect to voting behavior.
That's what these suggestions are designed to do
These proposals do not eliminate incentive to extract rewards from corrupting curation at all. They only tweak the returns a little.
Those incentives can - and should be - eliminated altogether.
I honestly think the only way to fix it at this point is to make actual curation more profitable than self voting, which requires a return to exponential curation curve and around a 50/50 split. I honestly don't know if anything will fix this shit at this point because like you said voting collusion trumps content quality for passive investment, but the fact that we literally have to build a second layer protocol to disregard the distribution of the first layer protocol because it is so fucked is simply hilarious to me. We've gone full circle to the point where we have no value proposition over someone creating their dapp and community on tron, eos, neo, or countless other blockchains. Three years of running a company like a personal piggy bank and then lobbing a hail mary by allowing people to circumvent the distribution is a joke at best.
This does not even impact curation for content quality at all. It merely changes the financial equations governing how profitable corrupting curation is.
We can eliminate financial incentive to corrupt curation altogether. We should.
Omg, this hurt to read due to the truth in building a second layer protocol to fix the first layer. I had never thought of it that clearly...
It's painful
This is a good idea in theory but in reality whales eventualy just vote for the same people because they think other whales will as well, in an attempt to maximize profits from curation. This ends up a few lucky authors getting most of the rewards as everyone piles in to get curation. We saw this happen when this platform first launched. In my opinion they are trying to solve a problem that may not be fundamentally solvable. I don't think stake based voting can work. Sounded good, but when money became involved, it changed everything.
Plus the post literally says...
Are people deliberately skipping this part?
I'd say most of us that have been around since the beginning realize that it's a bit "too little too late" considering the middle class on STEEM has basically been taken out to pasture and there's not much left besides mega-whales and plankton. If they made these changes years ago when plenty of us were asking for them and telling them explicitly that this (the situation we now find ourselves in) is where it leads, the warnings were not headed. You can't wait for a house to burn down and then realize you need to call the fire department if you want a chance to save anything of value inside.
We actually proposed something very similar 2 years ago as well:
https://steemit.com/steem/@steemitblog/details-on-proposed-comment-reward-curve
I hear you, I spent way too much time and energy trying to preach to people that their greed was destroying a potentially world changing technology, but none of them gave a shit as long as they could get theirs. There's no doubt why people left and it's not just the price. Whale games are whale games and if you didn't get in first or fuck people over for stake you ain't getting shit out of this place. Unless you live in Nigeria, Venezuela, or the Philippines (or somewhere equally impoverished) this technology isn't changing your life and empowering content creators. The greatest lie ever told on this platform was "bid-bots are for promotion." Why the hell would anyone come here and promote to a community of a couple thousand people that are pretty much smart enough to avoid the trending page unless they had no following to begin with. I literally don't even care to rant about it anymore, but I'm right there with you, we tried to stop this shit and the people that had the power to change it did what was in their best interest, just like they always do, and got rich off the illusion we (the community) were pedaling for them.
I hear you, but the problem isn't precisely greed, it's designing an economic system that's more resistant to game theoretical pitfalls like the prisoners dilemma or tragedy of the commons.
It's not that I don't understand if we all stopped and voted honestly, we're all better off, but individually I know that if i did that, everyone else would still just keep milking. So the options to me are join the milking or abstain and lose your share but not make any material difference to the failure of the platform overall. I bet many other stakeholders are in the same boat.
Good news is different economic incentives vary in how resilient they are against these game theoretical pitfalls. This is fixable :) I finally got Inc to listen to me, so that's a great start
@kevinwong said you were a smart guy. He"s right!
Posted using Partiko Android
I'm supporting a 50/50 split
Exponential curve is likely too strong because n^2 means someone who has 100 times more SP than you has a vote that carries 10,000 more weight than you. Now that people are more sophisticated, this perhaps would open up to even more abuse than the current system
But indeed some level of superliniearity is necessary. I like the convergent linear curve proposed by vandeberg as far back as 3 years ago. It has a superlinear head that forces all profitable spam into the light, and a linear tail so no collusive piling on between whales/bid bots
I also think a moderate amount of free downvotes are necessary. Basically every measure helps, but every measure has their own downsides if you tune them up too much. This is mostly an optimization problem.
I broadly agree with your points and think it's better late than never. I share your frustration as I've been proposing this for over a year, and I'm grateful that recently @justinw, @andrarchy @vandeberg and other inc members became receptive to these ideas
Sincerely, I hope it's not too late, but after years of being ignored and no delivery of promised updates it's hard to be enthusiastic about anything at this point. I guess time will tell. I agree we never need to revisit the old exponential curve, my thoughts were more slightly exponential, but I think the vandeberg model is fairly close to what I was thinking.
Yes, yourself, van and me are no foreigners to feeling that we have an answer but being told to go beat it for long periods of time
But now that we have Steemit Incs attention, it remains our best play. Incentivize the behavior we want. Each measure that does this (curation, free downvotes, superlinear) have their own drawbacks as you increase their intensity. So it's an opitmization problem of maximizing benefits - costs.
I believe this is a broadly fixable problem and yes, i feel it's overdue by a couple of years, but it might not be too late to turn this ship around
Isn't that behaviour development that imbues the underlying investment vehicle with greater value, creating capital gains? Let's do that instead of corrupting curation with ANY financial incentive through which funds intended to market Steem are instead gamed for profit.
Also - more curation rewards means delegation to curation pools (of which bit bots are one type) would become more profitable, not less. It is not immediately clear to me that the raising curation reward amount would have the desired effect, in fact the opposite seems very possible.
Precisely! It's very simple mathematics.
I shall digest everything later, as need to go to work. To look closely at the whole economic system is something I have said for over a year, but the 3 proposals are, yet again, tinkering with the most obvious parameters and ignoring some of the deeper problems within the core code. Returning to super-linear is a whale-feast, for example - what's the point in that? What are the newbies going to think? Gotta dash.... more later.... probably much more :-)
And you are right it wasn't perfect earlier on but it was certainly better back in 2017 than it is now in 2019.
I disagree outside of having more attention it didn't work any better except for a handful. A very small handful.
However, I do appreciate the opinion though.
I support the changes in this post, but I would also not agree necessarily that it was "better" in 2017. It was certainly "different" but there were other problems which were still very serious.
Since there is no suggestion to go back to the previous rules, I don't think it actualy matters. You have to evaluate this new proposed set of rules on its own merits, as they are significantly different from both the current rules and the 2017 rules.
At least in 2017 there was the incentive to create good content. Sure not all good content was being discovered but at least some was. Now there is no incentive. So yes in 2017 at least some content discovery and incentive existed even if flawed.
Now it's just bots and basically pay for popularity. This is also called pay to win.
And yes the current rules differ from 2017, but anything is better than to do nothing with the current proven broken economics.
Much of the problems I feel could be addressed by a better "pay for popularity" system that is part of the actual platform. The promoted tab failed in this and bidbots took its place. If we were to give some real thought into building an advertising economy into the platform, we could pull in top bloggers and vloggers and provide facilities to pay for popularity without diverting the whole reward pool through the bidbot economy.
If I had to score them out of 10
Hyperinflation - 1/10
n^2 - 2.5/10 (this appeared to have worked better because people weren't as sophisticated back then and abit and smooth were incurring some of the costs of making it work better than it otherwise would have)
our current system - 1.5/10
I think we can do maybe a 5.5-6/10 on our next attempt if we frame the problem correctly and identify both the benefits and costs of each of those measures to reach sensible numbers.
That was only for a relatively short time in order to demonstrate some of the negative effects of unrestrained n^2.
For most of the time it didn't really work at all, and people claiming that it did either weren't around at the time or were out of the loop and didn't know what was actually going on.
I agree @whatsup!
Check this idea out. Limit a person to 1 vote for 72 hours for each particular account.
This will make more people curate others as they can’t keep upvoting the same person all the time.
This will also tame bidbots, selfvoting and whale circle jerkin’ Action 😂
it's easy to circumvent by creating multiple accounts etc.