Make Steemit Great Again: Fork This Place!
Steemit, Inc. has another hard fork proposal near completion...and I have a proposal of my own. Let’s see how they stack up.
If you missed it, the original Hard Fork 17 proposal was released on January 10th. There were a bunch of changes stuffed into it, just like the last hard fork. One of the mantras of the post was a focus on simplicity – Keep It Simple Stupid, or K.I.S.S.
Unfortunately, it was obvious to most readers that some of the proposal was in fact making things more complex, particularly for new users who already have an issue with the STEEM/Steemit learning curve. I happen to think that some of the changes actually won’t resolve any of the underlying issues with STEEM/Steemit. Because of that, I have decided to make my own Hard Fork 17 proposal.
This is not an all-encompassing proposal – and hard forks shouldn’t be – but I think these ideas work well together and that they make sense for one fork.
My STEEM/Steemit Hard Fork 17 Proposal
These are a few changes to the code that I believe will improve many different aspects of the platform. I have stated some of these before in various comments, chat channels, or blog posts, but I want to lay them out here so that they can be considered together. If Steemit, Inc. can shove 20 different changes into each hard fork, then I think I can reasonably put a few of my own ideas into this proposal and have them considered as a whole. It’s only fair, right?
Here is my plan that simplifies rewards and increases incentives for non-bloggers, which is a vital aspect for STEEM Power accumulation and proper curating of content.
1 – Change the voting algorithm. This was mentioned in a few different posts lately (including my own). There is a problem with stake disparity and it affects the allocation of rewards on posts. The n^2 algorithm exacerbates these disparities and it should be replaced with something closer to linear, such as the n log(n) or the n^2 / (1 + n) algorithms that were mentioned in the recent post from @steemitblog. Changing the algorithm is something that most users seem to agree on and that is pretty rare to see. There’s no reason why it can’t be tried.
2 – Treat blog posts and comments the same. To accompany the change of the voting algorithm, parent and child posts should be treated exactly the same by the code. If there really is an issue with post engagement and it is being caused by a lack of incentive to vote on comments (even if this isn’t true, which I don’t believe it is), then treating all posts the same could increase the visibility/incentive to vote for comments and it would likely increase the average payout on those comments, especially if there was an accompanying tab for top comments in the Steemit interface. And there would be no need for a separate rewards pool to negatively impact intuitiveness and the already steep learning curve for new users.
3 – Increase curation rewards back to a 50/50 split with post rewards. In conjunction with 1 and 2, an increased incentive to vote for all posts should result in more voting power being used overall and a more equitable distribution of rewards to and from curators. Currently, post and curation rewards are split 75/25 in favor of posts. In reality, due to the reverse auction on early curating, the actual curation payouts are only around 12%. The rest of the 88% of rewards goes to authors. This throws the entire incentive structure of the platform out of balance and is the main reason why there is increased automation. Voters simply have very little incentive to curate manually and to curate better.
Curation rewards also happen to be practically the only reason to hold STEEM Power. Every time that incentive is reduced, the more likely it is that more users will power down and less will power up, which puts downward pressure on the STEEM price.
4 – Adjust the reverse auction to a much shorter time frame. Currently, every curator has to factor in the 30-minute reverse auction on a new post. If you’re not familiar with the reverse auction, it’s like this: If you vote on a post at 10 minutes, you can only receive a maximum of 33% of your total possible curation reward based on your STEEM Power and your voting power at the time of the vote. At 15 minutes, you can receive up to 50%. At 30 minutes, you can receive the full 100%. Any rewards that you forego under 30 minutes are given to the author of the post.
Waiting 30 minutes to vote on a post is unrealistic for active curators and for time-sensitive posts. Users should not be so severely punished for discovering good content quickly. It reduces both the quality of curation and the “fun factor” of discovering content first and being appropriately rewarded. A maximum auction time of 2 – 5 minutes would be more than sufficient.
5 – Send the auctioned curation rewards back into the curation pool. This is one idea of many that can be used to better distribute curation rewards among all voters and not just the first few large stakeholders. Rather than paying the curation rewards to the author, the auctioned rewards could be distributed to later voters. Another suggestion is to give a small curation “bonus” to early voters based on the amount of rewards that come from later ones. If an early curator’s vote is not corroborated by later voters, they don’t get much of a reward anyway. But if they were right in discovering valued content, then they should be rewarded a little more than the rest of the voters, but not so much that they’re earning 100 or 1000 times more in rewards with the same or similar stake.
There is a variety of possible solutions here, but the important factor is to increase the curation rewards pool as mentioned in number 3 and to keep the 50/50 rewards allocation in their respective pools so that we do not have a disparity like we currently have (88/12 instead of 75/25).
The “penalty” for later voters or smaller stakeholders should also be reduced so that the 10th or 150th vote on a post isn’t receiving nothing at all, while a whale voting around 25th can pull in 25% or 50% of the rewards. This needs to be flattened similar to the voting algorithm for post rewards distribution. Larger stakeholders will still receive a relatively big percentage of the rewards, but more users – particularly smaller stakeholders – should be able to claim a piece of the curation pie.
Never Underestimate the Power of Incentives
One of the main focuses of my above proposal is rewarding non-bloggers and incentivizing the work of good curators. The proposed comment rewards pool for the next hard fork would likely achieve the exact opposite. It eliminates the incentive for users to vote on comments. This is a huge mistake...and here’s why:
As many of you may know, @abit has been upvoting a lot of comments lately. And @abit has been receiving a lot of curation rewards as well. His upvotes on comments appear to be more valuable to him than upvotes on posts. The overwhelming majority of his votes are on comments. And who are the beneficiaries of those votes? The many commenters, of course.
But let’s not pretend that the curators aren’t at least partially motivated by this:
It doesn’t really matter what the personal motivations are. The fact that @abit is voting on comments and earning just as much or more rewards than other curators of blog posts isn’t something to be overlooked or dismissed. And the consequence of his voting is that more commenters are earning larger rewards for leaving comments – and many users are noticing.
So, there is clear evidence that curation incentives work. The only problems that I see are that comments and their parent posts are treated differently, that the curation rewards aren’t proportional enough to entice more smaller stakeholders, and that the user base simply isn’t large enough to generate more engagement in the first place. My proposal should correct the first two issues. A comment rewards pool will only create a larger fissure between parent and child posts and it completely removes the incentives to vote for others’ comments.
No More Gimmicks. No More Short-Sightedness. Stick to K.I.S.S.
The changes proposed here are meant for the long-term for STEEM/Steemit. Nothing will be ideal right now, but these changes should set up the platform and the different types of users for the future. If we are expecting big things to happen with STEEM/Steemit, then we need to propose and implement a system that can easily scale and resolve potential issues. A less complicated system is generally more suited for minor adjustments and scaling. The K.I.S.S. principle – as touted by Steemit, Inc. in their HF17 proposal – should actually be followed.
And we must never punish those who invest into the system by removing their ability to earn rewards/returns based on their earned or purchased stake.
As stated in my previous post, curation is one half of the two vital components on this platform. The rewards from curation are based on one’s STEEM Power. It’s one of the only reasons to purchase STEEM and/or to power up. Those who actually buy STEEM for the explicit purpose of curating and earning a return are the ones who actually provide the most value to other STEEM holders and to those who post content on the platform. Without buyers, STEEM is worthless.
Comment reward pools that reduce curation rewards by another 38% and other suggestions from users who want to eliminate curation rewards entirely are completely reckless. The constant complaints about what authors “deserve” and what constitutes a “decent reward” or “enough” rewards are nothing but subjective/arbitrary assessments. Those who say that every vote should be worth something – usually a penny is used as the minimum – simply don’t comprehend the system and why that’s not practical.
We need to stop focusing so much on rewards and how much any particular user is getting and remember that Steemit is only meant to be the onboarding mechanism for the STEEM currency and much bigger things to come (presumably). But we’re never going to get to those much bigger things if the code is constantly being changed to appease the relentless complaints from users who haven’t taken the time to understand the platform or who do not have a vision for the future. And the users aren’t entirely to blame. The platform’s leaders within the halls of Steemit, Inc. are a huge contributor to that lack of vision.
There needs to be more emphasis on simplifying the code for rewards/distributions, designing a marketable user interface that appeals to a wider demographic, managing expectations of all users, and keeping the user base focused on the long-term vision for STEEM/Steemit. (I will address more of these issues in follow-up posts.)
STEEM/Steemit will not see much investment if major (and even/often unnecessary) changes are continually made to the economics of the platform. Investors and developers who are trying to make apps and other platforms/interfaces on top of the STEEM blockchain need more stability.
On That Note – A Word on Witnesses and Hard Forks
The proposed changes in Hard Fork 17 are too much. There are too many changes all at once and many of them are unnecessary or even likely to be harmful, or at least likely to achieve the exact opposite of what is desired. Propose a couple of changes at a time and let the witnesses decide on them. Cramming a dozen or more changes in each fork is akin to politicians and their massive package-deal bills that are loaded with “pork barrel” spending. And I know how much @dan dislikes statism. Let’s make these proposals a little more frequent and a lot less bloated.
And please – let’s not just ram them through simply because Steemit, Inc. can stack the top-19 witness slots with those who are favorable to Steemit, Inc. and will approve anything that’s proposed. We need an honest consensus on platform changes. There are many witnesses who dislike these practices and have expressed their concerns repeatedly about them. Some of the current proposals are nowhere close to being a consensus within the community – including actual investors who have bought their stake – so there’s no reason why the changes should be forced on us.
If there are witnesses who oppose these changes and practices in general, let us know so that we as voters of witnesses can decide on who we want to support.
Let’s hear your thoughts about my hard fork proposal. Can we agree on it? Are they bad ideas? Let me know why you agree or disagree.
Follow me: @ats-david
I agree with a lot of what you've written here, and I wish I had more SP so my vote would have more impact on whether or not these things are actually heard by Steemit, Inc. I have a few specific comments about your proposal:
Definitely makes the most sense to me. I don't understand anything about this separate reward pool for comments and blog posts. The more I think about it, the more asinine it seems to me.
I slightly agree with 3, slightly disagree with 4, and wholeheartedly agree with 5. I really like 5. I'm answering them together because I believe they are all part of the same issue.
My perfect vision of curation rewards is something like this: Curation rewards should reward the smartest voters. If you stumble across an article by a brand-new author that's been sitting around for a while and it's spectacular but it hasn't received any upvotes, you as the first voter deserve a big chunk of its potential rewards. The curation rewards should absolutely be highest for the very first voter. Once it's gotten some traction, we can let the curation rewards level off some. But we should be willing to spend a pretty penny to pay the person who found it; we should reward the smartness of that voter.
On the other hand, if @krnl posts a new article, everybody knows it's going to join 6 others of his on the front page within minutes. Why do we pay curation rewards for his articles at all? I don't have to be smart to know that it's going to trend. My bot knows it will; @krnl's score in my bot's model is 242101. That's super high (mine, for comparison, is -2686). We don't need to waste the network's money to pay curation rewards for such articles.
Few people seem to understand this, but that's what the reverse auction is all about: not wasting money on paying curation for obvious winners.
To your 3 and 4: 3 (the fixed fraction) needs to be more flexible, and that's part of what the reverse auction does. I don't particularly think we're ready for 5 minutes, but 10 would probably be ok. From a bot's perspective, 10 minutes is an eternity. Ultimately I think we need a more dynamic way to allocate rewards (a smart algorithm that sends rewards where they're needed), and that's part of what your 5 would accomplish. If the reverse auction adequately removes rewards from obvious winners, then putting those back into the general curation pool would likely dramatically increase the amount of curation that's paid overall. It certainly would be a massive redistribution from authors to curators, and the management doesn't seem to like that direction right now.
And of course I'll include the disclaimer that I am talking in self-interest right now. If curation rewards doubled or tripled, I'd be able to make really decent money on the curation game. That increase would actually probably suffice to incentivize me to buy more steem and power it up. And isn't that what we all want?
5 is a great idea, but im pretty sure i ts not workable. Unless you just send it back to the entire curation pool, not just the pool for that post.
Because earlier voters have such a large edge over later voters, the RA proceeds that you split among the post-auction voters will be dominated not by the smartest voters, but by the voters that vote immediately after the 30 minute mark. So someone like wang could have Wang and Wang30. Wang votes 2 minutes in like he does now (and does OK) but then he also uses wang30 to vote precisely at the 30 minute mark, so wang 30 gets the lions share of the 20X bigger Reverse auction payment that wang generated with his vote.
Youre also creating a huge guaranteed rewards at the 30 minute mark for any post that has a decent amount of support, because there are so much RA funds up for grabs and the first people in will dominate their distribution.
Yes, that's the idea. To send it back to the curation pool in general.
The idea isn't to reward the first voter at the end of the auction. The rewards need to be spread further back on the vote list. The problem right now is
You can have voters 1-10 taking 25% or 50% of the curation rewards and voters 30 or 40 getting 0%, even with a decent amount of stake. Then, voter 80 or 100 can come in and take another 5 or 10%, if they have a much larger stake. The aim would be to make sure that voter 30 and 40 aren't always getting 0. Their vote is really no different in terms of discovery or ranking, and even though they beat half of the rest of the voters to the punch, they can come away with nothing at all while very late voters with a larger stake can still grab a piece of the pie.
I get the advantages of having a larger stake. But even with mine being 27,000+ SP, if I don't vote at the "right" time, my rewards can vary pretty significantly. It can be a matter of earning less than 0.1 SP or earning 5+ SP. And the time really doesn't even have much to do with it. It mostly has to do with which user voted before me, even if that other user's vote is only one block earlier (3 seconds or less).
See my reply to BP below. If you just send it back to the pool, it will end up distributed the same way as the pool currently is. If youre sacrificing 50% to the reverse auction now, and you send it allback to the pool, all youll do is just double everyone's curation rewards.
A big part of the problem is that even though the reverse auction effects the payout for early voters, it doesn't change their weight, so they still dominate the curation rewards distribution, they just give up most of the funds their domination entitles them to.
The real solution to this is to attenuate their effective voting stake in the auction, not their reward. That is to say, if you vote at the 90% RA mark (3.3 minutes in), then for the distribution of curaiton rewards, your vote should count likeyou had 10% of your SP. So a 100K SP voter in the third minute shouldbe treated for distribution purposes like a 10Ksp voter.
Yeah, I like that idea. Maybe this should be explored more. I'm sure we'll chat about it.
I maintain that's a design feature, not a problem. The purpose of the reverse auction is to reduce curation rewards for easy votes. Period. It must make obvious strategies self-defeating.
Because people who want to upvote themselves are going to do it regardless. It's not like there is another solution for them to cheat the system. The fact that we hardly see anyone upvoting themselves means its a non issue.
Why do you need a curve at all? Isn't voting power based on how much power someone has? There is no need for a curve which is why I suggested we remove it.
Whales could already upvote themselves now and make a significant amount of money why are none doing it?
Downvoting is not a hack, it is already built in. You just have to use the feature.
The algorithms happens to penalize almost every single posts on the platform, surely it can't be the best solution right..
30 or whatever still might get nothing, depending on how much SP he has, but he'll always do better for the whale coming in than he would have done otherwise.
That is to say, voter 30 never loses out when a whale comes in after him. (though if his stake is small enough, it might not be enough to actually get him a payout. but it will get him closer i think always). If he's not making money on a post with a faily high payout, its because he is dominated by the medium sized guys in postition 1-29.
Oh absolutely. I neglected to clarify that in my comment. The whole point of it is to send it to other posts. The reverse auction is there so we don't waste curation dollars on on obvious winners, so it wouldn't make any sense at all to shovel the money right back on top of the voters for that post.
Makes sense. I noticed when i was looking at curation rewards before that many of the "obvious winners" only get 5-7% in rewards.
The problem with giving it to curators on other posts is which other posts? If you split it up equally according to post payout, then it will mostly just go back to the biggest posts where its lost. If you only give it to posts with little money lost to the reverse auctions (that is to say, don't give it to the sure winners), then youll give a guaranteed payday to anyone voting on a non-voted (or low voted) post after 30 minutes.
That is to say, if a non-voted post has no votes after 30 minutes, its already a guaranteed 25% of however much your vote is worth by itself (which is a potential vector of abuse for whales anyway). But if its not just 25% of what your vote is worth by itself, but some share of a pretty significant RA redistribution pool, then there's even more.
Also, there is nothing stopping wang 30 (or any other bot) from just voting on everything with no rewards precisely at 30 minutes.
Essentially every new post would be eligible, subject to its own reverse auction. Same as now, only more curation rewards.
Well, roughly speaking, your last point is exactly the magic of it. It's designed to make good strategies self-defeating. If any of this ever creates a guaranteed payday, then people will start competing for that payday in the reverse auction and the payday will vanish. Just like the reverse auction currently takes away curation rewards from sure winners.
You're right, the exact details would have to be thought out well to make sure the redistribution works correctly. But notice that even the "vote every un-voted post after 30 minutes" strategy is essentially self-defeating, because if everybody did that on every post, their votes would all cancel out.
I haven't completely thhought it through like that, but my feeling is that it would create a giant snowball of reverse auction funds that get re re re re reversed but never awarded.
EDIT -- no it wouldn't, but it wouldn't change anything at all either. You would eventually get to the point where the curation rewards were increased in magnitude to the full 25%, but were distributed in exactly the same proportions as now.
So if we lost 50% of curation rewards now, it would just double all current awards.... which isnt necessarily bad.
Nesting limit.
Heh, that would be interesting. Maybe you'd need a relief valve to prevent overflows. Or, to increase the complexity even more (because KISS matters so much to these people), set it up dynamically so that the fuller the curation pool, the lower the curation/author split gets, essentially to push the unused curation rewards back into the author pool.
nah its not necessary, it would eventually reach equilibrium.
so like if you started at 10000, and lost 5K to the reverse auction, the numbers would look like
10000 -5k carried over
15000 - 7.5k carried over
17500 -- 8750 carried over
18750 -- 9375 carried over
19375 -- 19687 carried over.
it would approach 20K but i think never quite get there.
KISS is a dumb idea. Some problems just don't lend themselves to simple solutions.
Because I do not have very much knowledge about what is happening with the blockchain outside of steemit.com, my opinion is based solely on my experience here. There could be a master plan that I don’t fully understand.
But according to my observations of steemit.com, I would like to see a hard fork only address one issue at a time. By making several changes at once, it is impossible to tell which of the variables lead to a positive or negative result.
Personally, I think a flattening of the reward curve would be the best place to start. If more people felt emotionally invested in the platform, we would see more activity. It may seem silly, but if a new user could actually see their vote reward a post even .01 it would go a long way to creating this emotional investment. People like to feel like he/she matters. When a new user sees other people giving .01 for a vote yet theirs does nothing, it can be disheartening.
It would also encourage people to hold SP so that their vote would increase in value. Currently, there seems to be very little difference if one holds 500 Sp or 5,000 SP.
If more people control votes that matter, more content creators can get rewarded. If more content creators get rewarded, they will stay on the platform. They will create a variety of content that non-members may find interesting. Perhaps some of these people will choose to become members and create more content and continue this snowball effect.
After testing this flattened curve for a while and judging its success or failure, the next change could be tested. Perhaps that next change would be one from this list, one from the steemit inc list, or a new change inspired by the information gleaned from using a more flattened rewards curve.
I tend to agree. I'd rather see more forks with less changes than less forks with more changes. And I think most of the witnesses agree with this as well.
The biggest issue I have with the new fork proposal is that several of the changes just seem completely arbitrary and we won't know what changes made a difference, or if one change cancels out another. And I realize that my proposal consists of at least five changes, but those five are all inter-related and some of them can't be implemented on their own because they would just distort the rewards mechanisms further in the wrong direction.
I do agree that flattening the curves should be the first step, however. If we can do that and the results are good, then we can move on to the other options, if necessary. In any case, the curation rewards need to be improved and increased. It's the only purpose for holding STEEM/SP and it's the only non-speculative driver of demand, which STEEM is sorely lacking.
Many years ago, while learning how to fix aeroplane engines I was taught 'do one thing at a time'. If that didn't fix the problem you only have one thing to change back again, before you try the next fix. Can I suggest a similar approach here?. How long a time period is required to see if the fix worked?. How long have we got until we need to be perfect?.
Interesting approach. Less is more!
Well said. I agree in principle with all the ideas. Some of the details on #5 would obviously need to be worked out but the general concept of the rewards being too biased toward front runners I agree with. Every vote is important and valuable in contributing to the ranking and rewarding of the content, not only discovery, and should be recognized with rewards accordingly, not just the first (though the first can reasonably earn some extra for discovery).
Good point on ranking/discovery. Later votes are just as valuable for rankings and for essentially confirming the judgment of early voters. Without the later votes, the early votes don't mean much, so there's not much sense in paying early voters exponentially more for their vote.
Yes, but unless there is a significant penalty for later voting, there is a huge incentive to pile on.
For example, if curation rewards were 25%, and they were just divided up evenly according to stake weight without regard to who voted first, late voters would have a huge incentive to pile on to already successful posts for their share of a 25% of guaranteed moon.
This is a secondary consequence of extreme superlinearity. Taking the extreme opposite case of fully-linear rewards, even without any further adjustment, there would be no incentive to pile on. The curation rewards for being the last voter would be the same as voting on any new post.
I'm pretty sure that even if superlinearity in some form is kept we can in principle invert the superlinear function such the the last voter does not get a huge bonus from the superlinearity, but also doesn't get nothing.
Thats the problem with talking about changing curation rewards before we know exactly exactly where we're going w/ author rewards.. but yeah the piling on thing is gone if there's no superlinearity.
Of course, if there's no superlinearity and no significant reward for voting early, then it doesn't really matter that much what you vote on at all.
Significant is relative. I don't believe that the degree of order-penalty in the current system that essentially gives no reward at all to later voters is necessary. Earlier voters can still get a bonus. For example, perhaps 10% of each voter's reward could be paid to earlier voters as a finders fee. The first voter (getting fees from all of the rest) would earn by far the most (assuming several times as much total voting as that one vote), but even the last voter would earn 90% of the base reward.
In general I agree that the basic shape of rewards would need to be understood first, most likely.
I'm very new to steem and steemit, but I like your ideas... I'm just getting started so I'm a total newbie though, but I like the way you think so far. Great write!
Thank you!
My conclusion is: Do 1 and 2 this time but postpone 3, 4 and 5. Change the curve to a linear one, keep a single pool, see how things go. It's not good to suddenly change too many things. Sure we need more incentive for holding SP, but I don't like to add too much at once.
//Update: I need to make it clear that I meant to keep curation rewards for both root posts and replies.
I can live with that. The question is whether or not Steemit, Inc. can.
A comment reward pool will create a situation where the 2 metrics ( price and number of users) that should follow each other for sustainable growth will actually be going in opposite directions.
I don't have to tell you why this is bad..
So - I don't pretend to understand the full-spectrum economics of the system. But you've sold me on a couple of key principles. Creators of good content should be rewarded - because that's what makes the platform viable in the medium term. And commenting - and being rewarded for it - should be encouraged. Because that's what makes the community viable.
I'm in.
I agree with most of these.
Rewarding those who actually buy steem power is essential to increase the value of steem so it makes sense to reward curators more. https://steemit.com/steem/@snowflake/steem-inflation-a-tool-to-create-demand Cutting curation rewards by 38% is just insane especially if you are not going to create demand from somewhere else.
Regarding 1) I don't see the point of using any curve at all. This curve doesn't discourage self voting
https://steemit.com/steem/@snowflake/reward-curve-doesn-t-discourage-self-voting
A comment reward pool is a band aid solution, if everyone had a voice and rewards disparity wasn't so big people would vote a lot more for comments. Let's not beat around the bush and solve the problems at the root.
Overall this proposal sound good to me, I still think that it won't be enough to incentivize average joe to power up to gain influence because 99% of their votes will still be worth $0. However I think its a step in the right direction. I will support witnesses who want this.
I can't 100% sure but I'm reasonably confident your 99% is far off. With flattening the curves (both the stake curve and the time curve) and increasing the total curation pool, smaller votes will be worth more than $0 quite often. Working out some numbers here would be good. Consider that just restoring 50/50 from the current 88/12 and changing nothing else would increase all curation rewards (including the smallest) by over 4x. That's enough to push a whole lot of voters' rewards from $0 to >$0.
You are correct, it will probably be less than 99% and will improve power disparity which is why I am in favor of this proposal. But i still think the vast majority of users won't have any influence which really is an issue. The average joe is not going to spend tens of thousands to get influence on a social media site, he is going to pay a couple hundreds at most. A couple hundred bucks would make no difference regardless of any curve. Currently the system is out of touch with the reality, the whole thing should be scaled down and investors seperated from users.
Agree. Flattening the reward curve would help though. Is it enough? I don't know, but it should at least be tried.
Ofc it should be tried :) Actually it should just be removed completely.
The curve serves no purpose at all, it was meant to discourage self voting which is a non issue. Like I said already, self voting is like trolling, if someone were to do it repeatedly they would get downvoted and blacklisted. Reducing the large majority of users's rewards because of some issue that doesn't exist is ridiculous. I don't even understand why so many people want to 'flatten' it, just remove it already.
You keep saying this and I keep thinking you're missing something. How can you know it's a non-issue when the curve is currently dis-incentivizing it? If we made the curve linear (what you call "removing" the curve, which is also very confusing), a whale could come along and spam empty posts to the blockchain, upvote all of them, and receive a vastly higher reward than they would in the current system. Your answer is to downvote, but why hack the solution when we can bake it into the algorithms?
because of the reverse action it would be still not poor 50:50 !
So I make a step further and will demand the curators to get 62% and the authors 38% and use only one pool for rewarding posts and comments!
That's one approach. I'd prefer if the overall split weren't dependent on the auction/penalty. That can be done by giving it to later curators on the same post, as suggested in the OP, or I think better by keeping content and curation in separate pools according to the specified allocation (be it 50/50 or 75/25 or whatever). Any penalty on curation of one item would slightly increase the available funds for rewarding curation on others.
I'm going to assume these are suggestions for the upcoming HF17 and the the future.
This is surely coming in the future. I agree that this is an urgent matter though - there's no need to do a trial run with the Comments pool and wait for the next fork.
While this makes sense for Steemit or social blogging frontends, maybe Steemit Inc is looking at an opportunity for a Disqus-type competitor where the Steem blockchain would host comments on websites of various kinds. Splitting the comment and blog pools would be a necessary pre-requisite for such a solution. This would also explain the lack of curation rewards. Of course, this is just speculation.
Agreed on #4. Not sure about #3 and #5. #3 may be extreme and #5 may be contradictory to the system. While we are at it - I do feel the algorithm is too biased towards frontrunners, discouraging late voters. Also, the voting power decay is too biased towards accounts making hundreds of votes every day (bots). The decay should increase beyond a certain threshold instead of slowing down. I would guess both can be adjusted relatively easily.
Splitting comments and blog pool is not necessary. When creating a comment pool you achieve nothing, influence disparity will still be huge and you won't solve scalability either as you would need a whale on every single site using disqus ( steem version).
A website owner wanting to integrate steem version of disqus will look first how it is going to benefit his existing userbase. If everyone of its users need to buy $10 000 to be able to send a few cents to each other he is not going to bother implementing it.
Steem will have many different use cases, you can't just create features to accomodate each of them, what we need is a solid base where everyone can implement steem regardless of what product they want to build. In its current form steem doesn't make sense for websites to integrate as it won't benefits their user base.
Thanks, good points.
This was the reasoning from their initial proposal from @steemitblog:
To me, it simply appears that the focus is driving engagement on Steemit.com. And I would imagine that if they wanted to create a separate pool to be used on different sites strictly for commenting on non-STEEM/Steemit content, then they would probably need to create a side-chain for that, wouldn't they?
Aha, I remember that bit. Seems like a while ago! Fair enough, their stated intention is clear.
I don't understand the technical workings behind the scenes at all, but I do think they are working on something called "multi chain fabric". Not sure if that has anything to do with it?
The vote power decay is a very important and tricky tradeoff between mass voting (with slower/flatter decay and people feeling like it is a big deal every time they place a vote (due to rapidly declining vote power) and therefore need to be extremely careful with every single vote. I think the current tradeoff isn't terrible (it was reached after trying a faster decay and finding it problematic). It could perhaps be adjusted to be a little faster but shifting it too much risks turning voting from an enjoyable expression of opinion into a scarcity calculation, or even a perceived source of "punishment" for casual users.
Sorry, I didn't communicate that well. All I meant was "The decay should increase beyond a certain threshold". We can follow the same curve as it is up to a certain threshold. Let's say, 100-200 full strength votes - some amount which is clearly beyond regular curation behaviour. Beyond which, you only have vote spammers - whether bots or human - the voting power can decay fast. I'm sure you have seen there are bots which make 500-600 votes every day; this will penalize them without affecting the regular curators. In fact, they will have more R-shares allocated back to them, recovered from mass voters.
This is what I mean, if you excuse a layman's drawing -
Currently high frequency bots are all voting with arbitrary percentage of weights. Even if they can't vote with a percentage, actually it's not hard to get around a fixed limit: just split SP among several accounts and vote one by one (round-robin), or always vote with the account which has highest VP (best-first); when it's needed, vote with many or all accounts at same time.
Good point. Superlinear penalty probably would not work.
It's not about the number of votes, but voting power. So it doesn't matter if it's 100 votes at 100% or however many at 10%; after a certain threshold, the voting power would decay faster to penalize for vote spamming. When I said "100-200 full-strength votes" I really meant "100-200 full-strength votes or equivalent at other strengths". This may not reduce the number of votes - true - but it'll reduce influence/Rshares from high frequency bots.
Of course, you're right that they can just split SP among several accounts, but they are already doing that. This will just make it harder to do so. Today there might be people spamming with 500 votes at X% with 10 accounts taking their VP down to 20%; now they'll need to make 100 accounts. It'll be a deterrent.
High frequency bots don't have high influence/rshares. It's a trade off. You either have high frequency (voting on more posts), or high influence/rshares on less posts, but not both. There is only 20% VP regenerated everyday.
Since everyone can create accounts for free (registration fee is deposited into new account), number of accounts is not actually a barrier for bot runners -- they're already scripting everything. Sure it will need some work/efforts, but it won't stop them.
//Edit: forgot this: in current code, there is a penalty when calculating VP regeneration. The penalty is larger for voters with higher frequency.
[Comment tree limit]
Today you can choose between high frequency and high influence. There's little penalty for high frequency votes, or more accurately, extreme drainage of voting power. That's all I'm trying to say! By penalizing unreasonably high voting power drainage more, we can weed out the bots and vote spammers. It'll also encourage bots to improve their algorithms and vote wisely.
(Personally, I don't have any problem with bots, but it's a major issue for the community)
Imagine you decided to just say "fuck it" and started downvoting everything that came down the line. Or upvoting your own comments. Not like 40 like 200 or 400 or something in a day.
Yeah, you would pay on subsequent days down the line because it would take you 5 days regain your full power and start over. But on that one day where youre voting like a madman, you would have far more influence than you normally do.
someone like blacklist, for example, or asshole, who is only here to be a nuscience, can concentrate their nuscience-ness all into one day, by mortgaging their voting power on later days.
Im not sure im entirely convinced that any more penalty than the normal degradation of your power is necessary, but i think thats what liberosist is getting at.
Your reply is a bit confusing for me.
I understand that you want the system to be better, so do I. So we're discussing the "how".
My point is, high frequency bots already gave up influence. You want to punish them more by setting a limit on frequency? Then they'll split to several accounts then vote less frequently with each account (each achieve a "normal" frequency), but in total the frequency is same as before. So your "solution" will result in neither more nor less vote spamming, aka useless. I'm sure bots are always improving their algorithms, that strategy I just mentioned is a possible result of improvements. You keep saying that bots should vote less but vote wisely, but I'm afraid that the best algorithm they've got (after improvements) is to vote more. So, if we want to address the bot voting issue, we need to look for other solutions.
On the other hand, if we want to address the influence issue, better focus on the whales, the curve.
Yeah, you are probably right that some sort of increasing curve would work. I just want to avoid regular users experiencing severe or even easily perceptible penalties with normal use especially if voting on comments becomes more popular, or the UI is changed to encourage more voting (for example, by removing the visible delay). I offered the thought exercise once (when reducing the daily vote target was proposed) to consider how people would vote if the vote target were changed to 1000. I believe people would vote even more freely than now, and more comments would get votes. That would probably be a good thing, not a bad thing (ignoring unintended consequences).
IMO, this is the biggest problem with curation rewards at the moment. I don't think there's a good way to fix them without fixing this first.