You are viewing a single comment's thread from:

RE: Proposal: reduce Hive inflation by reducing curation rewards

in #hive-10293010 days ago

I feel if we drastically reduce curation we are going to eliminate any incentive to power up. The majority of users already power down and sell as fast as they earn it. Very few will lock up any significant amount to just vote witnesses and proposals.

You are trying to encourage manual curation, but this would reduce the incentive to do it. To manually curate with any decent amount of stake is literally a full time job unless you just drop large votes on anything half decent you see.

No matter how altruistic you are, it's very difficult to justify spending 4-12 hours a day reading and voting content you likely have zero interest in the first place. That's how long it would take to manually curate with a sizable stake.

Reducing or removing curation will encourage those with stake (or access to it via circle jerks) to create more shit content to take more of the author side. One of the reasons there is a significant reduction of spam and abuse is curation is more viable compared to shit post farms and you can be competitive without voting on your friends or your alts. I find a lot of people underestimate the value stake holders bring to the ecosystem. While creating content is time consuming, earning and buying stake is as well and not as easily doable by most users.

In another comment you mention this:

It's not really changing the incentives to invest in Hive. It's changing voting incentives. The amount of lost curation rewards is exactly balanced by the reduced inflation.

I don't think anyone outside of Hive really cares about voting incentives, and those are the people we need to encourage to invest in Hive. I do believe reducing inflation is valuable, and it is something that outsiders care about. I've suggested in the past we cut the reward pool in half until we can acquire more users. I think targeting curation only would drastically reduce any incentive to power up. In fact, as someone who has powered up since I have been here, I would struggle to justify it myself if curation was made to be pointless. Curation is already soul sucking, you are damned if you don't vote someone, you are damned if you do vote someone, you are damned if you downvote someone, you are damned if you don't. Not only that it takes an enormous amount of time, more so than any other crypto project I know of to just "keep up with the Joneses". The problem is made worse when you see garbage rewarded extremely well for little effort on a regular basis. Many stake holders will likely feel they are losing their percentage share unless they also become an author, and let's be honest most people here are not authors, so they will turn to easy solutions.

Let me turn this bus around and get back to the root of the problem.

auto voting is bad.

I disagree, I think shitty auto voting is bad. I mostly see myself supporting people who create good content and put in effort. I maintain an eye on these people I support and adjust accordingly. I also give significant votes to everyone I vote, $4-5 typically. Here is the real problem that I have brought up before. The majority of auto votes give a few cents, their vote is only there for curation rewards and nothing else. This is the bulk of the auto votes you see, these votes have zero concern about quality and effort and are purely for curation rewards. This is the real problem in my opinion. This also results in clumping where popular authors get the majority of the votes. This is also the result of very limited pool of "good" authors to vote.

The other problem is very few are downvoting to counter the shitty voting that goes on, which would be a balance to keep the auto voting in check.

Sort:  

I just dropped you a 100% upvote because, 1) you deserve it, and 2) that is how I manage to distribute my daily 20% recharge. I don't have but a couple of hours a day to dedicate to HIVE, so I give 100% upvotes on posts, and 50% on comments (with exceptions, like yours). I think it's also more appreciated on the receiving end too.

But get this: half of my votes usually go to comments, none of which ever reach $5.00. That's a huge disincentive to upvote comments (which should be just the opposite since comments are the life blood of a community - as you said, the vast majority will never be successful content creators). As incredible as it may seem, the opposite happens with posts receiving more than $5.00 of upvotes: we get unwanted curation trails as everyone piles on (sometimes piling on to what very well may not be exactly what we would call quality content).

In short, I'm rewarding your comment as I think it deserves, following what comes naturally, like we do everywhere else, IN SPITE OF BEING PENALIZED FOR IT (both of us).

That is what I believe needs to be fixed, and, not so ironically, I think that, by and large, by doing that, the curation trail issue would also be fixed in the process.

!ENGAGE 50

The majority of auto votes give a few cents, their vote is only there for curation rewards and nothing else. This is the bulk of the auto votes you see, these votes have zero concern about quality and effort and are purely for curation rewards.

This is the problem, this is what needs to be addressed not so much changing once again the reward pool system. If auto vote trail following required a minimum value then this would solve the entire issue. Start with a minimum of $0.750 value per vote, if a minnow account an alt account or a low HP account wants to follow with a zero value vote to bad they should be voting manually.

This will be the first Hard Fork for Hive, I for one am glad it is not Adjusting the vote/reward pool. It will also be the first HF I have seen since 2017 that did not include changing the reward pool/vote system.

I don't know if code can be written to limit vote trails/auto voting, but if it can it should be looked into.

Very limited is not the best words to use there. I know of at least two dozen, off of the top of my head, that refuse to post anything substantially qualified as 'quality' (which is also subjective to the eye of the beholder), simply because their posts consistently got overlooked and were never curated 'properly' as they should have been. It still happens. I used to write massive stories that took hours to compose between gathering proper graphics and making it 'pretty'. Sure a few reached 100+ dollars - but only because of bidbots. Not a single actual whale vote aside from grumpycat who thought I made too much on it. I recently wrote a story on my first fish, I thought it deserved at least 15$ but nope. Nothing. 6 minute read with over 1000 words.

Which, is a whole other ballgame there. There is a post right now on trending from @krnel that is twice as good in my mind (pun intended) as the one written by @acidyo about @ocd. Krnel's is worth half as much. The one from Krnel is also a shorter read than my fish story. What gets the most attention around here seems to mostly pertain to the blockchain and actual content creators, don't get near the credit they deserve. The 'whales' would rather just support the 'whales' so they can remain the 'whales'. Not all, but most. Voting habits and the blockchain doesn't lie. Some even sit at 100% VP/Mana for days until one of their 'whale' friends makes a post. And we won't even get into the biases from 'communities' and the separation that bred into the system.

Shitty auto-voting is bad, but I use it. Not necessarily for the curation, but to offer steady support because like you, I can't be here 12 hours a day. I am in agreement with you though (once again - kind of an odd thing, but hey, it's been happening more frequently the past few months), reducing the curation rewards is not the answer. The actual content creating community, all inclusive of Hive, was already pretty upset when it went to 50/50 and so many had to adapt. You take that away again and as you say, we will be hard pressed to attract any investors or even users, for that matter.

Very limited is not the best words to use there

I still stand by my words, even if every author was a "good author", we still have very few.

simply because their posts consistently got overlooked and were never curated 'properly' as they should have been

Really you mean as they feel it should have been. There is no "should have been" here. I've seen a duct tape meme get voted $943, and good posts make less than $1. You have to put in the work in any social media before you build an audience. That involves posting and engaging consistently. Social media is hard, yet Hive is still far easier than any I have seen do to the lack of competition, but it is far from fair.

Some even sit at 100% VP/Mana for days until one of their 'whale' friends makes a post.

The first half of this statement is a gift to the community, the second half is depressing.

The 'whales' would rather just support the 'whales' so they can remain the 'whales'. Sure a few reached 100+ dollars - but only because of bidbots. Not a single actual whale vote aside from grumpycat who thought I made too much on it.

I disagree, I don't do this, and most of the whales I know don't do this. There are not many of us. It isn't usually whales doing this, it's dolphins and orcas. A lot of the whales are in fact spreading their votes very thin to optimize curation or voting a group of favored authors (potentially friends). I don't see the behavior you describe often at the whale level.

I used to write massive stories that took hours to compose between gathering proper graphics and making it 'pretty'.

This is a systemic problem on Hive, people think because they do something they are owed something. Unless you have a contract, then it is up to the community to decide what it is worth. But I will agree 100% there is no consistency in voting based on quality here. This is a result of decentralization, favoritism, ulterior motives, and just poor content discovery options. EIP has done wonders for helping in this department, many more stakeholders are voting outside of their circles. It isn't perfect but the end result has been a vast improvement over Old Steem. No change we make will ever get us to where we want to be, but if we can get closer that is an improvement.

Not claiming to be entitled, but I did go to college for English Journalism. So, I can actually deem myself a decent judge as to what quality should be. It's not what you know or how much heart and soul you put into a post here. It's who you know. Simple as that, and the same goes for any centralized system as well. I've seen crapass posts worth well beyond their limits too. My point is, many a good and often a great thing gets overlooked. Which is one reason I feel auto-votes are a good thing. Take @snook, for example. She likes to make people smile, and that's one of the reasons I love her. She is in my autovoter because I don't care if she posts a video, 3 sentences or three chapters. I'm supporting HER. I know her. And maybe a few days later, I might have enough time to comment my thoughts on her post. Maybe it's 8 days later and I missed the deadline to even give her a vote! Or wait, no I didn't, she's on auto - man I love those things.

My point is, many a good and often a great thing gets overlooked. Which is one reason I feel auto-votes are a good thing.

I think they are a bad thing, because they make people too lazy to discover new posts or also honor efforts of former 'bad' authors to improve or just unknown writers.

It's easy to upvote 'good' authors automatically all the time but difficult real, hard work to find great posts.

People don't want to discovery "new posts" because it's a lot easier to farm on defi and do nothing, earning 10-20x yield returns even if it does collapse.

The Auto Voting is a serious problem if you want a social unicorn. On the flip side, poor returns from APR yields is a serious problem if you want a blockchain unicorn. It's not to say by any means you need any APR, or staking for that matter, it's just for this comparison since rewarding is still in plans along with inflation.

It needs to to go strongly in one direction or the other. Currently all the mechanics are in the middle.

New Investors:

And it's important to get new investors unfortunately, because that's how the entire mechanics work for any asset class, stock or crypto for that matter.

We need to support more experimenting like @blocktrades proposal and attempt more radical changes(a bit less focusing on "finding the sweet spots"). More extremes imho.

Not sure that's true either. My feed is full of them. I wouldn't follow someone if I didn't think they didn't do good work. What would be the point in following a terrible author?

I don't like the concept of 'good' authors, I prefer the idea of upvoting good posts! :)

Upvoting 'good' authors automatically means being too lazy to read and evaluate posts, means allwoing 'good' authors to be lazy too, as they receive upvotes independently from the quality of their posts, means not giving 'bad' authors the incentive to improve and have the chance of receiving upvotes when writing better posts than before, means to keep ignoring new, unknown users which leave the platform as fast as they came, because nobody makes the effort to seek for posts manually.

I see not a single reason why I should get less curation rewards when manually upvoting a two days old post which I had really read and evaluated than someone who didn't work at all but just let a programm do the job!

You may read more about my point of view here, here and here.

many a good and often a great thing gets overlooked.

I agree, and I'll quote my parent comment.

"This is a result of decentralization, favoritism, ulterior motives, and just poor content discovery options."

I'm supporting HER.

My original comment brings this up as well, I do the same thing. I use votes to support people, not always specific pieces of content. I like to support authors I believe that are creating good content and /or putting in effort. If I see something that stands out, I vote that as well. An important thing if you use votes to "support people" is reviewing those votes and those people as people tend to change when given automatic votes.

I do regularly just like you. I have my personal options and my community ones. Some are similar. But I'd say only half of, maybe less, haven't actually counted, of my votes are auto. Most of my personal is manual. I don't care if I get down in the 60% range, it's not just about the curation for me. Retaining users is a struggle itself. We wouldn't be much without them.

auto voting is bad.
I disagree, I think shitty auto voting is bad

@themarkymark please don't take this as an attack on you (I almost wrote auto attack xD).. But, I believe that auto voting on 4th minute with a stake large like yours is a bit shitty auto voting. It doesn't give time for manual curators to even read a post and possibly engage with author.

I am saying this because curation rewards are only incentive for manual curator, with other emotional advantages. So it would be good not to take that away from them (us).

I am saying this to you because I often see your votes when I am manually curating and I hope that you will give it a consideration to change it a bit.

By saying this, I also appreciate you voting on some small authors that are creating good content. So once again, I am not attacking u nor saying you need to change your thing.

I am just hoping you will consider what I said here.

Thanks.

I agree with this; autovoting is not the problem, shit autovoting is bad. While there is any incentive at all to autovote over manual vote people will do it. The key here is to align interests. TBH the 5 minute window has a lot to do with the way things are now. If we removed that, and let downvotes do their thing then we are getting closer. I would love to see the view of @theycallmedan on this. Hive is new; lets not meddle with to many things at once while the market is still getting confidence with us.

YES, all of this ^

Thank you for your engagement on this post, you have recieved ENGAGE tokens.

actually "bad curation" isn't as bad as it would seem:

https://peakd.com/hive-167922/@tobetada/how-bad-is-bad-curation

Honestly, auto voting is not too bad. It gives the simplicity to users or stakeholder who does not have time to curate and that is why there is a downvote option use for, to downvote the bad content or post.

My earnings:
Hive : 216,009 HP - Last 30 Days: 2,049.20 HP / $352.27
Steem: 172,800 SP - Last 30 Days: 4,954.97 SP / $841.98

Anyone with brain will leave this blockchain after 75% cut on curation

Anyone with a brain could figure out that a difference like that can only exist right now if fewer people are reaping those rewards, given the rewards systems are the same currently. That tends to point to a serious long term problem for Steem.

That's true, this is why I'm slowly powering down SP and converting earnings to HP but if proposal like this one gets funded it won't make any sense to me or any other investor

I get the feeling that there's no need to worry about it. I think too many people share your fear to try the experiment.

What I don't like about automatic voting is the abuse of the curve by programming to 5 minutes without giving time to manual curators. Eliminating the healing curve would be beneficial.

Another abuse of the automatic votes is when you use 100% of them and nothing to manual curation. The same group of authors will always be voted on, excluding others, and this discourages new users.

The solutions for many whales can be to get into a manual curation trail as you do with @cervantes and what you could do with others as @rutablockchain 😝

Now seriously, speaking of solutions. In the Whaleshares blockchain they did not have automatic votes, somehow they were not in the code, although I am not in favor of such a drastic solution. They also implemented (I think it would be at a frontend level) that to vote a post you had to open it and you couldn't directly from the feed.

@theycallmedan provided a solution to this and "delegated" their vote to several community leaders among whom I am, only following the vote of a new account with a fixed percentage. He relieved himself of curation and made it more decentralized, distributing that vote in several communities but keeping his stake for when he wants to vote. We as curators do this to help the community.

The whales must be integrated into the communities, so there are many ways to encourage manual curation.

Any proposal that goes through eliminating curation rewards will make it stop curation and as a consequence a stampede and abandonment of users. It is the content creators themselves who are bringing more people to HIVE, I don't see the whales in that. If there is only a ROI for maintaining HIVE, most of the community will leave and this project will no longer make sense.

What makes HIVE different from another blockchain where you earn ROI, is that here you can transform lives of people who express their talents, if you change that you destroy the magic

I've grown to dislike the 5min window. I've talked about this for awhile. It screws manual curators and makes most flee to auto bots so they can earn better curation rewards voting before manual curators. I would like to see a flat change here. When I have time today I'll jot down a few more ideas.

What I don't like about automatic voting is the abuse of the curve by programming to 5 minutes without giving time to manual curators. Eliminating the healing curve would be beneficial.

This seems like a simple solution to reduce auto-voting and give those seeking and rewarding content based on its perceived worth (and not the authors previous efforts) a fair share. If the CR was paid out purely based on stake voted, a % could be set aside and burnt?

One issue with doing flat rewards, even for a time, instead of the curve is that it opens up a new form of auto-voting. For example, if the rewards are proportional to stake for the first hour, then large stackholders can have bots watch which posts are getting first hour votes and pile in. The curves encourage voting before others, but not voting too early. Perhaps the early-voting-penalty time could be extended, and front-ends could build ways to delay manual votes for their users.

My thinking behind the CR based on stake is that there would be no financial benefit to piling on something already rewarded well. If your vote is worth $1, you get 50 cents in 7 days.

But yes, then we are back to potential self-voting of comments and buying votes.

It’s a tough one.

!ENGAGE 25

What you're saying is true, but I'm not sure it would be such a bad thing, from a quality curation perspective. Manual voters would identify the good posts, then auto-vote bots would just be magnifying the rewards to the authors of those posts. It does allow the bots to benefit from the work of the manual curators to some extent, but I think it would be better than the situation now.

Effectively, it would be allowing bots to do vote trailing based on more than one manual curator.

Thank you for your engagement on this post, you have recieved ENGAGE tokens.

It's good to know that community leader see the problem, trying to find the solution and asking for opinion (that's how I read your proposal). I bet we will have to get back to the discussion after the hardfork. There are many ideas in comments how to improve the system, here is mine:

  • anonymous downvotes
  • increase the downvote mana to 100% of upvote mana (both with same rebuilding rate)
  • after every downvote burn tokens instead of sending them back to reward pool

I also think that anonymous downvotes would be helpful. But while it may be possible to do it, there's no obvious way to do so. I will give it some thought though, since I need to solve a related problem for the reputation/rating system I want to develop: I want there to be a way for people to report negative ratings without being the potential subject of retaliation. I do have some preliminary ideas along those lines, but I'm not sure how easily they can be translated to Hive's reward system.

I think anonymous downvotes would dramatically improve this system. The reason auto curation on crap can happen is because non-whale users are never going to downvote a post that will harm some whale. If downvotes were anonymous curation could actually work the way it should, in both directions.

I agree on anon downvotes, would be a game changer. Not sure how it's possible.

I agree, reputation system needs to be replaced. Waiting for your ideas and post. The simplest/temporary fix would be the current system which will calculate account activity just for the latest 30 days

Steemians: "We centralized the chain, now we'll centralize distribution"

Steem is a shit show with a paragraph and image making $200+. If I had stake there and wanted curation rewards it would be a cake walk. I would trust anything there to have powered up between them potentially stealing it to the price going to zero when Justin Sun dumps his bags.

I feel if we drastically reduce curation we are going to eliminate any incentive to power up. The majority of users already power down and sell as fast as they earn it. Very few will lock up any significant amount to just vote witnesses and proposals.

I don't think it eliminates any incentive to power up, although I agree it reduces it. It would probably make sense to reduce the powerdown time to 30 days at the same time as a counter-incentive.

But what I don't agree with is that "disincentivizing staking" = incentivizing selling. Staking just to create scarcity is just playing a game with people's heads, IMO.

I disagree, I think shitty auto voting is bad.

I agree with that. I just haven't seen a solution that just eliminates bad auto-voting.

The other problem is very few are downvoting to counter the shitty voting that goes on, which would be a balance to keep the auto voting in check.

I've also yet to see a decent proposal for how to make downvoting work well socially, unfortunately.

But what I don't agree with is that "disincentivizing staking" = incentivizing selling. Staking just to create scarcity is just playing a game with people's heads, IMO.

That's not what I meant but I can see how you got that as I put two thoughts together. My point is we have a large portion of users powering down on a regular basis and selling. Many see Hive as a weekly paycheck, and some depend on it as such.

My point wasn't that it would incentive it, just that it is an existing condition. If there is less incentive to power up (or even just not power down), it will likely happen less.

My point wasn't that it would incentive it, just that it is an existing condition. If there is less incentive to power up (or even just not power down), it will likely happen less.

I'm not sure that's two different things. I feel like it's the same thing, just expressed two different ways. But maybe I'm still missing something.

The only way to eliminate auto voting is to reduce curation rewards if you constantly vote on the same author. For instance, you could add a curation multiplier (0-1) that is associated with time. If you haven't voted on an author for say 5 days your curation is multiplied by 1. If you have voted on the same author 10 times that day your multiplier is 0. The difference is burned or donated to the HPS.

Does that make sense? It overly complicated things, but gives you the voting behavior you are looking for. Although vote bots could be setup to get around that. That is always the case.

A sybil attack (voting with a rotating set of accounts) would be the easiest way to overcome this. So given the added complexity to implement the algorithm initially, and the relative ease of beating it, it's probably not worth doing. Plus it would incentivize a form of "bad" behavior (the sybil attack).

Split the stake between multiple accounts and factor in the curation (loss) multiplier when considering votes.
It would discourage large stakes from frequently dolloping large votes in the same place so they'll at least have to cast their net wider.

Agreed. I didn’t come to that conclusion until I was mostly done brainstorming the idea though :)