Auto-voting is very commonplace on Hive. Because curation rewards depend heavily on the timing of votes many stakeholders opt for sniping for maximum curation rewards by setting auto-votes on relatively small number of authors. Setting auto-votes is also a convenient way to make sure all of ones voting power gets used.
But does this negatively impact the effectiveness of curation on Hive?
I think it does. That is why I think it could be very beneficial to reduce the effect of recurring votes.
What the blockchain would have to do for each account is to maintain an array containing recently upvoted authors and a number indicating the frequency of that author having been upvoted (weighed by the voting weights used) in the recent past (say, during the last seven days). The larger the number, the less valuable that vote and the smaller the curation reward. Voting on the same author's posts twice a day would probably be sufficient to incur the maximum penalty, which could be maybe as high as 30-50%.
Yes, anything can be gamed, but I think such a heavy tax on recurring auto-votes would probably change the behavior of the majority of stakeholders for the better by motivating them to seek out authors whose posts one has not curated before. Content discovery is still a problem on this platform and I think a reform like this could improve it significantly.
What do you think? If this approach is any good in your opinion, how should be varied or fine-tuned?
I'm not so sure that we want to ban auto-voting. It's a great way to support people even when we're not able to be active, and it also helps individuals use their voting mana more effectively (i.e. not have it sit unused). The issue with auto-vote sniping is something that has come up, but that's as much of a problem with the rewards curve as it is with auto-voting.
Lowering the impact of auto-voting for individuals with a small circle of frequently-voted people means that systems that automate voting over a larger number of individuals gain dominance.
It means that a system like SBI, which is highly distributed, will see no penalty, while social users with five or six frequently upvoted people that they know will see a penalty. That's not necessarily a bad thing, but I'm not sure it's desirable.
Some questions that would definitely need to be answered:
For a start, I think that deciding how many people counts as being too frequently voted could be problematic. For instance, I follow a couple hundred people, but only a handful of them are frequently active. Some people post monthly or every other month, and some of them are legacy follows from way back when and don't post very often.
So, if we look at my auto-voting, I've got it set for probably three or four dozen people, but five or six get the lion's share of the votes because they're the most productive. I curate my auto-voting list to include people who make work I like, who I've worked with, or who I've got some other relationship with. Not all of them are equally active; some are very active, some are not.
One of the other issues that comes up is people not knowing that they're hurting themselves. This is one of the issues I see with the reward curve, but it's sort of a natural consequence of the way the system works and there's a benefit (hunting for cool things, upvoting them, and sharing them is rewarded).
In this case, you would need to make sure that every front-end notifies people about the reduction in vote value from over-voting accounts, and auto-voters could get around this easily either by just voting more (mitigating the results of the penalty), or gaming the system.
Another question here would be how to account for the penalty. For instance, if someone is offline for a while and then suddenly upvotes a bunch of content from a particular creator, do you assess a penalty then, or is it for posts based on time?
It creates a new set of rules about voting on or uploading content that people need to be familiar with. It's already better to spread out posts, which is good for most users but limits professional content creation (e.g. if a major news service were to try to set up on Hive, it would suffer a diminishing return on every article past its fourth). More disincentives to frequent chain interaction could really backfire.
I think the biggest problem with this would be that it would have the opposite desired effect. Instead of spreading out auto-votes or encouraging people to vote manually, it would make it a better option to toss massive votes on people who you know are going to have a great return. Someone who posts good content frequently but who appeals to a niche audience or isn't as established would get fewer votes because voting on them often diminishes the rewards for voters.
Seems like a good idea. But I think the first mission is to get them manually upvote and close the gap between whales and us, mortals.
But, I see we both are in the dreamland xD
What this proposal would do is at least incentivise users to spread out their votes wider. Doing that requires that one looks for users who create content likely to attract votes from others whose posts one does not usually vote on.
I see your good intentions. BUT... :D
If you want to fix something, you don't use bandages to try to heal ''the problem''.
In this case, it's not that good as it seems to propose this because you will not fix the basics of the problem and that is, IMO, manual curation that brings real engagement that brings a lot more with it.
This way of fixing things is a very common way in the real-world and one of the biggest reasons why we are not fixing things but only fix them for a couple of years while ''bandage'' holds on until it falls off and brings the same problem on the surface.
When the base is good, healthy, motivating,... , then we can build a better blockchain on it. Until then, it's all wishful, IMO.
BTW Don't get me wrong, I'm glad you think this way and trying to find solutions. But I just believe we need a solid, healthy base from where we can grow.
How is the base unhealthy in your opinion? What could realistically be done about it?
I have a different opinion about how tweaks like this could improve things. The Economic Improvement Proposal implemented in HF 21 on Steem had drastic and positive effect on curation on the platform. It killed the bid bot industry. The free downvotes made abuse prevention much cheaper. I personally know many authors who started earning much better.
The base is unhealthy in many ways, but sticking to this topic, it is unhealthy because engagement is one of the most important thing here, if not the most. And, when you see whales and others just auto upvoting shit posts, the whole value goes down and people start to think that it is easier to auto upvote and have more time in your life. If whales are doing it, why not me, why should I care about the value of the coin if whales don't. (I'm just saying what is general thought on this).
We need to kill auto upvote apps or whatever it is. NO MORE AUTO UPVOTES. Isn't this the way it should be? Yeah, I know it doesn't attract investors, but hey, it's not that we attracted so many of them because of the auto-upvote feature.
Once again, I am glad to see and hear that you are trying to fix the problem, but fixing one problem will leave the base rotten again that will lead to just another exploitation of this blockchain. We need to fix the base. Simply. IMO :)
There is no way "we" can kill anyone's auto-voting apps they run on their own servers. And it's really easy to write a simple script to auto-vote on your behalf. There is no way to even distinguish all auto-votes from manual votes. To the chain they all look similar.
If you don't like auto-votes what you can already do is never upvote any post that has got a lot of them or is likely to get a lot of them.
Changes in the rules can lead to drastic changes in the culture. HF 21 is a prime example of that.
I know there is no way, I joined the dark side and I'm not manually upvoting anymore and wasting my time. We could possibly kill it if whales start to act as an example, but that won't happen.
That is a really sad thing for me when I think about how I was before and I'm scared that most of the people are doing that - turning from manual to auto-upvoting, lowering the engagement, giving money into wrong hands (low-value posts sometimes/mostly), lowering the value of Hive, etc...
Yes, they can. I hope they will, but again I believe we need a healthy base.
Good talk, btw.
If I were to decide I would ditch auto votes and self votes but, you seem to have a solution to "tame autovotes" somehow. The reason why I don't like them is because they steal the organic side of this platform and it gives it too much of a robotic nature. I admit that I don't like robots too much. If curation would rely only on manual curation the level of engagement, traffic and better rewarded posts would be much bigger. The level of frustrated content creators that have left the platform over time would decrease. That's how I see it. A first step to fix this would be to implement your strategy. I doubt though that most of the whales would like that.
No one cannot be forced not to auto-vote.
I don't think most whales would necessarily object to this proposal. They did get the EIP implemented after all, which killed vote trading. There are many whales here who are motivated to improve the platform for long-term gain.
There are some that only curate manually as well.
Killed vote trading, brought downvotes on us, ... Now they can take their money from DAO since we cannot do bat shit there. :) #justathought
What did you expect from POS? I came to peace with the situation.
What you mean by "on us"? I have only been maliciously downvoted a handful of times. And I've done my part in taking rewards from obvious farmers. The EIP has worked fantastically well.
The DAO currently distributes only 1200 HBD or about 3700 HIVE per day. The daily total reward pool is about 120,000 HIVE per day. Even of the entire daily DAO funding were completely wasted (it isn't), it would still represent only about 3% of the entire reward pool.
We now have free downvotes (and I love that as I love EIP, btw). I'm not saying they did that to downvote people nor I'm saying this was planned to harm ''us'' or smaller users.
Just the fact that people who post can get downvoted but DAO can only get ''upvoted''. That would not be an interesting thing to think about if those who can now control DAO funding didn't get that much HP from shady activities before. OFC, not everyone.
And, I'm not saying we are now doing a bad job with DAO. But it can be exploited so easily and we know how people behave when they see a chance to exploit the system.
You can always vote on the Return Proposal. But yeah, one account one vote would work better under the current proposal voting system where votes don't decay or there is no voting power calculation.
Well see how that's going to work out.
It's an interesting proposal in connection of voting circles i guess. But discovery of the content you are looking is still very inefficient, inaccurate and time consuming. In general finding the content you want is very problematic and if we force people to spend hours and hours looking for that content could make it even worse. i.e. powering down the accounts and dumping Hive or simply upvoting totally random posts. So i guess we have to take this into consideration when implementing this change imho.
I do not want to sound very critic but honestly, now when we have communities, tags we are using on our posts are outdated. So let's say you are interested in some particular form of content, for example photography. You join the community and get all photography related posts. But then i want to dig in deeper and see images of puppies, That's where tags would come in handy, but no one will put a "puppy" tag on it as they are potentially missing the exposure of their post because of the way content is being discovered.
So no one will put tags for a more focused content discovery, no one will be able to find content efficiently degrading their UX, and no one will take tame to look for that quality content because it's so inefficient and time consuming. And penalties will make things even worse i'm afraid and may result in damage yet not perceivable.
Re-engineering how content is discovered on Hive requires major overhaul from this perspective and i do not see it being done anytime soon since it is so hardwired into the concept.
So i guess we are all hostages to the rules that have been created as a base content discovery mechanics, hence we have a current situation where looking for content is time consuming and no fun at all and instead we focus on personalities and rewards than a content itself.
Current search tab does not do any justice to the content discovery at all imho. And this is supposed to be a platform running based on content discovery.
To illustrate my competence on a topic to some extent i'd like to make an example. I'm running a photography startup BeScouted more of a community project since i'm out of funds and it's on life support currently. I know what fits photography might not fit other types of content but still.
I i enter curly i get all images with curly hair.
If i'm into lingerie for inspiration or whatever any other reasons it's as simple as typing "lingerie" in a search field.
Boom, i come in, find the content i want, i upvote it and i'm on my other activities that need my time.
The content discovery on Hive inherited from steemit is broken. Period. It could be solved with the help of front end solutions i guess. Since not all of the tags need to be recorded on chain but then it is not as decentralized as intended and very application dependant. And imho all kinds of rules and workarounds will not fix it. Unless you fixed the content discovery mechanism. But that's just my opinion. I might be wrong.
But despite BeScouted lacking any kind of financial incentives, community members come, consume and share a really quality content on the platform, because they like the speed and convenience of finding it.
Well that's true.
Can you elaborate on that?
I'm wondering if a front end that stored all the content on a centralized database and that stays in sync with the chain but that allows for much faster searching would solve this problem. The text content could still be verified to be the same using hashes calculated from the post on chain as well as its centralized counterpart.
Partially that would help, but then again, it is a centralized solution. Imagine this being so popular and some centralized solution gaining too much traction, they would probably monetize in some way. And if it get's shut down for any reason, the whole decentralized community would suffer.
It is implemented in exactly same way on BeScouted right now. And it works, since we had steem rewards implemented it took the best of the two worlds. But then again, i am struggling to maintain the website as it has operational cost and might take it down any day if i run into financial trouble.
Elaborating on that would be completely fictional as i used the word "not perceivable". That might be my bad English, but i mean that there might be consequences we might not consider before implementing it. Like a significant number of whales and orcas loosing interest since they lack tools for curating effectively and might just dump the whole stake as they see it not reasonable to stake it anymore in HP.
Or maybe when Dan was writing whitepaper, sorting posts in search results by upvote value seemed pretty resonable, best posts collect most rewards and trend on top. But maybe he did not consider that steem power will be concentrated in a few hands that would result in voting circles or bidding bots. Or maybe he did and made it on purpose to increase value of steem by giving incentive's to whales to stake and rape the reward pool and dumped his stake at a good price. It's hard to say now.
My point is that while your suggestions is worth of discussion and might work, we need to give a very good thought of possible implication, and some of them might really surprise us and do a lot of damage. Butterfly effect.
To add even more if you still reading this rant, why do we need to reinvent the wheel completely and make things complex. I am a very strong of proponent that simple rules create complex objects or systems. I mean there are pretty sophisticated and time proven content discovery mechanics and algorithms used by In Fb Yt and so on. Why not just take it, and put it on Hive blockchain, improving it just adding decentralization and incentive aspect.
Let's be honest, apart censorship and centralized profit taking those platforms are great, educational, entertaining. They do not go well with politics ofc cause platforms themselves are getting asses kicked i am sure so they need to take action. There are pretty powerful old men in some places as we all know.
First and foremost, we must create a platform where people would kill their time for no rewards. Then add a cherry of financial incentive on top and it will fly! Now we are trying to push a bit shitty product to the masses through different initiatives. People come, try it, see how it's working and leave. A few stay. Or even worth we attract different kinds of opportunists, that abuse the system by posting various stolen content just for financial incentive. All of it is because the platform is not fun to sue for masses, and most of that i am sure comes from the ways content is discovered on Hive. I kinda like the individual orientated approach, where you interact with the select number of individuals you like. But that's 50/50 content/personality and that does not work for masses as it does for me.
Sorry for this being so long, i did not intend it :)