A Simple Fix for the Problem of Low Effort Posts on Hive

in #hivelast year (edited)

Why do low effort posts feel so offensive?

When users make low effort posts knowing full well that they will get substantial auto-votes, it feels like they are just farming the Hive community without giving back. That is presumably happening, but it's also a somewhat counterproductive position. Social media is dominated by low-effort content - it is the norm. Nobody joins a social network in order to contribute effort, they do it to interact with their friends and family, and to find content that is of interest to them. Low effort content has value because it helps build a larger network that is more fun to be involved in.

And yet, if we allow low effort posts and low effort curation to run rampant, steemit.com is the result. The reward farming grows until it is all there is. Just browse that link for 2 minutes and it is plain to see.

image.png

Yet, this problem highlights a flaw in the design of the Hive rewards system. A social network that doesn't allow low effort content is not viable. The vast majority of the most viral content on Twitter, Instagram, Tiktok etc. is low effort content, and you never know what MS Paint edit will become the next mega-meme.

How to Allow Low Effort Posts without the excessive farming?

There are many possible ways to re-engineer the incentives on Hive, but with all the attempts to improve the tokenomics over the years, but attempts to tweak and adjust the tokenomics have uncertain results even in retrospect, let alone in advance. All the while requiring a major debate to reach consensus and often requiring substantial development work as well.

How about if we could fix this problem without any hard fork or any development for the core Hive code at all?

image.png

Because d.buzz is a short form content platform, they implement a default payout cap to discourage users from farming the rewards with short form content. Users can always set it higher or override it, but if you look at d.buzz trending page, most stick with the cap, and even when they don't, they typically raise it only to about 5-10 HBD. That's inclusive of curators, so a 10 HBD cap is usually just 5 HBD for themselves in reality.

image.png

Fair default?

I would suggest a fair default payout cap is 20 HBD. That's only 10 HBD for the author in practice.

If users do start raising their payout caps to $100 for their daily selfie, they are showing that they made a deliberate decision to farm the reward pool for more. In that case, have at them with downvotes, there is no argument that they are just using social media in a normal way. However the vast majority of well intentioned people who just want to use Hive like, you know, a social media platform, will be able to do so in peace.

Sort:  
Loading...

I really would like to see that experimented.
I find that an appealing suggestion. Leaving it up to the user actually seems to me to be a kind of inbuilt code of honour, where people are perhaps more likely to behave in a way that motivates them to give an accurate self-assessment of their published content. Rather than engaging in wishful thinking that mediocre/low effort content should be rewarded in a wonderful way. Although I would still say that, for example, visual material that is generally considered "valuable" because it is a contemporary testimony or other "low effort" products at first glance are not at second or third glance. The same applies the other way round: postings that appear elaborate at first glance can be quick and superficial products.

Thinking of widening/varying the functions here, I often thought about experiments with the "buttons", like for example to switch my user habit from the voting scale (preference method) into something different, the method of "least resistance".
On a scale of 0-10, one decides what resistance they experience towards a posting. Where one decides upon the least resistance, one gives a 0. Where one decides upon the most resistance (pain), one gives a 10. All values in between are given by feeling.

I think that would lead to significantly different results in the overall outcome of a posting and change as well the payout results (in the background, a 10 would correspond to a one to zero percent vote and a 0 resistance would correspond to a 100 % upvote).

Are you in the position to be able to code something like that and run it as an experiment?

I'm sure I could do it, but I've no idea what kind of effort it will take me to get familiar with Condenser and make such a change.

Do you mean something like finding the code for this element...

image.png

... and altering it to be something like this?

image.png

And then inverting the linear 100-10000 scale into a logarithmic 10000 - 100 in the background (technically a 100% upvote is 10000 weight and 1% is 100 weight).

To get any meaningful information out of it though you'd then need to test it on a bunch of people either by A/B testing it on Hive.blog or with a focus group.

Even then, in truth the big voters are auto-voting anyway, or delegating to curators who understand the underlying mechanics well enough.

Edit: After getting responses to this post, I no longer think we can fix the issue just with front end changes, as the cap would create problematic disincentives for curators. IMO, an upvote should have always had a recommended reward level attached, and the final reward should be based on the stake-weighted-median of recommended rewards. Curation rewards would also have to work differently in this case. However at this stage of the project I'm not sure if there's much value left in trying to continually readjust the tokenomics with hard forks either 😕

Thank you for your effort, appreciated.

Providing space for such playful experiments would be an innovation plus in my eyes. But you're right that you would need a very large number of participants who are keen to experiment. Or a completely new front end (high risk). What it would bring in the long term is uncertain.

With my idea, I wouldn't use emojis, but would actually work with numbers from 1-10. It would be exciting to compare the results, for example, if you set up a ranking and have results at the end of a quarter, for example, for a self-analysis, such as "which of my posts achieved the least resistance from the voters?" and also in comparison with other bloggers.

Curation rewards would also have to work differently in this case. However at this stage of the project I'm not sure if there's much value left in trying to continually readjust the tokenomics with hard forks either 😕

True. Still, it's fun to think of other than the usual functions and solutions. For whatever it might be useful.
Greetings :)

The use of emotes seems like a great idea.

Such a reward cap would need to be implemented on peakd first. Not sure if it is that easy.
And if voluntary, do you think that especially the shitposters would adhere to it? It is like with gun laws, the law abiding people are compliant, and the ones for which the laws are made, the criminals, they don´t care about them.

Think about it this way.

X = number of users just posting like normal on other social sites.
Y = number of users making low effort posts to farm rewards.

Z = number of users who need to be downvoted when their low effort posts get big upvotes. Without default cap, Z = X + Y.

If X don't bother to change their defaults or have only modest increases on default (the norm on d.buzz), then Z = Y, and Z < X + Y. That means it takes less effort to police Y, because there is no need to downvote X as well.

Y also lose plausible deniability, they can't claim they're being bullied off the platform for no reason.

We can ask @dbuzz, @chrisrice if implementing a default cap is easy. I expect it's actually just a few lines of code at most. Though it would need to be implemented by peakd.com, hive.blog and ecency at the least (three largest front ends).

I think PeakD.com already has a max-accepted-payout option.

The #1 issue related to max-accepted-payout is that curators get harmed on posts that earn more than the max.

That is why we switched the 0 setting to burn instead of max-accepted-payout. Needs blockchain level change.

Posted via D.Buzz

As far as I can see @peakd does not have this feature.

image.png

Interesting, I remember seeing it on some of the other front-ends.

That was over a year ago and it may have been Hive.Blog.

Posted via D.Buzz

Indeed Hive.blog does have it.

Cool! cool. Maybe I only saw it there and assumed all top social media platforms on #Hive had it.

Posted via D.Buzz

Hive front ends should consider adding a default payout cap, just like d.buzz has.

If a front-end adds this default-payout-cap option, they should perhaps also adjust their trending algorithms (e.g. to exclude posts that don't adhere to the cap) and/or add a filter so that users can set their own thresholds and filter out (from their feed) any posts that exceed a certain payout cap.


Ultimately, I think the key is having front-ends provide end-users greater info about posts / authors (e.g. circle-voting metrics) and let end-users easily filter out content that they don't want to see.


Also, I am curious as to what happens when curators vote above the author-chosen payout cap. Do they still get their curation rewards, and the excess author rewards get burned? Or does a curator who over-votes a capped post simply lose all or part of their curation rewards?

Maybe @chrisrice or @nathansenn know the answer to that question ...

The default ranking algorithm on hive.blog and peakd.com actually already account for this, as you can see if you look at d.buzz community on hive.blog or peakd.com.

image.png

Edit: misunderstood the first part of your comment. Filters are good in general, but I don't think they need to default filter based on the cap, because many users will have legitimate reasons to raise their cap (high effort posts, for example).

I think but can't say for certain that the curators end up splitting a smaller curation reward if the post exceeds the cap. I think basically all curators get cut in the same way. Curators will thus need to be more careful about how they vote, but IMO that is a positive thing. Lazy voting and auto voting should be discouraged. There will still be just as much curation rewards up for grabs as before, it will just take greater care to get your max rewards.

My understanding is that the excess remains in the pool.

Lazy voting and auto voting should be discouraged.

But if what you’re saying is true, that all curators get cut equally, then voting on such a post risks that some lazy voter after me will over vote and cost me an unknown (and perhaps large) portion of my curation rewards.

image.png

It looks like it does work that way. Aside from the one late voter, and some smaller votes that might be affected by rounding errors, everyone got the same rewards per rshare.

It's not ideal, because it does mean there's a risk of getting a lower reward per rshare when you vote on a post with a cap as opposed to an uncapped or very highly capped post.

There are ways to fix this but it goes straight back into hard fork territory 😕

Edit: A not too complex fix is to use current curation reward logic but apply rewards on a first-vote-first-paid basis. Thus only late voters are penalized when there is a cap. It's still a hard fork.

I will try and double check if my understanding is correct. I have an example here of one of my buzz's that would have gone over the cap, later in the day I'll look into it to see how it worked out for curators.

https://hiveblocks.com/hive-193084/@demotruk/1f7uq9p6yx7anq7ln61xgp

@trostparadox pointed out an issue we noticed in 2020/2021

We were interested in a change at the blockchain level that switched the algorithm for max-accepted-payout settings to burn author rewards above the max instead harming curators

@dbuzz may submit a PR for this. #HiveCore

Posted via D.Buzz

I like the idea. I figure this would be frontend-dependent Tweety-like posts and long-form posts can just continue as they are.

IMO the best approach is to have all the major frontends agree on a default cap, and then allow users to make a judgment when they want to adjust a cap. Thus, any change to the default is a statement of intent - "I worked hard on this and I think I deserve $X".

You have made such valid points dear friend, low effort postings can negatively affect the Hive Ecosystem but social media isn't supposed to be so stressful.

Usually I spend about an hour or two just to make a simple post because I wouldn't want a situation where someone down votes.

I think your solution with D. Buzz is also great. Can I make another suggestion though;

What if we can manage to create a TikTok version of Hive. Imagine telling people that they can earn the way they do on Hive by blogging but this time by posting short videos like on TikTok or when Vine used to be around.

It's just a suggestion though, I think database funding might even be a serious limitation.

Anyways, I loved your post the the idea behind it. You really nailed your point very well.

3speak is working on such a thing, you can check out 3speak shorts on their app.

Wow, I'm really happy you told me this, let me check that.. I'm sure it would be amazing.. thanks so much for replying 🥰

It's true that I was sitting around thinking the same thing a little while ago, that the posts are not good at all and they have two or three hundred dollars on them and they're on top of it. It's very heartbreaking for those users who work so hard and put their whole heart into creating a good post and when they post they don't get any votes at all. Yes, I myself have been working here for a year.

I think this is a great idea, very well presented. And it has generated some very informative comments about the mechanics.

If users do start raising their payout caps to $100 for their daily selfie, they are showing that they made a deliberate decision to farm the reward pool for more. In that case, have at them with downvotes,

Yes, except I would favour muting over downvoting.

Merry Christmas in advance 🎉🎉🎉. !AFIT


Hey @demotruk, you just received 10 AFIT tip from @caleb-marvel!
For more info about tipping AFIT tokens, check out this link

Hive la mejor red social del planeta...

Si eso es cierto, ¿por qué nadie se queda aquí? ¿Por qué Hive es tan oscuro e impopular?

Hive es lo máximo y dará impulso a las redes sociales descentralizados en un futuro a corto plazo dominaremos todas las redes, todos amarán ser Hivers.