Curation rewards are extremely low, only bots are willing to work for pennies/hour

in #curation8 years ago (edited)

Ok guys, there is something seriously wrong with the curation rewards. https://steemit.com/curation/@snowflake/why-curation-guilds-are-essential-to-the-future-of-steemit
I have been through hundred's of posts and the very large majority of them ( 98%+) only allocate about 3% of the payout amount to curation. This is what @pfunk told me

It changes the % allocated because curation reward is dependent on stake. If a good amount of stake donates curation rewards with votes before 30 minutes, it means more curation rewards in total will be donated to the post author.

This is my answer
What is the purpose of this rule? I don't get it.
Curation rewards should be a fixed percentage of the total payouts and people who find good content early should be rewarded but i really don't see the point of the rule you mentionned?
Good curators should be able to make a living of curation..with many post getting 3% curation rewards no one but bots is going to invest their time curating.

The curation rewards is the whole reason why people buy steem power, its also the only way for most people to actually participate in the system, why the heck would you pay curators only 3% of a payout's post?

Please someone clarify what is going on with curation rewards and have a look for yourself on posts before you say rewards are not 3%, because 98% of them are. No wonder why users are leaving, for some reason I always thought anyone could make some good money if they put some time and effort curating but this is absolutely not the case.

Sort:  

I've never found a good explanation of the 30 minute rule, and when I first found out about it I thought it was pretty dumb.

I've thought about it a lot, and now I'm pretty convinced that this rule is actually extremely important. Here's the practical effect of this rule: it incentivizes people to go find unpopular content. The 30 minute rule guarantees that essentially nobody gets any curation reward for voting for @charlieshrem or @dollarvigilante, as you've discovered. However, the curation rewards for posts like these can be surprisingly high:

The 30 minute rule is supposed to reallocate curation rewards away from popular content, and should disincentivize groupthink. One problem is that absolutely nobody has any idea how curation rewards actually work, so people still keep upvoting @charlieshrem thinking they'll get something for it.

ok that make sense

The 30 minute rule guarantees that essentially nobody gets any curation reward for voting for @charlieshrem or @dollarvigilante, as you've discovered

You know what? Im not even sure it does. Because popular content always gets a lot of money, 3% of $200 ( which is $6) is still better than 40% of $10 ( which is $4).

I really don't think reducing curation rewards when bots vote before 30 min is necessary to incentivizes people to find good unpopular content, the rule where if you upvote early you get more rewards is enough for people to focus on hidden unpopular content because that's where you will have the most chance to upvote first.

Now we end up in a situation where curators only receives 3% of the post's payout, that means people are disincentivize to participate in the system and to buy steem power.

Wouldn't a simple captcha get rid of the bots?

Wouldn't a simple captcha get rid of the bots?

Since the voting takes place at the blockchain API level, this would not work. There were a lot of discussions about it when the platform was starting to take off, and the general consensus was that there were going to be people who used bots whether we want them to or not.

With this in mind, the best approach is to find a way that bots and humans can at least 'get along' and even potentially complement/benefit each other.

The overall response from @biophil is totally correct though. Good curators (ones that want to make money) need to find ways to innovate and detect quality undiscovered content before others do. With the 30 minute limit in place, it forces them to look for things that are not already 'popular'.

Wouldn't a simple captcha get rid of the bots?

I don't think the point is to get rid of bots.

Now we end up in a situation where curators only receives 3% of the post's payout, that means people are disincentivize to participate in the system and to buy steem power.

Manual curators can earn 20% annual returns on their SP; automated curators can earn over 50%. Those are decently large numbers.

You know what? Im not even sure it does. Because popular content always gets a lot of money, 3% of $200 ( which is $6) is still better than 40% of $10 ( which is $4).

True. But $6 is a lot less than the $50 it would be otherwise. It all comes down to a philosophy of what curation rewards should do. The devs feel that curation rewards should favor exploration, and I tend to agree. Ask yourself this: why should people who vote for @charlieshrem get anything? Everybody knows that his posts are going to be popular, so it's a waste of money to pay people to "find his posts."

But that's just one philosophy, and the reason I adhere to it.

What do you think about authors deciding how much they want to give to curators?

The devs feel that curation rewards should favor exploration, and I tend to agree. Ask yourself this: why should people who vote for @charlieshrem get anything? Everybody knows that his posts are going to be popular, so it's a waste of money to pay people to "find his posts."

The problem is that decreasing rewards on popular post doesn't achieve the intended effect, people still upvote these posts and these post are still getting the bulk of the money. So now we end up with an even bigger problem which is that curation rewards are extremely low on the platform.
If devs want to favor exploration , letting authors decide how much they give to curators is the best way to do it imo.

The problem is that decreasing rewards on popular post doesn't achieve the intended effect, people still upvote these posts and these post are still getting the bulk of the money.

It's true. I've always been disappointed that the website gives absolutely no feedback to voters about how curation rewards work, so incentives only do something for people with the time and expertise to dig into the source code.

I see you are third in this list http://steemwhales.com/trending/?p=1&d=30&s=cr
Is this a list of the best curators? how are best curators calculated? how is curation score made? are you using a bot?

It's a list of the most profitable curators, normalized by SP. I think the curation score is calculated by this formula: (curation rewards earned in past X days) / (steem power).

So if you look on the 7-day list, the top three are currently @laonie (a whale), @biophil (a dolphin), and @philipnbrown (a minnow). They each curate at about the same efficiency, even though their SP amounts are vastly different.

Yes, I do run a bot. If you look through my account's voting history, you'll see that my bot never votes for popular posts - it specifically searches for posts that might have high rewards, but haven't received many votes yet.

If you'd like, you can send me your Posting Key and I can add you to my bot so you can rise up on the list of best curators as well. :)

Do you have an email address where I can send you my posting key? Thank you

Got your key, added it to my bot, and it all worked. You've already cast a vote or two. Results not guaranteed! :)

Yeah, you can grab my scraping-resistant email address here: http://www.ece.ucsb.edu/~pnbrown/

I have sent you an email with my key.
Let me know if you have it.
Thx

When people vote on things, the curators who vote first get a more than than the people who vote later. If there was no 30 minute delay, then the bots would all 'get there first', and human curators would be put at a disadvantage.

If you look at posts from lesser known authors (that do not have large bot followings), the curators from their posts usually vote later so the curation rewards are closer to 25%.

The problem with this is that posts with the biggest payouts will always have the lowest curation rewards because bots will always upvote them before the 30 min mark.

It's a game :)

I never understood the purpose of the 'timing' either.
seems to me it should be like a book store...if you see something you like, buy it. Then the author eventually get's paid.
Right now more time is spent on strategy and tactics attempting to 'game' the system than on actually (gasp) reading stuff.

I know that because I've made posts that would take a LONG time to view...and yet it got votes almost instantly.

I understand why they would put a 30 min delay because the earlier you curate the more rewards you get so it makes sense to put this delay so that bots can't be first every time, however I don't understand why would the curators rewards % decrease when bots upvote a post before the 30 min mark.

I always thought anyone could make some good money if they put some time and effort curating but this is absolutely not the case.

If you check the list of most effective curators you could see that humans still can compare with bots )

http://steemwhales.com/trending/?p=1&d=30&s=cr

Thanks I was looking for these stats. Do you know how the curation score is calculated ?

I think it's described here, but they might have made some changes
https://steemit.com/bots/@trogdor/building-better-bots-ranking-the-top-curators

The comments explaining the purpose and mechanism of the 30 minute rule are largely correct.

However, I do think you make a good point that the aggregate effect of these rules may be to result in in curation rewards that are too low, leaving little incentive to buy SP or to participate in curation. (This was the reason for my objection to cutting curation rewards more than in half from 50% to <25%)

It may also perversely incentivize voting for unpopular content that is actually crap. If @charlieshrem's posts are good (and I personally think they are, but that's obviously subjective), I'm not sure there should be an inbuilt mechanism encouraging people to not vote for them, and to vote for some mediocre poem instead, just because the latter doesn't have any votes yet.

I think @svamiva's idea is really good, it will incentivize new authors to be upvoted because they can put a larger curation rewards and it will also force bots to actually check the post curation details before upvoting.
I also think that late upvoters are important and that the degree to which early curators earn should be more even.
Do you think you can discuss this with the dantheman, ned or the devs team, I really think @svamiva is onto something here? Also do you know if these changes can be made at the blockchain level or only interface level?
I wrote a post about this, hopefully I get input from the steemit devs.
https://steemit.com/curation/@snowflake/let-authors-decide-how-much-rewards-they-want-to-allocate-to-curators
Btw thanks for upvoting my latest posts :)

I think the best way would be really let the author decide, how much curation reward he wants to give to curators. Someone who wants to make his post more visible might choose 1% of total reward for author and 99% for curators, someone like @charlieshrem might want to choose 99% for author 1% for curators option.

That's a good step but there are other issues, like the degree to which early curators get all the rewards, and later curators get little to nothing. In a very real sense, later curators votes are just as important and valuable, because they determine the amount of the reward. Yet they are incentivized to go find some crappy smoothie recipe and vote on that instead. I like your idea better than the current system though.

they are incentivized to go find some crappy smoothie recipe and vote on that instead

I doubt that anyone of those who are on the trending curators list by the curation score uses this method, vote for some random crap and hope for the better. Of course, it is an issue, but actually there's just a few dozen whales who are really incentivized to do so. But that problem could be solved through implementation of "internal rshares market", so any whale could just have a smoothie on the beach and collect interest for delegating her voting power without bothering with curation.

I doubt that anyone of those who are on the trending curators list by the curation score uses this method, vote for some random crap and hope for the better

It was hyperbole to illustrate a point. The algorithm favors unvoted content over voted content far in excess of the difference in quality or user preferences, or more to the point in large part regardless of the gap between current expected (relative) payout and (relative) quality. In fact top curators must use something in principle similar to the method I described, just a somewhat less extreme version of it.

the algorithm favors unvoted content

Just try to imagine the opposite

far in excess

Something highly subjective, isn't it ?

difference in quality

Even more subjective

From my personal experience I never suffered much because of need to choose between "low-quality/high expected payout" and "high quality/low expected payout" because of having enough "high quality/high expected payout" options ( and somehow restricted amount of voting power , I don't really want it to sunk below 80% , if I start to play in "casino" I think it would go below 50% )

Anyway, I was never able to figure out why rshares market idea is something to be ignored, is there something I don't know about ?

[lower reply]

Something highly subjective, isn't it ?

Not the way I'm using it. I'm taking the actual rewards produced by voting as the objective consensus of the user base as to the 'quality' or 'value' of the post (of course each voters' individual decisions are subjective). It is factually the case that later votes contribute to that value but receive little to no curation rewards. This directly incentivizes voting that diverges from objective consensus value. Larger payouts occur only because people ignore (lack of) curation rewards and vote anyway, but that doesn't change the fact that the curation rewards system directly discourages this and imposes a value system that demonstrably differs from that revealed by actual voter behavior.

Anyway, I was never able to figure out why rshares market idea is something to be ignored, is there something I don't know about ?

The team seems to prefer the curation guilds idea, which could be seen as a form of rshares market, just a somewhat different take on it. Both ideas do have their downsides, however. I don't have a link but there was a good post and discussion on the risks and downsides of curation guilds a while back. Most of the same arguments against that idea would apply to other forms of rshares markets as well.

well, yes, the system rewards those who discovered quality and value of some particular post earlier ( and I guess it's how it supposed to be by design ) but those who discovered it a bit later are probably to much penalized.
I think the "taxation" for early voting before 30 min shouldn't be paid to author but rather somehow distributed to later curators instead...but I guess you would say that alone is not enough ?

As for curation guilds I've seen a lot of discussions. For me the biggest downside is that it is sort of "bureaucratic machine", once established it's difficult to shut it off. If I don't have any long term plans, just want to curate some particular event ( like Steemfest for example) it would be nice to have an opportunity "to borrow" a few millions SP for a few weeks for this purpose on some sort of internal market. I guess guilds and rshares market could coexist, there's no need to stick for just one option.

A post is not truly curated until it has a good number of votes to signify its value

That's a brilliant idea.

Well, also on the long run I think that would work well, there might be some caveats.
Someone could decide, that instead for spending time trying to write one good quality post with 75% for author, instead one might just create 20 accounts and provide 100 crap posts every day with 10/90 author/curators ratio.

There might be another option as well, "your reputation = your author reward percentage "
So, 25% for a newbie, you and me should be happy with 55%, 66% for @timcliff and so on.

Sorry for two replies, but here is an example of a high paying post that had a larger share of curation rewards for the users. You and I had a good long conversation in this post if you remember ;)

https://steemit.com/steemit/@timcliff/attention-sharks-there-is-blood-in-the-water-are-you-ready-to-invest

yeah I remember of course :)

So does it means that your post wasn't upvoted by bots before the 30 min mark? if so why?

I really think they should solve the bot problem differently , because this is hurting real human curators. It's hurting everyone actually even bots. Curators should receive a fixed amount of the reward's payout.

Bots are only part of the picture. Anyone who votes with a lot of SP (whether it is a bot or human) is going to have a large influence on the curation rewards. For my post, most of the big SP votes came after the 30 minute mark.

I really think they should solve the bot problem differently , because this is hurting real human curators. It's hurting everyone actually even bots. Curators should receive a fixed amount of the reward's payout.

I think most of the curators are OK with the 30 minute limit. There is a game to be played, and there are still lots of routes to go that can earn good curation rewards. Voting for the top posts where the author gets tons of upvotes right away though is not the best place to do it though.

Also I have noticed that bots do not upvote posts exactly at the same time, some 2min,5min,10 min after post was published so I think it would be trivial to game the system, people could set up a timer to upvote post 30 min after it was published, Im sure many do this already.
The main question is why the heck do bots upvote before 30 min knowing that if they do so the curation payout will decrease and why does the curation payouts even decreases, what is the purpose of this? preventing bots to upvote before 30 min? it clearly isn't working..

It's an arms race between the different bots and real people. Voting earlier lowers the percentage reward for curation but getting in earlier than the other bots means that you get more of the curation payout. There is no perfect time to do it and people and bots are constantly fighting over this. Without the rule there would be no parameter to adjust and bots would win every time. It is not ideal but it actually improves the odds for real curators.

I get what you mean but i still don't understand why they vote before 30 min when they could all set up their bot to vote after 30 min and benefit from the large curation percentage.

Because if someone votes before them they get the majority of the reward, even if the reward is smaller. All it takes is for one to do it slightly earlier so maybe someone does it at 29 minutes and they get nearly the full reward and vote on it first. Then next time someone goes at 28 minutes and so on. Although the total curation reward goes down it may still be worth doing if you can vote first or second or third. Ultimately it becomes a race to the lowest level where it is still profitable there are so many potential variables that it may not be possible to have a perfect algorithm.
Further some people may actually want he majority of their rewards to go to the poster.

Do you know how the algorithm works, like how many percentage does the first,second,third,etc.. gets?

Further some people may actually want he majority of their rewards to go to the poster.

I see curating as the main way for people to make money on the platform.
As they say, only 1% of the users of a website actively create new content, while the other 99% of the participants consume it.

Curation is extremely important because of this, steemit will have to find a way for the 99% to make money and that's going to be through curation.
In the future I can easily see 10% go to authors and 90% go to curators because that's what the demographic tells us. If you want the 99% who consume content to buy steem power you need to make it worthwhile for them.

I seem to remember seeing a graph on it way back. I don't remember the exact figures but it goes down from the moment of posting to the 30 minute mark as a percentage. I don't know the exact figures for curation rewards percentage depending on where you vote in the queue. It is a function of when you vote AND the voting power you have (the stake you put in) so it is not a straightforward value.

It doesn't really matter to me that much I just upvote people whenever. Like someone else said it has the side effect of encouraging people to vote on unknown authors if they want good curation rewards. For me the most important priority is encouraging other users and rewarding good posts. If I get rewarded for curation great but I don't care much about it.

You don't care much because you receive some rewards from creating content, if you couldn't make a penny by posting I think you would be really happy to be able to make money curating. ( and that's about the majority of people in this platform)

I personnaly think we should get rid of the bots altogether with some captcha or something

You can make money if you curate intelligently. It takes had work though and I think that is the real problem.

Everyone can potentially make money posting but I understand that not everyone wants to put the time in that is required to do that. It is not something that everyone enjoys and some people aren't going to bother unless they make big rewards right away.

IMO that is just unrealistic and short sighted. You don't get something for nothing - or at least 99.99% of the time you don't.

The problem with trying to get rid of bots using captchas is that they can access the blockchain directly so it wouldn't work. We just have to accept the bots are there and will always be there.

Further if curation rewards are all you care about and you don't want to go to the trouble of curating new authors you can use the guidelines put forth by @laonie to configure your own voting bot.

This post has been linked to from another place on Steem.

Learn more about and upvote to support linkback bot v0.5. Flag this comment if you don't want the bot to continue posting linkbacks for your posts.

Built by @ontofractal

nice post!could we follow and upvote each other?

Coin Marketplace

STEEM 0.17
TRX 0.12
JST 0.027
BTC 55093.98
ETH 2894.76
USDT 1.00
SBD 2.22