Thanks for the research piece @miniature-tiger, it coincides with the equilibrium projected from Steem's current economic incentives/disincentives. To address some of your thoughts and concerns:-
As the analysis above shows, doubling curation rewards will only bring large accounts to around 42% curation - still a long way short of the rewards available from vote-selling. Although the proposal can be considered a step in the right direction, it is a considerable (and for many, controversial) change that does not look likely to solve the problem.
50/50 is indeed almost an impotent change by itself since curation % is always subject to a secondary market and self-voters are both authors/curators so they get it all anyway, which is why some of us have been proposing another 2 items besides the 50/50 as per highlighted in the post here. It's meant to be taken in tandem. Think of the proposal as a chord, it wouldn't work or sound half as well when played alone vs all-together.
The proposed approach of 50 / 50 curation (like the current 75 / 25 system) is too inflexible. Different Steem users have different priorities and any new system should reflect this, allowing Steem to cater to all needs.
Other than maybe (just guessing) computational complexity of the distribution mechanism, I'll copy and paste some responses here, which further address the points we're trying to talk about:-
There is a dirty little secret about curation %, it can be circumvented via a secondary market. So in theory people are free to kick back curation rewards (which some bid bots do) and author rewards and reach their own %.
In practice, with a certain level of free downvotes, the official curation % will likely prevail in that it'll determine economic behavior.
The idea behind all of this is to leave as much behind for the author as possible while using a combination of bribes and deterrence to get the stakeholders to actually vote on what they like rather than take their own vote rewards (either directly or through selling them). Curation, free downvotes and superlinear are just bribes and deterrences that are necessary, but we want as little of it as what's minimally sufficient, as they all have downsides/costs.
Superlinear makes it more difficult to place an exact value on a vote per SP as it's value is dependent on the future popularity of a post. This makes it more difficult to just vote on something that's shit, as you'll likely get more from curation if you vote on something that'll become more popular. More importantly, it also forces all profitable behavior into the light. You can't spam 5c micro votes across thousands of accounts using a bot and avoid detection. Well you can, but due to superlinear, you're doing it at a loss, because 50% of something popular is more profitable than 100%% (as here you're both curator and author) of something thats garbage.
And think of a hypothetical Steem economy with 100% author rewards and 0% curation rewards, 100/0. Voters will most likely just upvote themselves all the time. Broadly speaking, 100/0 is in effect, the economic equivalent of 0% author rewards and 100% curation rewards, 0/100, as voters just end up getting all the rewards voting anywhere. Nothing left for authors either way for any decent content. Looking at this symmetry, it stands to reason that 25/75 today is in effect like 75/25, and 40/60 is like 60/40. So 50/50 is in effect, just the same at 50/50. In fact I'd expect anything besides a close approximation of the 50/50 effect would end up collapsing into somewhere near the 0/100 or 100/0 region, after a long period of time. Again, just broadly speaking though. Not really debating 60/40 or 40/60 is better than 50/50, as long as it achieves the 50/50 effect in the end for every vote given out to something decent.
In short there'll be a secondary market at 50/50, or even the voting slider. While I think it's alright as well, seems like SMTs is a better fit for such things imo, as with other kinds of purposes or use-cases for Steem. Plus wouldn't a slider just cause most voters to go for anything 100% curation? And finally to address your other point below:-
The idea of "content discovery" does not really exist on Steem and the incentives of the current system are misaligned, discouraging voting on content as it accumulates votes.
Yup our proposal is solely focused on making Steem work as a content discovery and rewards platform. Everything else can be SMTs. It's currently misaligned and we wanna align it. Imo, I think it's a mistake conflating blockchain activities and incentives.
Thanks for the comment Kevin. Hopefully a few useful clarifications!
Taking a simplified example of how I think it would work, if my vote is worth 100m rshares, then if I set my curation slider to 50%, I get (the vests value of) 50m rshares and the author gets 50m rshares. If I set my curation slider to 0%, the author receives the full value of the 100m rshares and I get no curation. There is no interaction between the votes of different voters on a post, so the calculations should be much simpler than they are now. This should decrease computational overheads for the blockchain.
I think that if the voter sets the curation slider then there should be no need for a secondary market, since all voters can extract the desired value of their vote from the "primary market" and cannot achieve more in a secondary market than through the primary. Users who currently delegate to bid-bots could instead set their curation slider to their desired fixed percentage and trail "Trusted Curators" for the community / tags that interest them. This would direct visibility (and most likely some reward value) to the best content.
Perhaps. Although currently any user who wants 100% of the value of their vote can simply sell their delegation / vote on an alt / join a circle-vote etc. Pretty much all of these damage Steem, particularly bid-bots, as the end result is the promotion of poor quality content on trending. I expect (hope?) that the users who currently do not follow these approaches would retain a level of author reward distribution, perhaps 50/50, perhaps the current 75/25, but decided at their own discretion. And flexible.
Then hopefully bringing the passive investors (bid-bot delegators) back into manual curation or a "Trusted Curator" system would also bring back some author rewards (even if only the margins that currently go to bid-bots and vote buyers).
In the end I think that people are ingenious and will always circumvent whatever constraints are implemented to extract the full value of their votes if they really want to. The proposal tries to work with that, allowing all people to participate in good Steem behaviour (i.e. voting on the best content) whilst allowing each user to take a level of rewards they deem acceptable. And at the same time allowing complete flexibility so that different Steem users, dApps and businesses can distribute their rewards in line with their own specific needs.
There are some of issues I noticed with the current curation model and the response provided. Currently curators are incentivized to vote on posts that will "become popular" but these posts are not better quality content they simply have the predictability of receiving larger votes in the future due to big wallets or consistent bid bot purchases. Many of these top earning posts have a narrow window of larger curation earnings due to the undesirability of authors burning their large or paid for vote power during the initial 15 minutes. Most of the 100%+ curation earners are automatically voting with small amounts in less than 15 minutes on posts that will receive predictably larger votes in the future which has nothing to do with curating content and creates a mob mentality momentum for posts that will predictably land on trending/hot lists. I am curious to know how many actual accounts' (or better yet unique individuals) posts and votes have earned what percentage of curation. I think the number is very small and is mostly auto-bot voting on a small group of consistently overpaid authors within the initial 15 minutes.
Yes, I think that this is the main approach used by the 100+ club. Those predictable votes can come from bid-bots, from dApps with high SP such as DTube / Utopian / Oracle-D, or just from users that consistently receive high upvotes.
I think that there is a second system which involves making a huge number of utterly tiny upvotes on comments that are statistically likely to receive upvotes, but I need to do more research on this.
It is possible to follow the first approach described above using a list of authors that consistently produce good content. But I would agree that this has nothing to do with the "content discovery" idea of curation.
I think that overall the current curation system is broken. This is why I suggest a "flat curation" system - one without incentives so people are free to upvote whatever content they like and are not influenced by gamification.
I agree, thanks for shedding light on this issue. If you want to look more into dust vote curation bandits check out these accounts:
jadabug
hdu
ezravandi
imisstheoldkanye
votes4minnows
delabo
cheneats
elviento
accelerator
penghuren
...and many others.
I think that some of these accounts follow the second curation system I describe above. I need to do more research on it. But at first glance I don't think that it does any particular harm to Steem.
I can hardly blame them for trying to maximize ROI with the current system. Once I noticed them doing this to my posts I created my own accounts to front-run them with slightly higher votes to take back the curation rewards. For example if I vote with 0.002 immediately before they vote with 0.000 or 0.001 it prevents them from taking so much in curation rewards. The problem is there is a limited amount of rewards and time to front-run a large payout post so the more people that do this the less chance a small account has to receive significant rewards by naturally curating without auto bot voting.