RE: Improving Steem’s rankings to cater to diverse content preferences
Afaics, the problem can not be solved only with filtering the display by tag or grouping without changing the underlying computation of the rewards. Without my proposal, the voters have a mathematical incentive to vote on only the highest payout posts, in order to maximize their own curator rewards. Maybe it is not clear to everyone why I think that is the case. The reason is because I presume the voter has no way to predict which voters will vote on each blog post and thus which blog posts which be optimum for them to vote on to maximize their curator rewards constrained to their cluster. So I presume absent an a priori strategy to maximize their curator rewards, they will choose to vote honestly according their content quality preferences. Note this aspect of the game theory needs to be pondered carefully.
Additionally, I quote what I wrote in reply to @smooth:
I realize some people may be voting their conscience in spite of it not maximizing their curation rewards, but still my proposal benefits them because their conscience will then actually be more highly reflected in the ranking for the cluster that shares the same like-mindedness.
As the system is right now, perhaps he doesn't even need to "predict". He can vote, see whether others voted, and then if they did vote and raised money preserve the vote. If they didn't raise money, he can proceed to unvote and vote something else: https://steemit.com/steemit/@alexgr/curation-gamed-through-unvoting
This doesn't work. I'll explain why in a response to your post.
I am not concerned about the rewards. I think the rewards are working great. Popular content is getting the most money. Sure you can have content which gets a lot of votes but which doesn't get much money because not a lot of voting weigh is behind it but that will change over time as more people have Steem Power.
I don’t understand why you are not concerned that people are apparently not always voting for the content which they like the best, but instead for the content they think can give them the highest curator rewards? I am doubting that you have understood deeply enough the various concepts and impacts of this proposal.
The point is voters are being incentivized by the curator rewards to not vote on the most relevant content, but rather on the content that other whales have voted on. This apparently causes a swarm vortex effect where all the votes get sucked towards what ever got the early momentum, regardless if that blog post was not so much better in quality than all the others.
Granted I am not sure that everyone is voting to maximize their curator rewards. But if say two whales who have 3 million Steem Power voted for a blog post, then many others will vote for it, because not only their curator rewards will be higher for doing so than wasting their vote on another blog post, but also because that blog post being at the top of the listing will also cause it to be seen by EVERYBODY more often.
Hypothetically my proposal not only divvies up the curator rewards by like-mindedness, it also divvies up what people see at the top of the listings so not everybody sees the same top ranked blog posts.
In the abstract, my proposal is about adding more degrees-of-freedom to the ranking system in all aspects of it.