The increase in curation rewards should incentivize voters to vote for posts that will get more votes afterwards as opposed to just "voting for their own posts to capture author rewards". Also, note that voting for already popular posts leads to less curationrewards than voting for not yet popular posts, which in turn encourages 'finding' good posts before other people do. It's a sort of game to discover new 'potentially popular' content first.
There is no technical way to get rid of vote bots. But what can be done is modify the incentives to make vote bots less attractive. The new rules should decrease the attractiveness of vote bots by lowering the rewards for paying for their service (with the exception that paying for a vote bot vote to get visibility may be useful if it then results on a lot of 'follow-on' voting which would in turn mean that your post was interesting to other people).
There's also no clear way to prevent people form holding numerous accounts. But there's no reward advantage associated with distributing your SP among multiple accounts.
You can still vote for yourself and receive curation rewards, but the same vote could be used to vote for someone else's post before others vote for it, and the new rules would allow you to gain more curation rewards for voting for that post instead of your own post.
I think using curve instead of linear might reduce bidbot and self-voting or voting alts (of course this has to be combined with a separate flagging pool)
By curve, do you mean super-linear or sub-linear? I think a super-linear curve would encourage paying vote bots to upvote one's own posts to maximize the post's author rewards.
I understand the argument that this will make the post more "visible" and then it will get downvoted. But I'm not convinced this downvoting will actually occur.
Anyways, I consider superlinear curves a drastic solution, so I'd like to try other methods before moving to a solution that I think will introduce new problems.
I meant super-linear, don't need to be exactly n^1.3 as the proposal, basically anything more than n.
As for bidbots, my understanding is, under linear it's easier for bot owners to sell votes, as the voting amount of bots can be evenly distributed accordingly. It's more of a win-win for both vote buys and sellers.
While with super-linear, to split or distribute bot voting amount to vote buys is not as profitable (say, if 10 buys bid for a 500K sp bots with the same amount, they got 50K sp vote, but under superlinear, 50K is less than 1/10 that of 500K) which makes it less attractive for both owners and buys to use bots, meaning buyer might lose money by using bots, plus their content would get more "visible" which might bring potential flags. But yeah, gaming/last minute vote trading is still possible.
Ofc I might miss something here since some Steemians also told me super-linear curve would encourage paying vote, maybe we need some math proof.
Super-linear would make it more difficult for bot owners to estimate how much their vote is worth. But it also increases the value of "concentrated votes" and this is a much more powerful benefit for vote bots, which can easily make big votes due to their high SP. This benefit can quickly dominate over the relatively small issue of trying to exactly calculate the value of their product.
Now, it's possible that some sufficiently lower superlinear amount might not be such a huge benefit, but in such case it's also not going to affect calculation of the value of the vote much either.
I see your point, 50/50 curation would certainly be better, OK, if linear, then flagging will be quite important to do. Otherwise one can always vote alts instead of others and get more curation right?
There's no inherent reward advantage for voting for alts vs one's own account, in either the current rules or the proposed changes.
The only real reason nowadays to vote using alts versus one's own account under any of these rule sets is for "appearances". For example, by voting with alt accounts, you might fool other people into thinking your post had popular support (and conceivably lessen their desire to downvote the post for fear of "going against the crowd").