RE: HF20 Exploratory Data Analysis on Proposed PayOut Changes
Thanks a lot for the analysis, I was just wondering how to get some numbers concerning the impact of the change after reading the posts you mention yesterday.
For me, as you say, a lot of people will quickly adapt their voting behaviour. I know I will, but being dictated what I can or can not vote (I won't vote anything who had a vote in the first 15 minutes, if the change happens of course) is bad. And the goal is the voter giving a reward to the author, why should a fraction of my vote go to people I didn't upvote ? If I wanted that, I would have upvoted them. Curation is the same thing, but it's still about a post, it's not distributed across all top-paying posts, I'm giving a fraction of my vote to someone who upvoted the post before me and so potentially helped me discover the post.
It's quite simple, so why is Steemit Inc. messing with this principle ?
And I agree with @tcpolymath that not only voters will change their behaviours, authors will too to prevent early votes.
But I see some flaws in Steemit Inc. reflection about this :
- What will stop all "big actors" (large payouts) to vote everything in the first 15 minutes to get maximum rewards for themselves ? It could create kind of a mess. They don't have to vote everything, but imagine they vote everyone with a reputation > 50.
- If the real goal is to discourage self-voting, why not simply forbid it ? Why these complex changes in the rewards algorithm for something you could do with a line of code ?
I would love to know the position of big accounts : @utopian-io, @busy.org, etc... Do they want a fraction of their vote to be given back to top authors that have no link to the project ?
In the current system, a vote at post creation time is fully credited to the author with no curation rewards for the voter. In HF20, the author would end up with "only" 75% of this vote and still no curation rewards for the voter. I think the voting "sweet spot" will move from the current 20-25 mins towards 10-12 with HF20, but curation wise it's the same game.
It's super easy to create another account, delegate all your SP to there and let that one vote for you. There are reasons where a self-vote is OK. Forbidding it will not solve the problem IMO.
For the self-vote, OK, I spoke a bit too fast and it's clear that self-voters and circle-voters will be able to find a solution to continue their business anyway.
For the rewards, yes, it will change the optimal time for voting but still, if I understood correctly, if a post is voted 15 minutes after its publication, a part of the curation rewards (all votes that follow) you receive now will go to a pool then distributed to top-paid posts.
Maybe I didn't understand everything, but that's this principle I'm speaking about. To get a part on all curation rewards (all votes) on every post, they could upvote everything in that 15-minute window.
We probably mean the same thing, but just to be sure:
For both HF19 and HF20, 75% of the vote value directly go to the author, independent of the voting time.
In HF19, and additional share of the remaining 25% go to the author, if the vote is cast within the first 30 minutes. This has a linear decrease, with the full 25% going to the author for votes at post creation time and the full 25% to all curators for votes at or after 30 minutes.
In HF20, the author only get's the 75%. The remaining 25% either remain in the reward pool for votes at post creation time (e.g. are not payed out), or go to all curators for votes after 15 mins. This curator share also increases linearly from 0 to 15 mins.
Examples:
We mean the same thing. But I'm really looking to understand better the new system, and I thank you for your explanation.
But there's still (for me) some grey areas. For example when you say :
Shouldn't it be :
My question (and I think that's what I understood from tc's post) is that votes following this kind of vote (at 7.5 minutes) will increase the value "redirected to the reward pool" (like with curation, when a fraction of the value of your vote goes to previous voters) - the value redirected to the reward pool during the first 15 minutes will be considered as a "previous vote" gettin curation rewards for all votes after the 15 minutes.
I might be wrong, as I said, I'm still trying to grasp the principle of this change.
not really, because the remaining 1.25$ are simply not taken from the reward pool to be payed out, so they don't go anywhere. This is a strange situation, because the pending payout of such a post will show a value of $10, but only $8.75 will be payed out to authors and curators.
This is correct. The vote will not be eligible for curation reward if it's done immediately after post creation, but for any time later, also before 15 minutes, this vote will also earn curation rewards.
OK, thanks a lot, I'm beggining to understand the whole situation better.
So, to sum up :
If there's a vote before 15 minutes, a fraction of the rshares (linear between 0 and 15 minutes) won't go to the post (so increasing the payout value of all other posts - sames reward divided by less rshares), and the fraction of curation rewards of this "ghost" vote value (generated by all future votes) will also not be attributed to anyone (once again increasing value of all other posts).
Have I finally understood the thing or not ? :D