this change is a strict improvement
I don't think that's true at all. Certainly it isn't in the mathematical sense.
Anyway, I think this change is a necessary one.
Why is it better than making curation always 25% and giving nothing to the pool? Why is it better than doing away with the early voting penalty completely?
Even if you really want to solve the perceived problem with extra author rewards, both of those would be immensely better ways to do it.
Mathematics? What value function are you optimizing?
I'm coming from the point of view that the community should decide, not the author, so... Yes, it's about the perceived problem with extra author rewards at least in the context of self voting.
Making curation always 25% might be better. But I don't really have a strong feeling there. Fix reward assignment, and it "going to the pool" is just fine with me.
Getting so hyperfocused on one problem that you cause four new ones isn't helpful. You can wait until every clued-in author is working to make sure no one votes on their posts in the first fifteen minutes, and until comment-section lag becomes an obvious problem, or you can figure out that those are going to happen ahead of time and come up with a better solution.
Why would an author do that? The author gets the same regardless in the new rules. The curator voting incentives also don't change. The only thing is that authors are going to get 25% taken away guaranteed rather than getting a possible bonus for early votes.
I know why that's not so desirable from the perspective of distribution, and we can talk about that part, because I still think that I like the penalty for early self voters. I certainly was taking full advantage of that when I self voted and botted.
I'm now finding the thread between you and @timcliff so I might pop a few words there. That one will be less about the author and more about the effects of the early window penalty itself.
This won't be true. Both sophisticated voters and curation min-maxers will make decisions on which post to vote for based on how much value their vote can give to the post vs the pool, which will be lower for any post that has early votes.
Setting up so that voters have to make that decision also makes voting immensely more complicated.
It's definitely a big advantage for people who know how to do it. I push that hard myself. But if we're going to remove it, we should remove it by making it equally desirable to have votes at any time, not less desirable to have early votes.
I can get on board such a solution, that's for sure. But I don't see how.
Oh about the other part, I forgot the author is voting too (too focused on the not self voting part haha....). I'm not really that concerned about that extra stake coming from the author to begin with though.
How does the situation change for curation min maxers though? It seems only to change for the author.
(Edit: Oh yes of course. Those that wanted part of their vote to go to the author to begin with. Yes that is true.)
One way to do it is just assign the full 25% across the successful curators. You can keep the early voting penalty if you really want it. Take all the money you would be giving to the pool and add it to the curation reward pool of the post instead. This allows early votes to encourage later votes a little bit but not nearly as much.
Because the total curation reward for a post that has early votes will still be less, their curation rewards per vote will be substantially more on posts that get the full 25% because all the votes came in after fifteen. It doesn't really change a lot for them from the way it is now, but authors will want to encourage more curation rewards and thus more votes because the early votes don't benefit them anymore.
But I don't think people will change how early they vote because they aren't all going to agree to cast their votes all after fifteen minutes. The only thing that matters for a curator min maxxer is value currently on post, and expected total value of post. I suppose the latter might change a little bit, but it's hard to say exactly how.
(Re: equalizing, I think it should be more in that direction as well. Sqrt curve is pretty brutal)