I’m intrigued by voting & payout info moving to the Hivemind layer. I know it can be a contentious issue... but if the processing burdens are eased enough could this eliminate the 7 day window and facilitate “evergreen” content? I was always under the impression that aspect was tailored to fit what was technologically possible, and wasn’t a core tenet of the system itself. Is it becoming possible to scale past that limitation?
You are viewing a single comment's thread from:
The current changes only impact storage of non-active voting and payout info (old votes and payout information is no longer retained in hived after a post has paid out).
The actual calculation of payouts still needs to be done in the 1st layer code (the blockchain code itself), as you believed previously. The only way to move this calculation out of hived would be to pay the rewards in something other than HIVE/HBD.
As to the reasoning for the 7 day window, we haven't looked into why it was set to one week, although I was under the impression (seemingly like you) that it was to prevent too much load on hived (the core blockchain process).
Isn't the reason for the seven day window also to prevent delegating one's HP to several different accounts to be able to self-upvote own posts several times?
By the way, I think we should reconsider the five minutes curation window (directly after a post was written), because it clearly disadvantages manual curators who try to read and evaluate posts before they decide to upvote them or not. Until they have done that always lots of automatic upvotes have already come in. Also it makes 'curators' lazy (especially since there is 50 % curation reward): they simply upvote the same people again and again, while not enough users are seeking for new authors (respectively honor posts from old authors who aren't on their 'lists').
I see no reason to be rewarded less for a late manual upvote than for an early automated upvote.
Considering the high percentage of automated upvotes, I think it's simply a myth that manual curators find 'quality content' early, and then more upvotes follow and reward them.
I wrote two posts about that problem:
Here and here.
I don't think so, as the 7 day post reward window existed before delegation. But I think the delegation "return delay" is based off the 7 day post reward window, to prevent abuse of the type you mention. So it would likely need adjustment if the duration of the post reward window changed.
I agree that the current curation reward system is still not tuned very well. As you've mentioned, it has become obvious that the super-linear reward for early voters hasn't done one of the jobs it was designed for: to reward manual curators versus bots. But it was also designed to discourage "pile-on" voting where everyone just votes on the same trending posts, because later votes give less curation rewards. Again, it's possible it's not really doing this either, hard to say for sure.
But proper tuning is actually quite difficult, IMO, because the intersection of economics and human behavior is often more complicated than people expect. It may turn out that the only way to find something that works well is to keep making adjustments.
Cap all rewards at $20 hive.
That way, when a post maxes out, curators and bots will have to go searching for new authors.
The bots will just keep voting on the same maxed out posts which will give a distinct advantage to the humans!
I think that limit would be way too low to differentiate between 'great' and 'not so great' posts.
I'm not convinced that total payout is the most reliable indicator of a "great post".
The burnposts make tons of money for curators and ste.emcleeners (not exactly content that would appeal to the public at large) was the most consistent top earner.
What about a cap of $50 ?
The overwhelming majority of the posts I make and vote on make less than $0.25
Big payouts don't seem to attract readers or commentators, but they do seem to attract autovoters like flies.
Not always in reality, but in theory I think that if a post is 'great' (I know that's not easy to define) the author deserves 'quite some' (again it's not easy to name an exact number) reward. :)
I am not a fan of such strict limits, because actually I like the freedom to upvote whatever I want (also if already many other users liked the same post). I would prefer incentives to spread one's upvotes to many different authors, like for example 'diminishing returns', which wouldn't forbid to upvote the same users again and again but made it less beneficial.
Also the ideas I discussed with @blocktrades could lead to less vote accumulation on the posts of 'famous' users.
I agree with you that the rewards for many short posts from 'popular' authors (often witnesses) are way to high, considering the information they offer and their content quality.
Of course you are right! How time sometimes can blur the memory ... It's much more logical that the delegation "return delay" has been adapted to the post rewards window than reversed. :)
I wonder if curation rewards should solely depend on the weight of the upvote? So that neither the date of the upvote nor the number and strength of upvotes of other users would influence the curation rewards of a curator (but only his HP and the percentage of the upvote).
I know that would make self-votes on comments etc. more interesting again, however, at least for the author it would still be more beneficial to get more upvotes than only his own self-vote.
So as curve for the author rewards we could stick with convergent linear, but for curation rewards the simple formula that curation rewards would only depend on the vote weight of the curator could apply?
I know the matter is complicated and maybe my ideas have more weak points than I am aware at a first glance ...
Someone actually suggested your curation reward idea recently. It may be better than what we have now, but one potential problem is it might lead to more self voting. That could potentially be countered by downvoting, but so far all signs point that downvoting just doesn't happen much, no matter how the rules are changed.
Interesting (and good) to know that this idea is actually being discussed, and that some stakeholders are aware of the suboptimal curation situation (which can be quite frustrating for users who try to curate manually)!
Concerning my idea, right, the self-votes are a problem, the flags part of a possible solution.
Actually, I would modify my idea above in a way that reduced the benefits of self-voting:
I know that the probability to see this last step is rather low ...
Some time ago I had the idea of 'diminishing returns' (against self- and circle-voting) when upvoting the same accounts again and again (within a short timeframe), but it seems that can be circumvented by creating many alt accounts (even if it still should be suspicious and detectable if a certain group of accounts only upvotes each other).
... I will keep thinking about possible other solutions ...
I don't understand why "self-voting" is considered a crime.
And if it's so horrifying, can't you just restrict it in code?
I wouldn't call it a "crime", but in case everybody just upvoted themselves there wasn't any (financial) incentive anymore to produce 'quality content'.
There also wasn't any reason anymore for new users to join, start posting, tell their friends to join too (network effect), and thus increase the value of HIVE.
That's not that easy (even if there would be ideas) because of the possibility to create multiple accounts.
Cap all rewards at $20 hive.
That way, when a post maxes out, curators and bots will have to go searching for new authors.
The bots will just keep voting on the same maxed out posts which will give a distinct advantage to the humans!
Where the heck is this post ($40.55)? - https://hive.blog/hive-122315/@logiczombie/trust-cdc-gov
IMAGE SOURCE - https://hive.blog/payout
@jaki01 who do you think I should ask about this?
Do you still have the url of this post?
Then maybe you find it in hiveblocks?
Here's the url - https://hive.blog/hive-122315/@logiczombie/trust-cdc-gov
I'm just wondering what I'm missing, I've been lucky, the last few posts I've made received over $30.00 Hive and none of them showed up on the https://hive.blog/payout listing.
So the problem is only that it wasn't in the payout list? I wouldn't bother too much ... you got the money anyway. :)
However, if you want to open an issue you can do it there: https://gitlab.syncad.com/hive/condenser/-/issues
But maybe I didn't get what the problem is and you meant anything else?
When we spoke before, you suggested that higher payouts = higher ranking on the trending page and potentially higher views.
I don't really care about the payout, but what I do care about is views.
I just thought it was strange that none of my high-payout posts showed up in the rankings.
I'm not suggesting anything nefarious, I'm genuinely curious how this could happen, I mean, I've always heard that it's simply ranked by highest pending payout, but that doesn't appear to be true.
Is it perhaps because I received a rather large downvote on that post from @blocktrades? Do you know if that's somehow factored into the rankings?