A couple of things:
- reputation metric isn't currently part of consensus, just a tool for steemit.com
- reputation is cheatable at this point in time
- we are looking at some of the following options:
a. no curation rewards for voting on posts above a certain threshold (~$1000)
b. reduce the target level of voting activity (faster steem power decay)
c. implement bots that give extra power to those with good reputation by following their votes.
d. implement bots that grab the most profitable curation rewards from good authors minutes after they post (giving 90% to author) and discouraging others from jumping on the bandwagon simply for rewards.
There are many options out there and we are looking for things that are simple, scalable, and derived from first principles.
how does this solve the problem outlined in this post?
¨The current problem is that newbies may post very good content, they have very few chance to be noticed.¨
to a) if threshold is per post and not per user it doesnt help to solve the outlined problem
to b) vaster decay of voting power also doesnt help that voting is more spread, therefore is contra-productive to solving the above problem (that said out of other reasons im still in favor of it)
c) this also doesnt help newbies
d) also doesnt help newbies
I think a) actually will help the newbies. It goes like this:
Too complex. Why not just remove curation rewards?
I also feel that is the right idea. But then questions is how to encourage curators to work for the best of the platform, since its already hard to curate content. People will always upvote what they like, even if that doesn't bring them money. Just the discovery of new content will be harder, someone has to do that. If steemit would employ curation team to boost quality by giving some power to good posts and no power to bad ones and then present content to public, we would have a delay on new content but would certainly help things out along with the removal of curation rewards. Some may see this as some kind of censorship but curation is essential for this platform.
If you are trying for simple and scalable, why not impose a kind of "luxury tax" of say 15% on the top 15% of posts with a set timeframe, say 24 hrs (noon to noon).
-money set aside in pot
-given to top 15 posts with content deemed worthy (under $100 initial payout)
-voted on immediately after the announcement of the contenders for the payout (vote for 24 hrs)
-voted on by everyone, 1 vote per session (you can only vote for 1 post)
-highest vote getter gets receives the most reward, scaled down from 1 - 10
The best way I see to set up the secondary vote is an additional ONE VOTE button. This button would be used only to select the top 15 posts under $100, with payout percentage TBD (To Be Determined)
This will give you a vehicle to recognize minnows without having to change the voting structure. It also gives minnows a chance to have equal say in determining a payout, with the 1 vote rule to determine the 2nd chance winner. Thanks for your great work on Steemit.
1 ) I didn't know that
2 ) Tell me how in PM please ... (just kidding)
3 - a, b, c ) I see your team was already working on the solution before I even spotted the possible problem. I wouldn't expect less from you guys ;)
3 - d ) Woudn't it reward massively bloggers that already have good reputation ? If I had a bot following me and voting up after a minute anything I post, I would certainly post lots of low quality blogs every day.
I fear that bloggers that usually make a few hundreds of $ on their post continues to earn the same amounts (or even more) independently of the quality of their content. Only because bots and people expects them to continue earning this much.
b. It would encourage users to be more selective on what to vote and how much to vote in general. That is the best solution from the one's you've stated above. Also it's fitting to the current growth rate, a system like that I believe needs to be implemented at some point in the future.
Dan, is it theoretically possible to implement reputation into the consensus protocol?
I have some ideas that would make the reputation system almost un-gameable and bot resistant. I will post them soon, but they all require additions or modifications to the current consensus system.
I am excited to hear about the options you are considering. I really appreciate the thought you are putting into this. I think the voting/curation/reward system is one of the most important issues Steemit has at the moment. We have to get this right! Having the right incentive system could be the difference between Steemit becoming the #1 mainstream content site that everyone uses, or it not. I also find it impressive that you took the time to read and comment on this person's post. The level of caring that you and the other people working on Steemit have for the site gives me a lot of hope for the future of Steemit!
Hi Dan. I like (a) personally : )
However, I'm not sure it solves the issue, because the OP argues that quantity not quality will be rewarded for those with lots of followers, and that is what needs to change.
An issue I see with a. and b. is that they could still hurt a steemit minnow if he/she is having a rare successful post. I wouldn't want to see that happen (since I'm a minnow!)
c. and d. sound like better ideas to me. They are a good way to de-incentivize bad voting habits without hurting authors or good voters.
It's good that conversations are happening at a high level about tweaking the rewards system. It feels pretty rough right now in the sense that rewards don't match quality as much as it matches who wrote it. And as this post alludes to, much of the high quality minnow content goes unseen.
Why not gradually reduce the curation reward for posts that have already been curated much? I know there already is a large penalty in voting late, but the OP asserts that it is still profitable to blindly vote for trending posts. If that is true, simply add a factor to the curation reward formula that further reduces or eliminates rewards in proportion to the number of votes the post already has. Whether this be better done by count or by steem amount should be carefully thought through. It might be better to use count, to introduce a little bit of a "one steemer, one vote" element into an otherwise pretty "the rich get richer" themed scheme.