You are viewing a single comment's thread from:

RE: OPEN LETTER TO STEEMIT INC., THE WITNESSES, AND THE WHALES

in #abuse7 years ago (edited)

I agree that there is a problem but those who issued the hardforks have made it difficult to agree on where we made the mistake. The 4 post limit did prevent abuse, but not completely. It was prone to Sybil attacks as we saw from steemsports at the time who after having 4 posts trending per day only had to create another account to get to 6 posts trending a day.

@richardcrill and I could agree yesterday on the after party of steemit ramble with @shadowspub that it was one of the 2 main changes of the last hard-fork (10 vote soft limit or linear rewards) that should be rolled back but we disagree on which one. With each hard-fork having multiple rule changes we're left with no clear direction and steemit Inc refuse to even see that there is a problem. As far as they're concerned none of this will matter when they've built SMT's. Steem is just the prototype.

If I could fork steem myself the first thing that I would do is add vote negation, so that abusive voting can be prevented without having to make it a full time job and without causing collateral damage by downvoting the people the whale upvoted.

The next thing I would do is roll back the last hard-fork completely and take one of those changes at a time starting with voting power. Instead of losing 2% of your remaining VP with every 100% upvote, you would lose more voting power when voting on posts with higher payouts than you would voting on posts with lower rewards - discouraging the pile-ons that the exponential reward curve encouraged.

What this would mean is that if you have 100 accounts with equal SP that make up the size of a whale, each vote after the last would become more powerful in adding rewards to the payout (that was how exponential rewards worked). However, the accounts that voted first would lose less of their voting power than the accounts that voted at the end when the payout was already high. This means whale accounts voting at 100% would be slowed down more than minnows voting for posts with low payouts at the time of the vote.

This is because I believe the linear rewards curve was the wrong solution to a problem we were already having. Too much power in the hands of a few who could set up a bot to vote 40 times+ a day on the sure-hit content, leaving no incentive to vote for new members on the platform and completely dictating the distribution of the reward pool on the same authors all the time without regard for user experience.

Whatever this comment gets paid though, it was still a waste of my time because the CEO of steemit Inc is not listening.

Sort:  

I know you are always interested in new suggestions, and I wonder what you are thinking about my (already rather old) post concerning the reward curve: in case it wouldn't start linear, self-votes on 'empty' comments were less profitable. Therefore my suggestion is to let the reward curve start as n^2 and end linear. You can find the complete reasoning here.

(Very interesting is also this post from @clayop which I found later.)

Yes I absolutely agree that if it's feasibly possible then this is what we need to be implementing. The problem is steemit Inc no longer concern themselves with the economics of steem. They want to create a demand for SMT's by leaving steem as the broken prototype. I'd wonder how long it would take for somebody to develop steem cash, a fork of steem. I bet it could come years ahead of SMT's.

"Instead of losing 2% of your remaining VP with every 100% upvote, you would lose more voting power when voting on posts with higher payouts than you would voting on posts with lower rewards - discouraging the pile-ons that the exponential reward curve encouraged."

This is an astoundingly good idea that I unreservedly endorse. I already hesitate when I upvote posts with rewards over $100, and I feel this proposal is responsive to why I feel hesitant. So much good content languishes, that even when I am upvoting good content that is being significantly rewarded, I feel like I'm failing the platform, and those other good authors that haven't got the traction, for whatever reason.

Thanks!

I'm glad I'm not the only one who doesn't feel the need to always upvote content over $100, even if it is good. The few times I do it, I'm giving it the $0.10 I feel I would have given it if I were among the first there.

It's sad that with all the algorithms and hardforks that there isn't a way to incentivize human curation rather than trying to get there by discouraging the opposite. I don't have any better answers. It's just a lament that everything has to be looked at from the gamification point of view.

How about posts that have the "extra" of solving a captcha to prove they are "humanly curated" get an extra payout or something? Because I agree--"science" (behavioral psychology) tells us that rewarding behaviour is the most effective way to promote it. It's called positive re-enforcement. "Punishment;" namely, a method which reduces behaviour, is not as effective. In other words, "Kindness Wins--" or, at least that's the idea ;))

Your captcha idea is interesting. I wonder how many of those I would have to do before I got tired of it, though. Is there a way to do that where I don't have to constantly prove I'm not a bot, but where the bot fails the test everytime? :) In other words, something that puts all the work on the bot and not on me. :) Then, I'm in.

I don't think so... That's kinda part of it--proof-of-brain--every time I'd imagine. But I'm sure some smart person will come up with something :) . It's really not so bad, imo; especially the ones that you just have to check a box--take one second. At the end of the post, check the box, then it permits you to post.

I suppose. I can't remember now, exactly, but there was something I did a while back that required a series of captcha answers to get through and it was just plain annoying.

We're agreeing though that something that allows humans to differentiate from bots is a good thing. :) I'd prefer not having to do anything else because I'm the human and don't feel like I'm the one who has to prove I'm human, even if it is a simple and easy, fraction of a second box tick. :)

Here is how you fix it:
#1. No more auto-upvoting. If you want to upvote someone, do it live.
#2 Limit the amount you can upvote to your daily upvote value (weighted by your level) 3x for minnows, 2x for dolphins, 1x for whale. If your daily upvote is worth $3.00, that's the max you can upvote, either all at once or spread out over 300 pennies.
#3 Throttle back self-upvoting to no more than 25% of your daily upvote value.

The losing a higher percentage of VP when voting on higher rewarded posts sounds like an elegant solution, but wouldn't this serve a 'redistribution function' more effectively if it was combined with linear rather than exponential rewards?

Even if the CEO of Stinc isn't listening ATM, it's still useful to work through the alternatives in case he ever does, or for contributing to alternative platforms... this is still only a test after all!

No I think linear rewards have proven to create the wrong incentives for everybody. As suggested by @jaki01, something in which starts out linear and gradually becomes exponential is probably best. But I think we would all appreciate one change at a time so that we can properly evaluate the results before moving forward.

Whatever you do, please do not fork into a directly-competing "Steemit cash"-like because what such a thing does is devalue both branches. I contend that ideally we should insist and bring more and more compelling arguments until finally Ned listens

By then the competitors will have won. I wouldn't waste your time on somebody with all the invested interest to listen but with no sense when the advice comes from an ex partner.

I wasn't partner with Ned. I plan to go to Steemfest 3 (and I hope he will come, too). I'll seek him out and try to persuade him. If the arguments are good enough and repeated often enough maybe there's a chance he'll begin to consider ? I don't know but it's worth trying.

Yes there's a risk that by then the competitors will have won but what we should rather focus on is: would creating a "steem cash" - like fork do more good than harm or the opposite ? How does that risk (of a fork proving actually detrimental) compare to the risk of "competitors winning before Ned starts listening" ?

I don't claim I have the answer to the above but my gut feeling tells me that the latter is lower than the former; I don't know Ned's propensity to listen but the risk of diluting the steem brand seems so high to me that I wouldn't take it (as a shareholder who's invested above $15K of own hard-earned money in steem)