"Fixing" Curation Rewards -OR- How Much Is A Poor Person Worth?steemCreated with Sketch.

in #steemit8 years ago

TL;DR - DON'T change the n^2 voting algorithm.  Change it so that n varies with the square root of the voters Steem Power (instead of the Steem Power itself).

So . . . . we've come back around to the topic of curation awards.  For the most part, I agree strongly with ats-david's defense of them.  However, I strongly disagree with the belief that "piling on" is the true source of any problems.  Piling on is only a problem because predicting a whale vote is so lucrative.

Indeed, I would argue that the outrageously massive effects of whale curation is the ONLY currently obvious problem with curation awards.

A new user can come into SteemIt, spend about $15 and get 100 Steem Power.  With this, they have 0.00002 (1/50,000th) the voting power of the largest whales.  Users like @clayop and @smooth have noted that the n^2 voting algorithm makes this even worse (to the extent of effectively having 0.0000000004 the voting power) and have proposed changing the algorithm to n log(n) in order to fix the problem.  I argue that they are fixing the wrong problem and introducing unwanted side effects.

The correct solution is to use the square root of the voter's Steem Power instead of the full amount.  In this case, the new voter has 0.0045 (1/224th) the voting power of the largest whales.

While investment should be rewarded, there is really no reason why whales should be entitled to make such a killing with every vote and influence the community *so* heavily -- particularly since getting in early was the only real investment that many dolphins made.  Inequality of opportunity is a huge problem in the world and it is sad to see it replicated here when new users basically cannot get curation rewards except by front-running whales.

What do you say @ned and @dan?  Are you willing to trust the community (and take less money for each vote)?

Sort:  

Wouldn't it just encourage all the whales to split up their SP into smaller accounts to increase their voting power?

Sigh. Yes -- good point. BUT . . . .

  1. The limit of the effectiveness of this technique (as the number of accounts approaches infinity) is exactly what we have now -- so there is ZERO downside to doing this.
  2. There are several ways to reduce the incentive/ease of splitting
    a. Also include reputation in the calculation - I actually had this in there before but removed it for simplicity. I shouldn't have removed it because, again, it rewards very active posters.
    b. Require some sort of account validation to get curation rewards.

maybe some free post promotion with every powerup? Whales wouldn't bother with this while it would give incentive to new users.

These two ideas could go together well

The correct solution is to use the square root of the voter's Steem Power instead of the full amount.

Youre not supposed to use square roots on the blockchain. I know that because im super smart. Not because smooth told me yesterday. ANd also what tim said.

I'm not proposing to put square roots on the block-chain. Merely to calculate rewards with them.

yeah apparently you can't do that... i think it makes the blockchain turn evil or something.

Im all for equalizing influence, but it seems a little irrational to want a system where you can buy in for $15 and have a level of influence somehow comparable with a major stakeholder.

I don't call 0.0045 (1/224th) "somehow comparable" -- but what do you think is a fair ratio? Surely not the current 0.00002 (1/50,000th) . . . .

I think a stake weighted influence differential is fine. To me the guy that has 100 shares ought to have twice the say as the guy who has 50 shares.

Im willing to accept as a necessity that it probably has to be somewhat more superlinear than that.

So . . . . do you believe that Bill Gates should have thousands of times the votes for president that you have (based upon what he pays in income taxes and what you pay)?

No. Because america is a country. The president gets to make laws and do shit that has an effect on me even if i don't want anything to do with him. I can't sell back my stock in being an american citizen, wash my hands, and be done with it.

I believe he should have thousands (or millions or whatever) of times more votes for what product microsoft makes next... because microsoft is a company and he owns thousands or millions or whatever times more stock in microsoft than i do.

Ah . . . . but Bill Gates DOES have the equivalent of thousands of times the votes you do through massive campaign donations (if he wishes).

but Bill Gates DOES have the equivalent of thousands

The equivalent? No. He has considerably more influence than i do in some sense, but thats not the equivalent of having more votes than I do. Its something different but related (which is where the analogy started out, iirc)

There is a difference between 'similar' and the 'same'.

Hey, Mark. I just wanted to clarify something.

However, I strongly disagree with the belief that "piling on" is the true source of any problems.

I never stated that piling on was a source of any problem. If anything, it's just a consequence of voting habits, but it's not even a consequence of the curation rewards structure, in my opinion. I tend to think that a lot of the piling on is simply a result of what someone might call "lazy" curation. It likely happens because some voters simply curate the trending page or curate based on their list of trending authors that they have followed.

So, it's not necessarily a result of someone thinking, "If I vote for this person all the time, I'll at least get some rewards." Rather, it's probably something along the lines of, "This person seems popular, so I'll go ahead and vote for them because it's a safe vote...and I'll also get some kind of rewards."

In reality, the "piling on" from large stakeholders doesn't make much sense in terms of curation, particularly for those who are late to the party. Whales and most dolphins know this. It's much more lucrative to go out and vote for the content that hasn't been bombarded with votes already. Much of the piling on that we see is most likely due to these curators genuinely liking the content, thinking it deserves more rewards, or simply thinking that it's a safe vote for them to use.

My comments about piling on in my post had to do with the potential consequences of a revised algorithm that flattened rewards curves, which wouldn't necessarily prevent any sort of "lazy" curation, but would mitigate the effects of it by reducing the payouts as a result. It would force at least some of the curators to change their habits if their goals for curation were to receive a better return.

Sorry @ats-david . . . . I apparently phrased my intent poorly . . . .

Flattening the curve does reduce the payouts and spread the money further -- but I believe that it only hides the real problem.

I'd be willing to bet that reducing the impact of whales to a square root will also flatten the curve dramatically -- possibly even as much or more than altering the algorithm to n log n.

This sounds like it is worth a try.

If it doesn't work out, we can always switch it back.

The problem I see is that someone could get a few hundred bot accounts and wield a lot of influence. The whales don't use the full vote generally anyway. You'd have to buy $10ks of Steem to have much influence and most people won't do that. It could get interesting if some company decided to do that.

There was a reason why I had the $15 example . . . . :-) The problem could be mitigated somewhat if there was an account minimum for curation awards BUT . . . . honestly, it's pretty clear to me that, if Steem hits it big, there is going to have to be some sort of account validation for curation to prevent bot abuse. My minimalist bot easily makes over a 10% yearly return on investment and I could probably double that if I start throwing some of my AI expertise at it. The square root proposal plus account validation will dramatically reduce the incentive for a company to try to use a whale bot to harvest money from Steem.

I'm just following these ongoing curation discussions with some interest... the discussion I don't see so often is that of whether a simple upvote by a bot should be worth the same an upvote by a human who actually read the post and worth the same as a human upvote plus a meaningful comment left... where "meaningfulness" might be an extension of whether the comment was upvoted.

There is an interesting question as to whether you can reliably recognize a bot other than by the fact that they run 24 hours/day . . . . If someone wants to put up a decent wager that would make it worth my effort, I'd be willing to prove that you can't . . . .

Also, requiring a comment to be up-voted just isn't going to work.

Nice idea @mark-waser......i want to team up with you on Steem-pocalypse,you cool with that?

Unfortunately, Nathan Jones beat you to it. Maybe next time?

Maybe next time.Time difference btw Nigeria and Your side isn't helping me at all😢