Quote from: iamnotback on July 25, 2016, 11:51:55 PM

Quote from: iamnotback on July 25, 2016, 09:33:43 PM

The voting and reward algorithm in Steem seems to me to have a fundamental flaw.

If they made voting rewards linear, then all SP holders could simply vote for their own posts and recover their share of the dilution of the money supply. The game theory would be there is no financial incentive to vote for posts of others (although there might be an incentive for minnows who value the site functionality more than the tiny bit of rewards they control).

So instead they made rewards quadratic and even time incentivized so that you have to risk your vote on a post that you can't be sure will gain you the most rewards (since you don't know which post all the other users will vote on until later, which is why the votes are time incentivized to vote early). That also had the hype benefit of creating posts with exaggerated $50,000 rewards due these non-linear amplification algorithms.

But the quadratic and time incentivization game theory highly favors whales who can collude since the remove the risks the algorithms intended to create yet the colluding whales can take all the non-linear reward amplifications. I am not saying whales are colluding now, but for example if the media moguls can obtain significant stake and then accumulate 100% of the tokens over time by this game theory strategy thus controlling ranking of content on the site.

I believe the only solution to this is to let users control their own ranking algorithms (so whales can't predict rankings even if they collude on voting) and votes for computing rewards should always be linearly tallied. In other words, rankings need to be truly decentralized else the entire system is a clusterfuck back to centralized media control again. Meaning that if you only vote for yourself, you voting pattern make not align with like-mindedness with others thus you may have no influence on the ranking of your posts on the site, so you lose income from the votes of others. Meaning if you vote for yourself, you opt out of blogging.

Meaning I see no solution to the fact that it is impossible to incentivize whales to vote meaningfully and it is impossible to charge whales for the voting of minnows¹! Dan's communism fails.

I'll be reading over @theoretical's blogs to see if I have made any errors or incorrect assumptions. I'll update you if I find any.

¹ I do see one solution to this where whales are charged for voting but not allowed to vote but it can't work in the Dan's current design where STEEM are debased 50% yearly because obviously whales won't decide to hold STEEM.

I remember now that long ago (perhaps it was 2014) I already gone down this wild goose chase of voting from shared debasement and had realized it was fundamentally flawed and can't work.

Steemit is fucked. There is no way to fix this. Absolutely impossible. Fugetaboutit.

What they did was put the whales in control so they could over pay for blogs to generate the delusional to-da-moon groupthink.

I can't think of any way sure way to fix it. We might hope if we adopt my solution quoted above and make the forced voting small enough, that minnows will just vote their conscience (because it is a hassle otherwise perhaps), in that case maybe it works. But the whales will still opt out by voting themselves, so the minnows end up paying for all the blogging, which means it won't have sufficient funding. Power-law distribution rules.

If whales have a huge incentive to be restricted from voting (e.g. they don't have to hold for 2 years), then perhaps you can charge the debasement to them. But if the minnows are voting with the whales' share, then the minnows have more incentive to vote for themselves.

It just doesn't seem to work no matter how we design it.

Steemit's whales can justify throwing away money because they know otherwise they can't cash out. This is still a ponzi scheme and race to the exits before everyone realizes the voting system is totally dysfunctional. It will become more and more apparently over time.

They may be hoping they can get enough users into the system to begin some ecosystem work on microtransactions. Maybe the big lie in the voting algorithm is just a stopgap measure to onboard users. But I really can't see this group of developers ever succeeding because they seem to be always about fooling people.

Solid businesses are built on solid ethics.

I think maybe I have a design idea on how to structure debasement of voting that solves the problem above that I thought was mathematically insoluble!

I don't want to say I am certain, so I will think about this more (especially when rested).

If I am correct, then this changes my opinion about voting being worthless and insolubly dysfunctional.

Edit: hypothetically another advantage of my design concept is that the voting algorithm would not be a globalized calculation, which is a very important criteria if we want to be able to shard the blockchain for scaling purposes, which is something Graphene can't do.

My idea emphasizes the SP as "karma" and mathematically so.

P.S. I worked 18 hours today! My health is cooperating.