What are the Economic Goals ?
To get people to vote in a way that honestly reflects their own opinion of a post's appeal with as little negative side effects as possible. Honest voting is imperative for us to succeed as a content discovery and rewards social media platform.
What's the Current Problem?
The current economy is paying content indifferent voting behavior (self voting, vote selling to bid bots, etc.) 4x more than honest curation. This has resulted in an increasing number of stakeholders participating in the former over the latter over time. Some estimates have active SP participating in honest voting at lower than 25% currently. A single glance at trending shows you the gravity of the problem.
What's Being Proposed In The EIP?
The exact numbers have yet to be finalized, but we're likely gravitating towards 50% curation, 25% free downvotes and a convergent linear curve. You can read more about the curve on @vandeberg's post here.
Why Not Make One Change at a Time ?
Because every measure has negative side effects that increase as you turn up the dial. By using multiple measures which compliment each other, we can effectively minimize negative side effects while hopefully still overhaul the status quo voting behavior into one that generally reflects honest opinions on content.
Many before us have considered these measures independently. @kevinwong and I mainly contributed by being among the first to view this as a problem of optimizing economic policy - understanding how multiple measures compliment each other and anticipating a desirable set of behaviors at equilibrium while being sensitive to the extent of negative side effects. We made the recommendation of all 3 measures and their ballpark values (and received immense help from @justinw and @vandeberg for improving our initial curve) which is likely why you see our names associated with the EIP.
What Are the Negative Side Effects of Each Measure?
Curation - On paper, less money going to content creators. (At much higher levels, content creators actually get less in practice too)Free Downvotes - Modest increase in toxicity anticipated at the recommended amount. (At extreme levels, collusive profit driven downvote cabals and edge case of rewards pool being blocked or nearly blocked)
Convergent Linear - Introduces some level of inequality at the very low end of post rewards. (At extreme level it leads to collusive 'circle jerking' and piling on, as seen under n^2)
Under our recommendations we expect the negative side effects to be moderate (no where near the extremes). We've taken great care to minimize them while hopefully retaining sufficient incentive to compel most actors to vote honestly across the entire system.
How do you Expect These Measures To Work Together to Encourage Honest Voting?
By increasing curation to 50%, we've instantly decreased the profit gap between content indifferent voting and honest voting from 4x to 2x.
A moderate amount of free downvotes is likely further required to deter dishonest voting. At 25% free downvotes and 50% curation, for every abuser, it'll take exactly TWO same sized stakeholders to bring their votes down to average (mean) curation value. That means, we'll need two good actors for each bad actor to bring their rewards down to a point where they might as well be curating honestly (roughly speaking).
Convergent Linear forces all profitable voting behavior into the light to be assessed by voters by rendering posts or comments at the very low end of payouts less profitable. Otherwise, schemes revolving around hiding self votes in spam in an effort to evade downvotes would likely take place.
But Aren't You One of the Bad Actors Traf, Ya Cunt?
There are game theoretical reasons why individual good acts are deterred under a heavily flawed economic system. I can spend $6000 a month fighting abuse with the full knowledge that at least 75% of that money is just flowing into the pockets of others doing the exact same thing.
It's just too expensive and too futile to do anything 'good' in a broken economy. Knowing this, most of us capitulate and choose to just maximize our own stake to mitigate the negative price effects of a failing platform and in the process we make it worse. This is why we desperately need the EIP. I'll likely be switching sides once that's introduced.
We don't create a set of economic rules that rely on good people to be altruistic for the system to work. We create a set of economic rules that force bad people to do good things in order to be profitable.
Also, there's no need for that kind of language.
How Do You Think The EIP Will Affect Self Voters Like Yourself?
We'd be destroyed. That's the idea. And it'll likely have a positive cascading effect.
Currently, I know it'll cost someone $200 a day to take my rewards away only to have most of it go to others just like me. So I doubt they'll bother, if they do, I doubt they'll keep it up and if they do, I can always just sell to a bunch of bid bots and be largely immune from attack. This common knowledge of both the expensive cost and futility of fighting against dishonest voting paralyses the entire system from self correcting.
Under the EIP, it'll only take slightly more than two others of similar stake to my own to kick my rewards down to the point where I'd be making more curating honestly. And because I know it doesn't cost them a cent, they'll likely not relent. So the longer I keep it up, the more money I'd be wasting for not simply choosing to curate. Therefore, rationally, I would likely just curate honestly.
Now I wouldn't be too happy with not being able to self vote while others still do, so I'll likely use my free downvotes to combat other abusers. This also frees up the downvotes that were initially used on me to fight abuse elsewhere. The idea is to get a positive cascading effect and topple the current status quo of content indifferent voting practices.
Wouldn't a Higher Curation Rate Take More Money Away From Authors?
Highly unlikely at 50%, they'll likely see an significant increase in practice. Currently, an ever increasing lion's share of rewards are just going back to stakeholders either through self voting or vote selling. Basically if we're able to get people to mostly vote honestly, 50% of a large pie is much better than 75% of next to nothing for authors, which is what they're getting right now.
What Level of Increased Toxicity are We Looking At Due to 25% Free Downvotes?
There will always be disagreement and negative feelings are unavoidable when real money is involved on a platform like this. There will certainly be some downvotes that are used for purposes other than honest and legitimate reward disagreement, and even honest downvotes are not without contention.
However, the upside is a very real possibility of overhauling this entire platform and converting it into a largely functioning content discovery and rewards platform that it was always intended to be. Think about how many communities would flourish if 100m SP worth of votes flowed honestly towards content based on their general appeal. I believe some level of toxic downvotes are a reasonable price to pay for a real chance at a working platform instead of this shit show we have going on now.
Can You Guarantee The EIP Will Work?
No, and you don't want me to. As every measure here has negative side effects the more you crank it up, there's basically a trade off between how confident one can be of success and the negative side effects we'll incur. The initial numbers would likely need to be further adjusted and optimized over time.
That being said, I personally believe we'll likely see a very impactful and positive change in voting behavior once the EIP is introduced and the dust has settled, especially if it comes with a downvote pool that can be separately delegated from upvotes.
Ultimately, if it works, it was all me, if it doesn't, it was a team effort.
When will the EIP Come?
Hopefully the very next HF alongside SPS. The dev work is minimal and will likely not delay SPS.
.png)
