DONUTS: a case study in tokenized incentives

[s_ben]

@Eth_Man

Hey @s_ben and @cslarson reddit/r/Eth_Man and Maker MakerMan here.

Sup Eth/Maker_Man! I’m a little disappointed you’re not @Cred_Man, but still glad to have you If you opt in, you’ll start getting some Grain (redeemable for DAI (manually, no Uniswap, yet)). See this announcement on our Discord for deets.

@Eth_Man

I have been encouraging other governance systems to adopt carl’s Vote=Min(EconomicToken,GovernanceToken) model because I feel that governance should both require an economic stake of some kind that can trade and a governance token that does not move to be important requirements to measure governance participation (ala what vote).

I really like the simplicity and elegance of this. Easy for all participants to understand, which helps with gaming.

@Eth_Man

In this vein the link I posted above some time ago I was doing an analysis of what DONUT governance was doing and discovered how with the above min(D,C) formula governance was basically losing votes due to participants selling their DONUTs. This has implications for the functioning of governance generally based on voting needing to meet or exceed previous thresholds etc. Example from the above is at the time I wrote that voting power of the 80+M Donuts was down to approximately 46M.

This assumes an absolute threshold, not a relative one, right? The threshold could be something like, 30% of the available voting power, which can fluctuate.

@Eth_Man

One thing I asked in reddit of carl was whether there was a way we could see how DONUT distributions were determined so we could actually propose changes to the distribution model there to be more like what soureCRED was doing.

Would definitely be curious to see the algorithm. I assume they’re keeping the contribution scores secret, to avoid gaming. In the $MOONS distribution proposal that went up the other day, it says, “You can create a CSV with alternative contribution scores or propose changes to the algorithm used to calculate them from karma (as long as the changes can be implemented easily).” But do you know if they provide raw upvote/downvote, karma data to start with?

Re: what SourceCred is doing, as far as monetary policty (e.g. how much Grain is released each week), for now @decentralion, our TBD (Temporary Benevolent Dictator), basically determines this themselves, with input from the community. We have plans to progressively decentralize this, but have only theorized so far about how that might look. E.g., is Grain distributed in proportion to Cred created in a given week (i.e. as a proxy for work done)? Is Grain distributed by how much is invested via boosting (on roadmap for later this year)? Is it done in order to keep a prudent reserve and smooth out volatility for contributors? All of the above? Lots of open questions and ideas. Following projects like DONUTs closely in part to learn from real word outcomes.

Using min(CONTRIB, DONUT) to vote on issuance, as done for the recent DONUT halvening, seems like a promising solution. The trick, of course, is balancing the interests of contributors and investors. Or, on a macro scale, capitol and labor. I like the simplicity of having a single number. Semi-regular voting on Cred/Grain issuance could be a fair, efficient way for these two parties to negotiate and perform “price discovery”.

@Eth_Man

In the end these systems ideally should have as a goal to become self sustaining, self supporting but since this is highly coupled to reddit there needs to be some level of co-operation between reddit and their subreddit communities. Ideally some of the economic value of $DONUTS should also flow up to them to help cover the support and co-operation needed to improve these systems generally.

Flowing Cred and Grain to dependencies is part of SourceCred’s long-term vision. E.g. in the upcoming beta release, the default is that communities using SourceCred will by default flow 5% of their Cred and Grain to SourceCred. SourceCred is likewise looking to start flowing Cred and Grain to its dependencies, including Discourse, which is OSS.

As far as being tightly coupled with Reddit, that makes me a little nervous…Seems like Reddit could pull the plug on this at any time, for example if legal implications get too onerous for them. In which case, subreddits have the ERC-20 tokens (very valuable), but lack all the infrastructure to, for instance change the banner ad, charge advertisers, etc. Reddit founders seems into the idea of ‘self sovereign communities’, and I belive them. Just don’t trust shareholders of large centralized platforms generally…path to sovereignty is long.

@Eth_Man

One thing I saw you ask ben was whether there was a way to determine if there was cheating in the system. I was watching this show called “Connected” on Netflix where the host talked about use of Benford’s Law in detecting all kinds of ‘fraud’ in all kinds of systems.

Benford’s law looks very interesting! Will have to check out doc.

@Eth_Man

As to collecting data regarding what is going on with $DONUTS if I ever hook up with someone who can pull on-chain data and provide data sets for some of this then I will be able to start producing say weekly or monthly reports. Not just how how distributions are going (we have multiple csvs on that) but also what is going on with the vote over time as well as the banner, special memberships etc.

If you ever get a decent $DONUT budget, might know a guy;) He’s pretty busy, but analyzing Reddit is his hobby, so might be swayed.