Should Compound Retroactively Airdrop Tokens to Early Users?

Ideas
232

Okay, I think it’s time for a poll!

Summary of recent developments for folks just joining or returning to this discussion:

  • There appears to be consensus in this forum toward some kind of airdrop to early users with vesting.
  • Many good ideas about the distribution model have been shared, but the conversation stalled due to the absence of a curated early user list with sufficient data to inform a capital-weighted distribution.
  • I was inspired by @grasponcrypto’s efforts with Dune Analytics data to try my hand at curating a list of early-user interactions by directly querying on-chain data.
  • The community has helped refine this list, e.g. @grasponcrypto and @rleshner provided the guidance I needed to isolate and remove addresses that participated in an early Sybil attack on governance.

I am hearing two main objections to this proposal as posted:

  • Objection 1: the calculation of capital input should be an integral over C(t)m(t), rather than a sum over supply/borrow events of C(t_i)m(t_i), where C is the capital borrowed supplied and m is a multiplier that favors interactions that occurred earlier in the eligibility window (when Compound had a smaller TVL).
  • Objection 2: the 95% weight to the social distribution is too high.

The first objection is addressable but would require considerable additional effort to implement. Given the level of effort required, I would seek support from the Compound grants program to do this.

The second objection is easily addressable; I just happen to disagree with it. As evidence, here are plots of the distributions for the 95%-5% case and the 50%-50% case.

95-5 distribution
95-5 distribution800×400 17.9 KB

50-50 distribution
50-50 distribution800×400 18.1 KB

These distributions look almost identical, except the 100 or so addresses that could afford to plow multimillions (or more) USD equivalent into the protocol receive ~10x the rewards in the 50%-50% case. All that extra COMP comes from smaller users, who each end up with only half of the 20 COMP available in the 95%-5% case. I don’t see how that helps achieve the goal of promoting participation in governance by early users.

However, we should hear from more voices than @Andre1 and myself to understand where a consensus lies, if any.

What do you think?

Do you support the current proposed early user distribution list?
  • Yes, I support the currently posted list as-is, not interested in seeing how the more accurate block-by-block capital weights would change things
  • Yes, but I support @allthecolors applying for a Compound grant to implement the more accurate block-by-block capital weights and produce a report for the community
  • No, but I support @allthecolors applying for a Compound grant to implement the more accurate block-by-block capital weights and produce a report for the community
  • No, I don’t support the currently posted list, and the block-by-block capital weights wouldn’t change my mind
0 voters
9 Likes
233

About the " * Objection 1: the calculation of capital input should be an integral over C(t)m(t), rather than a sum over supply/borrow events of C(t_i)m(t_i), where C is the capital borrowed supplied and m is a multiplier that favors interactions that occurred earlier in the eligibility window (when Compound had a smaller TVL)."

Are you making a sum for one given adress of all the capital that was supplied even in case it was repaid, resupplied, repaid (the same money) ? meaning the whale who withdrew and supplied again next minut the same amount has his supply amount counting twice, three times, or ten times in your calculations for the capital weighted reward ? If yes it’s yet an other flaw of this approach.

If I get it right, let’s say Franklin : he deposited 100kusd on compound in 2018 which he locked for years.

Now you look at Anton : he would deposit 100kusd on compound too in 2019, but then he would play with aave and a dozen of other protocol back and forth, in a way he effectively played with only 100kusd, but deposited and withdrew it often from compound, with effectively dedicating its funds on compound a much smaller time than Anton. Assuming Anton would have done it a hundred times.

We end up with Anton received 100x the allocation that Franklin would get ? Despite Franklin was locking his funds to compound (the same amount) for a far larger time ?

Also I can tell you that the top 100-300 people in your ranking are far from millioneers, we are talking here about some who dedicated a dozen k

I would say there could be intermediate solution like 80-20, but even if keeping 95-5 most of all having that capital weighted allocation in opposition of the socialized, really capital weighted and less arbitrary as possible matching successful distribution as UNI

.I hope a compound grant can already be obtained for all the work you are providing, anyways I will support the proposition choosed by community and which is feasable, it seems people are voting for the current distribution

question : Would there be any easy way to extract that number per adress : Notional value of total interest paid + received ? Would then be a much fairer approach if that’s an easy path, but I don’t know if that number can be obtained

2 Likes
234

When the COMP Distribution began, it was expected that Governance would alter, improve, and rethink the program–its fantastic to see the community organizing, and doing the research to distribute COMP to early users.

There are a few questions to answer, and build consensus around:

  1. What is the purpose of distributing COMP to early users? Why is this a better use of COMP, than distributing it to current users, or to protocol improvements through development grants?
  2. How many COMP should be distributed to early users?
  3. What should the vesting schedule, mechanics, and implementation be?
  4. Which addresses should be eligible, and how should the quantity of COMP per user be calculated?

A lot of recent discussion has gone in to question 4, and I wanted to document my own views, and some of the historical issues surrounding this question.

The addresses that receive COMP should be early users that risked their time and capital to establish the protocol. This means excluding addresses that were deliberately created to manipulate the protocol; one example of this is the first community vote, which was “sybil” attacked with hundreds of addresses supplying ~0.01 Ether each.

Second, early should be defined as anybody that tested the protocol before the COMP token & distribution was announced on 2/26/20.

Third, users should be measured by usage of the protocol. For an interest rate market, usage is a function of capital over time; how other protocols have distributed tokens to users is irrelevant to Compound. This approach was piloted during the second community vote, and formed the basis of the COMP Distribution when it began for users. Luckily, in Compound’s case, this can be easily measured by interest earned, and interest paid.

Lastly, several members of the community feel that the distribution should exclude or limit “whales” in some way, even though these addresses took the most risk, and contributed the most value to the protocol. This is a political question – but could be implemented in a simple way, by allocating through the square root (or another power) of interest earned+paid, which would partially “flatten” the wealth effect.

I’m excited about the effort occurring here – and wanted to plug the Compound Grants program as a tool to help organize the research, smart contract development, auditing, and proposal work that this distribution will require.

23 Likes
235

Agree with Robert on all points here. To penalize large users for supplying big capital to a risky protocol (at the time) seems not just unfair, but borderline malicious. If the protocol had been compromised and they lost everything, would the community coalesce and chip in to make them whole? Successful risk taking should be rewarded, not punished.

5 Likes
236

How about smart contract wallets like Argent and Dharma? I prefer to use Argent to interact with Compound. However I did not see my address in the list.

4 Likes
237

@Compactivist see post 214 Should Compound Retroactively Airdrop Tokens to Early Users? - #214 by allthecolors

1 Like
238

I am not sure if the definition of early meaning before the token release in Feb 20 is the best fit. I’d recommend setting the cut off date somewhere when DeFi started gaining traction. I think the COMP distribution update was more meaningul in this regard than the original announcement COMP Governance & Distribution Update | by Robert Leshner | Compound | Medium

2 Likes
239

My first interaction with the protocol was on March 28th almost a month after the announcement and I wasn’t even aware of a COMP gov token.

Me and many early users interacted with the platform for the first time out of curiosity.

I discovered compound finance because of the integration with the Coinbase wallet.

So I made my first deposit with BAT then DAI then a monthly deposit and so on.

De-Fi wasn’t a thing until the summer, and I believe most early users weren’t aware of the future value of a government token.

I think @alive suggested date is good enough as an starting point and in the case of the Sybil attack, we could filter the wallets that interacted less than 2 times with the protocol as @blck suggested.

Personally, I believe this airdrop could help Compound in reach a better decentralization, as long as it doesn’t discriminate the early users who genuinely believe in the protocol.

3 Likes
240

If the protocol had been compromised and they lost everything, would the community coalesce and chip in to make them whole?

Already happened.

The rejection of this proposal is one of the reasons of why a Pro-rate airdrop will not work for decentralization.

It’s either a fixed amount for everyone as the social distribution of @allthecolors suggests or a minimum amount as other protocols did.

Making this process more complex it will take longer and longer to achieve, it’s been months since someone decided to take real action and started to share relevant data to finish this proposal.

2 Likes
241

I think the purpose of COMP distribution to early users should be to increase the degree of decentralization, because it is a COMP governance token (at least that’s what we call it). Early users risked their capital and interacted with the protocol during the bear market without the possibility of liquidity mining. On the other hand, most existing users are attracted to the COMP distribution, so the assumption is that they already have enough governance tokens (if they have sold them then their intention is clear). My opinion is that snapshot need to be taken on token launch (not on announcement).

The only fair distribution of COMP tokens is completely social (100%) so all eligible addreses get the same amount of COMP because it is a governance token after all. If we include capital distribution factor then the act itself has no purpose because the structure of the holders will remain the same (centralized).
Measured by usage is possible only before start of liquidity mining because last months only “whales” can afford to pay gas. Except that, usage metric is super relevant and it is a good idea.
Vesting period is necessary due to the possible DUMP on the market, + distributed and vested COMPs with some collateral feature (for accounts and platform stability, not for leverage).

agree, process will be faster that way

6 Likes
242

Time to reward those who deserve to be rewarded!

Thank you for your contribution Borovan

1 Like
243

And/Or a “thank you” gesture to all those who took part in the early days of the protocol, borovan shared his experience and feedback very frankly and if others feel the same it might be a good way to make amends. But other early large contributors may feel otherwise.

1 Like
244

I agree with you that Pro-rate airdrop doesnt change anything.

You mean early contributors or early investors? Early contributors were paid for that contribution and early investors obviously got their COMP share in which they invested. If I’m wrong, so the contributors worked for charity and the investors donated their funds to the protocol then the problem is in Etherscan data.
The purpose of COMP airdrop distribution should be protocol governance decentralization (in some degree) and treat each user equally regardless of the amount of capital he owns.
The fact that certain persons / funds treat the “DeFi” protocol as shares of the company creates a problem because “Defi” and “user governance” memes obviously serve as good marketing.
Borovan’s situation maybe is unfair, but I don’t believe he didn’t make money from the aforementioned liquidations.

3 Likes
245

It is clearly.
we hope see the proposal soon

246

You are totally right, it’s about Fairness. This means something totally different from one person to another.

Maybe his expectations were too high on “how compound would retroactively reward him if ever super successful”.

Something to put more thought into maybe: milestones programmatically enforced for community contributors (i.e. community does xyz and if the project reaches xyz target then the relevant members get an airdrop). Never really heard of this but it may help getting this early community members/contributors more excited, more fairly rewarded and remove implicit hopeful expectations.

1 Like
247

fwiw, UMA offers a solution that achieves exactly what @YouCompDolt is describing (so-called KPI options) with the “xyz target” being the total value locked in the protocol. It’s an interesting idea; at the same time, I see a lot of interest in this thread in keeping the early user distribution as simple and easy to understand as possible.

5 Likes
248

Looks like Sushiswap is now following UMA’s KPI Option model:

This is not about retroactive changes so we are getting off topic. But seems like a logical trend we might want to follow in our next strategy moves.

2 Likes
249

If the set kpi parameters are quantitative then we know what to expect - whale manipulation. I don’t know what qualitative indicators would be possible …
If the goal is to achieve greater protocol decentralization, then “KPI” is not exactly the best solution.
The UMA has targeted governance participation on its own and several other protocols. In the case of Compound, this is pointless because governance comes down to a few large accounts, and I don’t really believe it’s fair to any Compound user to reawrd accounts that didn’t use the protocol.
I will not comment.SushiSwap because they are not relevant.
@allthecolors I think the solution you offered (with 95% social weighted distribution) contributes the most to achieving the goal of airdrop. However, I assume that the goal is to decentralize governance and encourage early users to become more involved in working on protocol.

250

Itamar from Argent here.
@allthecolors, it seems you’re excluding all smart contract interactions including smart wallet users: Argent, Gnosis, etc…

Any rational for that? We could easily provide a list of Argent contract factories, or identify Argent wallets within your initial early users list if that helps.

We integrated Compound natively in Argent extremely early, in summer 2019, our users interacted with Compound before there was even 100m$ locked into the protocol so it wouldn’t be fair to exclude those and it would be a pretty quick fix to include them, we’re happy to help.

Itamar

4 Likes
251

@Itamar Yes, I provided a rationale in this post sharing the proposal list:

Please see the full post, but I’ll highlight this concluding section:

If integration teams want to advocate for their users to be included in this specific proposal, the way to make it practical for me to implement is make a pull request to the Github repo with a script that allows anyone to extract – exclusively from on-chain data – a full history (through block 10228172) of supply/withdraw/borrow/repay interactions with Compound for every EOA associated with your users, in the format shown in this csv file

It’s my goal not to press too heavily on the scales going forward, but I will reiterate my point in post #214 that projects whose users interacted with Compound in ways more complex than a simple proxy may not be able to provide a mapping like the one I described above, which is why I felt it is politically cleaner to separate contract-mediated and direct interactions into separate proposals.

8 Likes