The idea was not to remove early users who provided 100 dollar to the protocol from the socialized allocation, I think it was just meant to remove the sybil vote system attacker (who attacked the voting system) only, as an attack occured where the attackers created hundreds of adress and provided 0.1 eth to make a vote, I guess it should be easy and these will appear obviously in the list, not so many adresses voted at these time, I don’t doubt we will easily remove these from the list secondly. We will find easily later I have no doubt how to exclude them without excluding any legit user.
In fact the idea is an allocation for early users (pre comp distribution) 2018 2019 and begining of 2020 so the date taken should be JUNE 15 2020 i guess ?
For pushing the process faster, I would really agree to add a compensation (small and reasonable just to compensate for the time spent) in a following proposal for amazing people here (not myself) who would bring the technical tools permitting this to happen.
I would say the requirement should be more than 2 interaction to the protocol that contains 1 supply event and 1 withdrawal event, this would probably filter out the spam accounts
If someone supplied a large part of his stable portfolio for 2 years for example he would have only one deposit so one interaction, despite could have a large role in the protocol liquidity provision, so I think we should find something else or add something later, this should be easy.
adding to that the attackers might have withdrew their funds in a cheaper gas day so might not even remove him/them
So if I understood well rleshner described these sybil attacks as one user sending a very small amount of eth, 0.1eth (worth a few dollar at the time) to dozens or hundreds of adresses at the day of the vote and submitting his vote, maybe we could use several process to exclude these from the socialized part of the reward, and not only one process :
1)exclude dust amounts (less than 0.15 eth could be considered and would probably let in legit users)
2)exclude the adresses for which funds are locked for a too short period (less than 2 days of participation in the protocol)
3)etc
as a last solution eventually could try to sneak in the list and try to find manually obvious sybil attackers (as I said I guess not many people voted so should be clearly easy)
for example I guess these attacks happened on a specific day and would be very easy to see that if on average 1 new user is onboarded a day, (at early stage for example), the sudden spike of 600 new adresses in one day wouldn’t be natural
I think these matters are to be studied in a latter stage thought.
well plenty of possibility to avoid to exclude any legit user
Since the talks are about distribution to users, who used BEFORE governance token was introduced, and Compound protocol was pretty much the one who started all that governance token distributions, amount of accounts created specifically for purpose of airdrop is likely negligble. I believe it’s not really a problem at all.
Besides, since the goal isn’t really to give away free money to everybody, but rather reward longterm supporters, your Sybil worry is relatively easy to fix: As simple as adding a requirement to have equivalent amount of COMP in cCOMP for example to be able to claim that distribution. If user have less, he can claim proportionate amount. And of course, vesting for 3-4 years.
And majority of newcomers asking about airdrop should realise, that discussion is about users who interacted like almost a year ago or more. So people who discovered Compound recently, or lately last year after Uniswap drops are not eligeble anyway.
Actually, if seriously thinking about it…
Maybe it’s time to move over really. Maybe we should revamp that idea Completely. From airdropping to early users to something that everybody could benefit and what will benefit protocol at same time.
For example, we can introduce multiplier for early users, but let everybody participate. So anybody could buy COMP tokens, lock them in cComp for like 3 years and become eligible for claiming “airdrop” of COMP, which will be gradually distributed over same 3 years. Early supporters in that case can get either get double rewards or have half of the COMP locked capital requirements.
I have a friend who did exactly this. supplied a large amount during the early days and has just been watching it since. Would prob be best to go with option 2,3 so we dont leave out many who supplied once the same way.
Sirokko, your post goes too deep, in my opinion. The idea here is to keep it simple. The want here is to give back to early users, not generate a whole new game theory. Im not against that idea, but lets isolate it and put it on its own merit in a separate post. My 2 cents there.
@Andre1 - I agree, it shouldnt really be about how much one contributed and for how long. The idea should be that anyone who participated in good faith be included. The main idea should be inclusive not exclusive as at the heart that is the entire point of this proposal. You hit the nail on the head that the main goal would be discounting sybil attacks, and to simplify the proposal I think thats likely our best endgame. Something along the lines of:
before comp announce date.
non-malicious.
We could go back and forth about who deserves what and how to figure all that out, but for simplicity sake and to be able to push forward here, the above is, in my opinion, a great starting point. There may be a few accounts who did nothing major to warrant the benefit, and maybe they are not even aware of the proposal, but the other side of the double-edged knife would be attempting to be too exclusive and leave out those who were truly supportive in their own way and their own right.
Thus in my opinion the goal should be to include all users who acted to benefit comp in their own way with their own weight. Followed closely by excluding any and all MALICIOUS actors.
Once we go beyond the above it truly complicates things. There may be a handful of accounts who dont necessarily deserve it, there may be similar accounts who do; but in attempting to perfect, we may just bicker this proposal into attrition.
Currently I am spending many hours learning geth api calls to find a way to do some basic reports and i hope to get back within several days at least a high-level list of accounts. I’ll document what I did so it can be duplicated, discussed, and tuned, but I will start by limiting to all accounts before June 15, 2020 and then we can move on to exclude malicious actors and so forth.
I think the retro airdrop should be as inclusive as possible to bring people to the Compound governance.
Do it like UNI, don’t focus on whales or filter out small amounts. Have a substancial base airdrop amount and maybe reward more active users after that (like Uniswap rewarded LPs more than swap-only users)
But, unlike the UNI airdrop, include smart wallets (Argent, Dharma) from the start.
An airdrop is the one time where quadratic participation rewards can be emulated most accurately.
Disclaimer : I didn’t have a lot when I started in DeFi, and I deposited only small amounts using a smart wallet.
For Opspec, I used a fresh account for every withdrawl/deposit. Was not much gas back then and only my account and anothers.
I think 1 interaction per protocol, minimum 200$+. Would be better, keeps the spam accounts down and forces a actual contribution that the threshold can be as such that it filters out most spam amounts.
I think the two best airdrops to look at for guidance on how to go about this are UNI and TORN. Both of them distributed their tokens based on a snapshot ending several days before the announcement of the token, not the release, which is an important detail to think about here. COMP was announced on February 26, 2020, and to avoid giving to users who joined after the announcement merely in hopes of getting COMP, the snapshot should go up to around February 16, 2020.
Further, UNI had two streams of tokens given out to users (three if you count SOCKS tokens, but there’s no equivalent to that here): one amounting to 400 UNI per address that interacted with the protocol in any way, and another averaging 1,000 UNI per address that supplied liquidity to the protocol, allocated based on share of total liquidity on a per-second basis from the beginning of the protocol to the end of the snapshot.
A system like UNI’s would need very little modification to fit our case. We could have one stream give a smaller amount to every address that lent or borrowed from the protocol, and a second stream that gives a larger amount based on share of total deposited/borrowed over time.
There are of course details to work out, which will be easier to do once we have a ballpark estimate of how many addresses will be eligible. This is actually rather simple, since it’s a list of address that borrowed and lent before the cutoff date, and since you must first lend in order to borrow, it’s really just a list of addresses that borrowed. Note that addresses which only redeemed cTokens are not counted, similarly to Uniswap not counting address that only received from a swap but did not call the contracts. I’m going to suggest that the snapshot end at block 9500000.
So, the first item of business is to get that list! After that:
Determine which addresses (if any) are to be blacklisted as malicious and subtract them from the list.
Determine what sizes the streams should be.
Intuitively, I suggest 3.33% and 6.67% of total supply, respectively. This would come to 1,000,000 tokens total, or one-fifth of the total promised to the community.
Determine what the per-address amount allocated by the first stream will be. Something to consider: TORN had a multiplier based on age of deposit to allocate its tokens. Since our goal is to reward early participants, I think this is advisable here. The complete formula they used and a graph visualizing it can be seen at the announcement linked above.
Determine what the per-address amount allocated by the second stream will be.
This will require some intensive on-chain data querying. As far as I can see, we need to know the total amount supplied and borrowed on a per-second basis from the beginning of Compound to the end of the snapshot, the share of the total each user had in each second, the result of dividing the total Comp given by this stream by the number of seconds, and then divide that much between the addresses based on their share for each second.
Once all this is done, formalize it all into a format that comp holders can vote on.
I think this is a tool for identifying COMP farmers and assessing the risk of individual accounts. The tool does not provide data relevant to the airdrop.
This tool was used when the DAI liquidation took place. If I am wrong feel free to correct me.
I would just like to add that if airdrop occurs, comp farmers should not be excluded (note: I do not belong to that group) because these same Comp farmers inflated TVL and thus the current price of the COMP token. Far from it that this action is not good for protocol and such actions should be reduced.
And the duration of the Coinbase earn promotion should be avoided because it is spam in some way
that is also my opinion unfortunately, but I hope I will be wrong.
Airdrop would be good for everyone because it would equalize(or reduced the gap) governance power of early investors and users - the degree of decentralization would increase.
You mentionned very good points here
1)the announcement date could probably be the one we take for the reason you mentionned
2)we certainly wanna submit to the vote a socialized, but ALSO a capital weighted reward, however heavily we would weight toward the socialized reward, a la UNISWAP
Finally about your last point " Determine what the per-address amount allocated by the second stream will be." we can use instead the total amount of fees paid+received, it would match the current comp distribution, just an idea here, might be easier than the one you mention
the question is now, how do we get someone to commit to all this work ? clearly it seems neither the founder neither some users rushed to do it, so we might have to find some alternativ, compensation, how do you think ? Maybe we should attach to the proposal a call of DAI reserve to compensate the people making that possible technically to happen
I understand the concern of comp token being announced, but at that time airdrops were not something done, so I dont think that is much of a concern. Thats just, like, my opinion, man.
So i did some digging on dune analytics, its the only solution i’ve seen so far, on how to get some queries on addresses. There is not a way to export those, at least not without paying for a 400.00 pro account. However, they are free to peruse on the site. Here is a link to the dashboard, I have not done queries for all types, and included some baseline minimums just to keep out spam accounts - those are very small, like under 1 dollar or that was my intent anyway.
here is a link to those:
So far its only wbtc, eth, bat and usdt, but that is mostly because I didnt feel like spending more time on something that may not be even be useful. If you did use comp for one of those before June 15, 2020 and want to search for your address to see if it shows up, please do and let me know.
if anyone has an idea of other solutions for running these reports, let me know. I do not have 100 comp to open a CAP but I figure if we get all the data required, we might be that much closer to convincing someone who does to put one in!
If they want to work on something, let them work on something that users want but are not able to do (technical and capital limitation). However, this thread is the most numerous and has the most comments, and airdrop is needed to decentralize the protocol a bit.
