DAI Liquidation Event

Agreed, I have asked for some more details on how the oracle in the Compound Chain is going to operate (I know the node operators will be posting data, but I assume the reporter is still CB pro) but have not gotten a clear picture.

It would be a real shame if the lack of attention to fixing the oracle is simply because of the partnership you mention. I can’t imagine the community at large thinks that is a good idea. The silence is becoming rather frustrating, especially since I’m seeing DAI still prone to volatility on the CB pro market. Here is it today. Clearly this needs to be addressed ASAP.

Screen Shot 2021-01-10 at 4.38.28 PM

3 Likes

Compound looks like an abandoned project. Nobody seems to care. Luckily there are other options out there.

3 Likes

Another proof, that Coinbase is a really bad oracle source. Especially if it’s the only one.

3 Likes

Yeah, it’s painfully clear that Coinbase is an unreliable data source. The crashes, the price spikes etc. are becoming more and more frequent.

We need to swallow our pride and get integrated with Chainlink ASAP before the inevitable occurs.

4 Likes

Hello, I hope everyone is having a good 2021 so far. Let’s keep the ball rolling on fixing the issues that were highlighted in the thanksgiving false liquidation event, while the attack vector itself is being fixed.

Here I’m laying out a quick mock proposal for compensation for the victims of the exploit, which I wish to turn into a full proposal.

Specifically I’d like to get feedback from those that were against the previous proposal as this one seeks to address the issues that prevented its approval.

The two criticisms of the previous proposal were:

  1. The distribution of compensation was in COMP
  2. Too much of it was going to whales

Since it seems compensating in the assets liquidated is infeasible or unpopular, the first issue may be solved simply by drawing from DAI reserves which, based on conversation, seems to be an agreeable solution. It also may have positive secondary effects towards mitigating future exploits of the faulty oracle system assuming this proposal is merged before the attack vector is addressed.

The second issue may be solved using my previous computation of raw damages, found here:
script: https://pastebin.com/DcsXXxcJ 1
output: https://pastebin.com/y9s6997E 3

This is a more accurate computation of damages from the false liquidations than the previous proposal. A side effect of this is that same asset liquidations have drastically lowered damages as the manipulated DAI price was not a factor in these cases. As such the large whale receives a lower, but more accurate, proportion of the total compensation, whereas smaller users receive larger.

I believe this should be sufficient to address concern number 2. Although I appreciate @wario’s hard work so far, I do not think it makes sense or is necessary to modulate compensation amounts arbitrarily based on user behavior.

This proposal would raise the USD compensation amount from ~$6.8M to ~$7.5M if 100% of damages were compensated. If those that were against the previous proposal have issues with this amount, a simple percentage of the total amount should be suggested and agreed upon! (e.g. 90% would bring it down to approximately the previous compensation amount of $6.8M)

I plan to start working with the original proposal author, @kybx86, to draft a new proposal using this method. The next steps would be to use on-chain data to compute the actual numbers for damages then encode it into a proposal.

Let’s come together to make things right with Compounds users and improve the protocol!

5 Likes
  1. Too much of it was going to whales

I don’t think anyone in the previous compensation proposal stated that their reasoning for voting against it was because it was “going to whales”. To summarize the reasons stated, it was because most of the compensation would have gone to users egaged in recursive farming, and dumping of COMP rewards. As they were:
a) Gaming the COMP incentives
b) Not aligned with the long-term interests of the protocol, given the dumping

An example to be precise, the largest liquidated account (0x909b), one of such industrial farmers has a raw total damage of 3.8M, according to your computations. So that’s close to half of the total budgeted compensation.

Previous thread here: Compensation Proposal: Distribute COMP to Affected Users in the DAI Liquidations

3 Likes

Okay well I think compensation is warranted even if you just ignore the addresses doing “recursive farming”. Folks who used the platform as intended and got liquidated due to oracle issues definitely deserve reimbursement IMO.

Maybe just exclude that big address and that’s sufficient? I don’t think anyone wants to manually check the activity of all the addresses in that list to see what type of activity they were doing

3 Likes

It doesn’t sound reasonable to me to exclude compensation for accounts just because they “were big”. I proposed a criteria to determine people doing recursive farming and published the results here Compensation Proposal: Distribute COMP to Affected Users in the DAI Liquidations - #72 by MasterofNonce

I think this can be improved, but the basic idea is indeed to differentiate between accounts exploiting the incentives program and accounts using the protocol as intended.

Using such a criteria also results in a much lower compensation overall than what is proposed here and previously, with a total of 2.2M DAI.

1 Like

I don’t totally understand the formula you used, but it’s lowering my reimbursement by like 75%. And I wasn’t farming at all, I just didn’t want to swap my stables for DAI so I was borrowing it instead.

I should NOT have been liquidated under any realistic scenario, and should be entitled to full reimbursement.

I’m all for trying to reduce compensation for people not using the platform as intended, but I was and your formula indicates that I was not

2 Likes

In the proposal I shared, this is not the case. It is simply a side effect of computing real damages that the DAI-DAI farming liquidations will suffer less damages and thus require less compensation (only the 8% penalty for the false liquidation). There is no artificial limiting of compensation due to the size of a user. Since the large users were behaving this way, the compensation ends up skewed less towards large accounts naturally.

Sorry for not making that clear.

1 Like

No, that was clear. I was responding to @tob’s reply there.
However, yo do state that a criticism for the previous proposal was:

  1. Too much of it was going to whales

Which I don’t think is correct, as I commented previously. The criticisms were related to what these large accounts were doing.

1 Like

The formula as suggested was looking for accounts lending and borrowing stablecoins, which is common practice with these recursive farming operations. This formula, however, can be revised. I think the point to focus on is if there is agreement on determining compensation based on what the accounts were doing. That is to say, to not compensate recursive farmers for the reasons stated when voting against the previous proposal. If there is, then a second step is agreeing on the best way to determine that.

1 Like

Thanks for kicking this off!

IMO this looks like a good start. As far as I can tell the “real damage” is accurate, and I think compensating with Dai reserve seems fair, for both the effected users and the people in the community. The only thing is from what I can tell, there is only ~2M in Dai reserve right now, so the compensation may need to be distributed over time, or be discounted, or both.

I think it would be useful to get consensus on whether the compensation should be discounted and if so, what the % should be.

I also agree that it doesn’t make sense or is necessary to modulate compensation amounts arbitrarily based on user behavior. If people have issues with recursive yield farming, the right way to address this is to change the incentive so that yield farmers naturally move away; instead of randomly punish a subset of those people by not compensating in this or other similar events.

3 Likes

Great idea, because DAI reserve is on the low level now for fair compensation.

Every attempt to minimize compensation lead to new protocol weaknesses. If such behavior is not aligned with the long-term interest of protocol, why it was allowed?
Again, I am NOT “stablecoins” farmer but I understand users which are. Stablecoins farming was allowed for the short-term interest of protocol? For pumping TVL and market cap?
I think the short-term and long-term interest of protocol wasn’t aligned. and too much focus has been placed on stablecoins farming without giving users a tool with which to neutralize or mitigate such aggressive liquidation.
I contacted Defi Saver (asset management dapp), which have an automated way to repay debt in case of volatility. They argue that even their solution would not stop the Coinbase oracle error.
Kain from Synthetix had a great statement that wasn’t aligned interest of protocol users and COMP holders. But, in DeFi space we shouldn’t have separate groups in that way. In Compound case there is huge bug/error in decentralized economic incentives.

3 Likes

What is happening with this? I feel like it’s being ignored / forgotten

2 Likes

I think it isnt. Problem needs to be solved from root (COMP distribution, dumping COMP from exploitatiors,oracle fixing).
But if affected users will not be compensated due to obvious manpulation (some people said that there is not evidence for that - off course because Coinbase dont want to provide data) I will move all my funds to Aave.
@rleshner always speaks that users need to be in governance of protocol but in practice we are hostage of early investors.
One reason for drop out proposal 32 is that users will be compensated in COMP token. Am I crazy?

  1. statement - users need more governance power
  2. statement (from VC funds-not @rleshner ) - we dont like idea to compensate affected USERS in COMP token.
    Its obvious that affected addresses = USERS, because if you are not user of protocol you cannot be affected by “DAI manipulation”.
    Lots of contradictions and inconsistencies in the statements of the people who should keep the true users of the protocol. They really don’t do that.
    I expect the Gauntlet team and @rleshner not to abstain from voting next time because we want to know in which direction this protocol will develop.
3 Likes

So this is dead? Nothing is going to be done?

This really disappoints me, I don’t think I will use Compound again. Nothing has been done to address the exploit, and there’s just no accountability from Compound after providing an unsafe platform and causing folks to get liquidated even when using the “safe borrow percentage” recommended by Compound. Lending a stable and borrowing the “safe amount” of another stable should never lead to liquidation in the manner that occurred here.

I find it totally ridiculous that much smaller, less successful platforms like Harvest and Pickle fully or partially compensate users who experienced an exploit, while Compound, a leader in this space, chose to do nothing and ignore the issue. It would not have been that hard to issue vCOMP that vests over time. Or providing some other kind of restitution (I liked 0xb1’s proposal here https://twitter.com/0x_b1/status/1356328468284387331)

Considering how successful Compound has been, purely based off of the historical activity & support of the user base, this really just feels like a HUGE let down. I’m extremely disappointed.

@rleshner

4 Likes

I agree, users are not liquidated through their own fault, the error is by the protocol. Honestly, I am patiently waiting for the repair of the price feed oracle and next prosposal. If the losses of users (which have become very large with the increase in Ethereum prices) are not compensated, I will certainly transfer my funds elsewhere.
It is a shame that Coinbase gives a statement that everything worked properly and most WC funds (which have a protocol) support that.

3 Likes

is this real? When will the intention be explained in the forum in a normal way?

1 Like

Not sure, I was expecting it to get proposed by now

2 Likes