Maker MCD Ethereum System Liquidation report and Black Thursday Compensation Analysis

Greetings Maker community,

At long last the Maker MCD Ethereum Liquidation and Black Thursday Compensation Analysis reports are availble below.

Lets keep discussion on the details of the report, questions, comments, etc. focused in this thread.

Stay tuned for updates on when a conference call on this report will occur.


I guess the next step is to pick a cell from Table 8 and decide on an ETH valuation model.

I feel like there are too many options to decide without rank-choice voting in on-chain polls.


Inherently the biggest problem on the vault compensation report wasn’t doing 1 scenario but setting up to do a whole bunch of different ones. I was already looking at ETH DAI valuations to present one but gave up after a short time.

I think if there is some focus to discussion it will be to talk about the class of auctions we consider defective (1), and then (2) to decide what we think is a reasonable collateral return scenario under the market conditions where liquidation happened. My arbitrary (based on my business experience) split the difference between what someone could expect or want vs. what I could expect or want is in the right ballpark (10-12% collateral return) and this is what we saw as a rough average when additional keepers came on line.

So unless someone has a really strong arguement for more (perhaps an emotional compensation, since this was stressful for everyone), or less (because really under these circumstance expecting ANY collateral return is a pretty high expectation) landing right in the middle (give or take a few percent) honestly seems reasonable. Then the issue becomes the auction criterion selection process, but honestly even if one selects all with a collateral deficiency relative to the nominal value chosen the amount doesn’t more than double from left (0-bid) to right (ALL with a collateral deficit relative to chosen nominal).

The value of ETH in DAI - that is a whole different can of worms.


If this next poll is not to contradict the previous then it should ask the question of how to best achieve the benefits/objectives voted for already.

Below are the “Arguments for compensation” as presented to voters in the previous poll for reference:

The protocol performed in a sub-optimal manner. Users (i.e. vault holders) were injured. It is in the interest of Maker holders to try to make their users happy. The purpose of the MKR token is to insure against precisely this kind of failure.
No compensation may represent a distortion of symmetry and reciprocity i.e. the impression that we’re marketing the product and receiving the reward for growth of the platform, but leaving ecosystem/development risk entirely to the user. This could negatively impact reputation and future mass adoption.
Avoid bad press and acrimonious gestures.

Compensation done “down the middle” as suggested or any other fraction may appear reasonable, but will likely not achieve these objectives and in that sense be fruitless since you would have spent a great deal of DAI and still sit with the reputational damage, bad press, injured vault holders, negative impact on mass adoption etc. You would not have achieved the benefits voted for in the first poll.
Further defending a “reasonable” percentage in front of the press, injured users and potential legal proceedings could be difficult and more so being represented by a collective. The cleanest, quickest route ending the matter and focusing on the future could prove an investment especially considering you are not dealing with a single party you can shake hands with - all you have are your actions signalling intent and impacting brand and reputation.
For all these audiences you likely would have to be seen to have “done all you could” acting in the interest of users and that’s likely what will count in articles written on this issue. Considering the interest in the matter and competition in the space this could be defining for Maker.

Having said this presenting too many choices could be misleading and I propose we rather focus on fewer options framing the poll in a way that extends directly to desired outcomes. Otherwise you could be caught in a classic “decision by committee” trap diluting the question at hand to averages and missing the outcomes altogether.

A voter may think 12% is reasonable, but want the outcomes 24% would likely bring if the question was framed differently. This is not a strange concept - giving customers more than is reasonable is basically Amazon’s ethos.
The framing of the poll should focus on outcomes and provide limited options.


No amount of compensation is going to make those things go away, frankly.

Expecting 24% back is unreasonable in my view. Successful auctions on the day returned around at best 11% of collateral to Vaults. That gives us the best idea of what Vault Holders could have expected had there been enough (but not copious amounts) of keeper liquidity.

I’d actually like to split the difference between 0% and 11%, and return 5.5%. Given that responsibility for the outcome is shared between like 5 different entities (of which MKR Token Holders are only one) this still feels more than generous.

On another point those arguments for compensation represent a subset of the communities opinions. I didn’t agree with them as stated, but I still voted in favour. I just want to be clear they should not be considered the canonical source of truth on what the responsibilities of MKR Token Holders are, or what their reasons for voting in favor would be.

Regarding what the poll looks like, we’ll do our best to structure it to capture as much signal as possible, as efficiently as possible. We’ll also try to make sure it’s objective as possible. I’m sure full compensation of 24% will be on the poll. I’m also fairly sure that 0% compensation will be on the poll, given that we now have more information regarding auction effectiveness and the cost of compensation than we did previously, and that this new information may have caused people to change their minds. Probably we’ll see a few options between those extremes too.


I would really like to encourage the community to use this opportunity to get together and try to come up with a plan, or at least an outline of a plan, that clarifies how this process will work.

There are a significant amount of stakeholders that need to coordinate, a lot of people that need to be informed about the contents of this report, and not a trivial amount of code that needs to be written to support the intention of the previous poll. There are also auditing and custody issues that need to be addressed.

The success of this initiative will be far smoother if it begins to be treated as a real project, i.e. scoping, deliverables, scheduling, etc.

Moving immediately into more polls dedicated to best guesses at next steps is going to be a messy and potentially drawn out process.

Feels like there needs to be a working group here to pick up some of the coordination tasks…

Just my opinion, feel free to disregard.


We should also consider onboarding a legal domain team. I’m concerned if MakerDAO is not legally incorporated anywhere, it runs the risk of being treated like a general partnership.


Maybe this is this first thing to be decided then? I.E maybe we need to elect a working group to hash out the details of a plan here and lay out the roadmap.

Once we had that then the general beats of a plan laid out then we could start to distribute the work amongst people that might have the ability and expertise to execute the various tasks that would be required by that plan.


Personally I think 5.5% is still rather low considering Pre-Black Thursday and before the 0 bid auctions were taking place, the average vault holder was left with 17.7% after liquidation.

If you ask me 17.7% should be up for consideration as well.


Since there is some discussion around the nominal collateral return forming I realized I offered a limited set focused around an arbitrarily and possibly bias selected 10%. The nominal collateral return CAN be any number.

If I was going to suggest brackets on a poll (should the poll be constructed to fix the nominal collateral return) perhaps 0, 0-5%, 5-10%, 10-15% 15-20% 20-25% might be the best working poll choices to rank on. 0, being no compensation, 25% being full theoretical with everything else falling in between into kind of normalish catagories. 0-(no compensation at all), 0-5 (yes, but only a little), 5-10 (yes, but more than a little), 10-15(yes, and roughly 1/2 theoretical), (yes, more than 1/2 theoretical but distincly less than full theoretical), 20-25(yes, pretty much full theoretical or very close).

One could literally weigh the above MKR vote BY the compensation % chosen to create a weighted by MKR voting compensation % along with the winning ranked choice %.

Main goal of this post was to inform everyone this nominal collateral return % can be anything you want - just interpolate between matrix values for a good estimate of total.



First of all, I’m new to this forums so I wanted to say hello.
I’d like to thank you for this report, it’s greatly appreciated. However, I have a big concern here. This report says nothing about BAT vaults that got liquidated during Black Thrusday event. If you plan to reimburse all the victims, it’s only fair if you count BAT vaults as well. My BAT vault had almost 70k BAT in it and it got liquidated to 0 during Black Thursday event.

best regards,

1 Like

I’m fairly sure that we’ll end up applying the same level of compensation to BAT Vaults as we do to ETH Vaults. I think that from the perspective of analysing this data it was just easier to omit BAT. Don’t worry, no one’s forgotten you.

True, but normal circumstances are different from the sort of market event we saw on Black Thursday. Personally I don’t think it’s realistic to expect compensation based on the average good case in ideal circumstances. Others may well see it differently though. Like in my mind we have something like this:

24% - Theoretical best case, unlikely in practice unless ETH price appreciates after liquidation but before auctions end.
17.7% - Average good case in normal circumstances.
11% - Best case on Black Thursday with limited Keeper liquidity.
0% - Worst case given 0-bids.

Others may agree with you and suggest that 5.5% is low, and that 11% is already enough of a compromise. Difficult to say.

I don’t think election is necessary. If you are willing to form part of this working group then please make yourself known.

1 Like

Remember, we are not just talking about a technical problem here, but also of a false (or bad) advertising of the «13% liquidation penalty» that were only explained in the github in a not precise manner. This number was displayed everywhere, not only on UIs, but in every Medium article of ETH dev & experts. We can be sure that a big majority of the impacted vault holders were expecting 13% penalty applied on their loan.

Remember, if most of vault holders had known that in specific conditions they were going to receive 50% of what they should have received on a «normal day» (11% vs 24%, @LongForWisdom comment), they would have taken more care of their vault. (btw @LongForWisdom, I still don’t understand, as a User, where did you find the 11% and 17.7%?..).

Remember, a class action has been made, it should take months/years to get the final statement and a lot of human efforts. Vault holders that will accept the compensation will not be able to participate.

With all that points, me and other victims would appreciate a full compensation in ETH (~24.67% of total collateral). If not, we’ll feel like receiving a random compensation without any clear explanation: why 5.5% in DAI, why 11% in ETH, …


Which MKR Token Holders are not responsible for. As I said, I believe the responsibility for the loss is shared between like 5 entities. Imo these are:

MKR Token Holders - For setting insufficiently defensive auction parameters.
Various UI’s- For displaying 13% Liquidation price without explaining the details.
Maker Foundation - For failing to ensure robust Keeper participation and production of appropriate tools.
Ethereum Network - For failing to scale well.
Vault Holders - For getting liquidated.

I’m sure these are open to argument, but this is how I see the situation. The point is, MKR Token Holders are not fully responsible for the loss, why should they be fully responsible for any compensation.

From the report posted in the initial post of this thread. (Though I’ll note I only pulled out 11%, 17.7% came from @Vault2288)

As to the point as why DAI rather than ETH, that would be because the MKR Protocol can mint DAI and give it to you without anyone having to code or decide anything terribly complicated in order to source ETH. Compensation in DAI is likely to be a faster process than compensation in ETH.


Now eth is rising again, is this a factor to consider ? Lets say you take 100 dollar as piont off liquidation and base the compensation ( DAI) on that, it means the compensation would be peanuts.

Yeah, we started to prepare a poll for that (how does ETH price in DAI affect compensation),

However, Rich keeps saying that we need to pull together a team to plan this out better before we do more polls.

1 Like

I dont see a reason the a reason for pulling together a team. It feels we are wasting time and a lot of people when to close this chapter.


Yes, if we used the DAI value of the ETH at time of liquidation (or an average over Black Thursday) it would be less value than if we compensated in ETH (given the current ETH price).

We’ve yet to decide what ETH price we would use to determine the amount of DAI compensation though. We could use the current ETH price, although I don’t think that is necessarily a good idea either.

If I’m reading the report right (@MakerMan please correct me if I’m wrong) to compensate all ETH Vaults at 24% of collateral (~32k ETH) at a $200 ETH price comes to 6.4M DAI. Personally I think it’s doubtful that MKR Token Holders are going to incur that much more bad debt in order to compensate Vault Holders. But again, I could be wrong.

Mostly because things don’t just magically happen. The reason Rich wants a team is because previously the proposed polls promised things that there was no plan in place to implement.

Say for example we wanted to compensate in ETH, someone needs to figure out how to get the MKR Protocol to acquire that amount of ETH. The way this works is the Protocol can mint DAI, but then we’d have to use DEX’s to convert. Okay. But then we need to deal with slippage. Which DEX’s do we use? Who compiles the list of Vault Holders and how do we calculate how much we owe each one? Do Vault Holders need to be able to opt out? Ditto for BAT Vaults. If some Vaults do opt out, how does that process work? How does this interact with the class action lawsuit against the Foundation.

Furthermore, the people that are sorting all this stuff out need to get paid, because frankly, it’s all work. We’ve never paid anyone from the Maker Protocol before, so how does that work? Payment in advance? On delivery? By hour using a streaming solution?

It may sound simple, but it isn’t.

1 Like

Which is why it doesn’t make sense for it to be paid in DAI, but it in ETH.

It was ETH that was lost as a result of the 0 bid auctions

1 Like

From a process perspective it makes more sense to use DAI, because the Protocol has DAI. Exchanging DAI to ETH adds complexity and will delay any compensation you guys get.

Setting the amount of DAI to equal the price of ETH at the time of compensation is functionally the same value compensated as if it had been denominated in ETH, though admittedly it means that if the compensated parties want to hold ETH, they would need to exchange it themselves as they received it.