Polls to determine specific Black Thurs vault compensation

@MakerMan posted his liquidation report.

We need to write the on-chain polls to translate these data into a specific compensation level. In my opinion, there are too many options to efficiently decide using the current polling system. We really need rank-choice voting. Fortunately, @Derek Flossman predicted that rank-choice voting will be available by mid next week.

Therefore, I propose two polls:

  1. Pick a cell from Table 8. Let’s pick row from the 1.13 column. Every row will be a poll option (plus an option for zero compensation)
  2. Which DAI daily price do we use for ETH/BAT? Options: today’s price, Black Thursday’s price, or an average of today’s and Black Thursday’s price

Does this cover all the options that people want to vote on? Should the polls be conducted in a particular order or simultaneously?


Personally I’d prefer to use the “All / 1.14” column to make sure we don’t exclude any vaults that received less than the agreed target collateral return rate.

Also I think that instead of “today’s price” it could be “price at time compensation is made available”.


After further pondering, we should probably just pick an actual specific date when we write the poll to eliminate ambiguity.

I propose we run poll #2 first.

I see 1,5,8,9,10,11,12,13,15,20,25 all as % options on that table.

I would disregard 25% and make it 24.67%.

Additional important collateral percentages for taking into account might also include:

24.670% - maximum theoretical compensation
‭21.185‬% - the mid point between maximum compensation and PreBT average
17.700% - the average amount of collateral vault holders were left with before 0 bid auctions allowed

Re: Option 2 poll @monet-supply pointed out “price of time compensation is made available”
Could there perhaps be an option to set the DAI price of compensation to the ETH/BAT exchange rate at a fixed hour like 00:00 UTC on the day compensation is scheduled?

If that were to be consensus that would allow for compensation to be made in DAI that would equate quite closely to the same amount of ETH and BAT that Vault holders lost.


It’s complicated.

My heart says “100% compensation”.

My brain says “0% compensation”. If Maker decides to pay compensation I think we could possibly assume some legal responsibility for the outcomes of the Maker system. While we strive for making a robust system there should never be any legal guarantee for this. Note that this should open up a business opportunity for Vault liquedation insurance if anyone is interested.

Ugh, we can’t run poll #1 to pick a row from Table 8 because Table 8 doesn’t take liquidation price into account. See Maker MCD Ethereum System Liquidation report and Black Thursday Compensation Analysis

Nevermind. After further consideration, I think we’re good to go. I asked @MakerMan to add all the missing rows to Table 8 so we have 1% increments.

@MakerMan, to be fair, vaults like @Vault2288 that self-liquidated should be included as eligible for compensation along with other affected vaults. Can you update Table 8 with these vaults? Can you identify vaults like @Vault2288 or do we need to solicit owners to step forward?

Is there anything to discuss off-chain? Obviously vault holders would prefer the dollar value in present time since the price of ETH has roughly doubled. I believe this poll can be run next,

I think the argument that DAI is much simpler did not meet much resistance. If we refund in ETH/BAT then the mechanics become much more complex while the difference in compensation value is negligible. Do you advocate for compensation using ETH/BAT for some reason??

Yes, but how can we run a single poll that answers both the reference price date question and the percent collateral returned question at the same time?

1 Like

I think I’ve come around to your proposed sequence in light of not being able to come up with an alternative.


I can start adding the vault holders that step forward as a class. Vault2288 is pretty unique because the audit trail is pretty clean. What I want to see is ETH before and ETH after so we can see their collateral return percentage since we are using this as the primary discriminator for compensation.

Vault 2288 returned 585.18 on 5005ETH or collateral return 11.6919% I have looked at his audit trail but not verified it entirely. It took him a while to get the data out of defisaver so he knew how much ETH came back from his 5005ETH in his vault before Black Thursday.

I really want these people to step up themselves and provide their own verifiable audit trail.

BTW: Here is the 1.14 column you wanted broken down by %. I see a slight difference (no more than .02ETH) from report. I was re-establishing formulas to make sure certain lines were intact and saw this a very slight change in numbers.

1% 1125.21
2% 2250.41
3% 3376.08
4% 4530.08
5% 5694.70
6% 6868.85
7% 8083.00
8% 9312.78
9% 10551.40
10% 11796.35
11% 13048.64
12% 14308.02
13% 15575.06
14% 16858.14
15% 18171.76
16% 19495.66
17% 20860.15
18% 22231.58
19% 23603.22
20% 24978.50
21% 26367.80
22% 27765.89
23% 29189.14
24% 30617.51
25% 32047.09

1 Like

Are you okay with using whole percentages from MakerMan instead of these theoretically motivated levels (quoted)?

@Joshua_Pritikin Can understand the motivation to do so for simplicity. Perhaps it might still be worth acknowledging 24.67% and 17.7% as potentially relevant reference points in the details of the poll.

And for reference to @MakerMan comment here is my whole account history of transactions provided by DefiSaver:


1 Like

@Derek Any update on rank-choice voting?

Hi Joshua, development is complete and I am running through a few final tests to make sure everything is as expected. Some screenshots in the meantime:

User ability to submit multiple candidates/values. Subsequent available values can be added with the “add another choice” link. The vote once submitted will then show as current vote.

Progress will be shown in the format of a bar chart. The bar chart will surface green for 1st choice votes and grey for secondary choices that push that value up in ranking (not shown below) to determine the leader.
Screenshot 2020-05-07 at 17.34.08


I can add whatever % data points you guys want just give me a working list and I will punch out those numbers as well.