Summary
The Curve alETH/ETH exploit, after the return of funds by the exploiter, resulted in a hole of 3800 alETH borne by LPers of the exploited pool. The hole is due to the pool only being partially drained, allowing the remaining assets to be purchased/stolen at an extreme discount and disseminated through additional liquidity pools.
In some scenarios, purchasers of alETH used the Connext bridge to bridge to Optimism. The Optimism multisig disabled the ability of Connext contracts to mint alETH on Optimism to prevent further proliferation of the stolen alETH. This action led to all pending bridge transactions to fail, causing all in-transit alETH (420) to be lost. A majority of the in-transit alETH was purchased well below the backing price of alETH and was only available at that price due directly to the exploit.
In many scenarios, depegs, flash crashes, and market panic create impermanent loss for LPers who may then decide if they wish to remain in the pool or exit and take the loss. This is a known risk LPers accept when they decide to LP, and is par for the course in normal circumstances in the market. This was NOT the case however for alETH LPs during the exploit. In this case LPers in the pool did not consent to selling their assets under these conditions - **the assets were stolen and subsequently drained from the LPers as a result of an exploit that occurred entirely outside of the implicit conditions of LPing **. This loss to LPers is permanent, and the situation is zero-sum in that arbitrage profit from the sale of this alETH is directly tied to the loss to LPers.
Although arbitrage is a normal expected function of markets, many of the assets transacted in this case were stolen and arbitrageurs benefitted directly and indirectly from the loss to LPers, and therefore multiple stakeholders (LPers and stuck bridgers) may feel they have a claim on the same pool of assets. It should be noted that this proposal is not necessarily asserting that arbitrageurs intentionally did anything wrong, but a wrong was committed as LPers lost everything.
This proposal seeks to determine the best way to resolve the profits that bridgers would have realized from arbitrage had they been able to successfully bridge, versus the loss to LPers, which are directly correlated. This proposal represents only one of the many paths Alchemix is pursuing to maximize the reimbursement to alETH/ETH LPers, which includes the Alchemix Elixir AMO. The Alchemix DAO voters should weigh the full context of the events and the concerns of all stakeholders when deciding how these assets should be distributed.
Proposal
This proposal puts forth the following options to reimburse users that lost access to alETH in the paused Connext bridge on Sunday, July 30th. The total is approximately 420 alETH. The options differentiate between users that purchased alETH after the exploit and users that already owned alETH before the exploit.
Note, for all options below that depend on the price of alETH. The price of alETH will be determined as the maximum 0.733 (the minimum backing calculated in AIP 97 ), and the market price of alETH in the most liquid velodrome pool at the time of the vote ending.
For users that purchased alETH with stablecoins, they will be reimbursed based on the price of ETH when the proposal ends.
Option 1: [Full Reimbursement] Fully reimburse all alETH stuck in the Connext bridge regardless of if alETH was obtained from arbitrage in the aftermath of the exploit or through other means.
Option 2: [Arbers get ETH Principal + Minimum 10% Bounty on Arbitrage] For each bridger, add together the alETH owned before the exploit, the ETH spent to obtain alETH after the exploit, and 10% of the profit earned (alETH purchased minus ETH spent, times 10%). This is the amount of alETH the bridger is entitled to. Compare this to the amount of alETH the bridger kept on mainnet (total purchased alETH minus bridged alETH). If the amount kept exceeds the amount they are entitled to, there will be no refund. If the amount kept is less than what they are entitled to, the DAO will reimburse the difference in the form of alETH. The net effect is every arbitrageur will recieve their initial prinicpal back plus 10% of any profit they would have made (denominated in alETH).
Option 3: [Reimburse Principle + Arbitrage above 0.7 alETH per ETH] The rough loss in backing of alETH was about 30% assuming the exploiter did not return funds, leading to significant open market arbitrage around 0.7 alETH/ETH that is difficult to differentiate from the sale of the stolen alETH. For each bridger, determine the profit earned assuming alETH was purchased at 0.7 alETH/ETH. Reimburse this full amount as alETH, up to the amount bridged. This will result in a larger total refund to bridgers than Option 2.
***Option 4: [Reimburse spent ETH + 10%]: This option gives everyone what they spent to buy discounted alETH + 10%, denominated in alETH.
Option 5: [Delay]: Await further information such as additional fund recovery by other means before determining the best outcome for bridgers.
See this spreadsheet for an example of how each option plays out with different scenarios: https://docs.google.com/spreadsheets/d/1mx_Np8XlqSusZa9Um4eeRcQuMoPEetua2nBnptXg9ho/edit?usp=sharing
Details of Exploit
More details of the exploit will be shared in the post-mortem, but for now we will focus on a simplification that applies directly to the discussion on arbitrageurs, illustrating how their activities (intentional or unintentional) drained LPs of their holdings in the aftermath of the exploit.
- The exploiter removed all available ETH in the liquidity pool, as well as 4800 alETH.
- The remaining balance of the pool was approximately 3800 alETH and 0 ETH.
- An MEV bot purchased this 3800 alETH for a minuscule amount of ETH, leaving the pool with a low balance of alETH and ETH which was later drained.
- The MEV bot/arbitrageur sold the 3800 alETH into the sfrxETH/alETH curve pool for 500 ETH return.
- Other MEV bots/arbitrageurs then purchased the stolen depegged sfrxETH/alETH
- Various arbitrageurs and MEV bots continued to buy/sell the stolen alETH for profit
- In tandem, the crypto community became aware of the exploit and some alETH holders / LPs in other pools sold alETH or exited their LP positions.
The net result here is that the LPers have lost 3800 alETH, on top of what the exploiter took. The exploiter returned all the funds they stole, meaning the remaining hole is the 3800 alETH stolen in the aftermath.
If the original MEV bot had sold the 3800 alETH for 3800 ETH, then that would be the entirety of the missing funds. However, they sold at a steep discount (roughly 10 alETH per ETH). This meant that the net aggregate of the next purchase was many addresses buying 3800 alETH for 500 ETH. They then sold for a profit to other addresses, who did the same, until eventually the last buyer is found who is currently holding alETH. Therefore, the profit of every arbitrageur buying or selling the stolen alETH directly represents the loss to the LPers.
Proposed Solutions - Context
This proposal represents only one of the many paths Alchemix is pursuing to maximize the reimbursement to alETH/ETH LPers, which includes the Alchemix AMO.
Option 1 simply provides a full reimbursement to all stuck bridgers. The amount of reimbursement would effectively be a loss to LPers.
Option 2 seeks to calculate the amount of profit that each arbitrageur made, and return alETH as necessary so that each arbitrageur made at least a 10% profit on their actions based on the ETH they spent to purchase alETH. The intent here is to create an effective bounty to bridgers, regardless of intentions. This essentially creates a bounty for them securing funds as whitehats rather than them stealing from LPers. If a bridger made more than 10% profit external to bridging, then no refund is given.
Option 3 notes that while the DAO has direct control of bridger assets, many other arbitrageurs purchased both stolen and legitimate alETH on the market in order to speculate on recovery. It is very difficult to differentiate the behavior of consenting buyers/sellers dealing with legitimate alETH from the buying/selling of stolen alETH once the price began to stabilize. The approximate market driven price of this speculation, largely based on the calculated backing assuming no funds were returned, was 0.7 alETH/ETH. Therefore, to not hold bridgers to a higher standard than the rest of the market, the DAO would refund any profit earned assuming the purchase was made at the value of 0.7 alETH/ETH.
Option 4 applies the bounty to what was spent rather than what was earned.
Option 5 delays the vote if people feel the right solution has not yet been found.
Note that the rest of the market to a large degree cannot be held to any specific standard as the DAO has no direct control over their assets. Bridgers are a unique situation as they ceded control of their assets to the bridge contracts during/after the exploit.
Examples of Options
See this spreadsheet for an example of how each option plays out with different scenarios: https://docs.google.com/spreadsheets/d/1mx_Np8XlqSusZa9Um4eeRcQuMoPEetua2nBnptXg9ho/edit?usp=sharing
Individual Analysis
Here is the current analysis, with the source data in tab 2 and 3, with debank txns for most users as a secondary way of validating inputs: https://docs.google.com/spreadsheets/d/1q8JZJGkqpu5BsPv7Ud0o-SiYJ4OQj6MjtN-5H8IGwL0/edit?usp=sharing
Execution
The DAO will mint 420.53 alETH on Optimism. Addresses will be airdropped alETH per the winning option. The remaining alETH will be regained on Optimism as DAO owned, available as a resource when considering options to re-back alETH and refund LPers.
Anyone will have 7 days after the conclusion of the proposal to request re-evaluation of their txns if they believe an error has been made.
Voting Options
Voting will be ranked choice.
