To begin with, we want to make clear that our intention wasn’t to accuse the team of wrongdoing, but instead, to ensure understanding of the situation in a public setting, as we are committed to providing value to the whole ecosystem rather than operating behind closed doors. Having said that, we understand that this approach might be counterproductive and are committed to formulating standards for similar situations that may arise in the future to ensure that Arbitrum’s builder-focused ethos isn’t adversely affected and no unwanted negative externalities can arise. With this in mind, below, we’ve further elaborated on why the inquiry was raised and what is still unclear from our point of view.
Recap
The Ramses team has informed us that they were experimenting with a bot to capture MEV opportunities that would later be returned to token lockers and LPs. From the perspective of Ramses contributors, the extraction was a relatively small amount of money, and the negative externalities associated with their testing were inconsequential to LPs and token lockers. From our perspective, the ARB incentives were enough to offset the negative externalities of the extraction so that LPs/token lockers would not notice, and since there were no publicly announced plans to return this extracted value back to token holders, it looked like malicious behavior. Some of the funds extracted have seemingly been bridged through Across and deposited into Aave on Scroll, and although a small amount of money, adds ambiguity to whether or not the plan was always to return this extracted value to token lockers / LPs. In addition, the bot being operated is, to our knowledge, the only bot with the privilege to interact with these pools at 0 fees. To us, this nullifies the argument that the value would have been extracted regardless of the Ramses team operating the bot, as any other actor would be forced to pay the pool’s fees thus compensating token lockers and LPs through normal means.
Details
The main address in question is a Bot: 0x6e639af00851684e85913197a7d36bf3aa6b5a9a which we will refer to as “Wallet 1”.
Wallet 1 has privileged access to the Ramses DEX and is able to swap with effectively 0 fees. Though there are thousands of transactions to use as an example, a simple example can be found at hash 0x32308b2485fd22458492cd6b92c704fc1f1ae2b87476973adec17be246a93c80. When checking the logs of this transaction it is clear that the wallet is able to transact with Ramses pools in the privileged manner described. Wallet 1’s swap event flow within a txn lowers fees near zero for the Ramses pool, executes a swap, and then raises the fee back to the expected rate.
This practice allows Wallet 1 to capture “MEV” at 0 cost and effectively lift out liquidity from LPs without paying them to do so. Without the privileged access, an arbitrage capturing swap still creates fees to the LPs/Locked token holders—traditional arbitrage does not require fees set to zero. Even at small price changes that otherwise would not have allowed for MEV in a system with fees can be extracted with the zero-fee swaps. Additionally, transacting at near 0 fees allows the team to wash volume, which could impact the effectiveness of resulting data from the STIP Bridge program.
Ramses Response:
Ramses’ response starts with excerpts from a white paper describing a dynamic fee mechanism, which provide great context on their research into MEV. In addition to the whitepaper provided by Ramses, their response can be summarized into three key points:
- LPs on Ramses do not capture most fees, token lockers do.
- The team’s plan is to eventually return the extracted value to token lockers and LPs.
- The same arbitrage would be captured anyway by other bots.
Responding to the response
We appreciate Ramses’ quick response to the inquiry. That said, there are still a few aspects that are unclear to us. The mechanism deployed by Ramses is extracting value from LPs, while in many pools, the LPs’ main source of profits comes from ARB incentives. In other words, some of the value extraction from Ramses’ users could be seen as being supported by active ARB incentives.
Our main question relates to the plan to distribute this extracted value back to token lockers/LPs. If the redistribution was occurring in practice, we wouldn’t have brought up this inquiry.
It seems a portion of the extracted value so far has already been bridged off of Arbitrum via Across in hash: 0xcf50e1a05fa396737fba2ac7a4b6e3a3d3af03d954a1e51e2df1f5e12fedfcc9. The extracted value then appears to be deposited into Aave on Scroll 0x747c69b14ec96e2b1d0118ff5c67902078185001. Additionally, we haven’t been able to find any supporting documentation, announcements, or other publicly available resources covering this structure. It’s the lack of notice surrounding the program and the fact that funds haven’t been redistributed to users that we’d appreciate clarification on.
)
