A Light in the Ethereum Dark Forest: Reading the Game in the MEV Race
Originally written by Xinxin Cao, Hashkey Capital Research
Mev research group Flashbots reported in March that its third-party Ethereum client version, Mev-Geth, has attracted adoption from 12 mining pools with 58.8% of the network's hashing power. The rise in popularity of this version has been accompanied by a sharp drop in Ethereum Gas fees. Although the causal link has not yet been proven, it is enough to attract widespread attention and debate about the legitimacy of MEV in the community.
This paper hopes to clarify the definition of MEV and its impact on Ethereum ecology by analyzing the source of MEV, and discuss a possible state in the future based on the game relationship between stakeholders in the market order of MEV benefit distribution that Flashbots tries to establish.
Reference reading:
"Flashbots Transparency Report -- March 2021"
Over the years of mature operation, Ethereum has formed a stable pattern from producers (miners) to auction houses (mine pools) to bidders (users), regulating supply and demand of block space through the Priority Gas Fee Auction (PGA) mechanism and block incentives.
If transactions can be divided into price discovery, transaction matching and transaction execution, then at this stage, the core value of the block space is the decentralized transfer service (detrust of transaction execution) guaranteed by a series of core "components" of the blockchain, such as hash pointer, consensus mechanism and peer-to-peer network. It also further replaces the credit intermediation (detrust of dealmaking) that transactions rely on before they are completed by smart contracts, thus eliminating the frictional costs associated with "trust" in the whole transaction process and providing a critical infrastructure for the issuance and transaction of digital assets.
Reference reading:
"Deep analysis of the Ethereum block space market: the mining cycle will emerge a new pattern"
Drived by smart contracts, the Ethereum application layer derives the large-scale application innovation of DEFI business, which makes the block space begin to be filled with a part of complex transaction logic. The DEFI system built with LEGO bricks, such as overmortgage lending, AMM, income mining strategy, derivatives, insurance, and algorithm-based stable currency, has taken shape. The wealth effect of its composability has been discussed for some time, and the transparency and openness of block Spaces and unconfirmed trading pools have created new opportunities and challenges for participants in the price discovery process. MEV is a byproduct of this feature.
Price discovery is the decentralized market self-organizing development indispensable economic function, ordinary transaction user, arbitrage and clearing robot, economic system vulnerabilities hackers are even found that the price discovery of the participants (the following a general designation searcher), they search the gap between actual and expected, pay the cost of capital and transaction fees, to obtain the potential gains. Transparency and accessibility of information on the blockchain has provided new opportunities for efficient price discovery systems, and innovative lending models such as flash lending have facilitated price discovery.
But these searchers face new challenges. On the one hand, they are competing with the PGA for preferred trade opportunities. On the other, they face the prying eyes and predatory eyes of the front-runners. Competition between searchers highlights and drives up the time value of intra-block transactions, which is justifiably represented by higher fee costs than normal transactions. The discretencies of block generation time and the transparency of the unconfirmed trade pool allow the front-runner to have time and easy to monitor the front-runner transactions in the unconfirmed trade pool, participate in the PGA process together without paying the search cost, and close their own and the front-runner transactions in a specific time order by further raising the Gas fee. To achieve the purpose of getting something for nothing, so that the interests of the searchers are damaged by this free rider behavior. It is therefore suggested that we make a distinction between the nature of these two types of transactions when discussing the justification of MEV.
For example, after the lightning loan attackers manipulate the relative prices of AMM assets such as Curve, some arbitrageurs move the price gap quickly and obtain a large amount of arbitrage income, which is the legitimate income they help the market to respond quickly. The sandwich attack of an AMM transaction in their arbitrage process or the pre-empting of a direct copy of their arbitrage trade is a kind of free rider behavior. Similarly, a robot that tracks the ups and downs of the market and listens for price prognostics to update competing clearing opportunities performs the role of Keeper. A robot that listens for copies of trades sent by other clearing robots and acts as a clearing man first takes the labor of a pre-empted clearing trade. Some exotic trading opportunities require more effort to capture, only to be defeated by the front-runners (hackers involved in the Dodo security breach in March also had their hackles cut by the front-runners).
The opportunity brought by transaction transparency makes it impossible for searchers themselves to form technical barriers, and the MEV competitive market becomes a dark forest of leap-start robots. When highly publicised the capital market to attract various false start the robot, they themselves had to fight in the block space bidding mechanism, through the more rounds of PGA process to get the trade order priority, and has the right of block sorting mineral pools, endowed with the natural rights rent-seeking opportunities, mineral pool itself also enjoys an absolute advantage in the event of a false start. This represents an inherent flaw in the existing blockchain mechanism, and facilitates the process of distributing some of the searchers' profits to the front-runners, who then distribute them to the miners. I prefer to call the unallocated Maximum Extractable Value (MEV) that the runner gets from the searcher and the extra money the Miner gets from the PGA the Miner Extractable Value (MEV). Mev).
The existence of the MEV has led to a new kind of competition between searchers and jumpers, and provides miners with a source of revenue beyond miners' rewards and regular fees. Community has MEV would affect the decentralized characteristics of block chain, a kind of view is to extract MEV eventually link under the control of the work force of packaging block competition, have more work force of mine pool as flyers have a greater chance for success packaging MEV trading, will form a positive cycle motivation continuously attracting miners to the mine pool and expansion, This leads to the centralization of computing power, reducing the decentralization and security of blockchain.
But another view argue that because the mineral pools can't confirm whether they can compete on to the next package block, so the rational, mineral pool will search to MEV trading opportunities to sell to other mineral pools, to lock their extra income, MEV will share from the probability to participate in competition of MEV mining pool, improve its overall revenue.
Given the historical trend toward continuous specialization of labor, the skills and tools required to search for MEV transactions and computing mining vary widely, making it more likely that a division of labor will emerge between MEV searches and computing mining, and the solution proposed by Flashbots, described below, demonstrates that this division is taking hold.
A definite negative externality of MEV to the blockchain ecology is a pernicious effect on the Gas fee bidding mechanism. As long as the MEV is large enough, the first movers will be willing to pay hundreds of times more Gas than a normal transaction through cutthroat competition, resulting in a lot of failed and high fee transactions in the block space, resulting in a waste of the scarce resource of block space.
MEV has a certain antagonistic effect on ordinary transaction users and searchers. In traditional finance, regulators impose legal constraints on acts that disturb the order of market transactions, such as jump trading. In the decentralized financial ecology, the reform or optimization of the market mechanism is mainly driven by the game between stakeholders and the technological upgrading of the decentralized system.
In response to the negative impact of MEV on the blockchain ecology, Flashbots attempts to establish a market order for the distribution of MEV benefits for searchers and miners, in effect providing a way to take part of the PGA process down the chain and let mine pools bid on MEV. This does help alleviate the blockage problem caused by the MEV scramble.
But the new question arises, how to avoid the various evil possibilities of stakeholders in the system below the chain? This comes back to how to ensure the privacy of the transaction information before the link, otherwise, searchers' transaction information is still easy to be snatched up by miners. Currently, the Mev-Geth version of Flashbots uses a licensing system to regulate the behavior of its participants, while the latest proposal is to use a Sealed-Bid Auction mechanism based on SGX to achieve unlicensed access to the market and privacy of the transaction process. This is a topic worthy of further study and will not be explored in this article.


Figure 1. Flashbots design goals and the MEV-SGX architecture (from Flashbots MEV Roast 13)
The appearance of Flashbots changes the dark forest law of block space, and the search party can choose to compete for the priority right in an orderly market. The following analysis of the game of interested parties predicts that the new order is expected to eliminate the free-riding opportunity of the front-runners, and enable arbitrage and clearing robots to launch a fair PGA competition. The MEV allocated to miners evolved into a regular price-discovery premium with the privacy of the transaction protected. Whoever has the better strategy in the auction process will be able to implement the price discovery process at a relatively lower premium.

Figure 2. Two types of attacks to obtain MEV
Currently, the attack trade of the starter can be divided into two broad categories:
Trades aimed primarily at the AMM market allow the user to suffer a larger slippage loss (CS), the MEV gained by the front-runner. The user can control the CS by limiting the trading volume and setting the maximum slip point, so that 0≤MEV≤ CS. The first runner allocates a portion of the revenue MEV to the pool in the PGA.
Mainly for arbitrage, clearing transactions, jumping the gun, lightning lending attacks and other use of system loopholes to profit transactions can also be summarized into this category. The pre-emptive MEV obtained by the attacker is equal to the expected return of the pre-empted transaction, Re, making the pre-empted transaction fail or invalid, not only losing the expected return, but also paying the commission fee for nothing and in some cases incurring additional losses.
Back-running against prognosticator updates and large trades is actually a clever search activity that does no harm to the following trade, the searcher just grabs the profit opportunity before the other pursuers do.
In a pre-empted attack, the sender of the pre-empted transaction appears to be the most innocent victim. He puts his efforts into the hard-won search for a profit opportunity, and once it is published online, he is then given a free ride by other competitors. This free-rider phenomenon mismatches the labor and income of the price discovery process.
Flashbots provide new options for stakeholders.
AMM transaction users and searchers are extremely passive in the transparent and open mode without confirming the transaction, and their expected profits can be expressed as:
Re - Cgas - MEV 0 or higher
Where, RE is the expected revenue (minus the acceptable slip-point cost) during the price discovery process, CGAS is the common transaction fee, and MEV is the expected profit that may be snatched by the front-runner.
For the front-runners, their expected profit is expressed as:
MEV - MEV 0 or higher
Among them, MEV is the profit allocated to the mine pool in the PGA process.
The lower expected MEV that the front-runners are looking for is MEV. For AMM trading users, their strategy is to reduce trading volume and set a slip-point cap to constrains the MEV
In fact, when Flashbots becomes unlicensed, users can choose to directly participate in the mev-geth bidding process, bidding in the (MEV, MEV) range, and the biggest loss will be to pay the same amount of money as before, or the bidding fails and the deal is not made immediately, and the sandwich bots lose a lot of goals. The problem of sandwich attacks could also be alleviated if an auction market could be created in which transaction privacy is protected.
To be false start arbitrage trading, clearing, flyers can grab MEV more often than not, these transactions, search will be sent directly to the MEV - geth node in to protect the privacy of their trading at the same time also can have the advantage in the bidding (ruled out flyers competition, only compete with similar searchers, and bidding failure won't loss fee).
For the first movers, they will still be able to find pre-emptive trades without confirming the trade pool, and have an incentive to send transaction bundles to the MEV-Geth node. The reason is that the Mev-Geth node protects it from being further jumped by the pool, and in the off-chain bidding process, even if the transaction fails, it will not pack the up-chain and lose fees or more additional losses. As the computing power of the MEV-Geth node increases, it is also difficult for the stakeholders with the dual status of front-runner and mine pool to gain an advantage in the free ride. Imagine that the same pre-emptive transaction, which a normal pre-emptive robot sends to a mev-geth node with close to 60% of the computing power for packing, can finally compete with a single pool. As more and more searchers send arbitrage and clearing transactions to the Mev-Geth node in the form of privacy protection, the opportunities for the front-runners to hitch-ride will also decrease.
Finally, the interest distribution order established by Flashbots may form an auction market dominated by arbitrage and clearing robots and eliminate free-riding behavior. And the MEV nature of these bots' bids becomes a price discovery premium to gain transaction privacy against a jump start attack.
For the mine pool, there is theoretically a risk aversion preference. After all, the official client version of Ethereum has been running for many years, and its security has been verified. However, nearly 60% of the computing power is still switched to the third-party version, which is really worth paying attention to. Part of the reason for this may be that, driven by the EIP-1559, mine pools are stepping into the MEV market and faced with a choice of competing with MEV themselves or adopting the MEV-Geth version and following the market order established by Flashbots. For mines that do not act as searchers themselves, MEV-Geth can simply increase the income of miners and avoid the loss of miners; For a pool with a search capability, this depends on the opportunity cost given up by adopting MEV-Geth. For example, while MEV-Geth currently limits a block to one MEV transaction, it is possible for a pool to pool itself with one or more MEV transactions and make more money. However, searching for MEV requires the pool to acquire relevant skills and consume more resources, reducing its net income.
We are pleased to see that the blockchain infrastructure has set the wheels for a decentralized market, and the ecology is starting to iterate into various segments, giving us the good luck to look at interesting new markets like MEV.
MEV allows us to focus on the value of transaction data and the importance of block spatial sequencing. It also allows us to think about new roles beyond miners, pools, and users. Flashbots has many improvements, but it is a good example of how to provide order to this emerging trading market, to eliminate some of the factors that are not friendly to the overall economic development, and to establish the civilization of the cyber space. As a result, the extractable value of miners is expected to evolve into a regular price discovery premium with the privacy of the transaction protected.
Just like Flashbots trying to fork the unconfirmed trading pool module of the blockchain system, there is room for optimization in the mechanism design of the decentralized system, which is worth more thinking.
The original link
Welcome to join the official BlockBeats community:
Telegram Subscription Group: https://t.me/theblockbeats
Telegram Discussion Group: https://t.me/BlockBeats_App
Official Twitter Account: https://twitter.com/BlockBeatsAsia