While the crypto space is fairly welcoming to newcomers, sometimes it can get a little challenging when more technical terms like MEV are constantly being thrown around. Often cited as a primary feature of blockchains, as opposed to a bug that needs to be fixed, MEV, or Maximal Extractable Value, refers to the maximum value that can be extracted from block production beyond the standard block rewards and transaction fees by including, excluding, and changing the order of transactions in a block. In other words, MEV is a means of obtaining additional value by analyzing and immediately taking advantage of other users’ transactions before they are finalized.

As pending transactions are queued in the network’s mempool, or assets become mispriced across various liquidity pools in DeFi, network participants known as ‘searchers’ lie in wait to capitalize on these movements through MEV transactions. True to their name, these searchers will run a variety of complex algorithms on the transaction data to detect profitable MEV opportunities and have automated MEV bots do the heavy lifting by rapidly submitting those profitable transactions to the network.It is important to note that unlike wash trading, which is meant to generate false volumes in the DeFi and NFT space, MEV transactions are its antithesis, extracting additional real transaction volumes beyond conventional trading activity on decentralized networks in small yet meaningful ways. In this article, we will be jumping into the rabbit hole of the MEV industry, the various types of MEV transactions out there, how impactful these transactions can be, and how we can protect ourselves from malicious MEV attacks.

Types of MEVs

Before we get into the level of influence that MEV transactions have on any particular blockchain or network, it is important to recognize the various types of MEV techniques currently available. Some of the more common types of MEVs include arbitrage, sandwich attacks, liquidations, and just-in-time liquidity provision (JIT).

Broadly speaking, we can categorize these types of MEVs into two groups:

Most of the time, arbitrage and liquidations consist of a list of several actions, all performed simultaneously within a single transaction. These MEV techniques usually involve swapping assets between two or more liquidity pools. With DEX aggregators thrown into the mix, it becomes even more challenging to squeeze all the relevant token transfers into one transaction that fits on the network's block space.

On the other hand, JIT and sandwich attacks are performed over a series of separate transactions. However, they need to be executed within a set period of time to successfully complete the trade and capture a profit, adding a layer of difficulty in terms of timing the MEV opportunity. Although JIT and sandwich attacks similarly make use of multiple liquidity pools and protocols, the usage of numerous transactions is more integral to these types of MEVs.

Now that we've outlined the different kinds of MEV transactions, it's time to examine arbitrages, sandwich attacks, and liquidations in further detail. 

Arbitrage

Much like how arbitrage is typically performed in the traditional markets, MEV transactions of this nature take advantage of the different asset prices across various liquidity markets to purchase tokens at a lower price and sell them at a higher price on different exchanges to generate a profit. Thanks to the explosive growth of various DEXs, aggregators, and cross-chain bridges, there are now more liquidity pools than ever before, giving rise to even more opportunities for MEV bots to take advantage of any price discrepancies that might occur.

As a result, arbitrage MEVs form an essential part of DeFi, ensuring that asset prices remain consistent across various exchanges and ecosystems, further improving the efficiency of the overall markets. To explain how an arbitrage transaction might look like, let’s dive into how arbitrage searchers take advantage of different token prices across a variety of exchanges.

As a basic example, let’s assume that a searcher, Alice, will undertake two transactions involving two tokens between two liquidity pools by utilizing the difference in pricing of the tokens. Let’s say she finds the exchange rate of USDC for WETH in a liquidity pool - called LP1 - on Uniswap V3 is 1,350:1, which means 1,350 USDC for 1 WETH. Meanwhile, the exchange rate of the same pair in another liquidity pool - called LP2 - on UniSwap v3 is 1,400:1, which means 1 WETH is equivalent to 1,400 USDC. 

Alice sees an arbitrage opportunity here: the price of WETH for USDC in LP2 is lower than in LP1. Here is what Alice can do:

 Source: EigenPhi

Once the transactions have been completed, Alice will end up with 10.37 WETH, generating a gross profit of 0.37 ETH. However, Alice will also have to account for the transaction costs paid for performing the two transactions. Assuming that the total gas cost is 0.008 WETH, Alice will obtain a net profit of 0.362 ETH.

Next, let’s have a look at what a real-world transaction might look like.

Source: Etherscan

From the transaction above, we can tell that the searcher swapped USDC and WETH between 2 different liquidity pools on Uniswap V3 liquidity pools and acquired 1,104 USDC in gross profit. The cost of performing this single arbitrage transaction is 0.063 WETH (~$92). Ultimately, the searcher profited $1,012. You can visit this arbitrage’s page on EigenPhi to get a clearer Profit & Loss analysis and explore its token flow as below. 

It is important to note that arbitrage transactions are usually more complex than the examples above. Searchers tend to swap more than 2 tokens among 3 different avenues, which requires them to be informed of various price movements across a vast selection of liquidity pools.

In the world of Web3, arbitrage opportunities are not limited to just tokens but also from digital assets such as NFTs.  A searcher flipped an Otherdeed NFT for a profit of $1.6K. In another transaction, the searcher traded the ENS “collectbtc.eth” and made a profit of $11K, more than 10 times their initial cost of $1K.

Sandwich Attack

A sandwich attack, also known as front-running, involves ‘sandwiching’ the victim's transactions between two transactions initiated by the searchers/attackers, whose reordering of the transactions instantly inflicts an implicit loss on the affected user and possibly benefits the attackers. These types of attacks are nothing new, as the same methodologies are frequently applied in the world of high-frequency trading (HFT) in the traditional financial markets.

The beginning of a sandwich attack occurs in a network’s mempool, the waiting area for the transactions that have yet to be confirmed and added into a block by the block's miner or builder. In the event that a user sets a high slippage tolerance for the transaction, the searcher could take advantage of the opportunity by

But, how exactly does the attacker generate profits from this process? Here is an example:

 Source: EigenPhi

Once all the transactions are finalized, we can see that Alice generated an additional 2.5 WETH. Assuming that the gas fees and miner/builder incentives to reorder her transactions amount to 1.2 WETH, she would have made a net profit of 1.3 WETH. Meanwhile, Bob suffered a 700 USDC loss, as he could have gotten up to 700 USDC more if he had set the slippage at close to zero. As such, we can see why sandwich MEV transactions can be devastating to unsuspecting users, especially for large transactions.

Although it can be difficult to determine whether a transaction is part of a sandwich attack on Etherscan, tools like EigenPhi, Nansen, and Zeromev can help you identify them. EigenPhi’s Sandwich module provides an intuitive way to explore a sandwich’s Profit & Loss and its token flow chart. 

The key to executing a successful sandwich attack is to get the miner/builder to accept the reordered transactions. In the past, searchers used to guess how much gas fees they needed to pay to the miners/builders to guarantee that their transactions were still accepted while maintaining profitability. As such, searchers were forced to try and outdo one another, resulting in high gas prices, network congestion, and artificially scarce block space. 

These days, tools such as Flashbots Auctions have been designed to create a permissionless, transparent, and fair ecosystem for efficient MEV extraction, where searchers can submit their transaction bundles via a private auction mechanism. Yet, there are some concerns that the auction mechanism has increased centralization around builders and validators. 

Liquidations

For many crypto firms that have gone bust, such as Celsius, 3AC, and FTX, liquidations are an inevitable part of the bankruptcy proceedings, meant to protect creditors and allow them to retrieve any remaining funds that may otherwise be lost. However, the entire process could take months or even years, and most reclaimed assets will eventually end up with lawyers, leaving just a small share to the original debtees.

On the contrary, liquidations in DeFi can happen almost instantaneously. As the crypto markets continued to slide downwards in June 2022, 3AC was liquidated at $400 million without being interrupted by any attorneys. 

The lightning speed of DeFi’s liquidations comes from the smart contracts of lending protocols such as Aave and Compound, which are executed automatically once the conditions for liquidation are met. Moreover, MEV searchers, who act as the liquidators, are incentivized to trigger these liquidations, allowing them to obtain the borrower’s collateral at a discounted price, along with additional rewards, which are taken from the borrower as a penalty.

Searchers often compete in analyzing blockchain data quickly to identify borrowers that are eligible to be liquidated to rapidly submit a liquidation transaction and earn the additional incentives before the others. During periods of increased market volatility, it is not uncommon to see these fees being driven up even higher, further incentivizing liquidators to perform their tasks quickly. While it may seem unfortunate for the borrowers, liquidations are necessary to maintain the overall health of the lending platform, minimizing the possibility of incurring bad debt. 

Below is a simplified example of a liquidation process in DeFi:

Again, this is a much-simplified version of liquidation. An actual liquidation can be very complicated. There are about 200 lending protocols for the moment, each of which has its own set of parameters and risk-managing approach, resulting in increased complexity and higher risk for the liquidators. 

This can be seen from the token flow of a real-life liquidation, performed by a specially-made liquidation smart contract, as shown below.

It goes step by step like this: 

Note that since all the actions are carried out in a single transaction, the transaction can be executed without any errors since all of the relevant actions occur simultaneously. That’s why the liquidator receives the USDC in Step 1 first, even though it seems as if the liquidator had not repaid the loan until Step 3. 

Ultimately, the liquidator receives an additional $128.50 worth of USDC. After deducting gas costs and tips to block builders or miners totaling $18.70, the transaction generated a net profit of $109.80.

While lending protocols have continued to proliferate across DeFi, providing opportunities for more liquidators to become involved, the competition is incredibly ruthless. As time goes by, liquidation MEVs have become more and more centralized, with the top 3 liquidators accounting for 87% of total liquidation volumes among the top 10 liquidators. 

Just-in-time Liquidity Provision (JIT)

A relatively new type of MEV that has become increasingly popular is just-in-time liquidity provision, abbreviated as JIT. Similar to how sandwich attackers operate,  JIT searchers will monitor a network’s mempool for any incoming trades across various liquidity pools. When they detect a large pending transaction, the MEV bots will add liquidity to the targeted pool ‘just-in-time’ before the swap is confirmed. This way, the searcher can obtain a significant share of the trading fees and then quickly remove their liquidity immediately after that. With Uniswap V3's new concentrated liquidity feature, JIT bots play the role of highly active liquidity providers that extract additional swap fees away from other passive liquidity providers.

Similar to sandwich attacks, JIT requires several transactions to obtain a profit. However, unlike sandwich MEVs, JIT transactions actually benefit DEX users by providing larger amounts of liquidity, thereby lowering the slippage and providing better prices for traders.

Let’s look at a simple example. 

The image below shows the first three transactions of block 13601096. 

Source: Etherscan

In the first transaction, the searcher added dual-sided WETH and USDC liquidity to the Uniswap V3:USDC pool.

 Source: Etherscan

In the next transaction, a large-volume trader swaps 1.9M USDC for 409 WETH in the pool.

 Source: Etherscan

In the final transaction, the searcher then removes their liquidity from the Uniswap V3:USDC pool.

 Source: Etherscan

By adding liquidity before the large transaction and removing it right after, the searcher managed to earn 5309 USDC in trading fees.

In terms of  MEV revenues on Uniswap V3, JIT searchers are making far less compared to arbitrageurs and sandwich attackers, which makes sense, given that this MEV technique is still in its infancy. 

Insights and Impacts from MEV

Now that we’ve covered some of the common types of MEVs in crypto, you might be wondering just how big the MEV industry really is. Since the beginning of DeFi summer back in 2020, all the way to the NFT season of 2021, the MEV space has grown by leaps and bounds, with billions of dollars worth of MEV transactions happening daily. On Ethereum alone, the numbers below truly highlight the significance of MEV on the network. 

According to DeFiLlama, total trading volume across DEXs on Ethereum amounted to $664.83B in 2022, with EigenPhi recording $350.9B in arbitrage and sandwich transactions in the past year. MEV transactions made up 53% of total volumes across Ethereum-based DEXs. Unfortunately, sandwich attacks made up the majority of MEV volume, transacting over $309.9B in 2022 (88%).In 2022, 173,283 crypto users lost $129.9M to 457 searchers via 599,023 sandwiches. On average, each user lost approximately $750 to sandwich attacks. Yet, not all of this lost revenue is redirected back to the searchers. In fact, a large part of their income goes to miners (before The Merge) or block builders and validators (after the merge) as incentives for accepting the sandwich transactions. 

As to where these MEV searchers tend to operate, we can see that MEV activity thrived on the most popular DEXs on Ethereum, such as Uniswap and Curve. According to the most popular dashboard on Dune about Uniswap V3 MEV activity, the average daily percentage MEV volume, including arbitrages and sandwiches, reached 43.40% as of Jan 10, 2023. 

Source: Dune Analytics

On Uniswap V3 alone, over $236B worth of sandwich transactions were completed, with the top 3 most targetted pools being the Diamond Heart DAO pool, the WETH:USDC pool, and the USDC:USDT pool. Unsurprisingly, the leading decentralized exchange on Ethereum has also proven to be the most popular destination for arbitrageurs, with over $19B transacted.EigenPhi has also done detailed research on the impact of MEV on Curve, commissioned by the protocol itself. The report shows that, from Jun 1 to Oct 31, 2022, 

“The MEV volume crossed 20% of the total volume on most days. And there are some single days when the MEV volume is roaring to dominate the total volume”

Although Curve has seen much lower sandwich volumes compared to Uniswap, it has become the second most targeted protocol by MEV arbitrageurs. Curve’s multi-asset liquidity pools, such as the 3pool and Tricrypto pool, saw the most action, recording $2B and $1.7B in arbitrage volumes, respectively.

How to Prevent Bad MEVs?

After walking through these different MEVs and taking a glimpse into the numbers, it’s clear that bad MEVs, primarily sandwich attacks, have significantly impacted users in the crypto space. As such, it should be no surprise that crypto natives would want some level of protection against these attacks, lest they become other unwilling victims. According to the paper, SoK: MEV Countermeasures: Theory and Practice, there are four major categories of MEV countermeasures: 

Each category is not a one-size-fits-all solution, and they come with their own pros and cons. While we highly recommend exploring all the possible alternatives that are already available, it is always important to take the necessary precautions before interacting with these. platforms. Doing your own research and only testing with funds you are willing to lose will go a long way in protecting yourself in the permissionless world of crypto.

Conclusion

While MEV transactions may seem somewhat unethical, not all of them are equal, and some benefits come as a result of MEV searchers operating on the open market. MEV technology is constantly under scrutiny, and for some, the power of MEVs on a network that is decentralized and permissionless by nature may be a cause for concern. 

Some types of MEVs, like arbitrages, help increase market efficiency by ensuring asset prices are balanced between the multitude of markets in DeFi. At the same time, JITs serve to minimize slippage and promote better pricing for large-volume traders by increasing liquidity. On the other hand, sandwich MEVs can be detrimental to unsuspecting users, making them pay higher prices for a particular asset or receive back less than what they wanted. 

Regardless of the techniques used, the exponential growth of the crypto space will ensure that the MEV industry will continue to remain relevant in the near future. Subsequently, the crypto community should aim to curb destructive behavior from malicious MEVs by enhancing transparency and increasing awareness and knowledge about the importance of MEVs in DeFi, NFTs, and various other ecosystems.