Where MEV comes from
A block producer decides which pending transactions go into a block and in what order. ethereum.org defines MEV as the maximum value that can be extracted from block production in excess of the standard block reward and gas fees by including, excluding, and changing that order.[1]
The idea was first described under proof of work as miner extractable value. After The Merge, validators took over block production, and the term became maximal extractable value.[1]
Searchers are the main MEV hunters: bots and teams that watch the mempool, simulate outcomes, and submit transactions or bundles that profit only if they land in a specific position.[1]
MEV is possible because pending transactions are visible in the public mempool before they are final, and because whoever builds a block controls the order of transactions in it.[1]
ethereum.org also describes generalized front-runners: bots that copy any profitable pending transaction, replace the recipient address with their own, and submit it with a higher fee. They need no understanding of the underlying strategy, only a simulation showing that the copy makes money.[1]
The main types of MEV
DEX arbitrage is the most common. When the same pair trades at different prices on two exchanges, a searcher buys on one and sells on the other in a single transaction. ethereum.org cites a trade that turned 1,000 ETH into 1,045 ETH using an ETH/DAI price gap between Uniswap and Sushiswap.[1]
Liquidations are the second source. Lending protocols pay a fee to whoever repays an undercollateralized loan first, so searchers race to be first after a price update.[1]
Sandwich trading is the one that directly costs users. A searcher sees a large pending swap, buys the same asset just before it, and sells just after, capturing the price move the victim's own trade created.[1]
ethereum.org lists arbitrage and liquidations among the positive effects of MEV: DEX arbitrage helps users get correct prices across venues, and lending protocols rely on speedy liquidations. Sandwich trading, by contrast, comes directly out of another user's trade.[1]
There is a long tail as well. ethereum.org describes NFT MEV as an emergent category where searchers compete to snap up underpriced listings or be first into a popular mint.[1]
See also: Uniswap · Sushi · Curve Finance
Worked example: a sandwich trade
The numbers below are our own calculation, not data from a real pool. Take an illustrative pool with 1,000 ETH and 3,000,000 USDC, priced at 3,000 USDC per ETH, using the constant-product formula x times y equals k and ignoring pool fees. A user submits a swap of 300,000 USDC for ETH with 3% slippage tolerance.
Without interference, the user gets about 90.9 ETH: the pool moves to 3,300,000 USDC and roughly 909.1 ETH. That is an average of about 3,300 USDC per ETH.
A searcher front-runs with 100,000 USDC and receives about 32.3 ETH. The user's trade now executes at a worse price and receives about 85.4 ETH, which is 6.1% less than the 90.9 ETH quoted. With a 3% slippage limit the swap would revert, so this sandwich only works against a tolerance of about 6.1% or higher.
If the user did set that loose tolerance, the searcher then sells its 32.3 ETH back into the pool for about 119,900 USDC, a gross gain of about 19,900 USDC before gas and builder payments. The gap between 90.9 and 85.4 ETH is the value extracted from the user, split between searcher profit and fees paid to the builder.
The same arithmetic shows why trade size matters. On a 30,000 USDC swap in the same pool, the average price moves about 1% instead of 10%, leaving far less room for a sandwich to cover gas. Large trades against shallow pools are the main target, which is also why splitting a trade or routing through deeper pools lowers the risk.
Searchers, builders, and MEV-Boost
Since The Merge, many Ethereum validators do not assemble blocks themselves. They run MEV-Boost, open-source middleware built by Flashbots that lets a validator access blocks from a competitive market of builders through relays, which pass on the most profitable block.[2]
Searchers send bundles to builders, builders pack bundles and ordinary transactions into full blocks, relays check those blocks, and the validator proposes the most profitable one. ethereum.org names proposer-builder separation and the Builder API, of which MEV-Boost is an example, as the two solutions proposed so far to limit MEV's effect on decentralization and security.[1]
Under EIP-1559, the base fee is burned, so a validator's extra income from MEV comes from priority fees and direct payments by builders, not from the base fee.[4]
ethereum.org warns that MEV can accelerate validator centralization, because large staking pools can invest more in capturing MEV than solo stakers. MEV-Boost addresses this by letting any validator sell its block space to an open market of builders instead of running its own searching infrastructure.[1]
Proposer-builder separation shifts work rather than removing every problem. Validators depend on relays to pass on valid blocks, and ethereum.org describes combating MEV-related problems after The Merge as a core area of research, not a solved one.[1]
How to reduce MEV on your own trades
Use a tight slippage limit. It caps the price move a front-runner can force on you, as the example above shows.
Send trades through a private RPC. Flashbots Protect sends transactions to a private Flashbots mempool hidden from frontrunning and sandwich bots, and if a transaction creates MEV, the user can earn an MEV refund.[3]
Use DEX designs that batch orders or settle off the public mempool, such as intent-based swaps. They do not remove MEV from the chain, but they move it away from your order.
Check the effective price on a block explorer after large trades. If the executed price is consistently near your slippage limit rather than near the quoted price, your orders are being targeted, and a private route or tighter limit is the fix.
See also: CoW Swap · 1inch · Ethereum RPC endpoints
What MEV costs the network
MEV is not only a cost to individual traders. ethereum.org notes that the gas price auctions frontrunners run against each other cause network congestion and high gas prices for everyone else trying to run regular transactions.[1]
At the consensus layer, ethereum.org describes a risk that if the MEV in a block significantly exceeds the standard block reward, validators may be tempted to reorg blocks to capture it. It argues that proposer-builder separation reduces this threat of time-bandit attacks, because validators are no longer directly focused on optimizing MEV income.[1]
ethereum.org also credits Flashbots's MEV relay with reducing the effectiveness of generalized frontrunners and taking gas price auctions offchain, which lowered gas prices for ordinary users. The competition still exists, but much of it now happens in private auctions rather than as public bidding wars.[1]
For users, the practical cost shows up as worse execution on large or loose-slippage swaps. For the network, it shows up as pressure toward validator centralization, which is the concern ethereum.org lists under the effects of MEV after The Merge.[1]
See also: Best decentralized exchanges
The bottom line
MEV is not a bug that will be patched away: it follows from public mempools and a single party ordering each block. What you control is your exposure. On a 300,000 USDC trade in the example pool, a 3% slippage limit turned a 6.1% loss into a reverted transaction, and a private RPC removes the trade from the mempool before a searcher can see it.[3]
Educational content, not financial advice. Crypto assets are volatile; do your own research.