Originally defined as miner extractable value (MEV) in Flash Boys 2.0, MEV refers to the additional profit a miner can make by including, re-ordering, inserting, or omitting transactions from the blocks they produce. While the process of winning a block is fairly democratic, the winning block producer has significant control over the transactions that can be included in the block. This power is the underlying cause of MEV and should, in theory, only be used for forces of good. The Flashbots team advocates for rebranding “miner extractable value” to “maximal extractable value” to encompass all value that is extractable by privileged actors (e.g. validators in ETH2) beyond just what miners in PoW systems can extract.
Before an Ethereum transaction is included in a block, it goes to the Mempool, a publicly accessible staging area. This is where MEV bots search for transactions that they can use to capture some MEV, for example through arbitraging markets or liquidating newly undercollateralized loans.
For example, say there is a large Uniswap order to purchase SNX / WETH in the Mempool. An MEV bot will place a transaction immediately behind the large Uniswap order to arbitrage the price slippage in the Uniswap pool against the SNX / WETH price in another AMM pool like SushiSwap. This is considered benign or positive MEV, as it benefits the ecosystem.
However miners, as the ultimate gatekeepers of what is included on-chain, have a trump card over even MEV bots. Miners that have “generalized frontrunning” bots can copy the MEV bot’s arbitrage transaction and replace it with their own, thus capturing the profits from the arbitrage for themselves. Generalized frontrunning is considered a malignant MEV transaction, or a transaction that destabilizes the ecosystem. Censorship attacks and centralizing MEV extraction are two problems that have been identified by Ed Felten in regards to MEV auctions and pose significant dangers to the Ethereum community.
Time-bandit attacks are another MEV attack vector identified in the seminal Flash Boys 2.0 paper that have the potential to destabilize consensus. They pose a greater existential risk to blockchains with state, i.e. Ethereum. In this attack, a miner can “rewind” the chain and re-mine already executed blocks for their benefit, thus capturing MEV they otherwise wouldn’t have. The lower the block rewards are, the greater the incentive for miners to re-org blocks to capture historical MEV.