What MEV Actually Is, Without the Mystique

Most explanations of MEV start with a sandwich attack and stop there. That is a symptom, not the mechanism. To understand why extractable value exists at all, start with a duller observation: a blockchain is an auction where everyone can see the bids.

The mempool is a queue you can read

When you submit a transaction, it does not go straight into a block. It goes into the mempool — a shared waiting room that every node on the network can read. Your intent is public before it is final.

That gap between announced and settled is the entire opportunity. Anyone watching the mempool knows what you are about to do, and — because block builders order transactions by fee rather than by arrival time — they can pay to act on it first.

MEV is not a bug in a particular protocol. It is what happens when you run a sealed-bid market with the envelopes open.

Where the value actually comes from

Extractable value shows up in three broad shapes, and they are not morally equivalent:

• Arbitrage. The same asset trades at different prices on two venues. Buying the cheap one and selling the dear one moves them back together. This is the boring, useful kind — it is how prices stay consistent across a fragmented market.
• Liquidations. An undercollateralised loan needs closing. Protocols pay a bounty for whoever does it, because the alternative is bad debt.
• Sandwiching. A searcher sees your large swap, buys ahead of it, lets your order push the price up, and sells into it. Your slippage is their margin. This one is a genuine tax on the user.

The first two make markets work. The third is a transfer from you to someone faster. Conflating them is why the conversation goes nowhere.

A concrete example

Suppose ETH trades at 2,000 USDC on one pool and 2,006 on another. The opportunity is the spread, minus the cost of taking it:

const spread = (sell - buy) * size;          // gross edge
const cost   = gasPrice * gasUsed + fees;    // cost to capture it
const edge   = spread - cost;                // what is actually left

if (edge > MIN_EDGE) submit(bundle);

That subtraction is the whole game. The gross spread is visible to everyone watching the same two pools. What separates a profitable searcher from an unprofitable one is the second line — and the second line is mostly infrastructure.

Why latency decides the outcome

If ten actors see the same 6 USDC spread, the one who lands their bundle in the block captures it and the rest have burned gas on a reverted transaction. The edge is not the idea. The edge is arriving.

Nothing on that list is clever. It is all engineering — where your nodes sit, whether your state is warm, how many hops stand between the decision and the builder. Which is the uncomfortable conclusion of most MEV research: the strategy is usually public, and the moat is usually plumbing.

What this means if you are not a searcher

Two practical things.

First, your slippage tolerance is a bid. Setting it to 5% on a large swap advertises exactly how much someone may take from you. Tighten it and accept the occasional failed transaction.

Second, private routing removes the announcement. If your transaction never touches the public mempool and goes straight to a builder, there is no window in which to front-run it. That is not a clever trick; it is just declining to publish your intent early.

MEV is not going away, because the thing that produces it — a public, ordered, fee-priced ledger — is the thing that makes the ledger work. What can change is who is positioned to capture it, and whether the capture comes out of your trade or out of a price discrepancy that needed correcting anyway.