MEV & Sandwich Attacks: Why You Get Worse Prices Than You Should

You swap 1 ETH for USDC on Uniswap. The quote shows 3,400 USDC. The transaction confirms, but you end up with 3,362. No error message. Swap succeeded. You just got sandwiched.

It happens because every swap you broadcast is visible to bots before it confirms. There's a window — sometimes just a few seconds — where automated searchers can see your trade in the mempool, calculate how to profit from it, and submit their own transactions around yours. They make money. You get less than you were quoted. The bot's cut is called MEV — Maximal Extractable Value — and it's structural to how public blockchains work. You can't eliminate it, but you can reduce how much you give up.

How a sandwich attack works step by step

Your swap goes through a public mempool before it's included in a block. Bots monitor this mempool constantly. Here's the exact sequence when you get sandwiched:

• 1.You submit a swap — say, 1 ETH for USDC — with 1% slippage tolerance
• 2.A bot spots your transaction in the pending mempool
• 3.The bot submits its own buy for the same token with a higher gas fee, so it lands first in the block
• 4.Your transaction executes at a worse price because the bot's buy already moved it
• 5.The bot immediately sells what it just bought, capturing the price difference
• 6.You received fewer USDC than the original quote, within your slippage tolerance so the transaction succeeded

The attack is profitable on any trade large enough to move the price. On a $10,000 swap with 1% slippage, a bot can reliably extract $50–$100. Scaled across millions of daily transactions, this adds up fast.

Why slippage tolerance is your main lever

Slippage tolerance is how much worse than the quoted price you're willing to accept. It's meant to protect you from price movement between quote and execution — but it also sets the floor for how much a bot can extract from your trade.

• —0.1% slippage on stablecoins — makes most sandwich attacks unprofitable after gas costs
• —0.5% for liquid tokens like ETH, BTC wrapped variants — reasonable protection
• —1–3% for mid-cap tokens — a bot can comfortably extract this on large trades
• —5–10% (often the default) — this is free money for MEV bots

The problem is that low slippage means your transaction fails more often if the market moves. It's a tradeoff. For large trades, use low slippage and retry if it fails. For small trades, the gas cost of retrying often exceeds what you'd lose to MEV.

How to actually reduce MEV exposure

• —Flashbots Protect (Ethereum) — routes your transaction directly to validators, bypassing the public mempool. Bots can't see it before it confirms.
• —MEV Blocker RPC — similar approach, aggregates multiple private relays for better inclusion odds
• —CoW Protocol — batch settlement matching means there's no mempool transaction to front-run at all
• —Break large trades into smaller chunks — a $10k swap is a tempting target; four $2.5k swaps are less profitable to attack individually
• —Trade on DEXs with MEV protection built in — some aggregators like 1inch have MEV-resistant routing

Frequently asked questions