Sandwich Attack: MEV Explained & Defenses

A sandwich attack is an MEV (Maximal Extractable Value) tactic where an attacker “sandwiches” a victim’s on-chain trade: a front-run transaction before the victim, the victim’s transaction, then a back-run after. The attacker profits from the price movement caused by the victim’s trade.

Short version: attacker buys before you, you push price up, attacker sells after - attacker pockets the difference.

Step‑by‑Step (with UUID example) ​

1. A dApp submits a user order and logs an internal tracking id (UUID) off‑chain:
   victim_tx_uuid = "3fa85f64-5717-4562-b3fc-2c963f66afa6"
   Important: never expose such IDs on‑chain or in mempool‑visible metadata.
2. The victim’s transaction enters the public mempool: a market buy for 100 TOKEN.
3. An attacker watching the mempool sees the pending intent and issues a higher-priority buy (front‑run) mined just before the victim.
4. The victim’s buy runs next, pushing the token price up (slippage).
5. The attacker immediately sells (back‑run) into the higher price and pockets the difference (minus gas/priority fees).

---

Numeric Example - AMM (digit‑by‑digit) ​

This is a simplified illustrative AMM example.

Initial price per TOKEN = $1.00.
Attacker front‑runs and buys 100 TOKEN at $1.00 each.

Cost calculation (digit-by-digit):

• 100 × 1.00 = (100 × 1) + (100 × 0.00) = 100 + 0 = 100.00 → attacker pays $100.00.

Victim’s buy executes and moves price to $1.05 per TOKEN (≈ 5% impact).

Revenue calculation (digit-by-digit):

• 100 × 1.05 = (100 × 1) + (100 × 0.05) = 100 + 5 = 105.00 → attacker receives $105.00.

Profit calculation (digit-by-digit):

• Gross profit = 105.00 - 100.00 = 5.00.
• Assume gas/priority fees = 2.00.
• Net profit = 5.00 - 2.00 = 3.00 → $3.00 net.

Real AMM math uses constant product curves, slippage functions, and liquidity depth; this is simplified to show the concept clearly.

---

What Makes Victims Attractive? ​

• High slippage tolerance set by the user or dApp (gives attackers room to profit).
• Large trade size relative to pool liquidity (bigger price impact).
• Public mempool broadcasting (no privacy / private relay).
• Predictable fee/nonce behavior that allows precise front‑running placement.

---

How UUIDs (and Metadata) Factor In ​

• Bad: Placing UUIDs or verbose metadata on‑chain or in mempool‑visible fields makes it trivial for bots to correlate off‑chain high‑value orders with on‑chain txs. That’s an open invitation to MEV bots.
• Good: Keep UUIDs strictly off‑chain (server logs only) and never include them in calldata, memos, or other public fields.

---

Practical Defenses & Prioritized Mitigations ​

1. Private transaction relays (Flashbots / private RPCs) - bypass the public mempool so MEV bots can’t see pending txs.
2. Tighten slippage tolerances - require minimal acceptable price impact (e.g., 0.5%-1%) so attackers have less room.
3. Limit orders / batch auctions - remove first-seen priority and reduce front‑running opportunities.
4. Time‑weighted or randomized submission - avoid predictable fee/nonces used for exact placement.
5. UX slippage warnings - show estimated price impact and require explicit user acceptance.
6. Increase liquidity for important pairs - lower price impact per trade.
7. Avoid exposing identifiers on‑chain - never put internal UUIDs or rich metadata into publicly broadcast tx fields.
8. Monitoring & detection - log abnormal front/back‑run patterns and alert or pause suspicious trades.
9. Protocol solutions - support fair ordering (FO) or auctioned ordering where possible to remove MEV rent extraction.

Quick Tips for dApp Builders & Traders ​

• For builders: Treat UUIDs as internal secrets - use them only in server logs and communications. Don’t emit them in calldata or mempool-visible metadata.
• For traders: Use conservative slippage, split large orders, or use private execution when possible for big trades.

Final Summary - One Sentence ​

A sandwich attack is when an attacker front‑runs your trade, your trade raises the price, and the attacker back‑runs to sell into that higher price - profit comes at your expense. Protect with private relays, tight slippage, limit orders, and never leaking order IDs on‑chain.