Front-Running in Crypto and Blockchain

1. Front-running meaning (and origin)
2. Ethereum front-running — how it works on the blockchain
3. The Ethereum mempool and pending transactions
4. Front-running vs sandwich attacks
5. Front-running bots: how they detect opportunities
6. Is front-running legal in crypto?
7. How to protect yourself from front-running

Front-running meaning (and origin in traditional finance)

Front-running originated in equity markets. A broker who receives a large client order knows the order will move the stock price; if the broker buys for their own account first, then executes the client order, the broker profits from the price movement they caused. This is illegal under SEC rules (Rule 10b-5, market manipulation provisions) and ESMA equivalents in Europe — it's a breach of fiduciary duty and an abuse of non-public information.

The traditional-finance definition is narrow: it requires the front-runner to have a duty to the victim (broker-client, exchange-trader, insider-shareholder). In crypto, none of these duties exist — yet the same mechanic plays out at the protocol level because every pending transaction is broadcast publicly. The mechanic is identical; the legal framing is completely different.

Ethereum front-running — how it works on the blockchain

When you submit an Ethereum transaction — a Uniswap swap, an NFT purchase, a contract call — it doesn't go directly into a block. It first enters the mempool, where it waits for a block builder to include it. During those few seconds in the mempool, your transaction is visible to every Ethereum node on the network, including hundreds of front-running bots.

A front-running bot does the following in under 100 milliseconds:

1. Detects your pending swap in the mempool and decodes the calldata.
2. Simulates the swap against current pool reserves to predict the price impact.
3. Constructs its own swap in the same token pair, in the same direction as yours.
4. Submits its transaction to a block builder with a higher tip than yours.
5. The builder orders the bot's transaction before yours in the next block.

The bot's buy pushes the price up. Your buy then executes at the elevated price, getting fewer tokens than you would have without the front-runner. The price impact you cause is captured by the front-runner's tokens, which can now be sold for more than the bot paid.

The Ethereum mempool and pending transactions

The Ethereum mempool typically holds 30,000–80,000 pending transactions at any moment. Several thousand are DEX swaps, and each one is a potential front-running target. The mempool is global — every node sees every transaction — but propagation isn't instant. A transaction announced in Frankfurt reaches Tokyo a few hundred milliseconds later. That latency window is what determines who wins the front-run.

Sophisticated front-running bots run dedicated nodes in multiple geographic regions, sometimes peering directly with the busiest validators' nodes, to get pending transactions before they've propagated to the rest of the network. This is the latency arms race that determines profitability. A bot running on a generic cloud provider through a public RPC sees pending transactions 50–200ms after the top bots — way too late to outbid them.

Front-running vs sandwich attacks

Pure front-running places one transaction ahead of the victim's. The bot ends up holding the tokens it bought; selling them is a separate decision, exposed to whatever the price does after the victim's trade.

A sandwich attack adds a back-run on top: the bot front-runs the victim, then immediately sells the tokens at the new elevated price in the same block. The whole sequence — front-run + victim + back-run — is atomic: it either all lands or none of it lands. This eliminates the price risk of holding the front-run tokens.

Modern Ethereum MEV is almost entirely sandwich attacks, not pure front-running. Pure front-running survives in NFT mints (front-run a popular drop) and in tail strategies on illiquid chains, but on Ethereum mainnet, every front-run worth doing is wrapped into a sandwich.

Front-running bots: how they detect opportunities

A front-running bot evaluates every pending DEX transaction against the same checklist:

• Size. The victim's trade must move the price enough to cover gas + builder bribe + still leave profit.
• Slippage tolerance. Decoded from amountOutMin in the calldata. The bot can only push the price within this window before the victim's trade reverts.
• Token type. Standard ERC-20s only. Fee-on-transfer tokens, rebasing tokens, and tokens with transfer hooks are skipped — they break the sandwich math.
• Honeypot check. Simulated against a forked node to verify the token can be sold. Honeypots that allow buying but block selling are filtered out.
• Private flow flag. Transactions originating from Flashbots Protect, MEV-Share, CowSwap, or 1inch Fusion are skipped — they aren't public mempool transactions.

Production bots like JaredFromSubway evaluate thousands of pending transactions per minute, with the entire decision pipeline completing in single-digit milliseconds.

Is front-running legal in crypto?

In traditional securities markets, front-running is illegal under SEC Rule 10b-5 and ESMA market-abuse regulations — it requires a fiduciary duty (broker-client, exchange-trader) that crypto MEV bots don't have. On Ethereum, transactions are broadcast publicly to a permissionless network. Reading the mempool and submitting your own transaction is exactly how the protocol is designed to work — there is no non-public information being exploited, no hacking, no terms-of-service violation.

That said, regulation varies by jurisdiction, and US/EU regulators have signaled interest in extending market-abuse frameworks to crypto. The SEC's 2023 cases against centralized exchanges briefly touched MEV; the European MiCA framework includes provisions that may apply to MEV in the future. None of these have yet produced a definitive ruling on Ethereum MEV. Consult local legal counsel before operating commercially.

How to protect yourself from front-running

If you're a DEX user, four practical defenses exist:

• Flashbots Protect — submit your transaction privately to Flashbots, bypassing the public mempool. Front-runners can't see it before it lands.
• MEV-Share — Flashbots' newer system that selectively reveals parts of your transaction to searchers in exchange for a kickback if they extract value around it. You can opt to leak only minimal data.
• CowSwap and 1inch Fusion — intent-based DEXs that batch and match trades off-chain, executing the cleared trades atomically with no mempool exposure.
• Tight slippage — set your slippage tolerance to 0.1–0.5% on public-mempool swaps. Front-runners need slippage room to profit; tight slippage forces them to skip your trade.

For more on the MEV ecosystem, see the complete MEV bot guide or read about the largest operator on the JaredFromSubway page.

Further reading and references

• Flash Boys 2.0 (Daian et al., 2019) — Original academic treatment of blockchain front-running, including PGA mechanics.
• Flashbots architecture — How the Flashbots auction system replaced public-mempool front-running with sealed-bid bundle auctions.
• MEV-Share documentation — MEV-Share is Flashbots' protocol for sharing MEV with users instead of pure extraction.
• SEC Rule 10b-5 (market-manipulation prohibition) — The U.S. federal anti-fraud rule that frames how traditional securities front-running is treated; the enforcement model crypto regulators reference when extending market-abuse frameworks.
• EU MiCA Regulation (Article 91 — market abuse in crypto-assets) — The EU's Markets in Crypto-Assets Regulation extends market-abuse concepts (including front-running of orders) to crypto venues operating in the EU.