Every time you swap tokens on a decentralized exchange, your transaction sits in a public queue for a fraction of a second before it confirms. That window is enough for automated bots to read your trade, jump in front of it, and pocket the difference at your expense. This is the mechanics of Maximal Extractable Value (MEV), and it costs DEX traders real money on every unprotected swap.
MEV is not a niche problem. Sandwich attacks on Ethereum drained approximately $40 million from traders across 2025 alone, according to EigenPhi data published by Cointelegraph Research in December 2025, even as monthly extraction rates fell sharply from nearly $10 million in late 2024 to around $2.5 million by October 2025. The decline reflects growing adoption of protection tools, but the underlying vulnerability has not gone away. Attack frequency remained high at 60,000 to 90,000 incidents per month throughout the period.
This guide covers how MEV extraction works mechanically, which protection tools actually prevent it, how to configure slippage to minimize your exposure, and where CEX execution eliminates the problem entirely.
What MEV is and why it costs you money
Maximal Extractable Value is the profit available to anyone who can influence transaction ordering within a block. Originally called "Miner Extractable Value" when miners controlled block production, it was renamed after Ethereum's transition to proof of stake, where validators and block builders now control ordering. MEV exists because blockchain transactions are not processed in the order they are submitted. They sit in a public waiting area called the mempool, where anyone can see them before execution. Block builders choose which transactions to include and in what sequence, and that ordering power creates three extraction strategies that affect you directly.
Front-running. A bot sees your pending buy order and places its own buy ahead of yours, pushing the price up before your trade executes. You pay more, and the bot sells immediately after for a profit.
Back-running. A bot places a transaction immediately after yours to capture the price movement your trade created. This is less directly harmful to you but still extracts value from the ecosystem.
Sandwich attacks. The combination of both: a bot front-runs your trade by buying before you and pushing the price up, your trade executes at the inflated price, then the bot back-runs by selling at the higher price your buying pressure created. Your slippage tolerance effectively becomes the extraction budget for the bot.
The critical insight here is that your slippage setting is not just protection against normal price movement. It is the maximum amount a sandwich bot can extract from your trade. Setting 3% slippage on a $10,000 swap means you have announced to the network that you will accept up to $300 in value extraction.
How sandwich attacks work step by step
A sandwich attack exploits the transparency of the public mempool and the mechanics of AMM pricing. For a dedicated breakdown of the attack itself, see our guide to sandwich attack mechanics. Here we focus on the exact execution sequence, so you understand what you are defending against.
Traders executing large swaps on DEXs without MEV protection consistently receive worse fills than expected, while those who route through private mempools or use DEX aggregators with protection features report meaningfully better execution.
Step 1: Detection. You submit a swap of 5 ETH for USDC on Uniswap through the public mempool. Your transaction is visible to everyone before it executes. The bot's monitoring system identifies your trade, calculates the pool's reserve ratio, and estimates the price impact of your swap.
Step 2: Front-run transaction. The bot submits its own buy of 3 ETH from the same pool with a higher gas fee (priority fee), ensuring it executes before yours. This swap shifts the pool's reserve ratio, increasing the effective price of ETH in the pool.
Step 3: Your transaction executes. Your 5 ETH swap now executes against a pool where the price has already moved against you. If the bot's front-run pushed the price up 0.8% and your slippage tolerance is set at 1%, your transaction still succeeds but at a worse rate.
Step 4: Back-run. The bot immediately sells its 3 ETH back into the pool at the elevated price your trade pushed it to. The bot profits the spread between its buy price (before the front-run) and its sell price (after your trade), minus gas costs.
Profitability math. For the bot, a sandwich pays off when the value it can capture inside your slippage tolerance is greater than its costs: the gas for two transactions, plus the price impact of its own buy and sell. If what is left over is positive, your trade is worth attacking. On Ethereum mainnet, where base fees have spent most of 2025 and 2026 below 1 gwei (often around 0.05 gwei in mid-2026), gas is rarely what stops a sandwich, so bots can profitably target trades from roughly $1,000 upward depending on pool depth.
Despite the profitability compression seen in 2025, when average net profit per attack fell to around $3 according to EigenPhi, the volume of attacks remained high because bots compensate through quantity. In October 2025, 515 distinct bots were still operating on Ethereum, though only about 100 actively run sandwich trades in a typical month.
Slippage settings and MEV exposure
Your slippage tolerance directly controls your MEV exposure, and it is the single most important variable you can adjust without changing any tooling.
How slippage works
When you set 1% slippage on a swap, you are telling the smart contract to execute your trade only if the price has not moved more than 1% against you since you submitted it. If the price moves more than 1%, the transaction reverts, and you pay gas but get no fill.
The MEV connection
A sandwich bot calculates whether it can profitably push the price within your tolerance. If you set 5% slippage on a liquid pair, you have created a large window for extraction. If you set 0.3%, the bot's front-run would need to keep total price movement under 0.3%, which is often not profitable after gas costs.
Recommended settings by pair type:
Major pairs (ETH/USDC, BTC/WBTC): 0.3–0.5%
Mid-cap tokens with decent market liquidity: 0.5–1.0%
Low-cap or new listings: 1–3% (accept higher risk or use private routing)
Never exceed 5% unless you understand the token specifically requires it, as with rebase tokens or tokens with transfer taxes
The revert trap. Setting slippage too low causes transactions to revert during volatile periods. You pay gas but get no execution. The correct response is not to raise slippage but to use private transaction routing, which eliminates sandwich risk regardless of your tolerance setting.
Partial fill strategies. For large swaps, splitting into multiple smaller transactions reduces per-transaction profitability for bots. A single $50,000 swap with 1% slippage is a $500 extraction opportunity. Five $10,000 swaps each with 0.5% slippage present five $50 opportunities, each of which may fall below the bot's profitability threshold after gas.
Protection tool 1: Flashbots Protect
Flashbots Protect is a free RPC endpoint that routes your transactions through a private channel directly to block builders, bypassing the public mempool entirely. Your transaction is never visible to sandwich bots before execution.
How it works:
You add the Flashbots Protect RPC to your wallet. The current recommended endpoint is https://rpc.flashbots.net/fast, which shares your transaction with every registered builder for faster inclusion.
When you submit a transaction, it goes to the Flashbots relay instead of the public mempool.
The relay forwards your transaction to participating block builders.
Your transaction is either included in a block exactly as submitted or not included at all. It is never exposed publicly before execution.
Key properties:
No failed transaction fees. If your transaction would revert, it simply is not included. You do not pay gas for failed trades.
MEV refunds. If your transaction creates backrun opportunities (non-harmful MEV), Flashbots can capture that value and refund a portion to you.
Builder coverage. Flashbots Protect accesses the majority of Ethereum block space through multiple top builders.
No smart contract changes required. It is a drop-in RPC replacement. Switch the endpoint in MetaMask or your wallet of choice.
Limitations:
Ethereum mainnet only. Does not protect on L2s, Solana, or other chains.
Slightly slower inclusion. Because your transaction skips the public mempool, it may take 1 to 2 blocks longer to confirm during periods when Flashbots-connected builders are not winning blocks.
Does not protect against builder-level extraction. If a builder itself decides to sandwich you, Flashbots Protect does not prevent it, though reputational incentives discourage this.
As of October 2024, the most recent totals Flashbots has published, 2.1 million Ethereum accounts had used Protect since launch, protecting $43 billion in DEX volume and earning users 313 ETH in MEV refunds. Private mempools account for around half of all gas on Ethereum, a shift that took hold by 2024 and shows how broadly the ecosystem has moved toward protected transaction routing.
Protection tool 2: MEV Blocker and private mempools
MEV Blocker is a free RPC endpoint you add to your wallet, so it works at the user level much like Flashbots Protect.
How MEV Blocker works
When you route a transaction through it, MEV Blocker sends your order to a private network of searchers who are allowed to backrun it but never to front-run or sandwich it.
Because your transaction skips the public mempool and its slippage details are withheld, sandwiching is prevented by design rather than filtered out after the fact.
If a searcher wins a backrun bid, the value is rebated, with the bulk going back to you and a smaller share to the validator.
The trade-off is that you have to opt in by adding the RPC, and protection only covers the transactions you route through it.
Other private mempool solutions
CoW Swap batch auctions. CoW Protocol collects trade intents over approximately 30-second windows and processes them as batches. Within a batch, all trades of the same pair receive a uniform clearing price, and solvers compete to find the best execution route. Because intents stay private during collection and all users get the same price, sandwich attacks within a batch are structurally impossible.
1inch Fusion mode. Uses intent-based execution where independent resolvers (1inch's name for the market makers that fill your order) compete to execute it and settle it on-chain. Your trade never enters the public mempool as a raw transaction, which removes the detection vector for sandwich bots entirely.
Private RPCs from node providers. Services like QuickNode and Alchemy offer private transaction routing options that skip the public mempool, though coverage and guarantees vary by provider.
Choosing between tools. Flashbots Protect is the simplest single-user solution (change one RPC setting). CoW Swap offers the strongest structural guarantee, because its batch auction eliminates sandwiching by design rather than just by hiding transactions. MEV Blocker is another one-RPC option, about as easy to set up as Flashbots Protect, and it prevents sandwiching on any order you route through it.
CEX vs DEX: Where MEV cannot reach
On a centralized exchange, MEV as described above does not exist. The order book matching engine executes trades internally, in a closed system where no external party can observe or reorder pending orders. For a full breakdown of how the two venue types compare, see our guide to CEX versus DEX trade-offs.
Why CEX eliminates mempool MEV
No public mempool. Your order goes directly to the exchange's matching engine.
Deterministic execution. Price-time priority means your order fills in the sequence it was received.
No gas economics. There is no fee-based priority mechanism that lets others jump ahead of you.
No smart contract interaction. Execution happens off-chain in the exchange's internal systems.
CEX has its own extraction risks (not MEV but analogous)
Latency arbitrage. High-frequency trading firms colocate servers at exchange data centers to see and react to orders microseconds before retail traders can. This is functionally similar to front-running but operates through speed advantage rather than mempool visibility.
Market maker information advantage. Designated market makers on some exchanges see aggregate order flow before retail orders fill, potentially adjusting quotes.
Exchange front-running allegations. Historically, some exchanges have faced accusations of proprietary trading desks using knowledge of pending customer orders. Regulated exchanges have surveillance to prevent this.
The practical comparison for a $10,000 trade in ETH:
Venue | MEV risk | Pool/taker fee | Gas |
DEX without protection | $50–300 sandwich extraction | 0.3% | Yes |
DEX with Flashbots Protect | None | 0.3% | Yes |
CEX order book | None | 0.04–0.10% | None |
For a full breakdown of what trading on a CEX actually costs, see our guide to crypto trading fees. For traders who do not specifically need self-custody or permissionless token access, CEX execution remains the lowest-cost path for major pairs, and the MEV problem is exclusively a DEX and on-chain concern.
Quantifying your MEV cost
Before adopting protection tools, it is worth understanding how much MEV may already be costing you.
Transaction-level analysis. Compare your executed price to the market mid-price at the time of your swap submission. The gap between expected and received output reveals your total execution cost, which includes AMM price impact, pool fees, and any MEV extraction.
Tools for measuring MEV impact:
EigenPhi (eigenphi.io): Tracks sandwich attacks on Ethereum and identifies victim transactions. Search your address to see if your past trades were sandwiched.
Dune Analytics dashboards: Community-built dashboards tracking sandwich volume, average extraction per attack, and top bot addresses.
Benchmarks for typical retail impact:
Trades under $1,000 on liquid pairs: Rarely sandwiched, since the extraction would not cover the bot's gas and priority costs.
Trades $1,000–$10,000: Can be sandwiched on unprotected swaps. When it happens, extraction typically runs about 0.3–1.5% of trade value. Reliable public data on how often trades this size get hit is limited, so treat any single frequency figure as a rough estimate.
Trades above $10,000: The highest-value targets, because a larger swap moves the pool price more and widens the extraction window. Route these through private transactions rather than leaning on a slippage setting.
Monthly cost estimation. If you execute ten $5,000 DEX swaps per month without protection and 10% get sandwiched at an average 0.8% extraction: 10 trades x 10% hit rate x $5,000 x 0.8% = $40/month in MEV losses. Over a year, that is $480 in invisible cost that a free RPC change would eliminate.
MEV on Layer 2 and other chains
MEV is not an Ethereum-only problem. Every blockchain with a public mempool or observable transaction ordering faces some variant.
Ethereum Layer 2s (Arbitrum, Optimism, Base)
L2 sequencers process transactions in a first-come-first-served model with less transparent ordering. MEV on L2s exists but the infrastructure for extraction is less developed than on mainnet. Sequencer operators have the technical ability to extract MEV but face reputational and governance constraints. As L2s decentralize their sequencers, MEV dynamics will increasingly mirror mainnet conditions.
Solana
In March 2024, Jito Labs shut down the mempool service it had built for the Jito Block Engine, citing the negative impact of sandwich attacks on Solana users. When active, Jito tips had accounted for nearly 50% of Solana's Real Economic Value, and the mempool window had enabled front-running by providing a 200ms auction period for searchers to observe and outbid pending transactions.
Since the shutdown, Jito's infrastructure has grown substantially rather than retreating. By mid-2026, the Jito-Solana validator client runs on more than 95% of active stake, and Jito tips have become the dominant share of validators' non-base-fee revenue, consistently outpacing Solana's in-protocol priority fees by several times. In July 2025, Jito launched the Block Assembly Marketplace (BAM), a TEE-based block-building architecture that makes transaction sequencing transparent and verifiable while introducing privacy-preserving execution for users. You can read more about how this infrastructure works in our guide to Jito and Solana's MEV infrastructure. Solana's architecture still makes traditional public-mempool sandwich attacks harder than on Ethereum mainnet, but the validator-side MEV ecosystem has grown more sophisticated, not less.
BNB Chain
BNB Chain has a public mempool and is one of the more heavily sandwiched networks, so MEV exposure is real and bots are active. Protection is readily available, though: several major wallets and RPCs offer MEV-protected routing on BNB Chain, and the network has rolled out its own anti-sandwich measures. Turn on a protected RPC or wallet setting when you trade here.
Practical rule
If you are trading on any chain where your transaction is visible before execution, assume MEV bots are watching. Use private routing or intent-based execution regardless of which chain you are on.
Protection checklist for DEX traders
Before executing any DEX swap, run through this sequence:
Set slippage to 0.3–0.5% for major pairs. Only increase if the token specifically requires it.
Enable private transaction routing. Add Flashbots Protect RPC (Ethereum) or use CoW Swap / 1inch Fusion for intent-based execution.
Split large orders. Any swap above $10,000 should be broken into 2–5 smaller transactions.
Check pool depth before trading. If your trade exceeds 1% of pool reserves, expect significant price impact regardless of MEV protection.
Verify the token contract. Tokens with hidden trading fees or transfer restrictions can trap funds regardless of MEV settings.
Consider timing. Gas spikes during network congestion increase bot profitability thresholds but also increase your transaction costs. Off-peak hours (weekends, early UTC mornings) typically have lower gas but also thinner liquidity.
Use a DEX aggregator. Aggregators like 1inch, Paraswap, and CoW Swap route across multiple pools to minimize price impact and often include built-in MEV protection.
Frequently asked questions
What is MEV in simple terms?
MEV (Maximal Extractable Value) is profit that someone earns by controlling the order in which blockchain transactions are processed. When you submit a trade on a decentralized exchange, it enters a public waiting area where bots can see it before execution. These bots rearrange or insert their own transactions around yours to extract profit at your expense, most commonly through sandwich attacks that buy before you and sell after you, taking a cut of your trade within your slippage tolerance.
How much does MEV cost the average DEX trader?
For an unprotected trader making regular swaps above $1,000 on Ethereum mainnet, MEV extraction typically adds 0.3–1.5% in hidden costs per affected trade. Not every trade gets sandwiched, and reliable public data on exactly how often is limited, but larger trades (above $5,000) are the most attractive targets. According to EigenPhi data reported by Cointelegraph Research in December 2025, sandwich attacks drained approximately $40 million from Ethereum traders across 2025, even as monthly extraction rates fell sharply from nearly $10 million in late 2024 to around $2.5 million by October 2025.
Is Flashbots Protect safe to use?
Flashbots Protect is open source and widely adopted. As of October 2024, 2.1 million Ethereum accounts had used it since launch, protecting $43 billion in DEX volume. It does not require you to grant any special permissions or approvals. The service simply routes your transaction through a private channel to block builders instead of broadcasting it publicly, so your transaction either executes as submitted or is not included at all. You never pay gas for failed transactions. The primary trade-off is slightly slower inclusion times (1–2 additional blocks in some cases) because your transaction only reaches builders connected to the Flashbots relay.
Can MEV affect my trades on a centralized exchange?
Traditional MEV (sandwich attacks, front-running via mempool observation) does not exist on centralized exchanges because there is no public mempool. Your orders go directly to the matching engine in a closed system. However, centralized exchanges have analogous issues: high-frequency traders with speed advantages can react to your orders faster than you can modify them, and exchange market makers see aggregate flow data. These are different mechanisms from on-chain MEV but can produce similar outcomes of worse execution for retail traders.
Do I need MEV protection on Layer 2 networks?
Layer 2 MEV protection is becoming increasingly important. While L2s like Arbitrum and Base currently have centralized sequencers that process transactions in arrival order (reducing sandwich opportunities), MEV extraction infrastructure is developing on these networks. As L2s decentralize their sequencers, the MEV landscape will increasingly mirror Ethereum mainnet. Using private transaction routing on L2s where available provides proactive protection against an evolving threat.
Researched and written by the BloFin Academy editorial team with AI-assisted drafting. Primary sources include Flashbots Protect documentation for private transaction routing (Flashbots, https://docs.flashbots.net/flashbots-protect/overview); CoW Protocol MEV protection documentation for batch auction mechanics (CoW, https://docs.cow.fi/cow-protocol/concepts/benefits/mev-protection); EigenPhi MEV data for extraction statistics as reported by Cointelegraph Research (December 2025); Jito Labs for BAM launch details (July 2025); CoinDesk for Jito mempool shutdown reporting (March 2024). All facts independently verified against cited sources current as of July 2026.
This article is for informational purposes only and does not constitute financial advice. Cryptocurrency trading involves substantial risk of loss. Past performance does not guarantee future results. Always conduct your own research and consider your financial situation before trading. BloFin does not guarantee the accuracy of third-party data referenced herein.