Centralized exchanges (CEX) match trades through an order book and matching engine, executing in milliseconds with the exchange holding custody. Decentralized exchanges (DEX) execute on-chain through either automated market makers (pool-based pricing) or request-for-quote systems (dealer-quoted pricing), with users retaining custody of their private keys. The mechanism you trade through determines your fees, slippage exposure, settlement speed, and vulnerability to MEV extraction. This guide covers how each execution type works, where costs hide, and how to pick the right venue for your trade size and situation.
How CEX order books execute your trade
A centralized exchange order book is a real-time ledger of all open buy orders (bids) and sell orders (asks), matched by a matching engine using price-time priority. The buyer offering the highest price and the seller accepting the lowest price trade first, with ties broken by who posted earlier.
When you place a market buy for 70 BTC on a book showing 40 BTC offered at $60,100 and another 60 BTC at $60,200, the matching engine fills your first 40 at $60,100 and the remaining 30 at $60,200. Your average execution price lands at approximately $60,143, which is $43 above the best ask. That $43 is depth cost, one of the hidden fees invisible in the stated fee schedule.
The spread (gap between best bid and best ask) represents immediate crossing cost. On BTC/USDT during normal trading hours on major exchanges, spreads typically sit at 0.01-0.02%. During the May 2021 volatility spike, BTC spreads on some venues widened past 1%, multiplying execution costs tenfold for market orders.
Maker vs taker fees. Posting a limit order that rests on the book makes you a maker (typically 0.00-0.02% fee). Crossing the spread with a market order makes you a taker (typically 0.04-0.10% fee). The fee difference is the exchange paying you to add liquidity.
Why limit orders sit unfilled. If you place a limit buy at $60,000 and 50 BTC of existing bids already rest at that price, you are behind them in the queue. Price must reach $60,000 and those 50 BTC must fill before your order executes. In thin markets, large orders may fill only 70% at one level, leaving you with an unintended partial position and an open order that may never complete.
How AMMs price your swap
An automated market maker replaces the order book with a liquidity pool: a smart contract holding reserves of two tokens. The price is determined by the ratio of those reserves, not by other traders' posted orders. When you swap, you trade against the pool directly.
We see a recurring pattern where traders start on our centralized order book for speed and depth, then explore DEX trading once they want specific long-tail pairs unavailable on any single CEX.
Consider a pool holding 100 ETH and 300,000 USDC (implied price: $3,000/ETH). Swapping 1 ETH out costs you roughly 3,009 USDC after the 0.3% fee, with minimal price impact. Swapping 10 ETH shifts the reserve ratio significantly, and your effective price degrades to approximately $2,700 per ETH. The larger your trade relative to pool reserves, the worse your execution.
Three rules of AMM pricing:
Pool size determines impact. A 10 ETH swap in a 100 ETH pool moves price far more than in a 10,000 ETH pool.
Every trade moves the price. Unlike order books with resting levels, AMM price adjusts continuously with each swap.
Impact is calculable before submission. Given pool reserves, you can compute exact output prior to signing the transaction.
Slippage tolerance is your maximum acceptable deviation from the quoted price. Setting 0.3-0.5% for major pairs (ETH/USDC) provides protection without excessive revert risk. Setting it above 3% on liquid pairs invites sandwich attacks where bots front-run your trade, extract value within your tolerance window, and leave you with worse execution.
Gas and timing. Unlike CEX instant fills, AMM swaps require blockchain confirmation. Your transaction enters the public mempool, becomes visible to searchers, waits for block inclusion, and settles only after confirmation. During Ethereum congestion events in 2021, gas exceeded 200 gwei, making a simple swap cost $50+ in gas alone (source: Etherscan) Layer 2 networks now offer sub-$1 gas for the same operations.
Concentrated liquidity variants like Uniswap V3 let liquidity providers focus reserves within specific price ranges, improving capital efficiency for traders (tighter effective spreads) but requiring active management from LPs.
How RFQ systems get you a firm quote
Request-for-quote (RFQ) execution broadcasts your trade intent to a network of professional market makers (solvers) who respond with firm, executable price quotes. You compare quotes, accept the best one within its expiry window (typically under 30 seconds), and settlement occurs on-chain via smart contract.
The RFQ flow:
You submit a request specifying pair, size, and direction.
Multiple solvers compute and return binding quotes.
You accept the best quote before expiry.
On-chain settlement executes the accepted trade.
Where RFQ beats AMMs. For a $500,000 ETH/USDC trade, an AMM route might impose 50+ basis points of price impact across multiple pool hops. An RFQ solver can source liquidity from CEX inventory, OTC desks, and cross-chain venues, often delivering execution at 5-10 basis points better than the equivalent AMM path (source: 1inch). The advantage grows with trade size because AMM impact scales with size while dealer quotes do not.
Where AMMs beat RFQ. AMMs are permissionless and always available. RFQ depends on solver availability, and during extreme volatility, solvers may widen quotes or stop responding. For trades under $10,000 on liquid pairs, AMM execution is typically equivalent or better once you account for RFQ's additional complexity.
RFQ risk. Quote expiry means if you hesitate, the quote disappears. Solver availability varies by time of day and market conditions. Always use protocols with AMM fallback routing so your trade executes even if no solver responds.
The full cost stack across all three venues
Your total execution cost includes the stated fee, the spread or price impact you absorb, gas for on-chain settlement, and any value extracted by MEV bots. The effective cost formula is: (Expected Price minus Executed Price) plus Explicit Fees plus Gas. Most beginners compare only the stated fee and miss the other components, which often exceed the fee itself.
For a $10,000 ETH trade during normal conditions:
CEX order book: 0.10% taker fee = $10 total cost. No gas, no price impact on this size.
AMM (Ethereum L1): 0.30% pool fee + 0.1% price impact + $5 gas = $45 total cost.
AMM (L2): 0.30% pool fee + 0.1% price impact + $0.30 gas = $40 total cost.
RFQ: 0.05% effective spread + $0.30 gas = $5.30 total cost.
For a $200,000 ETH trade:
CEX order book: 0.10% fee + potential depth cost if book is thin = $200-400.
AMM: 0.30% fee + 0.5%+ impact + gas = $1,600+.
RFQ: Solver quote at 0.08% effective spread + gas = $160.
The cost comparison shifts dramatically with size. Below $5,000, venue choice barely matters on liquid pairs. Above $50,000, it can mean hundreds of dollars in execution quality difference.
MEV: The hidden tax on public mempools
Maximal Extractable Value (MEV) is profit extracted by reordering, inserting, or censoring transactions within a block. When you submit a DEX swap, your pending transaction is visible in the public mempool before confirmation, giving bots the opportunity to sandwich your trade. The most common form affecting retail traders exploits your slippage tolerance as an extraction budget.
How a sandwich works:
You submit a swap with 2% slippage tolerance.
A bot spots your pending transaction in the public mempool.
The bot front-runs you (buys the token, pushing price up).
Your trade executes at the elevated price (within your tolerance).
The bot back-runs you (sells at the price you pushed it to).
The bot profits the difference. You get worse execution.
Your slippage tolerance setting is the ceiling for how much a sandwich can extract. Setting 5% tolerance on a $10,000 swap means you have told the network you will accept up to $500 in value extraction.
Practical mitigations:
Keep slippage at 0.5% or below for liquid pairs.
Trade smaller sizes (less profitable for bots to attack).
Use private transaction submission (Flashbots Protect, MEV Blocker) to hide your trade from the public mempool (source: Flashbots).
Use L2 networks where MEV infrastructure is less developed.
Consider RFQ execution for larger trades, which avoids mempool exposure entirely.
Flashbots data shows over $680 million in cumulative MEV extracted on Ethereum mainnet since 2020 (source: Flashbots). This is not theoretical risk.
Choosing the right venue for your trade
No single venue is universally best. The right choice depends on four variables: your trade size relative to available liquidity, the token's listing availability across venues, your urgency of execution, and whether you prioritize the speed of centralized custody or the security of self-custody. Below is a practical decision framework organized by situation.
Use a CEX order book when:
Trading liquid majors (BTC, ETH) on the spot market in any size under $100,000. For a broader overview of how trading works across these venues, see what is crypto trading.
You need instant execution for risk management exits.
You want maker rebates on passive limit orders.
You are comfortable with exchange custody during the trade.
Use an AMM when:
Trading tokens not listed on centralized exchanges.
Your trade size is under $10,000 on liquid pairs.
You require permissionless access or want to maintain self-custody.
You are on an L2 where gas is negligible.
Use RFQ when:
Trade size exceeds $50,000 and AMM impact would be material.
You want price certainty before committing capital.
Pool depth is thin relative to your order size.
You prefer on-chain settlement without mempool exposure.
Three scenario examples:
- $500 ETH buy (liquid pair). CEX limit order: $0.50 fee, instant fill, zero impact. AMM on L2 also viable if funds are already on-chain ($0.30 gas + negligible impact). Venue choice barely matters at this size.
- $50,000 mid-cap altcoin purchase (thin CEX listing). RFQ or DEX aggregator: the AMM pool may hold only $200,000 in reserves, meaning a direct swap causes 25% price impact. Solvers can source across venues, reducing impact to under 1%. Always compare aggregator AMM route versus RFQ quote before executing.
- $200,000 ETH exit during volatility. CEX limit order or RFQ: AMM slippage spikes during volatility, gas surges, and sandwich risk increases. Set a CEX limit at your target price. If the book is thin, split across venues or use a TWAP approach to spread execution over time. For position sizing on large exits, ensure your exit does not exceed 10% of visible depth at the best level.
Pre-trade execution checklist
Before clicking buy or sell on any venue, running through a short verification sequence prevents the most common execution errors: trading into thin liquidity, accepting excessive slippage, or paying gas that exceeds the trade's expected edge. This checklist applies whether you are placing a CEX limit order, an AMM swap, or accepting an RFQ quote.
Check expected price vs mid-market. How far is your quoted execution from the theoretical fair price?
Verify depth or pool size. Is available liquidity at least 10x your trade size?
Set slippage appropriately (DEX). Under 0.5% for majors, under 1% for mid-caps. Never above 3%.
Estimate gas (DEX). Check current network conditions. If gas exceeds 1% of trade value, wait or use L2.
Compare venues. For trades above $10,000, check at least two execution paths before committing.
Verify token contract (DEX). Confirm the address via official project documentation, not search results.
After execution:
Compare expected fill versus actual fill.
Calculate total cost (fee + impact + gas).
Log venue performance in your trading journal for future reference.
Frequently asked questions
What is the simplest difference between CEX and DEX execution?
A CEX matches your order against other traders' orders using a centralized order book and matching engine, with the exchange holding custody during the trade. A DEX executes your swap on-chain against either pool reserves (AMM) or dealer quotes (RFQ), with you maintaining control of your private keys throughout. The trade-off is speed and simplicity on CEX versus self-custody and permissionless access on DEX.
Why might a DEX trade cost more than a CEX trade even without exchange fees?
DEX trades add gas costs for blockchain computation, price impact from shifting pool reserves, and potential MEV extraction from public mempool exposure. A $1,000 AMM swap on Ethereum L1 might cost $20 in gas plus 0.5% in price impact, totaling $25. The same trade on a CEX costs a 0.10% taker fee of $1. On L2 networks gas drops below $1, narrowing the gap significantly for smaller trades.
When should I use RFQ instead of swapping through an AMM?
Use RFQ when your trade size exceeds $50,000 and AMM price impact would degrade execution by more than 0.3%. RFQ solvers source liquidity across multiple venues and offer firm quotes without mempool exposure, which eliminates sandwich attack risk entirely. For trades under $10,000 on liquid pairs, AMMs are typically simpler and equivalent in cost. The break-even point where RFQ consistently outperforms depends on pool depth, but $50,000 is a practical threshold for most major pairs.
How does slippage tolerance protect me on a DEX, and when can it backfire?
Slippage tolerance sets the maximum price deviation you will accept before the transaction automatically reverts and fails. It protects against adverse price movement between submission and execution. It backfires when set too high because MEV bots treat your tolerance as an extraction budget. Setting 5% tolerance on a $10,000 trade tells searchers they can profitably sandwich you for up to $500 in extracted value.
What should I check before placing a large trade on any venue?
Verify that available liquidity (order book depth or pool reserves) is at least ten times your intended trade size. If it is not, your market order will walk through multiple price levels on a CEX or create severe price impact on an AMM. Compare execution quotes across at least two venues. For trades above $100,000, always check whether RFQ provides better pricing than direct AMM routing before committing.
Researched and written by the Blofin Academy editorial team with AI-assisted drafting. Primary sources include Uniswap V3 whitepaper and protocol documentation for AMM mechanics; BloFin exchange order book specifications for CEX execution; 1inch Network RFQ documentation for solver-based execution; Flashbots MEV research for extraction mechanics. All facts independently verified against cited documentation current as of April 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.