What Is Ethereum Arbitrage? A Plain-English Definition

Ethereum arbitrage is the practice of buying an asset on one venue and selling it on another at almost the same time to keep a small price difference. On Ethereum the venues are usually decentralised exchanges and liquidity pools, and the profit is whatever remains after gas and trading fees.

The core idea in one example

Imagine a token trades at 1.000 ETH on one pool and 1.010 ETH on another. If you could buy at the first and sell at the second instantly, you would earn about 0.010 ETH per token, before costs. That gap exists because each pool prices assets from its own reserves, and the pools do not talk to each other. Someone has to trade between them to bring prices back into line.

That someone is the arbitrageur. Their trades push the cheap pool up and the expensive pool down until the gap is too small to be worth the fees. In that sense arbitrage is a service: it keeps prices consistent across venues, and the profit is the payment for doing it quickly.

Key takeaway: arbitrage does not require a view on whether ETH will go up or down. It requires a gap that is larger than your total cost of closing it.

Why gaps appear on Ethereum

Automated market makers set prices by formula. When a large trade hits one pool, its price moves, and until someone corrects the others it differs from the rest of the market. Gaps also appear when new liquidity arrives, when a token is listed on a new venue, or when prices on centralised exchanges move first and on-chain pools follow a few seconds later.

Ethereum adds a twist: transactions wait in a queue and are included in blocks roughly every twelve seconds. Many searchers compete for the same opportunity, so the gap that existed when you saw it may already be taken by the time your transaction lands. The background is described in the ethereum.org overview of MEV, which treats arbitrage as one common form of value that block-building participants can capture.

The main types

Type	What you do	Typical difficulty
Cross-DEX arbitrage	Buy on one pool, sell on another	Moderate, widely competed
Triangular arbitrage	Cycle through three assets to end with more of the first	Moderate, see triangular arbitrage
CEX-DEX arbitrage	Trade between an exchange order book and an on-chain pool	Higher, needs funds on both sides, see CEX versus DEX arbitrage
Flash loan arbitrage	Borrow, trade and repay inside one transaction	Technical, see flash loan arbitrage explained

The vendor CryptoMEV Bot makes a similar split in its own educational material on crypto arbitrage trading, which describes cross-exchange, triangular and latency arbitrage and names market volatility, fees and slippage as the main hazards.

What it costs to capture a gap

Every arbitrage trade pays several costs, and each one eats into the gap:

1. Gas. Each swap consumes gas, and a two-leg trade uses more than a single swap. Gas is paid whether or not the trade is profitable, as long as the transaction is included.
2. Pool fees. Uniswap V2 pools charge 0.3 percent per swap, and Uniswap V3 offers several fee tiers. Two legs means two fees.
3. Slippage. Your own trade moves the price. Larger size means a worse average price.
4. Failed attempts. If a competitor lands first, your transaction may revert and you still lose the gas.

The upshot is that a visible 1 percent gap can be close to zero after costs. The article on gas fees and arbitrage profit shows how to run that arithmetic before you trade.

Where a bot fits in

Doing all of this by hand is impractical. A bot reads prices continuously, simulates the full path, checks that profit exceeds gas, and sends the transaction fast. CryptoMEV Bot’s Ethereum arbitrage page describes exactly that scope for Uniswap, Sushiswap, Curve and Balancer: scanning, simulation, gas-aware profit checks and private submission through Flashbots. Whether a given bot earns money is a different question, and the vendor page does not publish results, so the claim stays unverified.

If you are evaluating one, the sceptical questions are in how to choose an arbitrage bot and the risks of running one.

A single trade, step by step

It helps to walk through one cross-DEX opportunity from start to finish, because each step hides a cost or a risk.

1. Detect. The bot reads reserves from two pools holding the same pair and computes the implied price in each. A difference appears.
2. Size. The best trade size is not “as much as possible”. Because each pool’s price moves as you trade, profit rises with size at first and then falls. The bot looks for the size that maximises profit after fees.
3. Simulate. Before sending anything, the bot runs the whole path against the current chain state. If the simulated output is lower than the input plus gas, it stops.
4. Submit. The transaction is sent with a gas price that is competitive but not so high that it consumes the margin. Many searchers send privately rather than through the public mempool, a topic covered in Flashbots Protect and arbitrage.
5. Settle. If the transaction is included first and the prices have not moved, you keep the difference. If not, it reverts, or someone else took the gap, and you may have paid gas for nothing.

Seen this way, the profitable part is small and the cost-control part is large. That is why serious bots spend most of their logic on simulation and refusing bad trades. The practical question, covered in whether Ethereum arbitrage is still profitable, is how many gaps survive that filter.

Common misconceptions

• “Arbitrage is guaranteed.” Only the textbook version, with zero latency and zero fees, is risk-free. Real arbitrage has execution risk, smart-contract risk and competition.
• “A bot finds gaps nobody else sees.” Public price feeds and pool reserves are visible to everyone. Edge comes from speed, routing quality, private submission and cost control, not from secret information.
• “Bigger gaps are better.” A very large gap often signals a problem: a thin pool, a token with transfer taxes, a broken oracle or a honeypot contract. Verify before trusting an outlier.
• “Capital is the barrier.” Capital helps, but flash loans reduce the need for it. Technical skill and infrastructure are usually the harder barriers.
• “It is the same as MEV.” Arbitrage is one kind of MEV, but not the only one. The MEV bot versus arbitrage bot article draws the line.

For vendors, the equivalent misconception is that a feature list proves profitability. CryptoMEV Bot’s disclaimer states plainly that performance results are hypothetical and that users are responsible for monitoring their own trading, and that is the right frame for any tool you evaluate.

Who it suits, and who should skip it

Ethereum arbitrage suits people who are comfortable with wallets, smart-contract risk and monitoring software, and who accept that the learning curve costs money. It is a poor fit for anyone hoping for steady passive income. Competition from professional searchers is intense, and the profitable gaps are the ones that other bots have not already taken.

Summary

Ethereum arbitrage is trading price differences between venues, with profit equal to the gap minus gas, fees and slippage. It is conceptually simple and operationally hard. Learn the mechanics first, size the costs honestly, and treat any promise of guaranteed returns as a warning sign.

This is general information, not financial advice. Check current protocol fees and your local rules before trading.