Crypto Arbitrage Strategies: Types, Costs, and Execution Risks

What Is Crypto Arbitrage?

Crypto arbitrage is profiting from a temporary price difference for the same or economically equivalent asset across two or more markets or instruments, by simultaneously buying where it's cheaper and selling where it's more expensive (or otherwise structuring offsetting positions so the price difference is captured rather than left exposed to further movement).

The idea sounds close to risk-free: buy low here, sell high there, done. In practice, every real arbitrage trade carries execution risk, cost, and delay between the moment the price gap is observed and the moment it's actually captured — and several distinct strategies fall under the "arbitrage" label, each with a different mechanism and a different set of costs working against it.

Crypto markets are more prone to these price gaps than highly liquid, centralized markets like large-cap equities, because trading activity, liquidity, and price discovery are split across many separate exchanges rather than concentrated through a small number of tightly interconnected venues. That fragmentation is exactly what creates a supply of small, temporary discrepancies for arbitrage strategies to target — and exactly why those discrepancies tend to stay small and close quickly once enough participants are watching for them.

Cross-Exchange Arbitrage

Cross-exchange arbitrage is the most direct form: the same asset trades at different prices on two different exchanges at the same moment, and the trade buys on the cheaper venue and sells on the more expensive one.

Hypothetical example — for education only.

Bitcoin trades at $60,000 on Exchange A and $60,300 on Exchange B. The visible spread is $60,300 − $60,000 = $300, or $300 ÷ $60,000 = 0.5% of the purchase price. If total fees and slippage on both legs of the trade come to 0.35%, the potential margin left over is 0.5% − 0.35% = 0.15% — on a $60,000 trade, a gross gap of $300 against roughly $210 of costs (0.35% × $60,000), leaving about $90 (0.15% × $60,000) of potential profit before anything else goes wrong.

That 0.15% figure already assumes the asset can actually be moved from Exchange A to Exchange B — or that both exchanges already hold enough of the trader's capital that no transfer is needed at all. A transfer that takes minutes or hours gives the price gap time to close, narrow, or reverse before the second leg executes, which is exactly the risk covered later in this guide.

Triangular Arbitrage

Triangular arbitrage trades through three related pairs on the same venue rather than across two venues, when the quoted exchange rates between three assets become temporarily inconsistent with each other — for example, converting USDT to BTC, BTC to ETH, and ETH back to USDT. It's only profitable if the final amount of the starting asset exceeds the amount the trader began with, after every fee across all three legs is subtracted.

Hypothetical example — for education only.

Suppose the three quoted rates are: 1 BTC = $50,000 (USDT → BTC), 1 BTC = 25 ETH (BTC → ETH), and 1 ETH = $2,010 (ETH → USDT). The BTC price implied by going through ETH is 25 × $2,010 = $50,250 — $250 higher than the direct $50,000 quote, a mismatch that creates the opportunity: convert on the cheap direct route, then back through the richer ETH route.

Starting with $50,000: converting to BTC gives $50,000 ÷ $50,000 = 1 BTC; converting to ETH gives 1 × 25 = 25 ETH; converting back to USDT gives 25 × $2,010 = $50,250 — a gross profit of $250, or 0.5%, before any costs.

Now apply a 0.1% fee on each of the three legs. After the first conversion: 1 × (1 − 0.001) = 0.999 BTC. After the second: 0.999 × 25 × (1 − 0.001) = 24.950025 ETH. After the third: 24.950025 × $2,010 × (1 − 0.001) ≈ $50,099.40. Net profit is $50,099.40 − $50,000 = $99.40, or about 0.2% — roughly 0.3 percentage points of the original 0.5% gross gap consumed by three small fees, leaving a thin but positive margin in this example.

That result is sensitive to the fee assumption: a 0.2% fee per leg instead of 0.1% would push the three-leg cost close to the entire 0.5% gross gap, erasing most or all of the profit in this example. Triangular arbitrage margins are frequently this thin, which is part of why it's typically automated rather than executed by hand.

Funding-Rate Arbitrage

Funding-rate arbitrage holds a spot position and an offsetting perpetual-futures position at the same time — long one side, short the other, in equal size — so that a price move gains on one leg and loses a similar amount on the other, largely canceling directional exposure. What's left is the periodic funding payment exchanged between long and short perpetual holders, designed to keep a perpetual future's price tethered to the underlying spot price.

Hypothetical, illustrative figure — not a real historical rate.

If a perpetual future's funding rate is a hypothetical 0.01% charged every eight hours — three times a day — a trader long the spot asset and short an equal-sized perpetual position, positioned to receive that payment, would collect roughly 0.03% of the position's value per day from funding alone, before exchange fees, borrowing costs on either leg, and the risk that the funding rate changes or reverses sign entirely.

The directional cancellation is rarely perfect: the two legs can sit on different venues with different liquidation mechanics, margin requirements, and funding-payment schedules, and a large, fast price move can still create a mismatch between the two legs even when the position is nominally "hedged."

Cash-and-Carry Trading

Cash-and-carry trading buys the spot asset and simultaneously shorts a futures contract on the same asset that's trading at a premium, holding both positions until the futures contract expires. At expiry, a futures price converges to the spot price by construction, so the gap between the two prices at the time the trade was opened is captured as the futures position gains while the spot position's value tracks the underlying asset.

The premium being captured has to cover the cost of holding the spot position for that period — custody, any borrowing cost if the spot leg is financed, and trading fees on both legs — plus the risk that the futures exchange itself fails or restricts withdrawals before expiry. A wide premium that looks attractive gross can compress substantially once carrying costs for the full holding period are subtracted.

A futures contract trading above the spot price, as in this example, is described as being in contango. The reverse — futures trading below spot — is backwardation, and it flips the trade structure: instead of buying spot and shorting futures, a trader would short the spot asset and buy the discounted futures contract, a structure sometimes called a reverse cash-and-carry.

Statistical Arbitrage

Statistical arbitrage trades a temporary divergence in the pricing relationship between two or more correlated assets — for example, two tokens that have historically moved together — using a quantitative model to flag when that relationship has stretched further than its typical range and to size a position betting on it narrowing again. A common structure pairs a long position in one asset against a short position in a correlated peer, sized so that a broad market move affects both legs similarly, leaving the trade's profit or loss to depend mainly on the relationship between the two assets narrowing or widening rather than on the market's overall direction.

Unlike cross-exchange or triangular arbitrage, statistical arbitrage isn't locking in a mathematically guaranteed profit; it's a probabilistic bet that a historical relationship reasserts itself, which means it inherits the same core weakness as any mean-reversion approach: the relationship can keep diverging instead of narrowing, and a model built on a historical lookback window offers no guarantee that the correlation it measured still holds once market conditions change. A stretched relationship can be noise, or it can be the sign of something structural — a project-specific development affecting only one of the two assets — that never reverts.

The Full Cost List That Erodes Apparent Arbitrage Profit

Every strategy above shares the same vulnerability: a gross price gap is not a net return until a long list of costs and frictions has been subtracted from it.

• Trading fees on every leg of every trade, on every venue involved.
• Withdrawal fees charged by an exchange to move an asset off the platform.
• Network fees paid to the underlying blockchain to process a transfer.
• Slippage from executing against a limited order book rather than at the last quoted price — see the crypto liquidity and slippage guide for how this compounds across order-book depth, which matters directly for arbitrage cost accounting rather than being a general aside.
• Transfer time for an asset to move from one venue to another.
• Deposit confirmation time — the blockchain confirmations required before a receiving exchange credits the deposit and makes it tradable.
• Funding costs on any perpetual-futures leg of the trade.
• Borrowing costs if any leg of the trade is financed rather than fully funded with owned capital.
• Liquidity constraints that widen slippage or cap how much size can be executed at the quoted price.
• Withdrawal limits that can cap how much capital can actually move within the window the opportunity exists.
• Exchange risk — the platform holding funds mid-arbitrage could restrict withdrawals, freeze accounts, or fail outright; see the crypto exchange and custody risk guide for the full range of platform failure modes, which apply directly to the cross-exchange counterparty exposure in this strategy rather than being a generic risk footnote.
• Tax consequences — each leg of an arbitrage trade can be a separate taxable event depending on jurisdiction, which is a cost of doing the trade even when the trade itself is profitable before tax.

None of these costs is exotic on its own. The mistake is treating the visible spread as if it were already net of all of them, when in most real arbitrage attempts it isn't close.

Why "The Spread You See" Is Never "The Return You Get"

A quoted price gap is a snapshot of two prices at one instant. The return an arbitrage trade actually realizes depends on everything that happens between that snapshot and the moment both legs are fully closed out — and for a cross-exchange trade requiring an actual transfer of the asset, that window can be the difference between a real profit and a real loss.

A trader who buys on the cheaper exchange, initiates a transfer, and waits for network confirmations and exchange crediting before selling on the more expensive venue is exposed to the price on both exchanges moving during that entire window — not just at the moment the opportunity was first spotted. A gap that looked like 0.5% when observed can be 0.1%, zero, or negative by the time the second leg actually executes, and the capital is not available for anything else while it's in transit.

Who Does Arbitrage Suit — and Who Should Avoid It?

Beginners in particular tend to underestimate the cost list above, overestimate execution speed, and underestimate how much capital sits immobilized mid-transfer, unable to be redeployed, while an opportunity is being chased.

Common Mistakes

• Treating the visible spread as the expected return. Fees, slippage, and transfer time stand between the two.
• Underestimating transfer and confirmation time. A spread that exists at the moment of observation can close entirely before a cross-exchange transfer settles.
• Ignoring smaller fee categories. Withdrawal fees, network fees, and funding or borrowing costs are individually small but add up across every leg of the trade.
• Assuming a hedged position is fully hedged. Funding-rate and cash-and-carry positions can still carry basis, liquidation, and venue-mismatch risk despite looking directionally neutral on paper.
• Concentrating capital on one exchange to chase a spread. That capital is now exposed to a single platform's exchange risk for the full duration of the trade.
• Sizing a triangular or cross-exchange trade without confirming the fee schedule on every leg first. A single higher-than-assumed fee can turn a thin positive margin into a loss.
• Overlooking tax treatment on each leg. Some jurisdictions treat crypto-to-crypto conversions — such as the BTC-to-ETH leg of a triangular trade — as a taxable disposal, which is easy to miss when a trader models only the trading-fee cost and not the tax cost.

Crypto Arbitrage FAQs

Related Guides

• Crypto trading strategies — the pillar guide this page is part of.
• Range trading and grid strategies — a different way to profit from price oscillation, without needing two markets to disagree.
• DCA and position trading — a longer-horizon approach with no dependence on a temporary price gap.
• Swing trading — capturing multi-day directional moves rather than pricing discrepancies.
• Day trading vs. scalping — other fast-execution approaches with their own cost sensitivities.
• Trend following and momentum — a directional alternative to arbitrage's market-neutral approach.
• Backtesting a crypto strategy — testing whether an arbitrage or statistical-arbitrage model's edge is credible.
• Crypto exchange and custody risk — the counterparty risk behind holding capital on multiple venues.
• Crypto liquidity and slippage — how order-book depth erodes execution on both legs of an arbitrage trade.
• Crypto order types — the order mechanics used to execute each leg of an arbitrage trade.
• Crypto position-size calculator — sizing any directional residual left after a hedge.