Arbitrage Trading: Profit from Price Gaps
Arbitrage seeks to capture a price difference between equivalent assets or linked transactions. A quoted difference is only a starting point: the relevant prices must be executable for your quantity, and the proceeds must exceed all costs. Delays, incomplete executions, and restrictions can turn an apparent opportunity into a loss.
Use LuxAlgo’s native charts and Quant, our coding agent, to investigate price relationships and explicit research rules. Confirm actual opportunities using the appropriate execution venues and their live bid/ask data. A historical chart comparison is not proof that both sides of an arbitrage could have traded.
What makes a price difference tradable?
Compare the price at which you can buy with the price at which you can sell. That generally means an available ask on the purchase side and an available bid on the sale side, not two last-traded prices.
- Asset identity: check currency, contract specifications, share class, token network, and conversion ratios.
- Timing: compare contemporaneous quotes and account for delivery delays.
- Quantity: verify depth for the intended size rather than assuming the best quote covers the whole order.
- Access: confirm account eligibility, funding, borrowing, trading permissions, and withdrawal restrictions.
- Total cost: include fees, adverse execution, financing, conversion, and inventory rebalancing.
A similarly named stock listing or wrapped token may not be immediately interchangeable with another instrument. Persistent differences can reflect genuine conversion costs, settlement frictions, or restrictions rather than overlooked profit.
A cross-venue example after costs
Assume 100 identical units are available to buy at $100.00 on venue A and sell at $100.50 on venue B. Assume sufficient funded access to both transactions, including inventory or an approved borrowing arrangement on the sale side.
| Component | Illustrative amount |
|---|---|
| Purchase: 100 × $100.00 | $10,000.00 |
| Sale: 100 × $100.50 | $10,050.00 |
| Gross difference | $50.00 |
| Purchase fee at an assumed 0.1% | −$10.00 |
| Sale fee at an assumed 0.1% | −$10.05 |
| Adverse-execution allowance | −$10.00 |
| Inventory-rebalancing allowance | −$5.00 |
| Estimated net result | $14.95 |
These fees and allowances are hypothetical, not a standard market tariff. If the available sale price falls to $100.20 before execution, the gross difference becomes $20.00. With $20.02 in the assumed trading fees and the same $15 allowances, the estimate becomes a $15.02 loss.
Report return against the capital actually committed across the arrangement, including inventory and collateral. Dividing the profit by only one small margin deposit can conceal the strategy’s exposure.
Common arbitrage approaches
Cross-exchange trading
Buying on one venue and selling on another requires both legs to complete under acceptable conditions. Pre-positioned funds or inventory can reduce dependence on an immediate transfer, but leave capital exposed to the venues and create later rebalancing needs.
Buying first and waiting for a transfer before selling exposes you to price changes during that interval. Kraken’s arbitrage guide describes restrictions, wallet maintenance, liquidity, and execution delays that can prevent an observed difference from being captured. A high displayed premium may be a warning about inaccessible funds or difficult withdrawals.
Triangular currency arbitrage
A triangular sequence converts through three currencies and ends in the starting currency. State every rate’s units and use the executable side of each conversion.
$1,000 × 0.90 EUR per USD = €900
€900 × 0.86 GBP per EUR = £774
£774 × 1.30 USD per GBP = $1,006.20
The gross difference is $6.20. If each conversion deducts a hypothetical 0.1% fee from its output, the final amount is $1,006.20 × 0.999³, approximately $1,003.18—a $3.18 gain before other costs. Depth, rounding, and adverse execution can remove it.
Reversing the route requires the relevant opposite-side quotes; it is not generally valid to invert midpoint rates. Three orders sent to a venue are not automatically one guaranteed all-or-nothing transaction.
Statistical and basis strategies
Statistical arbitrage trades an estimated relationship among assets. The relationship may fail to converge, so this is not the same as locking in an executable price discrepancy. Historical correlation alone does not establish a dependable hedge.
Spot-versus-derivative strategies also involve contract, collateral, and funding assumptions. Coinbase’s perpetual-futures overview explains how funding helps align perpetual and spot prices. Funding rates can change, and price exposure is not necessarily eliminated perfectly throughout a trade.
Execution controls that matter
| Failure mode | Control to define before trading |
|---|---|
| Only one side executes | Maximum unhedged quantity and an explicit hedge or unwind procedure |
| Quote becomes stale | Age limits, refreshed depth, and a rule for rejecting the opportunity |
| Order response is uncertain | Reconcile authoritative order state before resending |
| Price limits prevent completion | Handling for remaining quantity and existing exposure |
| Transfers are delayed | Inventory limits and a realistic rebalancing plan |
| Venue becomes unavailable | Capital exposure limits and tested recovery procedures |
A limit order controls price but may not fill. Sending two orders at nearly the same time does not guarantee paired execution. Splitting orders can reduce immediate size but adds timing and incomplete-execution risk.
An emergency stop may prevent new submissions and request cancellations; it does not undo completed trades or guarantee that all outstanding orders disappear instantly. Check positions and confirmed executions before assuming the system is flat.
Crypto and decentralized markets
For centralized exchanges, inspect fees, supported networks, minimum order sizes, account restrictions, and deposit-crediting times. For decentralized venues, include network fees, liquidity-dependent price impact, smart-contract behavior, and transaction ordering.
A transaction that groups several on-chain operations can have different execution guarantees from separate transactions across venues. Do not assume that a decentralized transaction can make a centralized-exchange order atomic with it. Bridges and wrapped assets introduce additional assumptions about conversion and access.
Specialist flash-loan strategies require protocol-specific repayment and transaction rules. They are not a substitute for capital planning or code review, and a reverted transaction can still incur network costs. Research the exact protocol before treating a theoretical route as executable.
Using LuxAlgo for research
LuxAlgo’s native charts let you review selected markets in one workspace. Confirm each symbol’s exchange, currency, and timeframe. Candle closes from different feeds are useful research observations, but do not establish simultaneous executable bids and asks.
Consult the data documentation when interpreting coverage and aggregation. Executed-volume tools are not a replacement for a venue’s current order book.
Ask Quant to express a research rule with explicit units, timing, and costs. Review the generated code and run it on supported historical data. Use strategy settings and trade logs to inspect what the simulation actually assumes.
A chart strategy does not automatically reproduce multi-venue funding, paired executions, transfer delays, or historical borrowing availability. Quant’s code generation and simulation should not be described as a guaranteed arbitrage scanner or live execution service.
Video: arbitrage basics
The following introduction explains the basic price-difference idea. Apply the cost, access, and execution qualifications above when evaluating any example.
A practical research checklist
Identify the exact instruments, capture timestamped executable quotes, calculate the full cost for the intended quantity, and define what happens if either leg fails. Test assumptions against actual market access and keep a record of rejected opportunities. The aim is to distinguish a displayed price difference from a transaction that can plausibly complete at a positive net result.
Read next