How to Calculate FX Cross Rates: Bid, Ask and False Arbitrage Signals

A cross rate is easy to calculate at mid-market prices—and easy to misread as a trading opportunity. The useful question is not whether two calculator results differ. It is how much currency you would actually receive on one route, how much you would pay on the other, and whether both routes can be executed for the same amount and settlement date. For executable cross rates, the bid–ask sides matter as much as the multiplication or division.

Start with what the quote is buying

EUR/USD means dollars per euro. The euro is the base currency; buying this pair means buying euros with dollars. At a dealer quote of 1.1000/1.1002, you sell euros at the dealer's 1.1000 bid and buy euros at its 1.1002 ask. Calling the lower number your buying rate reverses the trade and can manufacture a profit that does not exist.

All rates below are hypothetical, not current market observations. Assume firm quotes available to the same participant at the same instant, sufficient size, matching spot settlement dates and no additional commissions initially. These assumptions isolate the arithmetic; removing them can change the executable result.

EUR/JPY: follow the cash, then multiply

To sell €10,000 for yen through dollars, first sell euros at 1.1000 and receive $11,000. Then sell those dollars at 150.00 and receive ¥1,650,000. The synthetic EUR/JPY bid is therefore 1.1000 × 150.00 = 165.000000.

The reverse task uses both asks. Buying €10,000 costs $11,002 at 1.1002. Buying those dollars with yen costs 11,002 × 150.04 = ¥1,650,740.08. The synthetic ask is 1.1002 × 150.04 = 165.074008. Its extra decimals are calculation precision, not a claim that a trading venue accepts that tick size.

The ¥740.08 difference is the two spreads expressed in yen for this amount, before any separately charged fees. Multiplying midpoints would hide it. The unit check is equally useful: USD per EUR multiplied by JPY per USD leaves JPY per EUR. If your units do not cancel, the quote direction is wrong.

EUR/GBP: division changes which side you need

EUR/USD and GBP/USD both quote dollars per unit of the base currency. To turn euros into pounds, sell euros for dollars at the EUR/USD bid, then buy pounds at the GBP/USD ask. The synthetic EUR/GBP bid is 1.1000 ÷ 1.2503 ≈ 0.87978885. To buy euros with pounds, the relevant ask is 1.1002 ÷ 1.2500 = 0.88016.

Thus, for A/C and B/C, the A/B bid is the first bid divided by the second ask; its ask is the first ask divided by the second bid. Do not divide bid by bid merely because the labels match. Writing down the two currency exchanges is more reliable than memorising a rule without its direction.

Inverting a pair also switches sides. USD/EUR is 1/1.1002 ≈ 0.90892565 bid and 1/1.1000 ≈ 0.90909091 ask. Taking reciprocals without swapping the columns produces a bid above the ask—an immediate warning that the calculation, not the market, is crossed.

A better rate is not automatically triangular arbitrage

Now suppose direct EUR/JPY is 164.99/165.08. The dollar route offers 165.000000/165.074008: slightly better on both sides under our assumptions. That may favour one conversion route before additional fees. It does not create a profitable round trip.

Sell €10,000 through dollars to receive ¥1,650,000, then repurchase euros at the direct ask of 165.08. You finish with €9,995.15, about €4.85 less than you started with, even before commissions. Comparing the two bids—or their midpoints—would miss the price of closing the loop.

A potential gross discrepancy requires the direct bid to exceed the synthetic ask, or the synthetic bid to exceed the direct ask. The comparison must use matching amounts and settlement dates. A net opportunity additionally requires the difference to survive all fees and to be executable across every leg; a screenshot of asynchronous quotes proves neither.

Where the spreadsheet stops matching the trade

Top-of-book prices may cover only a small amount. A larger order can sweep several price levels, so the correct input becomes the amount-weighted execution price, not the best displayed quote. An indicative feed, a delayed quote or a price restricted to a different credit relationship is not interchangeable with a firm price available to your account.

There is also leg risk. If the euro sale fills but the dollar-to-yen leg does not, the account holds dollars, not the intended yen exposure. A limit price controls an acceptable execution price, not the certainty of completion; FastBull's guide to limit orders and partial fills explains that distinction. Different currency holidays or value dates can add funding requirements, while futures and cash-settled derivatives cannot simply be substituted for deliverable spot conversions.

A practical cross-rate audit

1. Write the units beside every quote and state which currency you own and which you need.
2. Capture both sides, timestamps, available sizes and value dates from prices accessible to the same account.
3. Calculate each leg as an actual purchase or sale. Keep intermediate precision; apply the relevant currency and execution rounding rules at the end.
4. Convert commissions, minimum fees and any financing costs into one reporting currency. Do not add pips directly to cash amounts.
5. Test the return leg at its executable opposite side, then consider whether all required quantities can fill without leaving an unwanted balance.
6. Reconcile the estimate to confirmations and final currency balances. If the inputs were midpoints or non-firm indications, label the result theoretical.

The decisive test is a completed currency path, not an attractive-looking cross-rate difference. That distinction turns a formula into a useful tool for checking conversion costs—and prevents a pricing identity from being mistaken for a trading edge.