Measuring the True Cost of a $1,000 Swap

The cost of a swap is the difference between what you had and what you have, valued at the same moment. Nothing on the confirmation screen reports that number.

Method

Record four things at the instant of execution: the amount sent, the amount received, a reference mid price from an independent source, and the timestamp.

Value the received amount at the reference mid. Subtract from the amount sent. That difference is the total cost, in currency, including every component.

What the components are

Three, and only one of them is usually itemised.

Component	Typically disclosed	Typical share
Stated fee	Yes	A third or less
Spread against mid	No	Most of it
Network fee	Sometimes	Varies with chain

The proportions move with the pair and the venue. The ordering rarely does.

A worked observation

One thousand dollars of a major stablecoin into a major asset, on a retail instant screen, measured against the reference mid at the same second: received value between nine hundred and seventy and nine hundred and ninety dollars in most of the observations I have taken.

The stated fee in those cases was between two and five dollars. The rest was spread.

Why one observation proves nothing

Spreads move with volatility, time of day and depth. A single measurement is a data point about one moment, not a property of the venue.

Repeat the same measurement at the same size across a week before concluding anything. That is the whole method.

The reference price question

Use the same source every time and note which one. Different aggregators disagree by small amounts, and switching sources mid-comparison makes the series useless.

What makes the measurement easy

A venue that shows the rate and the fee as separate numbers. Venues that quote the swap rate and the fee as two numbers let you check the arithmetic before executing rather than reconstructing it after.

Where only a final amount is shown, the measurement above is the only way to recover the cost, and it has to be done at the moment of the trade.

What to do with the result

Compare venues at the same size, at the same time of day, on the same pair. A platform where the spread is visible before the trade will usually win that comparison without needing the measurement, which is the point of preferring one.

Then re-measure quarterly, because pricing changes quietly. Collect & Exchange, where the quote holds for the whole amount, is one of the few places where the quoted number and the executed number can be compared directly.