Comparing broker spreads: why the lowest headline number fails in volatile markets

Spread tables lie when volatility is low

The spread on a broker's marketing page is usually the minimum spread, achieved during the London–New York overlap when liquidity is deepest. Take mid-October: Brent crude recovered from 100.22 dollars on 5 October to 106.06 dollars on 14 October, and spot gold briefly lost 4,100 dollars intraday with a daily range in the hundreds of dollars. In that environment spreads and slippage widen together, while the number on the marketing page does not update.

The three metrics worth comparing

The first is the time-weighted average spread, not the minimum. Sample the Asian, European and New York sessions separately and weight them by the hours you actually trade; only then do you approach the real cost.

The second is the peak spread during data releases. At the instant US CPI, payrolls or PPI are published, spreads on the major pairs can widen to three to five times normal. If your strategy includes event trading, this matters more than the average.

The third is the slippage distribution, especially on stop orders. Record the difference between the executed price and the requested price for your last twenty stop fills, then take the median and the maximum. The median describes the normal case; the maximum describes tail risk.

Overnight financing is a cost too

With front-end dollar rates high and the market pricing a December hike, the overnight financing cost of being short dollar and long a non-dollar currency rises noticeably. Holding a position for a week rather than a day can cost more than the spread itself. This is routinely ignored when brokers are compared.

How to run a comparison that means something

Fix one instrument (for example EUR/USD and XAU/USD), fix one time window (for example 20:00 to 23:00 Beijing time) and fix one sample size (for example twenty consecutive trading days), then record the three metrics on that consistent basis. Comparisons with small samples produce unstable conclusions.

Cost is not the only factor that determines profit and loss, but it is the only one that can be determined in advance and that occurs on every single trade.