Weekend Gaps: What the Data Says
Gaps by instrument
| Instrument | Unit | Median | 95th | Weekend 95th | Largest |
|---|---|---|---|---|---|
| Gold (XAU/USD) | pips (0.1) | 0.27 | 28.14 | 134.54 | 1119.37 |
| GBP/JPY | pips | 0.2 | 32.94 | 64.91 | 314.7 |
| EUR/JPY | pips | 0.1 | 27.2 | 59.4 | 270.1 |
| USD/JPY | pips | 0.1 | 20.6 | 36.66 | 316.5 |
| GBP/USD | pips | 0.1 | 18.43 | 30.9 | 254.1 |
| AUD/CAD | pips | 0.1 | 16.59 | 26.08 | 97.7 |
| EUR/USD | pips | 0.1 | 12.9 | 23.44 | 163.5 |
| AUD/USD | pips | 0.1 | 11.07 | 22.36 | 158.5 |
| USD/CAD | pips | 0.1 | 12.34 | 20.4 | 209.5 |
| EUR/GBP | pips | 0.1 | 11.99 | 16.95 | 71.3 |
The median is not the risk
Half of all gaps are smaller than a spread, which is why gap risk feels theoretical until it is not. The number that matters is the tail: one open in twenty on Gold (XAU/USD) jumps 134.54 or more, and the largest in this sample was 1119.37. A stop placed inside that distance was not honoured at its price.
What a backtest does with a gap
Most replay a stop as filled exactly where it sat. Reality fills at the first available price after the open, which on a gapped Monday can be well beyond it. The difference is not noise — it is the difference between a strategy that survives a bad weekend and one that only survives in the record.
The practical correction is to test what happens when the worst gap in this table lands on an open position, rather than to assume it will not.
FXAbsolute free backtester
Test the finding on the chart—not just on the page
Open FXAbsolute and replay USD/JPY candle by candle. Turn the observation into explicit entry, stop, target and exit rules, then measure the result yourself.
Questions
How big is a typical forex weekend gap?
Small — GBP/USD's median close-to-open gap is 0.1 pips. The risk is in the tail: the 95th percentile weekend gap is 30.9 pips and the largest in the sample was 254.1.
Which instrument gaps the most?
Of those measured here, Gold (XAU/USD), with a 95th-percentile weekend gap of 134.54 pips (0.1).
Do stop losses work over the weekend?
A stop is an instruction to exit at the first available price once the level trades, not a guarantee of that price. When the market reopens beyond it, the fill is beyond it too. That is exactly what most backtests do not model.
How were these gaps measured?
As the absolute difference between one trading day's closing price and the next trading day's opening price, across 1,594 such transitions per instrument.