At 8:29 on a CPI morning, EURUSD's spread is 0.6 pips and the order book is thick. At 8:30:01 the spread can be 8 pips, the book is a ghost town, and price can travel 40 pips in under a second. Any position you hold through that second is no longer the trade you planned. The stop you placed 25 pips away is not a promise, it's a request, and requests get filled at whatever price exists when liquidity comes back.
What actually happens to your stop
A stop loss is a market order that triggers at a price. In normal conditions on a major pair it fills within a fraction of a pip of the trigger. During a top-tier release it can fill 5, 10, or 20 pips through the level, because the prices between trigger and fill simply never traded. Concrete version: $10,000 account, 0.40 lots of EURUSD, stop 25 pips away, planned risk $100. Nonfarm payrolls print hot, price gaps through your level and the stop fills 12 pips late. The loss is 37 pips × $4 per pip, $148. You planned 1R and paid 1.5R, and execution did nothing wrong. The mistake was being there, at size, at 8:30.
The pre-news decision tree
This risk is unusually polite: it publishes its schedule. CPI and nonfarm payrolls hit at 8:30am New York time, FOMC statements at 2:00pm, central bank rate decisions on dated mornings. Marking the day's red-flag times takes two minutes in the pre-session routine. Then, ten minutes before any top-tier release touching your instrument, run one decision in order of preference:
- Flatten (the default): if the trade is under plus 1R, or the release directly prices your instrument (payrolls for dollar pairs, inventories for oil, FOMC for nearly everything), take the profit or the small loss now. A scratch beats a coin flip with bad odds and slippage attached.
- Reduce: if the trade is well in profit with a stop behind real structure, cutting to half size halves the slippage exposure while keeping a stake in a favorable resolution.
- Hold, rarely: only when the position is at least plus 2R, the stop sits beyond the pre-news structure, and you accept in advance that it may fill a full R worse than placed.
Notice what is not in the tree: your opinion about the number. Guessing the print and positioning for it is a different and much harder game. This checklist is about positions you already hold, and it works precisely because it ignores forecasts.
The minutes after the release
Spreads on majors usually normalize within one to five minutes, and the first violent print frequently reverses once the algorithms are done repricing. That argues for patience twice over. If you flattened, you lost nothing but a spread: re-entry on fresh structure five minutes later is always available, now with information instead of exposure. And never widen a stop 'to survive the news'. That choice combines the worst of every branch: full slippage exposure, extra planned risk, and a decision made for emotional reasons at the exact moment decisions are worst. If the stop is in danger of being hit on the release, the honest conclusion is that the position should not be open.