Separate jobs from people and interpret a mixed employment release. Compare stop, stop-limit and opposing orders, with a worked example of loss beyond a planned trigger.
Separate two tasks: identify what each US employment statistic measures, and understand how your orders execute when quotes change sharply. Headline payroll growth alone does not bridge the gap between information and an actual fill. “Strong data” does not establish that the dollar must rise.
This guide provides report-reading and order-checking worksheets, not a recommendation to trade the release. You can complete the observation exercise without a position.
The report combines establishment and household surveys. Payroll employment concerns jobs; the household survey supplies measures including unemployment. Multiple jobholding helps explain why job and person counts need not match. Check the BLS methods guide for coverage.
| Item | Record | Do not infer automatically |
|---|---|---|
| Nonfarm payroll change | Period, unit and initial/revised vintage | Every meaning of “newly employed people” |
| Unemployment rate | Rate and survey basis | That a lower rate proves every labour condition improved |
| Participation | Read alongside unemployment | That everyone outside the labour force is unemployed |
| Average hourly earnings | Covered group, monthly/annual basis and revisions | The raise received by an unchanged individual worker |
| Hours and industries | Composition and changes | Identical conditions across all industries |
| Revisions | Original/latest vintage and difference | Information that was already available at the initial release |
Do not subtract unlike series to invent an employment “gap”. If values conflict, first check periods, adjustment, units and versions rather than alleging fabrication.
Suppose initial payroll growth is +180 thousand jobs against one provider's +150 thousand consensus: a +30 thousand difference. Last month is revised from +160 to +120 thousand: −40 thousand. Unemployment changes from 4.0% to 4.1%, while monthly earnings growth is 0.2% versus a provider forecast of 0.3%.
These are teaching inputs, not a real release. Payrolls beat that forecast, history is revised down, unemployment rises and earnings miss that forecast. Keep each observation separate. Simple addition cannot transform the package into a certain trading direction. A 0.1-percentage-point unemployment-rate change is not “0.1% more unemployed people”.
Initial and later vintages are not equally real-time information. The BLS FAQ explains revisions and survey differences. Historical event analysis needs the version released at the time; today's revised series can introduce information unavailable then.
First open the actual BLS schedule and confirm date and local time. Next freeze the forecast source and known previous-value snapshot. List open quantities, pending orders, margin rules and platform clocks. Then establish quote-side, stop and cancellation terms.
Clicking cancel is not proof of cancellation. Requests, acceptance, triggering, fills and cancellation can be separate records. Preserve the complete timeline instead of relying on a later price screenshot.
| Situation | What can happen | Records to check |
|---|---|---|
| Ordinary stop | Triggering leads to execution under applicable rules, potentially at another price | Quote side, trigger time, execution price, quantity and terms |
| Stop-limit | A price limit applies after triggering, but execution may not occur | Limit, validity, partial fills and remainder |
| Opposing breakout orders | A fast reversal can trigger both sides; cancelling the other may be too late | Both IDs, sequence, OCO rules and cancellation acknowledgement |
Opposite directions do not mean zero combined loss. Hedged positions, netting, margin offsets and OCO are implementation-specific. If guaranteed stops are offered, check eligible products, fees, distances and exclusions. Do not treat an ordinary stop as a guaranteed price.
For a fictional EUR10,000 long, entry 1.1000 and ordinary stop trigger 1.0980, assume USD1 per pip. Ignoring fees, the planned trigger distance is 20 pips or USD20. An actual exit at 1.0965 produces a 35-pip fill-to-fill loss, USD35: USD15 beyond that plan.
This is not a forecast of future slippage. It illustrates why planned and realised numbers differ. With a separate USD0.70 round-trip commission, the total is −USD35.70. Actual buy/sell fills already incorporate their execution-price difference; do not mechanically subtract another spread.
Halving quantity reduces the dollar effect of a given move with other inputs unchanged. It does not remove gaps, connection problems, uncertain orders or the possibility of exceeding a budget. Lower notional exposure is not automatically suitable for every person's circumstances.
There is no universal 15-minute safety boundary across products, sessions and platforms. Record spreads against your existing comparable sample, quote continuity, clear request states and margin changes. These observations still do not guarantee safety.
Write observable reasons not to add orders: stale quote timestamps, an unconfirmed cancellation, costs beyond the prior budget, or an unread original release. That is more auditable than “wait for stability”. Completing the checklist does not establish that a trade is worthwhile.
Keep the original release, pre-release forecast snapshot, actuals/revisions, platform times/zones, bid/ask quotes and full requests/fills. Preserve unfilled and rejected requests and analyse them separately. Use the execution-evidence guide and trading journal.
Why can prices move against an apparent beat? A headline does not encompass components, revisions, previous pricing and simultaneous news. Do not invent a single explanation without evidence.
Do opposing pending orders avoid choosing direction? They alter the trigger design, not execution risk. Both may trigger, with slippage and cancellation races.
Is sitting out necessarily a missed opportunity? Observation needs no position. Without a tested rule and tolerable conditions, not participating is a complete decision.