NFP breakout after the jobs report: a 32-release test on real FX quotes

“NFP” commonly means the monthly US Employment Situation release. The payroll number comes from a survey of employers. Unemployment comes from a different household survey. Wages and revisions arrive in the same release; the first market response is not a clean vote on one headline number.

The Bureau of Labor Statistics normally publishes at 08:30 New York time. We matched its actual 32 release dates from 5 January 2024 to 4 September 2026 with 15-minute EUR/USD and USD/JPY bid candles. The dates include Thursday, Tuesday and Wednesday releases. The October 2025 report was never published after the US government funding lapse; inventing a first-Friday event there would contaminate this test.

We are testing a specific delayed breakout. We do not claim that buying or selling exactly at 08:30 can be reconstructed from 15-minute candles.

Take the high and low of the 08:30–08:45 ET candle as a fixed opening range. Starting with the 08:45 candle, wait for the first 15-minute close strictly above its high or below its low. Accept such a close no later than 10:15 ET. No close outside means no trade.

Enter at the next candle’s open in the breakout direction. Put the stop at the opposite edge of the original range and the target at a distance equal to the entry-to-stop distance (1R). Close a still-open position at the 12:00 ET boundary. When a 15-minute candle touches both stop and target, assume the stop happened first. Take at most one entry per pair and publication.

Compute gross R from the bid-price candles, then deduct a hypothetical 1.2 pips for the full round trip. Also recalculate at 3 pips. These are comparison assumptions, not observed NFP spreads or guaranteed executable prices. There is no leverage, compounding or position-size optimisation in the result.

The first EUR/USD candle averaged 54.0 pips across 31 usable reports; USD/JPY averaged 97.3 pips across 32. At the same clock time one week before each release, the comparison averages were 14.2 pips (30 days) and 23.2 pips (31 days), respectively. This is a descriptive reference point, not a causal experiment: other news could occur in the comparison week.

A larger first candle also makes its high and low harder to exceed and pushes the opposite-edge stop further away. On EUR/USD, just 16 of 31 usable events triggered our closing-price breakout. USD/JPY triggered 22 of 32. Market excitement at 08:30 therefore does not automatically translate into a later trade under this rule.

After the assumed 1.2-pip allowance, EUR/USD had 16 entries, 7 positive results and a sum of −1.763R. USD/JPY had 22 entries, 8 positive results and −2.236R. The target was reached twice in EUR/USD and once in USD/JPY; most positions were still open at the midday exit. These are R units added across hypothetical independent trades, not a percentage return on an account.

Under the separate 3-pip sensitivity, EUR/USD fell to −2.314R and USD/JPY to −2.801R. This model does not show a robust advantage. The sample is small, the two pairs react to the same announcements, and changing the range, entry cut-off, exit or target after seeing the data would create a new hypothesis rather than validate this one.

On 5 June 2026, both series used the same BLS publication at 08:30 ET. EUR/USD broke down and our next-candle entry was 1.15864; its opposite-edge stop was 1.16352. The 1R target was reached: +0.975R after the 1.2-pip allowance.

USD/JPY broke upward that day. Its next-candle entry was 160.216 and its stop 159.885. The stop was touched: −1.036R after the same allowance. A single economic report produced a modelled gain in one pair and a loss in the other. Cherry-picking either chart would hide that difference.

On 16 December 2025, when the jobs report followed a disrupted government calendar, EUR/USD also signalled a long and hit its stop: −1.032R after the allowance. The time conversions come from the official New York release clock; the report was not a normal first-Friday entry.

The yearly view does not rescue a universal rule. EUR/USD totalled +0.729R in 2024, −2.028R in 2025 and −0.464R in 2026 through September. USD/JPY totalled +0.141R, +0.312R and −2.689R in those periods. These subdivisions contain only three to eight trades apiece and should not be promoted as reliable year-specific regimes.

A no-entry release remains in the denominator of available days when we report signal frequency, but contributes no R. EUR/USD has no 15-minute candles for 6 December 2024 in the saved export, so that date is marked missing instead of interpolated. USD/JPY retains all 32 report dates.

The quote archive contains BID OHLC candles only. It does not show contemporaneous ask, market depth, spread, queue position, broker rejection or actual order fills. A stop and a target touched in the same 15-minute candle have no observable order; our stop-first rule chooses the adverse outcome. Gaps through a stop, fast slippage and variable NFP spreads can make a real result worse than either fixed allowance.

The source export ends 18 September 2026 for EUR/USD and 23 September 2026 for USD/JPY; the final studied release is 4 September. The EUR/USD gap on 6 December 2024 was excluded before summary statistics. We did not select days according to the payroll surprise, earnings or unemployment, and we did not retrieve market forecasts as they stood before each release. This is a test of a price rule after a scheduled report, not a test of predicting the report.

A small, historically selected window does not establish the expected future return. The 1.2- and 3-pip values are inputs for sensitivity, not claims about typical NFP execution. Verify spread, fills and risk on the actual broker account before treating any similar setup as executable.