The payroll number released on the first Friday of each month is, by construction, an estimate drawn from an incomplete sample. August 2026 is a good example: the first print came in at 16.2 (in units of 10k persons), and that figure will still go through two routine revisions over the following two months. The market prices the first print immediately; the revisions land quietly weeks later, long after the initial move has run its course — which is why their effect on gold is so often ignored.
For anyone trading gold, reading only the first print means reading half the dataset. Understanding how the number drifts from first print to current value is the first step in judging how much to trust a payroll release.
This site keeps a dedicated revision tracker that records the full path of each payroll print from its first release to its current value. The tracker covers six print months as of 2026-09-20, and the pattern is informative: the average absolute drift is 0.9, with two upward revisions and no downward revisions, but the individual months are not alike.
| Print month | First print | Current value | Drift |
|---|---|---|---|
| 2026-03 | 21.4 | 21.4 | 0.0 |
| 2026-04 | 14.8 | 14.8 | 0.0 |
| 2026-05 | 6.3 | 6.3 | 0.0 |
| 2026-06 | 2.0 | 3.1 | +1.1 |
| 2026-07 | −2.3 | 2.1 | +4.4 |
| 2026-08 | 16.2 | 16.2 | 0.0 |
Values in 10k persons. 2026-08 was published 2026-09-04 and is still awaiting its first routine revision; the next release is due 2026-10-02.
The July 2026 case is the clearest illustration: an initial print of −2.3 that is now +2.1 is a swing of 4.4 — enough to turn a headline about falling employment into one about a modest gain. When a print's first value keeps diverging from its final value, something in the sample or the seasonal adjustment is off, and the immediate market impact of that first print deserves a discount. This kind of tracking is work most gold sites never do, and it is what separates data noise from a trend signal.
The inflation backdrop amplifies the risk. CPI is running at 3.4% year-over-year (2026-08), PPI at 5.44% (2026-08) and core PCE at 3.38% (2026-Q2) — all elevated. The higher inflation runs, the more the statistical quality of employment data is exposed to re-estimation of wages and hours worked, and the wider the uncertainty around the size of any revision.
If the 16.2 first print for August 2026 is revised substantially up or down, the market's verdict on whether employment is cooling or proving resilient can flip outright, and with it the expected path for the Fed. Watching the drift is watching the source of expectation revisions. The next scheduled release, due 2026-10-02, carries an estimate of ≈+48k (range −2k to +98k) with probabilities of 17% negative, 68% for 0–100k and 15% above 100k — a range wide enough that the revision path will matter as much as the headline.
Payroll revisions matter to gold because they change an input to Fed decisions. The upper bound of the federal funds target range is 4.0% (2026-09-17), the next decision falls on 2026-10-29, and market pricing on Polymarket currently reads hike 100.0% · hold 0.0% · cut 0.0%. The site's own event estimate for that meeting is a hold at 4.00%, a reminder that model estimates and market pricing are labelled separately here and never conflated.
In a near-unanimous tightening price like this, a revision that points to a weakening labour market can shake the 100.0% hike pricing and become an event-level catalyst for gold. The rate backdrop raises the stakes: the 10Y Treasury yield has climbed to 4.99%, up 29bp over 20 days and sitting in the 99th percentile of its 52-week range, while the dollar index stands at 100.217. High nominal rates make gold more sensitive to every correction in the employment data.
COMEX gold's latest daily close is 4415.9 USD/oz (2026-09-18) and its live tick is 4418.56 USD/oz, +36.96 on the day (+0.84%). Shanghai gold Au99.99 closed its night session at 951.12 CNY/g, up 0.52%.
| Driver factor | Score | Window |
|---|---|---|
| Treasury yield trend | −100 | 20d |
| Oil→CPI transmission | −100 | 30d |
| Policy pricing | −60 | Sep FOMC |
| US Dollar Index (DXY) | −58 | 20d |
| Central-bank gold buying | +45 | structural |
| US debt burden | +39 | structural |
| Risk sentiment (VIX) | −5 | 20d |
| Inflation direction | −4 | 2mo |
| ETF holdings (SPDR) | +4 | 1d |
In the site's driver model the factors weighing on gold lead the table: the Treasury yield trend at −100 and oil-to-CPI transmission at −100, with Brent up 25.9% over 30 days to 103.87 USD/barrel. Central-bank gold buying is the only major positive factor at +45, with IMF-reporting central banks net-buying +281 t over the last 12 months (2025-08 to 2026-07). Add the long-run backdrop of federal debt at 122.6% of GDP (2026-Q1), and the rate-expectation swings that payroll revisions generate are a core variable in why gold is trading around these levels.
For gold traders, the tracker's drift record is worth building into the routine. First, treat the day of a first print as directional observation only, and wait for both routine revisions to land before acting with size. Second, check the print against the inflation readings — CPI at 3.4%, PPI at 5.44%, core PCE at 3.38% — since the wider the contradiction between jobs and price pressure, the more reason to be wary of later revisions. Third, before the 2026-10-29 FOMC decision, remember that any payroll revision can rewrite the 100.0% hike pricing, so plan the response in advance.
Knowing how much a number can be trusted is itself a form of edge. For the drift tracker alongside the full daily readings, go back to the daily dashboard.
It varies by month and cannot be generalised. The revision tracker covers six print months: the average absolute drift is 0.9 (in units of 10k persons), with two upward revisions and no downward ones so far. The individual cases differ sharply — 2026-07 moved from an initial −2.3 to 2.1, a drift of +4.4, and 2026-06 from 2.0 to 3.1, a drift of +1.1, while the three months before that have not been revised at all.
Because revisions change how the market prices the Fed path. Market pricing on Polymarket currently puts hike odds at 100.0%, hold at 0.0% and cut at 0.0% for the next decision on 2026-10-29, while the site's own event estimate is a hold at 4.00%. If revisions show the labour market clearly weakening, that one-sided pricing can loosen, and the pressure on gold from a 10Y yield of 4.99% changes with it. A first print moves the tape; a revision moves the trend.
In the site's driver model the leading factors are the Treasury yield trend at −100 (20-day window) and oil-to-CPI transmission at −100 (Brent up 25.9% over 30 days to 103.87 USD/barrel), followed by policy pricing at −60 and the dollar index at −58. Central-bank gold buying is the only large positive factor at +45, with +281 t of net purchases over the last 12 months, and the US debt burden scores +39. Gold's live tick is 4418.56 USD/oz, +0.84% on the day.
Three steps. First, place the current first print in the context of the ten-month payroll series, which currently runs from 4.1 for 2025-11 to 16.2 for 2026-08. Second, watch the direction of subsequent revisions, since sustained upward or downward drift is a trend signal rather than noise. Third, cross-check the data against the inflation readings — CPI at 3.4%, core PCE at 3.38% and PPI at 5.44% — because the wider the contradiction, the more reason to distrust the first print.
The readings snapshot is dated 2026-09-20 11:00 JST. Gold's latest daily close is 4415.9 USD/oz (2026-09-18) with a live tick of 4418.56 USD/oz, Shanghai gold closed at 951.12 CNY/g, and the inflation and payroll data run through 2026-08 and 2026-Q2 respectively. Everything here is a structured account of data facts, not investment advice; any trading decision is your own risk.
Source: Gold Data Reading · XAU Daily — https://xaudaily.com/ · All figures above are taken verbatim from the site's English readings payload (snapshot 2026-09-20 11:00 JST) and match the daily dashboard item by item. These are daily readings, not investment advice.