XAUUSD signal: buy the $4,030 retest, target $4,092

Gold spent two sessions rejecting $3,970 and one morning destroying it. Spot was $4,060.56 at the 20:00 UTC cut, +3.36% against Tuesday's $3,928.50 close, after a session that ran from $3,920.00 in the 00:00 UTC bar to $4,070.42 at 18:00 — a $150.42 range, 2.03 × ATR(14). Kitco had spot at $4,059.20 at 14:04 ET, CNBC quoted $4,061.30 and metalcharts $4,062.07, so the band at the cut is $4,058–4,062 (Kitco, CNBC, metalcharts). The session is still open: those are the numbers as they stand, not a close.

The dollar broke before gold did

Democrats were projected to take both chambers of Congress at 02:09 UTC, and the dollar was sold from the open (Vantage). The dollar index traded 101.17–101.19 through the Asian and European sessions, half a percent lower and its weakest in more than five weeks (Reuters via Kitco). Bloomberg called it the dollar's largest one-day fall in two months (Bloomberg Law).

The bond market did the rest. The 10-year yield fell about five basis points to 4.90%, with a later Bloomberg read at 4.93% after a two-basis-point drop, and the 2-year dropped to 4.5458%, a near two-month low (Guardian, Bloomberg Law, Business Recorder). The Treasury also signalled it may cut longer-dated issuance in 2027, which is a supply story stacked on top of a political one.

This is a dollar trade, not a haven trade: equities reversed the midterm rally, gold and silver did not. Silver added about 5% to $52.01–52.22 and the gold/silver ratio sits near 77.8, barely moved from the desk's own 78.89 on 30 October (goldstocklive, metalcharts). Both metals repriced against the dollar together. Nothing was hedged.

Good data did not stop it

The morning was not dovish. ADP put October private payrolls at +86,000 against a 75,000 estimate, unchanged on a downwardly revised September, with job-changers' pay up 4.8% (CNBC). ISM services came in at 54.5 against 54.8 expected, with new orders falling to 55.9 from 59.8 and business activity to 53.8 from 56.5, while the employment index rose to 51.4 and prices paid stayed at 73.0 (ISM release, Reuters via Investing.com).

Gold's second leg started with that print: the 15:00 bar closed $4,048.95 and the 16:00 bar ran to $4,067.76. The market took the soft orders component and ignored the firm employment one. Friday's payrolls are the real test — 13:30 UTC, +65,000 consensus after +29,000 in September, unemployment seen at 4.2% — and a December hike was still priced at roughly 76–87% into today's rally (Taurex, Reuters via Investing.com). A 3.4% day on a labour-market beat is not a rate-cut trade. It is a dollar trade that has not yet met payrolls.

Three rejections, then a hundred-dollar break

The ceiling was measured, not imagined. Gold printed $3,970.45 Monday afternoon, $3,970.31 in Monday's 16:00 bar and $3,969.96 in Tuesday's 13:00 bar — three highs inside 49 cents, and Tuesday's plan sold the $3,945 shelf into it (yesterday's plan).

Today did not test it a fourth time. The 05:00 UTC bar opened $3,951.98 and printed $4,009.62, then closed $3,998.11 — $39 through the ceiling in sixty minutes, with no retest of $3,970 afterwards. The 06:00 bar reached $4,027.94. From there the market built a shelf: the 12:00, 13:00, 14:00 and 15:00 bars all bottomed between $4,029.33 and $4,032.62, a $3.29 band touched four times, and the 16:00 bar broke up from it to $4,067.76.

The high is interesting for the wrong reason. $4,070.42 stalls inside the October supply shelf — $4,065.87 on 27 October, $4,072.65 on 29 October, $4,082.79 on 28 October. Gold rallying 3.4% and stopping at the underside of the old highs is not a coincidence, and it is why this plan buys a retest rather than chasing the breakout.

Where the levels actually are

• $4,082.79 — 28 October high, the top of the supply shelf. TP1 clears it by $14.
• $4,070.42 — today's high, 18:00 bar.
• $4,060.56 — the 20:00 UTC cut.
• $4,030 — the four-touch shelf from 12:00 to 15:00. This plan's entry.
• $4,021.49 / $4,015.30 — the 08:00 and 07:00 bar lows, the first step under the shelf.
• $3,996.42 — the 06:00 bar low, the floor of the entire post-breakout range.
• $3,970.45 / $3,969.96 — the triple ceiling that broke today, now support.
• $3,963 — this plan's stop, 0.10 × ATR below that ceiling.
• $3,920.00 — today's low, the 00:00 bar.
• $4,097 / $4,164 — TP1 and TP2.

The plan

• Instrument: XAUUSD spot, MT5 gold
• Session: Thursday 5 November — Wednesday 23:00 UTC reopen (07:00 SGT) to the 20:00 UTC time stop (04:00 SGT Friday)
• Bias: higher while $3,970 holds as support; two hourly closes back under $3,963 end the sequence
• Entry: $4,030 buy limit
• Stop loss: $3,963
• Take profit 1: $4,097 — 1.0R, close half
• Take profit 2: $4,164 — 2.0R, close the rest
• Risk-reward: 1:1 to TP1, 1:2 to TP2
• Do not take it if: gold opens above $4,065 and holds there through the first two hours without trading $4,030 — the order is void, do not chase it — or if spot trades under $3,963 before the fill, in which case the breakout failed and there is no trade.
• Time stop: flat or unfilled by 20:00 UTC Thursday

The arithmetic, so it can be checked

• Stop distance $67 = 0.906 × ATR(14) $73.97, measured on fourteen completed daily bars from 15 October to 3 November. The 0.7 floor is $51.78; the stop clears it by $15.22. 30 October's $156.96 range is still inside the window, which is why ATR is nearly $74 and why a 1R target sits $67 away.
• The stop is placed on the invalidation, not on a round number. $3,963 is $7.45 — 0.10 × ATR — below the $3,970.45 high that capped this market three times. If gold gives the whole breakout back and trades under the old ceiling, the impulse is dead and the plan should be too. A stop at $4,000 would have been tighter and more comfortable; it would also have been 0.4 × ATR, inside the noise band that has decided four of this desk's last six verdicts.
• Entry is $30.56 under the cut — 0.413 × ATR, inside the half-ATR rule of $36.99. The last three sessions all printed their low before 10:00 UTC: 2 November at the $3,875.21 open, 3 November at $3,895.59 in the 09:00 bar, 4 November at $3,920.00 in the 00:00 bar. A $30 pullback into the Asian session is the normal shape of this market this week.
• TP1 at $4,097 is 1R, $26.58 above today's high, and the first price that clears the October supply shelf rather than stalling under it. TP2 at $4,164 is the 19–21 October shelf where the market broke down.
• Break-even win rate: the simple version — whole position out at $4,097, no runner — needs 50%. Splitting 50/50 between TP1 and TP2 needs 40%. A runner scratched at breakeven after TP1 needs 67%, which is why the runner is not scratched.

Yesterday's order, and what it cost

Tuesday's $3,945 sell limit filled in Wednesday's 01:00 UTC bar and its $4,004 stop was traded through by the 05:00 bar at $4,009.62 — before the ceiling was ever retested. That is a loss, and the grading pass books it at 21:15 UTC tonight; this article does not pre-write the verdict, because the track record is the only place a verdict belongs.

The lesson is specific. The stop was $32.85 above a level that had held three times, placed on the assumption it would hold a fourth — and a political event took it out in an hour without the level being tested again. Levels do not survive events. The plan above puts its stop where the breakout itself would be wrong, not where a rejection is expected, and it buys the retest instead of shorting the strength. Everything on this board is now a bet on whether a midterm-driven dollar move survives Friday's payrolls, and the dollar hub is the place to track that (dollar index).

Execution: the spread runs $0.30–0.50 at the 23:00 reopen and is widest for the first fifteen minutes, so a $67 stop is not $67 of realised risk on a thin fill — size from the fill, not the plan. Investing.com is the reference feed for the stop; a $5 disagreement between feeds on a stop this wide is an execution error, not a market outcome, and $3,963 is deliberately below the level where any plausible feed print diverges. A fill before the London open is a pullback inside an uptrend; a fill after 15:00 UTC means sellers are being handed the bid rather than buyers defending a shelf, so take TP1 at the first offer and do not carry the runner into payrolls week.

Sources: CNBC — gold and the midterm result, Reuters via Kitco — gold gains 3%, Bloomberg Law — bonds rally on the sweep, Guardian — markets live, Vantage — dollar index and 10-year, CNBC — ADP 86,000, ISM — services PMI release, Reuters via Investing.com — ISM and payrolls consensus, Taurex — week ahead, Kitco — spot. Levels, ranges and ATR are the desk's own calculations from Investing.com daily and hourly bars.

Not financial advice. This is a published plan with a published stop, not a recommendation to trade. Leverage on XAUUSD can lose more than the margin posted, and Friday's payrolls at 13:30 UTC can gap this market either way.