XAUUSD signal: sell the $3,903 bounce, target $3,843

Gold settled Friday at $3,875.44 on Investing.com's XAU/USD series, down $146.08 or 3.63%, after opening $4,023.94, printing the session high $4,027.13 in the first hour and the low $3,870.17 in the 19:00 UTC bar. The other feeds sit within a dollar: Kitco had $3,875.05 and a weekly range of $4,082.50–$3,870.25, Trading Economics $3,875.84. Call the band $3,875–$3,876.

That is the lowest close in thirteen months — since 2 October 2025. October finished 6.78% lower, and spot is 30.7% below the $5,595.46 high of 29 January on the same series.

The easy explanation writes itself. The tape does not support it: on Friday the 10-year yield fell.

Both of gold's drivers were bid and gold fell anyway

• Yields went down, not up. The 10-year slipped about 6 basis points to 4.929% (FXPremiere); a second compilation has 4.94%, down 4bp (Yield Curves Today). Feeds differ on the basis points; they agree on the direction.
• The dollar barely moved. DXY closed at 101.964, down 0.07% on the day and up 0.15% on the week, still under 102 (FXPremiere, Moneycontrol).
• The real yield has been drifting at 2.63%, down a basis point, on data through 29 October (GoldTrack).

A weaker dollar and a lower 10-year, and bullion loses 3.6%. That is not a rate story. That is a seller who needed cash.

Doubts over AI capital spending triggered a rout in tech stocks and broad deleveraging across asset classes; precious metals were collateral, not a view (Moneycontrol). The tell is silver: −19% in October, through $50 for the first time in eleven months, −9% on Friday alone, and the gold/silver ratio at roughly 79 — an eleven-month high. When silver collapses and gold follows it down, the trade is liquidity, not forecast.

Positioning had not capitulated. Managed money held a net long of about 105,600 gold contracts as of 27 October, up 1,400 on the week — a report that predates Friday (GoldTrack). The next COT is the first honest measure of how much of that was flushed, and it is the number that decides whether this is a low or a pause.

What actually broke on Friday

The last shelf that mattered was $3,955.32, the high of the 13:00 UTC bar — the bar that took gold through $3,950. From there the hourlies made lower highs for seven straight hours: $3,921.13, $3,923.84, $3,902.96, $3,896.77, $3,883.99, $3,885.33. Nothing retested the shelf.

Beneath that, two references matter more than any Fib level:

• $3,886.49 — the swing low of 28 October 2025, the anniversary shelf that launched last year's run to $5,595. Friday closed $11 under it.
• $4,099.27 — the 20-day average of closes, 5.46% above spot. Until gold closes back above that line, a rally and a bounce are the same thing.

Below $3,870 the desk's touch data gets thin: across thirteen months only one daily extreme printed in each $10 bucket between $3,810 and $3,860. There is no wall down here, which cuts both ways for the target.

Why $3,903, and not the shelf at $3,955

The desk knows this failure by name. Three of its last five sell limits were parked above a falling market and never filled: $4,144 on 22 October (never within $17.80), $4,054 on 23 October (never within $20), and $4,048 on 30 October — $29.16 above a session high of $4,018.84, while gold fell $146 and cleared both targets. Friday's plan called the direction, cleared $4,000 and $3,952, and earned nothing. The track record books it expired at 0R.

An order that sits a full ATR above spot is a spectator. So today's entry is 0.38 × ATR above the close, not 0.5 to 1.0, and it is a price the tape actually printed on Friday: $3,903 is the high of the 16:00 UTC bar, the first hour after the $3,920 shelf gave way. A retrace to it is a $28 bounce, the kind of move a Monday reopen produces routinely.

The trade-off is honest: a closer entry is a weaker level, and the wider stop is the price. If gold never gets back to $3,903, the day is 0R and the desk will write it down as one.

The plan

• Instrument: XAUUSD spot, MT5 gold
• Session: Monday 2 November — Sunday 23:00 UTC reopen (07:00 SGT) to the 20:00 UTC cut (04:00 SGT Tuesday)
• Bias: lower while $3,955 caps it; an hourly close above $3,963 ends the sequence
• Entry: $3,903 sell limit
• Stop loss: $3,963
• Take profit 1: $3,843 — 1.0R, close half
• Take profit 2: $3,783 — 2.0R, close the rest
• Risk-reward: 1:1 to TP1, 1:2 to TP2 on the full plan
• Do not take it if: gold opens above $3,905 or below $3,868 at the 23:00 UTC reopen. A gap through the top leaves the fill sitting inside the stop; a gap under $3,868 means the market is past the level and you would be selling a break, not a bounce.
• Time stop: flat or unfilled by 20:00 UTC Monday

The arithmetic, so it can be checked rather than trusted:

• Stop distance $60 = 0.831 × ATR(14) $72.23, computed on 14 completed daily bars. The desk's floor is 0.70 × ATR, or $50.56; a stop at the floor would sit $0.60 above Friday's 13:00 high and be decided by a spread.
• $3,963 sits $7.68 above the $3,955.32 structural high — a 0.106 × ATR buffer, priced off Investing.com's hourly series. Kitco's spot reports run $5–8 below the aggregate feeds, so the buffer is measured against the higher of the two, not the friendlier one.
• Entry is $27.56 above spot, or 0.38 × ATR — inside the half-ATR limit.
• Break-even win rate: 40% if both targets print (0.5R at TP1 plus 1.0R on the runner, against a 1R loss), and 66.7% if the runner is scratched at entry. A 1R first target against a noise-aware stop needs 50%; this one needs 40%.
• TP1 is $27 below Friday's low, in the thin zone under a thirteen-month floor, which is why it takes half off. TP2 needs gold at its lowest since August 2025 — a range resolution, not something Friday's close guarantees.

Execution: the gold spread widens to $0.30–$0.50 at the 23:00 UTC reopen and through the 15:00 UTC print — at $3,903 a $0.40 spread is 0.7% of the stop distance, material only on an opening-tick fill. Size off the $60 stop.

The one release that matters on Monday

Monday is thin. The only scheduled US print is ISM Manufacturing at 15:00 UTC (10:00 EST), forecast 54.7 against 54.5 prior (GoldTrack, FXPremiere). With the Fed holding at 3.75–4.00% on 28 October on a 10–2 vote where both dissents wanted a hike (CNBC), a beat supports the December hike and is worth more to the short side than a miss. A clear miss softens the hike case and can lift gold through $3,955. The stop prices that outcome; it does not argue with it.

Behind it: JOLTS and the midterms on Tuesday, payrolls on Friday. Neither is Monday's problem.

The other side of the trade

Kitco's weekly survey turned: 10 of 16 Wall Street analysts expect lower, one expects higher, five sideways, while 53% of 184 retail votes still expect gains (FinanceFeeds). A crowded short is its own hazard. The bull case is not stupid: Friday's drop had no macro cause — yields fell, the dollar was flat, the real yield slipped — positioning was stretched and got liquidated, and $3,886 is the exact shelf that launched the last run.

Weekend headlines cut both ways. Iran offered a plan to reopen the Strait of Hormuz within 48 hours while Washington said talks were moving, after a weekend of carrier deployments and threats (Why Gold Moved). Oil down means less hike pressure — gold-positive — and less war premium — gold-negative. Either impulse can gap the Sunday reopen by more than the median weekend gap ($4.52 across the last seventeen), which is why the plan is conditional on where the market opens, not on what a headline said.

If gold reclaims $3,955 and holds an hourly close above it, the whole of Friday's drop is back in play and the first objective is $4,000, the level Kitco's survey put on the bullish side. The plan has no position in that world; it is stopped out, and the October ledger records what a caught fade costs.

Sources

• Reuters via Business Recorder — gold at a four-month low on hike fears, crude and inflation
• Moneycontrol — metals brace for payrolls and tech earnings as crude tracks geopolitical risk
• FXPremiere — XAU/USD week ahead: the 3,879 floor and Monday's ISM
• FinanceFeeds — gold at $3,875 into the midterms and payrolls
• GoldTrack — week ahead: midterms and the jobs report follow a sharp drop
• Why Gold Moved — 1 November 2026 headlines
• CNBC — Fed holds rates steady, October 2026
• Price data: Investing.com XAU/USD daily and hourly series, pulled for this session

Not financial advice. Levels are structural references, not predictions, and a stop is a price at which the idea is wrong — not a guarantee of the fill. Gold has moved 3% or more in a session four times in the last three weeks; size accordingly.

Planning a gold position? The signals board carries every plan the desk publishes, and session levels groups them by the levels they were built on.