Three things to know this week
- Volatility alone makes a test of the 52-week low likely. With our own 20-day realised volatility of 20.7%, a no-direction model puts the chance of trading at or below 3,879.50 before Friday’s close at about 69%, and at or above 4,000.00 at about 39%. These are experimental model estimates of how far price may travel, not which way.
- Positioning gives no directional edge at this level. In today’s 20–40th percentile band gold was higher 13 weeks later in 51% of weeks, against 62% for all weeks since 2006. With only about 11 independent windows in that band, the difference is noise. What low positioning does tell us: there is less forced long selling left than in February and March.
- Friday’s jobs report is the main scheduled event. Since 2016, gold’s median daily range on first-Friday payroll days was 1.4%, against 1.2% on other days (123 payroll days).
Where gold stands
Over 52 weeks gold is 2.5% lower. Past drawdowns this deep (2008 and 2013) are only two episodes, too few to give a reliable base rate, so we do not publish one.
The driver scorecard
Each row is a channel that bank research desks and the World Gold Council’s return-attribution work use to explain gold. The right column is our reading for gold’s direction, based on the data, with the date of each reading.
| Driver | Reading | Recent change | Context | For gold |
|---|---|---|---|---|
| US 10-yr real yield | 2.60% (30 Oct) | −32 bp in 4 wks · +13 bp in 13 wks | 95th pctl of 3 yrs; 3-yr high 2.92% in Oct | ▼ Headwind level high, momentum easing |
| US dollar index | 101.9 (30 Oct) | −0.1% in 4 wks · +2.1% in 13 wks | up from a 3-yr low of 96.9 | ▼ Headwind |
| Fed path | Effective fed funds 3.88% (daily) | hiked in September, held in October | 2-yr yield 4.56%, 68 bp above fed funds: markets price more tightening | ▼ Headwind |
| Oil and inflation hedging | WTI ≈ 98–104 dollars | +23% in 13 wks | 10-yr breakeven inflation 2.34% | ◆ Mixed hedge demand vs a hawkish Fed |
| Hedge funds (managed money) | 105.6k net long (27 Oct) | −14.8k in 4 wks | 37th pctl of 3 yrs; longs 122,508 vs shorts 16,946 | ● Neutral not crowded |
| Swap dealers (bullion banks) | −206k net (27 Oct) | +14.7k in 1 wk (short covering) | open interest 396,810, 13th pctl: low participation | ● Neutral |
| Central banks | 289 t bought in Q2 2026 | +62% y/y (World Gold Council) | structural buyer | ▲ Tailwind |
| Seasonality (2000–2025) | November up in 58% of years | median +0.7%; December up 65% | 26 years: weak evidence | ● Neutral |
How much do rates and the dollar explain?
Over 156 weeks, weekly moves in real yields and the dollar explain only 12% of weekly gold moves (R², n = 156 weeks). Over the last 52 weeks that rises to 18% (n = 52). In that window, each +10 bp in real yields went with about −1.2% on gold, and each +1% on the dollar with about −1.1%. Macro sets the background, but most weekly moves come from flows and positioning, which is why this briefing tracks them together.
Who is long, who is short
Hedge funds cut their net long from 144.7k on 25 August to 105.6k, but gross shorts are only 16,946 contracts. Open interest is in the 13th percentile of three years, so fewer participants are in the market.
Does positioning predict the next quarter?
| Positioning band | Higher after 4 wks | Higher after 13 wks | Median 13 wks | n (weeks) | ≈ independent (13 wk) |
|---|---|---|---|---|---|
| 0–20th | 58% | 60% | +1.9% | 237 | 18 |
| 20–40th · now | 51% | 51% | +0.1% | 146 | 11 |
| 40–60th | 53% | 54% | +0.7% | 160 | 12 |
| 60–80th | 55% | 74% | +4.1% | 160 | 12 |
| 80–100th | 58% | 65% | +3.5% | 192 | 15 |
| All weeks (baseline) | 56% | 62% | +2.4% | 1,051 | 81 |
Bands use the managed-money net long as a share of open interest, ranked against the previous three years, over 1,064 weeks of CFTC data (2006–2026). Windows overlap, so the independent sample is roughly n ÷ 13 for the 13-week columns. We have not yet computed bootstrap intervals; until we do, read every band as “not clearly different from the baseline” unless the gap is large.
Where the stops are likely sitting
| Side | Level | Price (rounded) | Formed | Status |
|---|---|---|---|---|
| Buy-side | October high | 4,259.00 | 2 Oct | resting |
| Buy-side | Prior-week high | 4,097.00 | 28 Oct | resting |
| Buy-side | Round number | 4,000.00 | – | resting |
| Buy-side | Monday high | 3,981.00 | 2 Nov | resting |
| Sell-side | June–July equal lows | 3,959.00 | Jun–Jul | taken 30 Oct, price back below |
| Sell-side | Monday low | 3,889.00 | 2 Nov | resting |
| Sell-side | 52-week low | 3,879.50 | 30 Oct | resting |
| Sell-side | Round number | 3,800.00 | – | resting |
See the same idea on spot XAU/USD, updated through the day, on the free Gold Liquidity Map, and the research behind it: previous day high and low study and Asian range study.
What can happen into Friday
The odds below are experimental model estimates: a driftless volatility model using our own 20-day realised volatility (20.7%), from the 3 Nov snapshot to the Friday 6 Nov close. They measure reach, not direction, and have not yet been calibrated against outcomes. The conditions under each scenario come from the driver scorecard.
Sell-side extension
Price takes the 3,879.50 area and holds below it. The next pool is the 3,800.00 round number.
More likely if payrolls and wages run hot and real yields and the dollar rise.
Reach 3,800.00 by Friday: ≈ 21%
Sweep and reclaim
Price takes the 3,879.50 area, then closes back above the 3,889–3,959 shelf.
More likely if data is soft or yields keep easing (already 32 bp off the peak).
Reach 3,879.50 by Friday: ≈ 69%
Relief squeeze
Price goes straight up through 3,981.00 and 4,000.00 toward the 4,097.00 prior-week high. Swap-dealer short covering is already under way.
More likely after weak payrolls or a dovish surprise.
Reach 4,000.00: ≈ 39% · reach 4,097.00: ≈ 7%
A reclaim is not a setup on its own. In our method, a sweep of a level is followed by a CHoCH (a candle body closing beyond the external swing; a wick does not count), then a retest of the broken level or the order block of the CHoCH leg. Only then is the idea confirmed, with the order block of the CHoCH leg as the zone to study. No break and retest, no trade. This briefing gives no entries, stops or targets.
Longer base rates (context, not a forecast)
- Across all weeks since 2006, gold was higher 13 weeks later in 62% of cases (median +2.4%, n = 1,051 overlapping weeks).
- On days when gold sat 20% or more below its 52-week high it was higher 63 trading days later in 62% of cases (n = 334 days). Most of those days come from just two episodes, so this is weak evidence.
- Over the last 52 weeks the median weekly range was $224.00.
Event risk this week
| Day | Time ET / PKT | Event | Why gold cares |
|---|---|---|---|
| Tue 3 Nov | all day | US midterm elections, results overnight | fiscal and policy uncertainty; dollar reaction |
| Wed 4 Nov | 10:00 / 20:00 | ISM services PMI | growth and the prices paid sub-index |
| Thu 5 Nov | 08:30 / 18:30 | Weekly jobless claims | labour-market trend |
| Fri 6 Nov | 08:30 / 18:30 | US jobs report: payrolls, unemployment, wages | usually the biggest scheduled move of the week |
| Fri 6 Nov | 15:30 / 01:30 Sat | CFTC Commitments of Traders (positions of 3 Nov) | next positioning update |
US clocks moved back on 1 November, so New York is now UTC−5 and Pakistan time is New York + 10 hours. CFTC has announced that its positions report will move to twice a week by the end of 2026, so the release day in this table will change.
What we will check next week
Each briefing logs measurable estimates before the week, and the next edition grades them (Brier score against a 50% baseline), right or wrong. Grading rule: COMEX gold front-month (December) high, low and settlement for Tuesday 3 Nov to Friday 6 Nov, with levels as written below. Model: driftless realised-volatility model, version 1.
| ID | By the Friday 6 Nov close, gold futures… | Model estimate |
|---|---|---|
| E1 | trades at or below 3,879.50 (52-week low area) | 69% |
| E2 | trades at or above 4,000.00 (round number) | 39% |
| E3 | trades at or below 3,800.00 (round number) | 21% |
| E4 | trades at or above 4,097.00 (prior-week high) | 7% |
| E5 | settles above 4,000.00 on Friday | 20% |
How this briefing is made
- Positions: CFTC Commitments of Traders, disaggregated, futures only, COMEX gold. Positions as of Tuesday 27 Oct, released Friday 30 Oct.
- Rates and the dollar: official US Treasury and Federal Reserve daily series, readings dated in the scorecard.
- Prices: COMEX gold futures (December), a delayed snapshot during the 3 Nov session. Every price is rounded to $0.50 and indicative. We publish derived figures only, never price history.
- Volatility: our own 20-day realised volatility from daily closes. Implied-volatility indices are not shown on this page.
- Base rates: XAU Setups’ own analysis, with sample sizes. Overlapping windows are flagged.
- Review: drafted from frozen inputs, checked and approved by a person before publishing. Corrections are noted on the page, as in our editorial policy.