Home › Gold forecasts › Saturday 31 October 2026
Gold closed the week at $3,875.84, down $118 (−2.96%), and $145 on Friday alone (−3.6%) — the lowest close in a year and gold’s worst October since 2008. The weekly call closed inside the first box, the one we ranked first — the first time in four weeks the weekly order was right — but its confirming number was not met, so the verdict is right by luck. Friday’s daily call was wrong: the price fell $34 below the lowest box, the second time in eight days the liquidity chain ran past our boxes. And the gram? EGP 6,180 — down only 110 pounds on the week (−1.75%), because the trade dollar rose to 56.68. Before writing next week’s call — the week of the US midterm elections (Tuesday) and the jobs report (Friday) — we held an advisory council: two research rounds across analysts on five continents, and a third round looking only for the strongest case against the majority.
Friday close $3,876 · Week −$118 (−2.96%) · Trade dollar EGP 56.68 · banks 55.74 ·
The weekly call (published Saturday 24 October): 40% the liquidation runs on below $4,000 ($3,830–$3,980) · 35% the pause after the liquidation ($3,980–$4,080) · 25% the bounce ($4,080–$4,200). The close was $3,875.84 — inside the first. Right range and right order, for the first time in four weeks.
The first box’s confirming number was “Brent at $118 or above, or the S&P 500 at 7,540 or below”. Brent: the December contract closed at $117.64 (36 cents short), the January contract at $113.43. The S&P 500 closed at 7,580.07 (−1.8%) — 40 points above the line. Both conditions came very close and neither was met. On the reading harshest to us: right by luck.
Friday’s daily call: we published 40% the move continues ($3,910–$3,970, a box already widened to $60 under correction six) · 35% the pause ($3,970–$4,030) · 25% the bounce ($4,030–$4,090). The close was $3,875.84 — $34 below the lowest box. The employment cost index printed 0.8% (the first box’s condition was met) and the direction was right, but the price ran through the whole box. Verdict: wrong. We will not call it “close” — the same rule we applied on Thursday 22.
⚖️ A note on the reading: the close from four sources: Trading Economics $3,875.84 · Kitco $3,874.40 and $3,875.05 · Kitco’s evening report $3,879.90. All four sit in the same weekly box and below the same daily box, so neither verdict changes.
Record: 55 published · 53 graded · 48 correct · 91% (the new weekly call is open). Friday’s map → · How grading works →
| Forecast | Close | Range | Order | Confirming number | Verdict |
|---|---|---|---|---|---|
| Mon 26 | 4,016.72 | ✅ first | ✅ | ✅ the quiet rule | Right for the reason |
| Tue 27 | 4,060.20 | ✅ second | ❌ by 20 cents | ❌ Brent 113.9, not 113.1 | Right by luck |
| Wed 28 | 4,067.78 | ✅ first | ✅ | ❌ hike dissents (10–2) | Right by luck |
| Thu 29 | 4,021.31 | ✅ first | ✅ | ✅ core PCE 3.1% | Right for the reason |
| Fri 30 | 3,875.84 | ❌ $34 below the lowest box | ➖ direction right | ✅ ECI 0.8% | Wrong |
| Weekly (24 Oct) | 3,875.84 | ✅ first | ✅ | ❌ Brent 117.64, S&P 7,580 | Right by luck |
Compared with last week:
🟢 Order: 4 of 6 (last week 1 of 6). The biggest improvement in the table, and it came from one rule:
“rank on behaviour, not logic”.
🟢 Fully right: 2 of 6 (last week none).
⚫ Range: 5 of 6 — the same as last week. And the one miss was the same kind: a liquidity day, with the price running
past the lowest box.
⚫ Confirming number: 3 of 6 — the same as last week.
So this week we learned to rank. We have not yet learned to measure the size of a liquidity day. That is what correction seven, below, is about.
The problem in numbers: correction six said “on liquidity days, the first box is $60 wide”. On Thursday 22 the price fell $116 with a $50 box. On Friday 30 it fell $145 with a $60 box. So correction six pointed the right way, but it was too small. And widening the box every time is not the answer — a very wide box says nothing.
The new rule: when the liquidity signature is showing (equities, the yield and gold falling together), we add a fourth box, open at the bottom — the “tail box” — with an explicit probability (usually 10%) and its own condition. If the price lands in it, the range counts as right but the order counts as wrong (because it is ranked last). That way we admit in advance that liquidity days can run further than any box, without stretching the other boxes until they mean nothing.
Correction six stays as it is (the first box $60 daily and $150 weekly). The earlier rules: 1 thresholds read “at or above” · 2 a 0.12 band on the yield · 3 calm is confirmed by the size of the move · 4 on a scheduled release, the condition is the number plus the close, and it sits on the surprise · 5 a threshold sits at least 1% (0.06 on the yield) from the price · 6 the liquidity box.
| Now | Reference | |
|---|---|---|
| Weekly close, per ounce | $3,876 | −$118 on last Friday (−2.96%) |
| 💱 Trade dollar (implied) | EGP 56.68 | +0.70 in a week (from 55.98) |
| 24K, per gram, we sell | EGP 7,062 | |
| 21K, per gram, we sell | EGP 6,180 | −110 on the week (−1.75%) |
| 18K, per gram, we sell | EGP 5,297 |
➡️ 21K at EGP 6,180 — from today’s counter update (2:08 PM). Last Saturday it
was 6,290, so −110 pounds (−1.75%) in a week when the ounce fell 2.96%. In detail: the ounce’s fall was worth
about −186 pounds on the gram, and the trade dollar’s rise about +76. The official price is at
today’s prices.
💱 The Souq El Dahab app (today): a bank-dollar average of 55.74 — down from
56.10 on Tuesday and 56.23 on Monday — a market trade dollar of 56.55, a market average for 21K of EGP 6,170 (selling),
and the ounce on screen at $3,877.70. So the premium is back to about 1.45% in the market and 1.7% at our counter. Tuesday’s
rule says a premium above 1% is a warning: either the banks rise again, or the gram falls as the trade catches up with the bank.
📊 Dollar pressure — the week turned at the end: Monday an outflow of $663 million · Tuesday 619 ·
Wednesday 391 · Thursday an inflow of $368 million (net buying by foreigners and Arabs — Hapi Journal, from
Exchange data). The first large inflow day since 19 October. Net for the week: an outflow of about $1.3 billion, after $1.84
billion the week before. And the banks answered the inflow day: the dollar is down 36 piastres since Tuesday.
Next week is very sensitive: the US midterm elections on Tuesday, two central banks expected to hike (Australia on Tuesday and England on Thursday, a quarter point each), and the US jobs report on Friday. So before writing a number we read: in the first round, five tracks in parallel (North America · Europe · Asia, the Middle East, Africa, Australia and Latin America · the calendar and positioning · Friday’s closes); in the second, an independent check of the figures against other sources, plus a deliberate search for the strongest case against the majority. Here is what came out, by name and date — and what we did not find:
🔁 The contrarian round — the strongest case against the majority: (1) Main Street 53% bullish while professionals are 63% bearish — that split has historically appeared near lows, though it does not time the week. (2) BCA (26 October) said the deviation from the 200-day average reached July’s level, which was followed by a short-lived low. (3) Central banks and China are still buying. (4) Commerzbank calls the fall “overdone”. And the strongest case for further falls: speculators’ net long on COMEX is still 186,451 contracts (down only 13,000 to 27 October) — so there is still a lot that can be sold, and most of the break happened after the report date.
The council’s conclusion: the large majority sees next week as lower, or a low being built below 3,900, with the recurring downside levels 3,850 · 3,800 · 3,750, and further out 3,500. Resistance: 3,900–3,910 · 4,000 · 4,015–4,040. None of the big names sees a strong bounce before yields or oil come down.
A crowded, sensitive week: Monday the ISM manufacturing index · Tuesday the US elections + the Reserve Bank of Australia (hike expected) · Wednesday ISM services and ADP · Thursday the Bank of England (hike to 4% expected) and Hammack (who dissented for a Fed hike) · Friday the jobs report (65,000 expected, after 29,000). The liquidity signature has been showing since Friday, so this call is written under correction six (a $150 first box) and correction seven (the tail box).
⚠️ Ranked on behaviour: Friday closed near the day’s low, the liquidity signature is showing, professionals are 63% bearish, and speculators still hold large longs. So continuation first. Last week this way of ranking was right 4 times out of 6.
🇪🇬 21K at the end of next week — on three trade-dollar assumptions:
Steady at 56.68: EGP 5,900 – 6,538 (from $3,700 to $4,100), most likely 5,900 – 6,139
Up one pound (57.68): EGP 6,004 – 6,653, most likely 6,004 – 6,247
Down one pound (55.68): EGP 5,796 – 6,422, most likely 5,796 – 6,031
The premium is above 1% (trade 56.55 against banks 55.74 in the market), and foreigners came back to buy on Thursday — so the third
assumption (the dollar easing) is not far-fetched this week, for the first time in weeks.
✏️ Every threshold carries its distance from the price (correction five), the first box is $150 wide (correction six), and the tail box is published with its probability (correction seven). If you find a threshold close to the price or a narrow box, write to us.
If the ounce moves $50, the gram moves about 80 pounds. If the trade dollar moves one pound, the gram moves about 109 pounds. This week the ounce weighed more (−186 pounds) and the dollar softened it (+76). Next week both could fall together for the first time: the ounce under pressure, and the dollar with a reason to ease (a 1.7% premium plus Thursday’s foreign inflow).
| Ounce ↓ · Dollar → | 55.5 | 56.5 | 57.5 |
|---|---|---|---|
| $3,700 (the tail-box line) | 5,777 | 5,881 | 5,985 |
| $3,850 (the first dividing line) | 6,011 | 6,119 | 6,228 |
| $3,876 (the weekly close) | 6,052 | 6,161 | 6,270 |
| $3,950 (the second dividing line) | 6,167 | 6,278 | 6,389 |
| $4,100 (the call’s ceiling) | 6,401 | 6,517 | 6,632 |
📍 Our gram at 6,180 sits just above the “3,876 × 56.5” cell. If gold falls to $3,800 and the dollar eases to 55.5 (the banks), the gram is about 5,933 — down 247. If gold falls to $3,800 and the dollar rises to 57.5, about 6,147 — down only 33. Next week the dollar could make a 200-pound difference. Raw metal before making charge, hallmarking and tax.
Monday 26: Brent fell on talk of “a deal or strikes on Iran after the elections”, and gold circled before the Fed,
closing at 4,017. The banks crossed 56.
Tuesday 27: strong US data (durable goods 1.8%, consumer confidence 91.7); gold closed at 4,060.20 — 20 cents above a dividing
line. The trade dollar met the bank (56.07 against 56.10).
Wednesday 28: the Fed held 10–2, with two dissents for a hike (Cook and Hammack). Gold dipped to 4,035 on the statement and
closed at 4,068 — the news was priced.
Thursday 29: core PCE 3.1% as expected, GDP 3.4% (above expectations). Gold closed at 4,021. Foreigners came back to Treasury
bills with $368 million.
Friday 30: equities −1.8%, Brent (December) at 117.64 on an attack in Hormuz, the employment cost index 0.8% as expected. Gold
−3.6% to 3,876 — the lowest close in a year. Silver broke below $50.
Three weeks ago we wrote “the order is our weak point”. This week the order was right 4 times out of 6 — including the weekly call, for the first time in a month. That came from one rule: rank on behaviour, not logic. When the market is falling and everything is being sold, continuation first. When the market is waiting for a big event, calm first. When the expected number prints as expected, calm first.
But the one range miss was the same miss for the second week: a liquidity day, and the price ran further than any box. Thursday 22 fell 116; Friday 30 fell 145. Correction six widened the box to 60, and it still was not enough. So instead of widening until the box means nothing, we added the tail box with its explicit probability.
And the lesson for you in Egypt: the ounce fell 3% and the gram fell 1.75%. The dollar absorbed about a third of the fall. Next week that could flip: the banks have started to ease (55.74) and foreigners came back to buy on Thursday. If both continue, the gram will feel the ounce’s fall in full — or more.
| $4,100 | The ceiling of the call |
| $4,015–$4,040 | Kitco resistance |
| $4,000 | The psychological number |
| $3,950 | The dividing line between the low and the bounce |
| $3,890–$3,910 | The first resistance |
| $3,876 | The weekly close |
| $3,870 | Friday’s low |
| $3,850 | The dividing line between the liquidation and the low |
| $3,780–$3,800 | Kitco support |
| $3,750 | Chandler’s floor |
| $3,700 | The tail-box line |
| $3,500 | Colombo’s support |
| EGP 56.68 | The implied trade dollar |
| EGP 56.55 | The market trade dollar |
| EGP 55.74 | The bank-dollar average (easing) — a premium of 1.45–1.7% |
Next week: the S&P 7,500 or Brent 115 means $3,700–$3,850. A move under 2% means $3,850–$3,950. Jobs at 40,000 or fewer, or the yield at 4.85, means above $3,950. The S&P 7,350 means below $3,700. And the gram answers to both the ounce and the dollar — and both could fall together.
Monday 2: ISM manufacturing 5:00 PM (54.7 expected) · Goolsbee speaks 6:15 PM.
Tuesday 3: the Reserve Bank of Australia 5:30 AM (a hike to 4.85% expected) · the US elections
all day, results in the early hours of Wednesday · JOLTS job openings 5:00 PM.
Wednesday 4: ADP 3:15 PM (65,000 expected) · ISM services 5:00 PM (54.8 expected).
Thursday 5: the Bank of England 2:00 PM (a hike to 4% expected) · jobless claims 3:30 PM ·
Hammack 3:50 PM and Barkin 4:10 PM.
Friday 6: the jobs report 3:30 PM (65,000 expected, unemployment 4.2%, wages 0.3%) · Michigan 5:00 PM ·
Williams, Musalem and Bowman · midnight: the verdict on this call.
The US is back on winter time, so US releases are now 3:30 PM Cairo time instead of 2:30.
Friday’s closes: gold ($3,875.84), the yield, the dollar, the employment cost index and the Chicago index from Trading Economics; gold ($3,874.40 / $3,875.05 / $3,879.90), Brent December ($117.64) and the S&P 500 (7,580.07) from Kitco; Brent January ($113.4) from OilPrice. The advisory council: the Kitco weekly survey and closing report (30 October); TD Securities, State Street, Sucden, BCA and Metals Focus via Kitco (26–29 October); Saxo Bank’s Commodity Weekly (30 October); BullionVault (30 October); Mining.com (30 October, including Commerzbank, TD and Asian demand); Hapi Journal (29–30 October); the World Gold Council’s “Gold and Midterms” (22 October); FXStreet (Radomski). Calendar: Trading Economics. Positioning: the CFTC COMEX report dated 27 October. Egypt: foreign and Arab trading in government debt ($368 million bought on Thursday) from Exchange data via Hapi Journal; the bank-dollar average (55.74), the market trade dollar (56.55) and the market 21K average (6,170) from the Souq El Dahab app, 31 October; the implied trade dollar (56.68) computed from our counter price and Friday’s close; counter prices from our own published price list, updated 31 October at 2:08 PM. What we could not read: Reuters, MarketWatch, CNBC and The Hindu BusinessLine (closed), and the World Gold Council’s Q3 report (not yet out).
This report is the English edition. The Arabic edition is primary and was published first, at the same date’s Arabic page. Every figure here appears there.