Home › Gold forecasts › Thursday 24 September 2026
The US 10-year yield reached 5.1%, its highest since 2007. Gold lost $77 in one session and closed Wednesday at $4,287, below our weekly floor. It trades at $4,280 this morning. One session left before the verdict.
Spot now $4,280 · Wednesday’s close $4,287 · 21K in Suez EGP 6,240 ·
We published three scenarios: 45% holding the averages ($4,300–$4,360), 30% breaking the cluster ($4,250–$4,300), and 25% a rebound ($4,360–$4,400). The close was $4,287.39 — inside the second.
The condition we attached to that scenario read: stronger-than-expected PMIs confirm the hawkish case and the dollar extends — $4,307 goes and $4,300 with it. The manufacturing PMI came in at 57.0 against 53.5 expected, services at 58.7 against 56.0, the fastest US business activity in more than five years. And $4,300 went.
By our rule that is a correct call. By the same rule, the ranking was wrong for the second day running: the consolidation we put first at 45% happened neither on Tuesday nor on Wednesday. What we see in ourselves, said plainly: we are leaning toward “consolidation” because it is our weekly call — and the market is under no obligation to our weekly call.
Record: 23 published · 19 correct · 95% (19 of 20 graded; today’s is still open). How grading works →
The weekly call — the one that disagreed with a Kitco survey of 16 bullish analysts out of 16 — settles on the Friday 25 September close and nothing else. The range we weighted at 45% is $4,300–$4,420.
Gold closed Wednesday at $4,287 and is at $4,280 this morning — about twenty dollars below the floor of our range for now. For the weekly call to be right, gold has to get back above $4,300 and close there on Friday.
The rule is the same either way. If Friday closes above $4,300, we will write that the call was right. If it closes below, we will write that we were right about direction against Wall Street (they said up, and it went down) and wrong about the bucket. Just as plainly, on Saturday.
| Now | Reference | |
|---|---|---|
| Spot gold, per ounce | $4,280 | −$7 on the day |
| Wednesday’s close | $4,287 | inside our second scenario |
| 24K, per gram, we sell | EGP 7,131 | was 7,155 |
| 22K, per gram, we sell | EGP 6,537 | was 6,559 |
| 21K, per gram, we sell | EGP 6,240 | was 6,260 |
| 18K, per gram, we sell | EGP 5,348 | was 5,365 |
📉 21K came down EGP 20 a gram, reset at 11:40 AM today. This is the mirror image of yesterday: the ounce is down about $37 since yesterday’s reset, but the gram fell only EGP 20, because the implied dollar rate inside our counter price rose from about 51.5 to 51.8. Had it held, the gram would be near EGP 6,207 rather than 6,240. Yesterday the currency pulled the gram down; today it held it up. Live prices are always at today’s prices.
Covering the 24 September close through the morning of Friday 25 September
🇪🇬 21K in Suez by tomorrow morning: EGP 6,123 – 6,357, most likely EGP 6,196 – 6,284, assuming the currency holds.
✏️ Why the weights changed this time: two days running we put consolidation first at 45% and it did not happen. So we lowered it to 40% and raised the downside to 35%. Not to cover ourselves — because the yield is at a nineteen-year high and the market prices a second hike in October at about 70%, and that is real pressure.
On Monday 21 September we wrote: if the dollar firms and the yield goes back above 5%, we test the 100-day average at $4,320 and may break it. On Wednesday we wrote: stronger-than-expected PMIs … $4,307 goes and $4,300 with it.
Both conditions arrived, in that order. The yield crossed 5%, the PMIs came in stronger, and the levels broke one after another.
But notice the difference: the condition was right, and the weights were wrong. We had written the correct path in detail and put only 30% on it. That is the hardest mistake in analysis: to see the risk, write it down, and still not give it the weight it deserves because it runs against your original view.
That is why we keep saying: watch the condition, not the percentage. The percentage is our opinion; the condition is something you can monitor yourself.
| If the ounce is | 21K per gram works out near |
|---|---|
| $4,200 | EGP 6,123 |
| $4,230 (the neckline) | EGP 6,167 |
| $4,250 | EGP 6,196 |
| $4,280 (now) | EGP 6,240 |
| $4,300 (floor of our range) | EGP 6,269 |
| $4,360 | EGP 6,357 |
📍 Raw metal at today’s local conversion factor (1.458), before making charge, hallmarking and tax.
| $4,420 | Top of our weekly range |
| $4,359 | 100-day exponential moving average |
| $4,316 | 50% retracement |
| $4,300 | Floor of our weekly range — the number that decides the call |
| $4,287 | Wednesday’s close |
| $4,230 | The neckline — 7 August and 16 September lows |
| $4,227 | 61.8% retracement |
The number that decides the week is $4,300: a Friday close above it and our weekly call is right; below it, wrong. If $4,230 breaks on a close, the whole technical picture turns toward a deeper decline. Settled on the Friday 25 September close.
Thursday 10:30 AM — the Swiss National Bank held its rate at zero. (Done.)
Thursday 3:30 PM — US jobless claims, expected around 196,000. A lower number
means a tight labour market and pressure on gold.
Thursday, during the day — Trump and Xi meet at the White House. Rare earths and
technology restrictions are expected on the agenda.
Friday — the weekly close our call is graded on.
The verdict on the weekly call goes up on Saturday 26 September, whichever way it lands.
Spot price, session closes, the 10-year yield and October hike odds from Trading Economics; flash PMI figures from S&P Global via Kitco; technical levels and RSI from FXStreet; the Swiss National Bank decision from CNBC; local prices from our own published price list, reset at 11:40 AM on 24 September.
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.