The previous edition called 4,000 overhead supply and said the weekend's three developments were net bearish for the open. The open was bearish. The close was not, and the reason is that the oil channel that was doing the damage reversed sign in a single session.

Evening edition, published at 00:25 Beijing on Tuesday October 27, which is 12:25 in New York on Monday October 26. The Asian session and the European session are complete; the New York cash session opened at 09:30 ET and this edition was written about three hours into it, so no 16:00 ET settlement exists for October 26 and none is reported here. The last price available is the site's own structured quote file at 00:17:37 Beijing: spot gold 4,019.70 dollars an ounce, up 21.99 dollars or 0.55 percent on that file's own reference close of 3,997.71, with a session range of 3,978 to 4,026.68. The range was set at the extremes: the low was printed in the Asian morning and the high in the New York hour, and the metal ran 35.26 dollars from the 09:30 ET cash open at 3,991.42 to that high before fading 3.27 dollars into the last print of this window. Two United States releases that fall inside this window are included with their release times and their consensus figures, because they were published before the cutoff. The Chicago Fed National Activity Index for September was released at 08:30 ET, or 20:30 Beijing, at 0.07 against a consensus of minus 0.1 and a prior of minus 0.08 that was revised down from minus 0.04; the Dallas Fed manufacturing survey for October was released at 10:30 ET, or 22:30 Beijing, at 13.4 against a consensus of 8.8 on one desk and 7.0 on another, from a prior of 9.8. Both were stronger than expected, and the December rate path was marked higher rather than lower during the same session in which gold rose. Three corrections are made in this edition and they are named rather than quietly applied. The 114.50 dollar Brent print that the 23:05 version of this file carried as the January contract is the December 2026 contract, and the January 2027 contract was 109.62 dollars; the split of today's ten-year move into breakeven and real rate cannot be computed from published data at this hour, because the Treasury's inflation-protected curve for October 26 is not published yet, and the 23:05 version stated that split as fact; and the range of December hike prices quoted by other desks was restated from 78 to 83 percent to 77 to 80 percent, which is the range the named boards actually carry. What is deliberately absent is anything that does not yet exist. There is no New York settlement price, no October 26 close for the Shanghai night session, no CFTC report covering this week's liquidation, and no SPDR tonnage figure that has been unambiguously dated for October 26. Every price on this page is a live or last-published print with its own timestamp, and where two boards disagree the difference is disclosed rather than averaged away.

The single most useful fact on this page is that the December path and the gold price moved in the same direction today, and that no Federal Reserve official spoke at all. The Federal Reserve has been in blackout since the weekend of October 17 and decides on Wednesday October 28 at 14:00 ET with Chair Warsh speaking at 14:30 ET, so the only new information in this coverage window is data, market pricing, and one Treasury announcement that has nothing to do with policy. On the site's own CME FedWatch composite at 05:57 Beijing, the October meeting was priced at a 90 percent chance of no change and December at 23.8 percent unchanged against 68.9 percent for a cumulative 25 basis point hike and 7.3 percent for 50 or more, which is 76.2 percent for at least one hike against 74.2 percent on the October 23 close. That is the opposite of the direction the previous edition recorded, when the same measure had been cut from 93.6 percent to 74.2 percent in three sessions. Other desks quoted the December measure between 77 and 80 percent, and those are different strike dates and different feeds and are stated rather than averaged; the 23:05 version of this page put the range at 78 to 83 percent, which no named board carries. Two United States releases landed inside the window and both were stronger than expected: Chicago Fed national activity at 0.07 against minus 0.1 expected with the prior revised down from minus 0.08, released at 08:30 ET, and Dallas Fed manufacturing at 13.4 against 7.0 to 8.8 expected from a prior of 9.8, released at 10:30 ET. The constructive reading comes from the energy side: Brent's December contract down nearly 2 percent to about 114.48 dollars after Qatari mediators were reported in Tehran means the inflation impulse is easing while activity holds up, which is what lets the Fed hold without hiking. For a metal that has repeatedly failed to respond to the rate path, the important consequence is that today's rally was bought with oil, and with a dollar that contributed nothing rather than a lower discount rate. The 23:05 version of this page stated that the whole nominal decline came out of breakeven inflation while the real yield was unchanged near 2.65 percent; that claim is withdrawn, because the Treasury publishes its daily inflation-protected curve only after the US close and no October 26 real yield existed at that hour. Alan Blinder, a former vice chair and not a Fed voice, argued on October 26 for forward guidance that hints the Fed is pausing while standing ready to raise again, and the reported projection summary still carries four members wanting a further 50 basis points of tightening by year end, so a unanimous statement is not available. Wednesday's statement treatment of the energy shock is the item that matters, not the 25 basis points everyone has already agreed to withhold.

The most important currency fact of October 26 is that the dollar contributed nothing to gold's rally, which removes the mechanism the previous edition used to explain the metal's weakness, and the boards do not even agree on which way the dollar moved. The CNGOLD closing board carries the dollar index at 101.6062, down 0.2002 percent against a previous New York close of 101.8100, and the Reuters wire had it little changed at 101.65, while the ICE dollar-index continuous contract traded at 101.91 and then 101.89, up 0.10 and 0.07 percent against that feed's own previous reference level. Those are two different calibers of the same index, and the honest reading is that the dollar was flat to marginally higher on the day, not falling. That matters for the argument on this page: if the dollar had fallen, gold's recovery would have had a second cause, and the oil leg would be doing half the work. On the contract feed it is doing all of it. The European session had the euro at 1.1223 and sterling at 1.3385, both helped by relief that France escaped a Moody's downgrade with the rating affirmed at Aa3 and a negative outlook despite social unrest; Reuters' Hong Kong print had the euro at 1.1218, and by the site's own quote file at 00:17:37 Beijing the euro was back at 1.1177 and sterling at 1.3347, so the European relief rally did not survive the New York open. The exceptions are the two that carry information: the dollar was 0.8237 against the Swiss franc, so the franc firmed while equities and gold both rose, which is not the behaviour of a dollar-denominated margin call, and the dollar was 1.4001 against the Canadian dollar, essentially unchanged, even though Brent's December contract fell nearly 2 percent to about 114.48 dollars. A market that does not buy the Canadian dollar on a 2 percent fall in crude is a market that does not treat 114 dollars as an inflation shock, which is precisely the condition that lets breakeven inflation compress while the real yield does not move. The yen was the outlier and it weakened, 157.575 against its 156.818 reference close after holding a one-month low of 156.08 on Friday, which is what a funding currency does when the ten-year nominal yield falls to about 4.97 percent. Offshore the renminbi was 6.6702, and the onshore gold contracts closed lower while offshore gold rose, continuing the pattern the previous edition identified, that Shanghai falls faster than the international price. The one domestic fact that cuts the other way is the listed equity market, which closed up 4.16 percent in ChiNext while the gold names in it fell 6 to 9 percent.

The single most consequential number published since the previous edition is that the global physical-gold ETF complex is at its highest level in ounces since mid-2022, and it reached it during a week in which the metal fell about 4 percent. Fourteen consecutive weeks of inflows took holdings to close to 102 million ounces, and Kitco put the number at 101.4 million ounces after a 177,400 ounce addition on Thursday, with year-to-date net purchases of 2.43 million ounces. The World Gold Council put September's global fund inflows at 10 billion dollars, with Chinese gold exchange-traded funds adding 11 tonnes, and China's official gold holdings rising from 76.73 million fine troy ounces in August to 77.47 million in September, an increase of about 23 tonnes that the Council described as the People's Bank's largest addition since October 2023. Against that, the one vehicle the previous edition used as its proof that the ETF bid had stopped reacting to the tape, the US-listed SPDR trust, took a 1.56 billion dollar one-day outflow, its largest since March 4, leaving assets at 136.1 billion, their lowest since August 4, and Saxo counted gold ETF holdings down 12.8 tonnes into the end of the week. Those two facts are not contradictory once the geography is specified, and the geography is the story: the money that kept arriving was in products sold outside the United States, including the domestic Chinese funds that took inflows for a thirteenth consecutive session, while the American wrapper was sold. The previous edition's framing that the marginal buyer had changed identity was one step from being wrong; it had changed address. The official bid is also accelerating rather than merely present: Goldman Sachs estimates global central bank purchases at 79 tonnes in August, more than four times the pre-2022 monthly average of about 17 tonnes, with China the largest identifiable buyer at roughly 25 tonnes, and the World Gold Council has September global gold ETF holdings up 67 tonnes while CFTC managed-money net longs fell about 84 tonnes. And the position that would settle the argument is not published until Friday October 30. The last CFTC report covers the week ended October 20 and shows managed money net long at 104,206 hands, down 2,883, with total open interest up 5,934 to 402,279, the long leg 3,777 hands lower against a short leg 894 hands lower, calculated net length at 25.90 percent of open interest against 27.02 percent a week earlier, and the book 32,565 contracts below its September 1 level after eight straight weekly cuts, so the tape has run ahead of the positioning data for a fifth session running. Against all of that sits a named central bank seller, Kazakhstan, taking roughly three percent of 386 tonnes to rotate into Treasuries, and the oil premium that produced today's rally is worth roughly 20 dollars a barrel against mainstream model fair value of 94 dollars and can be withdrawn by a single headline, because the Houthi attacks on Saudi energy facilities continued over the weekend. One further note on caliber, which this edition has corrected rather than inherited. The 23:05 version of this page carried the October 23 ten-year inflation-protected real yield at about 2.76 percent next to a 4.995 percent nominal close and a 2.32 percent breakeven, and those three cannot be reconciled with one another, since 4.995 less 2.76 is 2.235 rather than 2.32. It then replaced them with a different single-source set, 4.9945 percent nominal, 2.6415 percent real and roughly 2.353 percent breakeven, and computed every decomposition in this report from that set alone. Both calibers are recorded here, the Treasury's own par curve and the market-close set, and neither is an October 26 number, because the October 26 inflation-protected curve is not published until after the US close.

The map from the weekend edition did not survive its own test, and the way it failed is worth stating precisely because it is the same failure mode the previous edition had already diagnosed in the metal itself. That edition said 4,000 had become overhead supply, 46 dollars above the settlement, that it was the nearest thing to overhead supply anyone trading this metal on Monday would look at first, and that the first hour would establish which side of 4,000 the market believed it was trading on. It also said 3,970 was the only bid that had not been traded, which was what made it both important and fragile. What Monday actually did was open at 3,995.14, print 3,978.01 in the Asian morning, and then reclaim 4,000 for the rest of the session, running as high as 4,026.68 in the New York hour with 3,970 never tested. The level that was written as a ceiling was defended on the way up, and the bid that was written as untested was not reached, and the reason is the same reason the metal itself has been hard to pin down for a month: a level that was simultaneously a support and a trigger cannot be used as either. The pivot itself has moved 24 dollars higher, from the 3,995 band to 4,019.70, which is where the structural criticism now applies, because that new pivot was set by a live print rather than by a settlement and therefore carries no volume under it, exactly as 3,970 did not. Above, the first rung is now a number the market has already traded once today, 4,033.58 is the October 23 high, the January 2027 COMEX contract at 4,049.40 and the 4,060.40 reference carried on October 22 sit above it, and the futures-implied zone sits inside the Chandler band at 4,065 to 4,075, which is where a continuation would have to go to become a confirmation on the Saxo scale of 4,230. TD Securities puts the level at which systematic trend followers could start covering shorts at 4,092, which is above every one of those rungs and is the number the auction of the next leg of this move would have to clear. Below, the sequence is 3,978, then 3,970 untested, then the October 23 low at 3,975.70, the June low at 3,960 to 3,940 and the 3,942.43 low of the year. The new structural level is the SPDR trust's 17 July low of 999.02 tonnes, which the FXEmpire desk treats as the point at which its 3,560 dollar scenario becomes the relevant one. In Shanghai, the main contract closed at 862.58 and the T+D at 859.90 while offshore gold rose, and the night session then went the other way to 862.88 at 22:23, so the domestic bid arrived after the offshore move, and the listed gold equities in that market fell 6 to 9 percent on a day the broader A-share market rose 4.16 percent in ChiNext. The one timing fact to carry into the close is that this session's low was made eleven hours before its high, in Asia rather than in New York, so the risk was taken in the Asian hours and the money arrived in the American ones.