Gold bounced off a low near $4,110. The bounce is a fact. The end of the selloff is not. On Monday, September 28, 2026, spot and futures feeds showed the metal trading down into the low $4,110s. On Tuesday, September 29, buyers lifted it. How far they lifted it depends on the feed and the hour. The lift did not put the price back where it closed on Friday. A rebound that does not repair the break is a pause. A pause is not a verdict.
This piece has one theme. A gold rebound answers where the price was. It does not answer why it fell. The selloff was driven by the real return on cash, by a rejected truce that failed to give gold a war bid, and by Chinese futures length coming off before a holiday shutdown. None of those three was repealed by a bounce of a few tens of dollars. Until the cause turns, the question in the headline stays open. The selloff can be over. It can also be early. Tuesday did not choose.
This is news and context. It is not investment advice. It is not a gold price forecast you should trade. It is not a call to buy or sell gold, gold stocks, or any miner. A low near a round number attracts stories. Stories are not bids.
What the low was, and what the bounce was
Start with the tape, because the tape is checkable and the feeds do not agree. One futures series showed Monday’s range with a low of $4,111.57 and a close near $4,127, after an open near $4,260. Another spot series put Monday’s low near $4,115 and the close between about $4,115 and $4,133. A third description of Monday’s close used $4,114.99 and a drop of about 4 percent from Friday’s close near $4,285. The headline’s $4,110 is the round number of that low. Use the round number as a label. Use the feed when you need a print. Do not average them into a cleaner story than the market printed.
Tuesday’s gold prices today were a bounce, and the size of the bounce was modest. Early in the Asian day, one report had the spot gold price nearly flat near $4,118. Later, a physical-market note put spot near $4,163, up a little more than 1 percent. Other live feeds in the same window were still in the mid-$4,140s. A bounce of roughly $30 to $50 off the low is real. It is also small next to Monday’s damage and small next to Friday. Friday’s close, on the series that used $4,285, is still about $140 to $170 above the Tuesday prints in that cluster. A gold price recovery that leaves Friday overhead is a gold rebound. It is not a gold rally that has repaired the week.
The larger map matters more than the Tuesday tick. One close series put the January 29, 2026, all-time closing high near $5,405. Monday’s close near $4,115 sat about 24 percent under that high, and about 5.8 percent lower for 2026 to date, inside a 52-week range of roughly $3,825 to $5,405. A metal can be far above where it traded a few years ago and still be in a gold price decline this year. Both sentences are allowed. Using only the flattering one is how a bounce gets sold as a turn.
Silver tells you the bounce was not a broad rush back into precious metals. On the same Monday close series, silver fell about 5.7 percent to near $60.64. The gold-silver ratio finished near 68. Tuesday notes had silver near $61 while gold did more of the lifting. If buyers were fleeing every metal, silver would not have been the one that fell harder on Monday and lagged on Tuesday. The day looked more like a gold repricing than a panic out of every rock. A repricing can still have further to go. It is a different animal from a crash in the whole complex.
One theme
The theme is the fee. Gold pays no interest. Cash, after inflation, does. When that real return rises, holding the bar has a cost. Goldman Sachs precious-metals specialist Adam Gillard put it on Monday without decoration. Gold can hedge inflation, fiscal concerns, and geopolitical risk. When cash suddenly offers a very large positive real return, the opportunity cost becomes difficult to ignore.
Tuesday’s bounce did not shrink that fee. Front-end real rates were the part of Monday that did not reverse when oil headlines did. Published accounts on Tuesday still had the dollar a little firmer, with one reading of the dollar index near 101.3, and Treasury yields still elevated. Market-implied odds of a Federal Reserve rate hike at the October meeting were still described around 70 percent. Fed interest rates are the pipe. The real yield is the water. The water was still running against a metal that pays nothing.
So the honest split is this. Can gold rebound after the $4,110 low? It already did, for a session, by tens of dollars. Is the gold selloff just beginning? Nobody gets to answer that from a bounce. The selloff is a process that ends when the fee falls, or when the sellers who must leave have left. A higher print on Tuesday shows that someone bought the low. It does not show that the fee has changed its mind.
What Monday broke
Monday was the third-biggest daily drop since the leverage washout at the end of January. That comparison is about borrowed futures, not about a new monetary era. A gold price correction built on leverage comes out faster than a correction built on bars in a vault. Open interest and margin are the vocabulary. Civilization is not.
The chart damage was specific. On the note that tracked Monday’s break, spot near $4,135 sat under a 50-day average near $4,266 and under a 100-day average near $4,319. The 200-day average was still below the price, near $4,046. Gold moving averages are memories of old closes. Breaking the two shorter ones says the last two to four months of buyers are under water. Holding the long one says the last forty weeks have not been surrendered. A market can be damaged and not broken. Monday was damaged. Tuesday, still well under the 50-day, did not undamage it.
That is gold chart analysis in one paragraph, and it should stay that plain. Gold technical analysis names levels. It does not cause them. The cause was the fee, plus two catalysts that arrived on time. President Trump rejected Iran’s proposal for a seven-day truce. Chinese traders cut gold length into a holiday. Oil whipped around on Hormuz headlines and did not drag gold back and forth with it. Gold went one way. One-way traffic means the driver did not reverse when the headline did.
The China cut was measurable. Aggregate open interest on the Shanghai open fell by about 11,000 contracts, about 2.6 percent, in the day session. Absolute volume across the Shanghai Futures Exchange and the Shanghai Gold Exchange together was not a record panic. Gillard added the correction a sloppy reading misses. An onshore position that is long one Chinese venue and short the other can inflate futures open interest. The speculative long is probably overstated. It is still a long. Elevated length into a shutdown is length with a deadline.
The deadline is close. The Shanghai Futures Exchange is shut from October 1 through October 7. Today is September 29. The door closes in two days. Gillard did not think traders wanted to hold that length over a week of Iran headlines and live rates. He expected more liquidation, and the note said follow-through was already printing outside China. A Tuesday bounce in London and New York can happen in the same week as a further cut in Shanghai. Both can be true. The bounce is a price. The holiday is a clock. Clocks do not care about a $40 rally.
People also asked
Is the gold selloff just beginning?
Is the gold selloff just beginning? It might be. It might not. The useful version of the question is narrower. Has the cause stopped? The cause was a two-year real rate back around its highest level in more than two years, with inflation expectations not rising enough to offset it. Add a rejected truce that failed to produce gold safe haven demand on the day it should have been most valuable. Add Chinese length that still has a holiday exit in front of it. If those three are still in force after Tuesday’s bounce, then the selloff has a right to continue. Beginning is the wrong word for a move that is already 24 percent under the January closing high. Continuing is the live word.
What would make “just beginning” false? A real yield that gives back the last few weeks. A truce that sticks and lets oil settle, so the hike fear tied to oil cools. A Chinese exit that finishes in an orderly way and stops, because the true speculative pile was smaller than the chart. Any one of those can turn a gold market correction into a gold price recovery. None of them was delivered by the Tuesday print. Until one is delivered, treat “the low is in” as a hope. Hopes are allowed. They are not levels.
Can gold rebound after the $4,110 low?
Can gold rebound after the $4,110 low? Yes. It did. Markets rebound off round numbers all the time, because orders live there, and because a low that looks like a floor invites the first brave bid. The rebound does not need a new theory of money. It needs a seller who is done for the hour and a buyer who wants a cheaper hedge than Friday’s price. That trade can last a day. It can last a week. It becomes a gold trend only if the fee stops rising and the holiday selling does not resume on the screens that stay open.
There is a harder version of the same answer. Gold can rebound and still be in a gold breakdown of the shorter averages. Price can rise $50 and remain $130 under the 50-day. Gold momentum on a one-day chart can look up while gold momentum on a one-month chart looks down. Both charts are honest. They are honest about different clocks. Traders who pick the clock that flatters the bounce will call a turn. Traders who pick the clock that flatters the damage will call a trap. The adult reading uses both clocks and refuses to let either one issue a gold price prediction.
Interest rates are still the boss
Interest rates and gold are an old couple, and this month they are not getting along. Higher Fed interest rates raise the yield you skip by holding a bar. The market is not waiting for the October meeting to discover this. It is positioning before the meeting. Accounts on Tuesday put Fed rate hike expectations for October around 70 percent. Odds move. A 70 percent probability is not a decision. It is a price on a decision. The Fed policy outlook that matters for the metal is the path of real yields, not the adjective in a speech.
Federal Reserve gold is not a program. The Fed does not set the gold price. It sets a policy rate, and the market sets the real rate by subtracting what inflation is expected to do. Inflation and gold rhyme in the textbooks when expected inflation jumps and the real yield falls. That is not this week’s rhyme. Expected inflation did not jump enough. The real yield rose. The textbook hedge was outbid by cash. Anyone selling a gold market forecast that ignores that outbidding is selling last year’s chapter.
Oil is the bridge, and the bridge is still crowded. A physical-market note on Tuesday said crude was still above $100 after Washington rejected Iran’s offer to reopen the Strait of Hormuz. Dear oil keeps the inflation scare alive. It also keeps the hike scare alive. Gold wants the first scare and hates the second. On Monday the second won. On Tuesday oil did not hand the metal a clean de-escalation. August had been the friendly mix: oil down, hike fear down, haven demand still alive, gold up off the lows. This week has not reprinted August. A gold outlook 2026 that copies August onto September is a collage.
Global economic uncertainty is real and still not a blank check. A strait, a shutdown in Shanghai, a central bank meeting, and a dollar that is firm rather than collapsing are four different uncertainties. Only some of them bid gold. Uncertainty about rates bids cash. Uncertainty about a border can bid gold if it does not also bid the hike. Lump them together and you will buy the costume of the headline and miss the fee. The gold market outlook runs through the fee first and the headline second. A gold price outlook that starts with the headline and never reaches the fee is a story about the wrong boss.
The safe haven that did not pay on the war day
Gold safe haven demand is the story people reach for when a truce fails. It failed the test on Monday. The rejection of a seven-day truce was the sort of line that is supposed to put a bid under the metal. It did not. Oil moved both ways on the headline stack. Gold moved down. A haven that only works when the shock is not also a reason for tighter policy is a part-time haven. Part-time is allowed. Calling it full-time on a day it was off duty is marketing.
The slow bid is a different fact, and it should not be erased to make Monday neater. A physical-market note, citing the World Gold Council, said the People’s Bank of China added 20.2 tonnes in August, the largest month since October 2023. The same note, citing customs data, said mainland gold imports were about 1,000 tonnes in the first seven months of 2026, up about 78 percent from a year earlier, even with local prices much higher. Those are structural claims about official and import demand. They are not a description of Monday’s futures tape. Official buyers do not have to show up in the same hour as a margin call. They also do not cancel the margin call. Hold both. A gold bull market, if the phrase is going to be used at all, has to survive a year that is down from a January high and a week that broke the 50-day. Survival is not the same as a guarantee that the next low is the last one.
Physical desks reporting steady coin and bar interest on the dip is also a real category, and it is easy to overread. Coin buyers are a slow bid. Futures sellers are a fast offer. On a day of gold price volatility, the fast offer sets the print and the slow bid sets the argument about the next quarter. Tuesday’s bounce can be the slow bid showing up for an hour. It can also be futures covering. The print does not tell you which. The next inventory and positioning reports might. Until they do, do not promote a coin queue into a turn in the gold trend.
Levels that matter, and levels that only look like they do
Gold key levels are few, and they should be written in pencil. The Monday low, near $4,110 to $4,115 depending on the feed, is now a shelf because it was touched and it bounced. A shelf is a memory. It becomes support only if the next seller fails there. Gold support and resistance are not property lines. They are places where orders clustered once. The next cluster can be somewhere else if the fee moves.
Above the price, the resistance that matters is not a mysterious line. It is the damage. Friday’s close near $4,285, on the series that used it. The 50-day near $4,266. The 100-day near $4,319. Gold resistance levels are those three, in plain English. A gold breakout that does not clear them is a bounce inside a break. A bounce inside a break can feel like a gold price recovery in the afternoon and look like noise on Friday. Gold technical outlook, if it is going to be worth reading, says what must be reclaimed and what happens if it is not. Reclaim the 50-day and the conversation changes. Fail under it and Tuesday was a rest.
Below the price, the 200-day near $4,046 is the level that separates a damaged uptrend from a gold breakdown of the longer average. It is about $70 to $120 under the Tuesday cluster, depending on the print. That is not a forecast that price goes there. It is the next memory on the chart if the holiday liquidation Gillard expected actually arrives, and if real yields push again. Gold chart analysis that hides the 200-day because the bounce feels better is analysis with a thumb on the scale.
There is no gold price target in this article. A bank, a chart reader, or a loud account may publish one. Their target is their model. A target that does not say what real yield it assumes is a number without a cause. If you need a conditional sentence instead of a target, use this one. The bounce holds if the Monday low holds and the fee stops rising. The bounce fails if Shanghai’s leftover length hits the screens that stay open between October 1 and October 7, or if the two-year real rate makes another high. That is a gold market forecast only in the sense that it names the switches. It is not a gold price prediction of the next print.
What the bounce does to miners
Gold stocks are a lever on the metal, and a lever works both ways. A $50 bounce in gold can move a miner by more than $50 worth of mood, because the market prices the next year of metal, not the last hour. It can also do nothing, if the market thinks the bounce is a holiday pause and the fee is still rising. Gold mining stocks outlook is not “gold bounced, so the shares must.” It is “gold bounced, and the fee is unchanged, so the shares are a louder version of an unfinished argument.”
Gold mining companies that already pour metal have a buffer the explorer does not. Their costs are the gap. A producer whose all-in cost sits far under $4,100 still earns a wide margin at the Monday low. The margin is smaller than it was at $5,400, and it is still wide next to any normal decade. That width is why a selloff in the metal is not the same event as a crisis at the mine. It is also why the share can fall even while the mine prints cash. Shares price the change in the margin and the fear that the change continues. Cash flow prices the margin itself. Do not confuse the two when you look at a red quote.
Junior gold stocks and gold exploration stocks do not have that buffer. A junior lives on the next hole and the next placement. A bounce of $40 does not fill a treasury. A further slide toward the 200-day does not empty a good deposit, but it can empty the financing window. Canadian mining stocks and Canadian gold stocks include both kinds of company, on the Toronto exchange and on the venture market. TSXV gold stocks are, on average, closer to the financing window and further from a mill. A gold rebound that does not change the fee does not change that difference. Treat them as different instruments that share a word.
Gold stocks to watch, if the phrase is going to mean anything this week, are not a list of buys. They are names with a dated test. Does the company say what it costs to produce an ounce, and is that cost still comfortable under $4,100? Does it need to sell shares before the October Fed meeting? Does its largest asset depend on a gold price target that assumed January’s high? Those questions are a watchlist. They are not a recommendation. This desk does not name a winner. A winner is a cost, a balance sheet, and a country, checked one at a time. A headline about $4,110 does not do the checking.
Sentiment is a crowd, not a signal
Gold market sentiment on a bounce day splits into two crowds that both sound certain. One crowd says the low is in because the price turned. The other says the low is a trap because the cause remains. Certainty is the tell that someone has picked a clock and thrown the other clock away. Gold price volatility is what you get when both crowds are large and the holiday clock is two days from striking. Volatility is not a view. It is the argument, priced.
The January shadow should keep the certainty humble. That washout was leverage leaving. Monday rhymed with it without matching it. If Shanghai’s remaining length comes out into a shut domestic market, the rhyme can get closer. If the length was overstated and the rest of it can sit still, the rhyme fades and the bounce gets room. Positioning reports will age this paragraph. Until they do, gold market sentiment is a weather report. You do not build a house on a weather report. You decide whether to carry leverage through a storm.
A useful mood check is the question a holder can answer without a chart. What does cash pay me, after inflation, to not hold this hedge? If the answer is “a lot, and more than a month ago,” the bounce has not changed the personal math. If the answer is “I do not know,” the holder does not know the fee Gillard said had become hard to ignore. Not knowing the fee is how people experience a gold selloff as a betrayal instead of as a comparison. The metal did not betray the hedge. Cash won the comparison for a week. Comparisons can flip. They do not flip because a round number bounced.
Three ways the next ten days can look
These are paths, not predictions. The October 1 shutdown and the October Fed meeting sit inside them. So does the fee.
One. The Monday low holds, Shanghai’s exit is smaller than feared, and the real yield stops rising. Then Tuesday’s bounce was the start of a repair, and the work of repair is still the climb back through $4,266 and $4,285. That climb can take time. A gold rally that reclaims those levels would change the gold technical outlook. It would not, by itself, make January’s high the next stop. The fee would still have to cooperate. This path is the one the bounce crowd is already spending. It is possible. It is not yet paid for.
Two. The low breaks after the Shanghai door shuts, because the selling that still wants out has to print in London and New York. Real yields push, or oil keeps the hike odds pinned near 70 percent. Then the bounce was a pause before a test of the 200-day near $4,046. That test can fail, and it can hold. Holding it would be a different, uglier kind of rebound. Breaking it would be the gold price correction turning into a break of the longer average. This path is the one Gillard’s holiday warning points at. A warning is not a fate. It is a reason to know the date.
Three. The price chops between the low and the 50-day while everyone waits for the Fed. Chop is the path that fools both crowds. The bounce crowd declares a turn every afternoon. The breakdown crowd declares a crash every morning. The fee does not move enough to choose. Gold price volatility stays high and the net change is small. This path is common after a shock, and it is the path that punishes oversized bets in both directions. Small is a strategy in a chop. Loud is a cost.
What would end the selloff, and what would extend it
The selloff ends, in the sense that matters, when the fee falls or the forced seller is gone. A stuck truce that lowers oil and lowers hike fear would be a candidate. A softer Fed path than the 70 percent October odds would be a candidate. A finished Chinese exit, visible in open interest that stops falling, would be a candidate. Two of the three would be stronger than one. Price making a higher low near $4,110 is a candidate only if it is accompanied by one of those. A higher low alone is a picture. Pictures need a cause.
The selloff extends if the fee makes a new high, or if the holiday exit is larger than the Tuesday bid. It also extends, more slowly, if nothing dramatic happens and the 50-day simply stays overhead while rallies are sold. Extension does not require a crash. It requires the absence of a repair. A market can grind. Grinds are how a gold market correction becomes a season rather than a headline. Seasons are harder to write about and easier to lose money in, because they do not feel like the day you were warned about.
Official demand can cushion an extension without ending it. If the August pace of official buying is still in force, dips have a slow buyer. Slow buyers do not stop a futures week. They change the price at which a futures week becomes interesting to them. Confusing a cushion with a floor is how people hold leverage through a holiday and call it conviction. Conviction that cannot survive a margin call is a story about size, not about geology.
How to read the next print
This is a filter, not a plan.
Read the real two-year before you read the gold headline. If the real yield is up and gold is up, the bounce is fighting the fee and may be a squeeze of shorts. If the real yield is down and gold is up, the fee is helping and the bounce has a cause. If both are down, you are watching something else, and the something else may be the holiday. Different causes, different lifetimes. A fee can last months. A holiday exit expires on October 7. Do not give them the same shelf life.
Read who can trade. From October 1 to October 7, Shanghai futures are shut. A gold move in that week is a move without the market that cut length on Monday. It can be quieter. It can be sharper. Gillard’s bias was the sharper one. A bias is not a fact until it prints. The shutdown is the fact. Put the dates on the card. Gold key levels without dates are decorations.
Read the averages as a scoreboard, not as a prophecy. Under the 50-day and above the 200-day is the current score. A close back over the 50-day changes the score. A close under the Monday low changes it the other way. Gold technical analysis that updates the score is useful. Gold technical analysis that promises the next quarter from one candle is a costume.
Read miners only after the metal’s cause is clear, and then read their costs. A producer with a wide margin under $4,100 is a different risk from a venture listing that needs a placement. Canadian gold stocks are not one trade. Gold exploration stocks are not a producer in waiting unless the treasury and the drill results say so. Gold stocks to watch means names that can answer those questions in a filing. It does not mean names that moved the most on the bounce.
What this does not mean
It does not mean the low is in. A bounce off $4,110 is necessary for a low and nowhere close to sufficient. Necessary is the part the Tuesday crowd will repeat. Sufficient is the part they will skip. Skip it and you have a slogan.
It does not mean the selloff must continue in a straight line. Markets do not owe Gillard’s warning a second down day. They also do not owe the dip buyer a refund of January. Both debts are imaginary. The live facts are the fee, the holiday clock, and the averages that were broken and not reclaimed.
It does not mean gold has failed as a hedge for a decade. A part-time haven can be the right holding for a person who has sized it as a hedge and knows the fee. It is a bad holding for a person who bought it as a substitute for cash on the week cash started paying more. The error is the job description, not the metal. Inflation and gold can rhyme again when expected inflation outruns yields. This week they did not rhyme. A week is not a decade. A decade is not an excuse for misreading the week.
It does not mean a miner is cheap because the metal dipped. Cheap is future cash against today’s share price. The dip changes the metal assumption. It does not audit the mine. Gold mining stocks that fell with the metal may be pricing more damage than the mine will feel, or less. You cannot know which from a gold rebound headline. You can know that leverage cuts both ways, and that venture shares add a second leverage called dilution.
It does not mean this desk has a gold price target, a gold price forecast, or a gold price prediction. The conditional sentence above is the whole forecast: the bounce holds if the low holds and the fee stops rising; it fails if the holiday selling or a new high in real yields arrives. Anyone who turns that sentence into a number has left the evidence.
The close
Gold rebounded from a low near $4,110. On Monday, September 28, feeds showed a low in the low $4,110s and a close near $4,115 to $4,130, after Friday near $4,285 on one series. On Tuesday, September 29, the spot gold price was higher, somewhere from the high $4,110s early to the mid-$4,160s in later notes, with other feeds still in the mid-$4,140s. The bounce is real. It is tens of dollars. It is not Friday. It is not the 50-day near $4,266. It is not the 100-day near $4,319. The 200-day near $4,046 was not tested. The January closing high near $5,405 is still about a quarter above the low.
The cause of the drop did not bounce. Real yields at the front end were still the fee Gillard named. October hike odds were still described around 70 percent. The dollar was still a little firm. Crude was still dear after a rejected offer to reopen Hormuz. Shanghai still faces a shutdown from October 1 to October 7, with speculative length that is probably overstated and still, in his view, too large to hold comfortably. Official buying and import demand can be real and still be too slow to set Tuesday’s print.
Is the gold selloff over or just beginning? Over, only if the fee or the forced seller has finished. Just beginning, only if you ignore that a quarter of the January high is already gone. The accurate sentence is duller, and it is the theme. The rebound is a location. The selloff is a comparison between a metal that pays nothing and cash that, this month, pays a real return. Tuesday did not flip the comparison. It only showed that $4,110 found a bid for a while. A bid is not a pardon. The fee is still the boss.
Important information
This article is for information and education only. It is not investment advice, tax advice, or legal advice. It is not a recommendation to buy, sell, or hold gold, silver, gold stocks, Canadian mining stocks, junior gold stocks, TSXV gold stocks, or any other asset. Gold stocks to watch, here, means names with a dated public test. It is not a buy list. There is no gold price target in this piece. Commodity prices are volatile. Leveraged futures can cost more than the cash posted. Miners face cost, political, operational, and dilution risk. A bounce can fail. Past prices do not indicate future results.
Price levels differ by feed and by hour. Monday’s low near $4,110 to $4,115, closes near $4,115 to $4,133, Friday references near $4,285, the January 29, 2026 closing high near $5,405, the roughly 24 percent drawdown, the year-to-date figure near minus 5.8 percent, and the 52-week range near $3,825 to $5,405 are drawn from published spot and futures series on September 28 and 29, 2026, which did not match to the dollar. Tuesday prints cited here range from about $4,118 in an early report to about $4,163 in a later physical-market note. Moving-average levels near $4,266, $4,319, and $4,046, the third-biggest drop since late January, the Shanghai open-interest drop of about 11,000 contracts or 2.6 percent, and the October 1 to October 7 closure follow a September 28 market note citing Adam Gillard of Goldman Sachs. October hike odds around 70 percent, a dollar index near 101.3, and crude above $100 are from published market accounts on September 29 and can change. Official-sector and import figures are attributed to a physical-market note citing the World Gold Council and customs data. They are not findings of this desk. Readers should check live quotes and primary sources. This note does not consider any person’s goals or finances.