VIEWTHE SHORT VERSION — FULL ANALYSIS FOR THE WORKINGS
79Au197.0
Gold- O-Matic
Eight forces. One gauge. No PhD required.
Most gold coverage tells you the price. This shows you why. Eight macro forces, weighed against each other, give you a single read on where the pressure is and which way it's pointing.
PAST 7 DAYS · computed from the site’s own daily record.
Refreshed twice daily · weekly · quarterly
LATEST · WHAT MOVED
This week's headline
The single line that best explains today's price
Today's read
Central banks are carrying this market close to alone. The haven premium is bleeding out, and real rates have stopped mattering.
Gold closed August at $4,563 an ounce, up more than a third over twelve months and back within reach of February's record. It has done that with the real 10-year yield at 2.42% — a level that would once have been a clear headwind, but which sits around the 60th percentile of the 1971–2026 record, so this dial scores it as neither help nor hindrance. The lifting is being done by one buyer: central banks took 289 tonnes in the second quarter, 62% more than a year earlier. Note what is not helping. Exchange-traded funds sold 45 tonnes over the same quarter, jewellery demand fell to its weakest quarter on record, and the Iran-Ukraine premium is fading rather than building. Official-sector buying is now 46% of the total pull on this gauge. One force is doing nearly half the work, and that is worth watching precisely because it has not been tested by a reversal.
written 1 Sept 2026
REGIMEStructural / haven driven
What to watchNEXT 14 DAYS · 3 EVENTS
September FOMC · Sep 15–16
Decision plus a fresh . The Warsh Fed has now held at 3.50–3.75% for five consecutive meetings, and July's vote was 9–3 — with all three dissents preferring a hike, not a cut. The dot plot is where that pressure becomes visible.
Fed independence · ongoing
The administration's move against Governor Lisa Cook puts the Fed's autonomy in front of the courts. This gauge has no factor for central-bank credibility, so it would register only indirectly — through official-sector buying, or through real rates if the path of policy actually changes.
August CPI · Sep 11
Released 8:30 am ET. Feeds breakevens and the real-yield path into the September meeting. A hot print raises the inflation factor and the Fed's resistance at the same time; the two partly cancel, which is why inflation carries only 12% here.
THE READING
Pressure gauge & tug-of-war
Composite score and who is pulling hardest
WEIGHTED BY:
PRESSURE GAUGEcomposite reading · −30 to +30
Higher = more upward pressure on gold
THE THREE WEIGHTINGS SPAN +2 TO +7
TUG-OF-WARwho's pulling, how hard
Pulling Down
Tug-of-war on price8 forces · 7 pulling · 1 monitoring
Pulling Up
Real yields · FOMC
Jewelry · India/China
Production · costs
Au
Official sector
Risk · haven flows
CPI · breakevens
Investor positioning
Monitoring · tracked but not pulling right now
US Dollar118.1 broad dollar index · as of 2026-09-04
0.30total down-pull
+0.52net
0.82total up-pull
HOW THE NUMBER IS BUILTeach force's contribution, summing to the needle
−30−150+15+30
·Baseline
0
0.0
▲Central Banks
4.4
+4.4
▲Geopolitics & Haven
6.2
+1.8
▼Real Rates & Fed
4.6
-1.6
▲Inflation
5.8
+1.2
▲ETF Flows
6.6
+0.8
▼Physical Demand
5.8
-0.8
▼Mine Supply
5.2
-0.6
–US Dollar
5.2
0.0
=Needle
5.2
5.2
Starts at 0 · each force adds or subtracts gauge points · sums to the needle
THE UNDERLYING QUESTION
What actually moves the price?
Why no formula catches gold, and what to watch instead
Gold produces nothing
A share of a company pays you a dividend. A bond pays interest. A building pays rent. For all of those you can add up the money they throw off, argue about it, and reach a number that means something — a value the price can be measured against.
Gold pays nothing. It has never paid anything and it never will. An ounce in a vault produces exactly as much income as an ounce in the ground, which is none.
That has one consequence, and everything else on this page follows from it: there is no fair value for gold. There is nothing to calculate. There is only what someone is willing to pay today.
So the price is set by who is buying, and why
If there's no anchor, then "what moves gold" always resolves into a simpler question: who is buying right now, and what are they trying to achieve?
That's not a dodge. It's the actual mechanism. Someone moving money out of a currency they've stopped trusting behaves nothing like someone hedging an inflation forecast, who behaves nothing like a family buying a wedding necklace, who behaves nothing like a fund manager rotating out of bonds. They buy different amounts, at different speeds, for different reasons, and they stop for different reasons.
The price is where all of that meets. Not a formula — a crowd, with a changing composition.
And the answer rotates
Here's the part that catches people out. The dominant buyer changes, and when it changes, so does the rule that appears to govern the price.
For roughly forty years, the dominant buyer was an investor weighing gold's zero yield against what a government bond paid after inflation. When bonds paid well, gold was expensive to hold and it fell. When they paid nothing, gold cost nothing to hold and it rose. That relationship was so reliable for so long that people stopped treating it as a description and started treating it as physics.
Since 2022 the dominant buyer has been central banks — buying at roughly double their long-run pace, moving national reserves out of dollars, with a fear premium layered on top. They don't care what bonds yield. They aren't trading. They buy, and the gold goes into a vault, and it doesn't come out.
Same metal. Different buyer. Different rule. And gold rose through four years of high real yields, doing precisely what the textbook said it couldn't.
Go back further and it's stranger still. Before 1933 gold was money in the United States, fixed at $20.67 an ounce. Roosevelt made private ownership illegal and revalued it to $35. Bretton Woods pinned the world to that $35 in 1944. Nixon cut the link in 1971 and gold floated freely for the first time in modern history. It ran to $850 by January 1980, then fell for twenty years while central banks were net sellers. Exchange-traded funds arrived in 2004 and brought a new kind of buyer entirely.
Six different answers to "what sets the price," and in two of those eras the answer was Congress.
190019331971198020012022now
1900–1933
Fixed by law at $20.67. Gold was money.
1933–1971
Fixed by law at $35. Private ownership banned in the US until 1974.
1971–1980
Floating for the first time. Inflation and crisis. $35 to $850.
1980–2001
Trust restored. Twenty-year decline, central banks net sellers.
2001–2022
Real rates and the arrival of ETFs.
2022–now
Central banks, and the risk that reserves can be frozen.
Six eras, six different answers. In two of them, the answer was Congress. Pre-1971 line is the statutory price, not a market price.
What the buyer is really buying is distrust
Gold is the only major asset that is nobody's promise. That single fact explains why the buyer keeps changing.
Which is why this page looks the way it does
If the rule changes when the buyer changes, then a single equation can only ever describe one era. It will look brilliant while that era lasts and then quietly stop working, usually without announcing it.
That's why this page gives you eight forces rather than one number, and three competing weightings rather than one answer. The eight are the motives that show up repeatedly. The three regimes are three theories about which motive is currently in charge.
It's also why the backtest on this page shows no forward signal, and why that finding is printed rather than buried. Averaging eight motives is the wrong operation when only one or two are ever driving. The composite describes conditions honestly and forecasts badly, and now you know why — it isn't a weak model, it's the wrong shape of question. What it's good for is telling you who's pushing, how hard, and in which direction. Not where the price lands.
You don't need a PhD for this. You need to know who's buying, and what they've stopped trusting.
Fun facts regarding gold
CONTRIBUTION
Who's doing the work?
Regime-weighted stack of the eight forces
A is the market's ruling story — which force is actually in charge of gold's price right now. For forty years the textbook story was that real interest rates ran the show: rates up, gold down. Since 2022 that stopped working: central banks became relentless buyers and war added a fear premium, and gold rose despite brutal rates. Same eight forces, different boss. Switching regimes re-weights the eight forces according to each story — and how far the needle moves tells you how much the reading depends on which story you believe. And it changes everything: switch the story here and every number on this page is re-scored under it.
The same needle as the top of the page. Switching regimes re-scores everything on this page — the hero dial, every factor card, the tug-of-war — not just this section.
Today's regime: central banks and geopolitics are doing the work. Real rates matter less than they used to.
SHARE OF TOTAL PULL
Official sector39%
Risk · haven flows16%
Real yields · FOMC14%
CPI · breakevens11%
Investor positioning7%
Jewelry · India/China7%
Production · costs5%
Bullish pull Bearish pull Neutral / no pull
DRIVERS · THE EIGHT FORCES
Primary, secondary, and background
Click any card for the deeper read on that force
WEIGHTED BY:
The eight forces compose to this score
IPRIMARYdoing most of the pulling
Central Banks
Official sector
Quarterly · 2026Q1
UP
22%
steady
have been buying gold by the for four years straight — and they aren't price-sensitive. That's the floor under this whole bull market.
2016–2026 · qtly tonnes
Geopolitics & Haven
Risk · haven flows
Weekly · as of 8 Sep 2026
UP
18%
fading
When investors are scared, they buy gold. Right now there's a war premium baked into the price — and it's slowly bleeding out as the ceasefire holds.
2016–2026 · GPR index
Real Rates & Fed
Real yields · FOMC
Daily · as of 8 Sep 2026
DOWN
16%
risingAbove the backtest range
When you can earn around 2.4% real on Treasuries, holding gold (which pays nothing) costs you something. That cost is real — but measured against fifty-five years of rates rather than the last decade, 2.4% is an ordinary number, so this dial now scores it as neither help nor hindrance.
2016–2026 · %
Inflation
CPI · breakevens
Daily · as of 9 Sep 2026
UP
12%
rising
Sticky inflation makes gold attractive as a store of value. But it also keeps — so the two effects partially cancel.
2016–2026 · % (10Y breakeven)
IISECONDARYmeaningful, not dominant
US Dollar
Broad dollar
Weekly · as of 4 Sep 2026
FLAT
10%
steady
A strong dollar makes gold pricier in every other currency, so the two usually pull against each other; when the dollar slips, gold breathes easier. Where it sits right now is scored live from the broad dollar index.
2016–2026 · index
ETF Flows
Investor positioning
Quarterly · 2026Q1
UP
8%
steady
investors are gold's fast money — they pile in and out faster than any other buyer here. The score follows the most recent quarter's net flow, whichever way it points.
2016–2026 · qtly tonnes
IIIBACKGROUNDslow-moving, sets the floor
Physical Demand
Jewelry · India/China
Quarterly · 2026Q1
DOWN
8%
steady
When gold gets this expensive, ordinary buyers (jewelry, gifts) pull back. That's a small drag on demand — but the floor is still surprisingly firm in India.
2016–2026 · qtly tonnes
Mine Supply
Production · costs
Quarterly · 2026Q1
DOWN
6%
steady
Mines can't ramp up quickly no matter the price, so supply moves slowly and rarely drives the market on its own. Its score tracks the latest quarter's output against recent norms.
2016–2026 · qtly tonnes
TIMINGSome of this changes tomorrow. Some takes a decade.Now vs. Soon vs. Structuraldon't confuse imminent volatility with the slow tectonic shifts
Now· days
August CPISep 11
September FOMCSep 15–16
Fed independenceongoing
Iran ceasefireongoing
Soon· weeks
October FOMCOct 27–28
December FOMCDec 8–9
ECB · BOJ meetingsSep–Oct
Indian festival demandOct–Nov
Q4 CPI/PCE cycleOct–Dec
Russia–Ukraineweeks
Structural· months → years
CB buying regimeongoing 2022→
Dollar de-anchoringmulti-year
US fiscal trajectorymulti-year
BRICS settlementmulti-year
THE TWO HALVES, OVER TIME
Two lines, one scale. The gold line scores the four fast forces together; the sage dashed line scores the four slow ones. Zero means that half is balanced — above it, that half is pushing gold up. The faint line behind them is the gold price, on its own scale, there to show shape only.
What to look for is the gap. When the two lines sit together, the quick half of the market and the slow half agree. When they pull apart — the shaded band — they are telling different stories about gold. Since 2022 the slow half has stayed bullish while the fast half turned hostile, and gold rose anyway. That disagreement is the central claim of this page: the old rate-driven rule stopped being the whole story.
This is not a forecast, and the price line is not proof. Both tracks are scored from the forces themselves — neither is fitted to price, and the price line is not evidence that they predict it.
FAST
real rates · inflation · dollar · geopolitics
SLOW
central banks · physical · ETF flows · mine supply
GOLD PRICE
own scale · shape only, not fitted
COMPUTED · 2026Q1
Both tracks have stayed within 8 points of each other for 4 quarters — fast 10.0, slow 8.6. The gap currently reads 1.4 points.
COMPOSITE SCORE · −30 TO +30
Markers: the March 2020 liquidity panic, and Q1 2022 — when the Fed began hiking and Russian reserves were frozen in the same quarter. The tracks part here.
2026Q1
Scrubbing replays history — it does not change the live gauge above, and readings are scored with today's weightings.
regime aligned · run 4q · gap 1.4 · trailing IC -0.067
FORCES IN ORBITThe eight forces, planetarium viewcloser to the sun = heavier weight in the current regime
Closer to the sun = heavier weight in the current regime. Each planet glows in its direction-color.
SCENARIOS · STRESS-TEST THE READ
What if?
Click any scenario to commit, hover to preview the projected score
WEIGHTED BY:
CURRENT5leans UPhover a scenario for preview
EXPERIMENT · SET THE CONDITIONS
Try it yourself
Set each factor's real-world number and the model scores it exactly as it scores the live data. The gold pip on each slider marks today's actual reading.
WEIGHTED BY:
Right now, switching between the three weightings alone moves the reading 5.4 points — from Real-rate driven (classic) at 1.8 to Crisis / risk-off at 7.2. You haven't touched a single slider yet.
Move any slider to begin
YOUR COMPOSITE
5leans UP
Set the conditions — the model scores the pressure. Each slider is the real-world number itself; the dial only moves when your setting crosses one of the model's scoring thresholds.
Central Bankstoday 240 t
official sellingheavy official buying
240 tBULLISH
Geopolitics & Haventoday 140
calm worldcrisis headlines
140BULLISH
Real Rates & Fedtoday 2.45%
deeply negativepunishingly high
2.45%BEARISH
Inflationtoday 2.35%
low expectationshot expectations
2.35%BULLISH
US Dollartoday 118.0
weak dollarstrong dollar
118.0NEUTRAL
ETF Flowstoday 60 t
heavy outflowsheavy inflows
60 tBULLISH
Physical Demandtoday 770 t
weak demandstrong demand
770 tBEARISH
Mine Supplytoday 880 t
thin mining yearheavy mine output
880 tBEARISH
TRACK RECORD · TEN YEARS · METHOD
Track record & method
The same rules running on the live site, run against ten years of history. Five views: does the track gold, which factors do the work, what fifty-five years of real rates say, how the model actually scores, and what would prove it wrong.
GOLD PRICE Δ, 10Y
+277%
$1,221 → $4,608 (2016Q1 – 2026Q1)
MEAN COMPOSITE
4.5
flat territory for most of the decade
OBSERVED RANGE
-2–11
directional read, not a forecast
TIME IN "LEANS UP"
37%
composite > +5
Backtest covers 2016Q1 to 2026Q1 — the last quarter in which all eight factors have published data. Four of them come from World Gold Council Gold Demand Trends, which publishes quarterly, so the series ends one to two quarters behind the live reading at the top of this page. Gold has moved since this chart ends.
Leaning up · above +5Balanced · −5 to +5Leaning down · below −5
201620182020202220242026
2026Q1. The composite read 9.4 and gold stood at $4,608.35. There is not yet a full year of data after this quarter to compare against.
Reading the chart honestly. The composite is a directional read — a snapshot of which way the balance of forces is pulling gold at each moment — not a price forecast. Its range across the decade is narrow (-2–11), because the underlying forces oscillate; meanwhile gold went from $1221 to $4608.35. So comparing composite level to price level is the wrong test (they're built on different scales). What the chart does show is that the composite's peaks — COVID 2020, Ukraine 2022, and the current central-bank buying stretch — all coincide with periods of clear gold strength. The compressions in late 2018 (Fed hikes) and 2022 Q3-Q4 (real rates spiking) show up as expected too. That's the useful part: the composite reads the room. It doesn't tell you tomorrow's price; it tells you what's currently doing the pulling.
OUT OF SAMPLE · FOUR FACTORS · 2006Q1–2026Q3
The out-of-sample test, and what it says
CORRELATION, 4 QUARTERS AHEAD
−0.57
p < .001 · 79 overlapping windows
RANK CORRELATION
−0.59
the sign is not an outlier artefact
DIRECTIONAL HIT RATE
64%
49 of 77 calls, gold rose in most of them anyway
PURE OUT-OF-SAMPLE 2006Q1–2015Q4
−0.64
years the thresholds never saw · p < .001
Everything else on this tab is scored over the decade the thresholds were built on. This block is not. Four of the eight forces have public data reaching back well before the calibration window, so they can be scored over years the ruleset never saw: real rates, inflation expectations, the dollar, and geopolitical risk. Their weights are renormalised over the four, and the site's own composite function does the arithmetic — same code, longer record.
The result is the strongest honest finding on this site, and it is not a flattering one. The four-factor composite's relationship with the following year's gold return is not weak. It is strongly negative. Over this longer record, the readings this instrument would have called bullish came before the worse returns, and the readings it would have called bearish came before the better ones.
Take that seriously before taking it literally. An inverted relationship is not a hidden trading signal, and reading the gauge backwards would be a mistake: the sign is unstable across weighting schemes and across horizons, the windows overlap so the significance is overstated, and four forces are not eight. What it does establish is that the agreement you can see on the ten-year chart above does not survive contact with data the thresholds were not fitted to.
This is the test the site would most like to have passed. It is published here because a track record that only covers the years you calibrated on is not a track record.
Correlation between the four-factor composite and the gold return over the following one, two and four quarters, under each of the three weighting schemes. Weights are renormalised over the four factors that have data this far back — under the default scheme that is real rates 0.286, inflation 0.214, dollar 0.179, geopolitics 0.321. Windows overlap, so the p-values are optimistic.
Weighting
1Q
2Q
4Q
Hit (4Q)
Structural / haven driven (default)
−0.45
−0.49
−0.57
64%
Real-rate driven (classic)
−0.48
−0.63
−0.79
63%
Crisis / risk-off
−0.28
−0.21
−0.23
65%
Real rates before 2003 are not used here — this backtest starts at 2006Q1, inside the TIPS era, so no spliced series enters it. The fifty-five-year spliced record is on the , where the join is drawn.
Why the composite reads flat when its parts do not
The composite's forward correlation is small and unstable, and for most of this site's life it sat at zero. That is easy to read as “none of this measures anything.” The arithmetic says something more specific, and the table below prints it rather than asserting it.
Taken one at a time, several of the eight forces track forward returns quite strongly over this record. The trouble is the direction. Across these quarters, a bullish real-rates reading and a bullish dollar reading each preceded lower returns, not higher ones. The composite adds every force with a positive weight, so those pull against the others and a good deal of the signal cancels before it reaches the needle.
That is the 2022 break showing up in the arithmetic rather than in an argument. Real yields rose sharply that year and gold rose anyway, so across this particular decade the quarters that scored bullish on rates were the quiet ones and the quarters that scored bearish were the boom. Whether that reverses, or whether the old relationship is simply gone, is the open question underneath this whole instrument.
One thing to hold onto while reading the composite row. The real-rate thresholds were rebuilt in August 2026, after the author had already seen how the previous ones scored on exactly this table. That change lifted the composite's correlation here from roughly zero to mildly positive. A number that improves after you have seen the answer is not evidence, and it should not be read as any.
One further caution. Gold roughly quadrupled across this record, so forward returns correlate strongly with the calendar itself. Remove a simple time trend and most of these relationships shrink sharply — the composite's own falls to almost nothing. A decade that goes almost entirely one way is a poor place to learn what predicts anything.
Correlation of each force’s signed reading with the following four quarters’ gold return, across 37 overlapping windows. “Trend-removed” regresses both series on the quarter index first.
Force
Raw
Trend-removed
Sign
Central Banks
+0.724
+0.574
as added
Geopolitics & Haven
+0.374
+0.149
as added
Real Rates & Fed
−0.789
−0.704
opposite to how the composite adds it
Inflation
+0.318
−0.118
as added
US Dollar
−0.737
−0.516
opposite to how the composite adds it
ETF Flows
−0.234
−0.074
opposite to how the composite adds it
Physical Demand
+0.148
+0.134
as added
Mine Supply
−0.251
−0.164
opposite to how the composite adds it
Composite (current weights)
+0.262
+0.059
the eight, added up
Running opposite to the composite’s own arithmetic: Real Rates & Fed, US Dollar, ETF Flows, Mine Supply.
CROWDING · POSITIONING READ
Crowding
One number, built from four measures of how much buying has already happened: price momentum, ETF flows, physical demand, and speculative . It tells you how much room is left, not which way the price goes. Separate from the gauge — this is not a second pressure reading.
WEIGHTED BY:
GAUGE
5
pressure on the price
−30 to +30
CROWDING
61
how much has already happened
percentile, 0 to 100
Gauge flat or down, crowding hot — a rally running on belief rather than conditions.
Different scales, different questions — the gauge is direction, crowding is how much has already happened.
coldneutralwarmhot
few buyers in yet — room to runmost already in — less room, sharper reactions
WARMpercentile 61
Conviction is building. Enough is already priced that a surprise the other way would sting, though this is short of the extremes that usually precede a sharp unwind.
Price momentum (12m)
+23.7% over twelve months, higher than 70% of readings since 2016
70
ETF flows (quarterly)
+62 t last quarter, below the trailing four-quarter average of +200 t
55
Physical demand (jewellery + bar & coin)
773 t in 2026Q1, ranked like-for-like against every prior Q1 since 2016. Inverted: weak physical demand means the price is being held up by investment demand alone, a more crowded state.
60
Speculative positioning (CFTC)
136.8k contracts, higher than 58% of the last three years
58
Forward calendar · next ~16 weeks, plus what just happened
Sep 15–16
FOMC — decision plus dot plot
Oct 27–28
FOMC — rate decision, no new forecasts
Dec 8–9
FOMC — decision plus dot plot
Recently concluded
Aug 21
OFAC Hormuz waiver expires
OFAC's General License X, which had authorized Iranian oil exports through Aug 21, 2026, was already revoked on July 7 after attacks on tankers in the Strait of Hormuz, with a wind-down deadline of July 17; no waiver was reinstated or extended, so the Aug 21 expiration date passed without any active authorization in place.
The FOMC voted 9-3 to hold the federal funds rate steady at 3.50%-3.75%, marking the fifth consecutive meeting without a change. Cleveland's Beth Hammack, Minneapolis's Neel Kashkari, and Dallas's Lorie Logan dissented, preferring a 25-basis-point rate hike.
Bar length is how hard this page expects the event to push the gauge — two bars, meaningful; three bars, a likely repricing. That is a judgement call, not a forecast.
Historical playbookhow gold has behaved in episodes like this
1970s stagflation
Gold ~+2,300% over the decade
Persistent inflation + loss of confidence = gold's golden era.
Gulf War 1990
+7.5% in 6 months
Wars in oil regions reliably move gold.
9/11 / 2001
+5.9% in 1 month
Crisis flight to safety is fast and binary.
2008–2011 bull
+170% over 3 years
QE and real-yield crush created the post-crisis bull.
Russia invasion 2022
+8.2% in 1 month, then the regime change
The reserve-freeze rewrote the central bank playbook.
LEARN THE MECHANICS6 quick explainersno jargon, 90 seconds each
What does the Fed have to do with gold?
Rates → real yields → opportunity cost → price
Read
Why are central banks suddenly buying so much gold?
Because in 2022, the U.S. froze Russia's reserves.
Read
What's a "real yield" and why do I keep hearing about it?
Nominal yield minus inflation = your actual return.
Read
Physical gold vs. an ETF like GLD — what's the difference?
Same exposure, very different logistics.
Read
Why do wars and crises move gold?
Money looks for somewhere safe that nobody else can revoke.
Read
Why do some forces move gold this week and others take years?
Fast forces set the price. Slow forces set the floor.
Read
PEOPLE ARE ASKING · ASK THE GAUGE
People are asking
Ask about the forces on this page and get an answer drawn from the same data the gauge runs on. If the honest answer is that nobody knows, it says so.
Ask something
Answers are drawn from this site’s own eight factors, their current readings, and ten years of tested history. Ask what is pressing on gold and it will tell you at length. Ask it to pick a direction and it will decline, because nothing here has been shown to predict one.
Answers appear in the list below. Questions are screened automatically before they publish.
If the gauge doesn't predict the price, what is it actually for?
It tells you what is pressing on gold right now, and which forces are doing the pressing. That is a different job from forecasting, and a more honest one. When we tested the composite against what gold did next across ten years, high readings did not reliably precede gains — so the page says so plainly rather than implying a power it does not have. Read it the way you would read a barometer: useful for understanding present conditions, not a promise about next week.
answered 2026 Q3
Why does the dollar matter if gold is already priced in dollars?
Because most of the world does not buy in dollars. When the dollar strengthens, gold gets more expensive for buyers in rupees, yuan and euros — so demand from the largest physical markets softens even though the dollar price has not moved. The dollar is therefore less a driver of gold's value than a tax on everyone else's ability to buy it.
answered 2026 Q3
Central banks have been buying heavily. What happens if they stop?
It would remove the single largest weight in the current model — official-sector buying carries 22% and has been reading strongly bullish. But the honest answer is that we do not know how gold would respond, because the last decade has never tested it. Central bank demand went from roughly 100 tonnes a quarter before 2022 to over 400 at its peak. If that reverses, the gauge will register it within one published quarter, and the page will show a very different reading. That is worth watching precisely because it has not happened yet.
answered 2026 Q3
Why is inflation only worth 12% when everyone calls gold an inflation hedge?
Because over this particular decade it has not behaved like one. Gold's largest run came while inflation expectations sat in a narrow band around 2.3%, and our own testing found the inflation factor indistinguishable from noise against subsequent price moves. The weights reflect what has been doing the work recently, not what the textbook says should. If that changes, the weights should change with it — and the page should say when they did.
answered 2026 Q3
Answers are AI-generated from this site’s data and methodology. They describe conditions; they are not financial advice and not a forecast. The author holds gold-tracking ETFs — see the disclosure at the foot of the page.
How investors actually get gold exposure
Physical (coins, bars)
via Bullion dealers, mints
+No counterparty risk
−Storage, insurance, wide spreads
GLD
via Brokerage account
+Most liquid; tracks spot well
−0.40% expense ratio (higher than peers)
IAU
via Brokerage account
+Lower expense ratio (0.25%); same exposure as GLD
−Slightly less liquid
GLDM, BAR, SGOL
via Brokerage account
+Lowest expense ratios (~0.10–0.17%)
−Younger, less institutional flow
Miners (GDX, GDXJ)
via Brokerage account
+Operating leverage to gold price
−Equity risk; can underperform gold
Futures (GC)
via Futures broker
+Leverage, no storage
−Margin, expiration, sophistication required
Not a recommendation of any specific vehicle — just a map. Most cost-conscious investors use the lower-fee ETFs (IAU, GLDM, BAR, SGOL); most who want the metal in hand use physical.
BEYOND GOLD
What about everything else?
Descriptive only — this gauge does not read them
Silver gets talked about as gold's little brother, and it does share some of gold's habits. People buy coins and bars of it, it sits in ETFs, it tends to catch a bid when people are nervous. But more than half of it gets used up. The Silver Institute's World Silver Survey put industrial demand at a record 680.5 million ounces in 2024, against total demand of about 1.16 billion. Solar panels, electrical contacts, brazing alloys, and increasingly the electronics inside data centres. Gold's industrial use is a rounding error next to its other demand. Silver's is the main event.
That changes what moves the price. A factory order book matters to silver in a way it simply does not to gold. When manufacturing slows, a large slice of silver demand slows with it, possibly at the same moment the fear trade is pushing the other way. The two can cancel. That is part of why silver's price behaviour looks erratic next to gold's rather than merely more energetic.
Supply is stranger still. The USGS notes that silver is mostly recovered as a by-product of lead-zinc, copper and gold mining, with dedicated primary silver mines accounting for under 28% of output in 2024. So the people digging up most of the world's silver are not making decisions based on the silver price at all. They are chasing copper, or zinc, or gold. A high silver price does far less to bring on new supply than intuition suggests.
And the silver market is much smaller than gold's, so the same amount of money arriving moves it further. That is the honest version of "silver has more upside". It has more of everything, in both directions.
Platinum is quoted in ounces and sold in bars, which makes it look like a monetary metal. It is not one. No central bank holds it as a reserve asset. Its demand comes from catalytic converters, chemical and petroleum refining, glassmaking, jewellery, and increasingly hydrogen electrolysers. The World Platinum Investment Council puts automotive demand near three million ounces against total demand of roughly 7.8 million, which makes cars the single largest slice.
So platinum tracks the car industry more closely than it tracks anything financial. Which engines are being built, what emissions rules are being written, and how platinum's price compares to palladium, which can substitute for it in petrol catalysts, all matter more than real yields or the dollar.
Its supply is the most concentrated of the three metals. USGS figures put South Africa at roughly 70% of world platinum mine production. A power cut, a flooded shaft or a labour dispute in one country is therefore a global supply event. Platinum has spent recent years in deficit, with the gap covered from above-ground stocks.
One more thing worth knowing, because it comes up. Platinum traded above gold for long stretches of the twentieth century and into the 2000s, and then it stopped doing that. Relationships like that one are descriptions of a period, not laws.
The nickname does a lot of work. Bitcoin does share one property with gold, which is a supply no government can expand. Its issuance is capped at twenty-one million coins by the rules of the software. That is a genuinely different kind of scarcity from gold's, where the limit is geological and mine supply responds to price, slowly.
After that the resemblance thins quickly. The official sector treats them nothing alike. Central banks have been buying gold in size and hold it as a reserve asset, and Bitcoin has no equivalent official bid, which removes the single heaviest force on this gauge outright. Gold has thousands of years of monetary history behind it. Bitcoin has existed since 2009, so every backtest anyone quotes covers about seventeen years. That is not enough history to settle an argument about how something behaves in a crisis.
And crisis behaviour is the part most worth knowing. Bitcoin is often sold as an inflation hedge, but in the worst weeks it has generally traded like a risk asset, falling alongside technology stocks rather than rising against them. Research covering the period since the spot ETFs launched in January 2024 finds its correlation with equities has increased rather than decreased, while its correlation with gold sits near zero. Two things that both rise over time are not thereby the same thing.
One parallel is real and worth watching. Gold ETFs arrived in 2004 and changed who could own gold and how fast money could move in and out of it. Bitcoin's spot ETFs arrived in January 2024. If you want a precedent for what that does to an asset, the gold story is the closest one available.
This site takes no view on whether anyone should own any of it. The author holds gold ETFs, Ethereum and a Nasdaq index fund. That is stated here rather than left to the footer, because a Nasdaq position is a position in the very thing Bitcoin has been tracking.
The rest of the metals live on their own page
Lithium, cobalt, nickel and the rest of the battery complex raise the same questions this site asks about gold. Where does it come from, who controls the supply, and how is the price actually set. Answering those properly takes more room than a gold site ought to give them.
So they have their own page. Copper, lithium, cobalt, nickel, graphite, rare earths, titanium, uranium, helium and the rest, each answered the same way — where it comes from, who controls it, and how the price is really set. Some of the answers are strange. One of them is the only thing on the list that genuinely runs out, and one of them is a part of the gold story this site would rather not print.
None of it feeds the gauge. It is there because it is interesting.
The eight forces on this gauge were chosen and weighted for gold, and most of them have no counterpart in anything above. Central banks buy gold. They do not buy silver, platinum, lithium or cobalt. Gold's mine supply responds sluggishly to price because a mine takes a decade to build; silver's and cobalt's respond sluggishly because most of both arrives as somebody else's by-product; lithium's responds sluggishly because the bottleneck is a chemical plant rather than a hole. Same word, three unrelated mechanisms.
The deepest difference is what happens after somebody buys the stuff. Almost all the gold ever mined still exists, which is why the above-ground stock is the market. The battery metals exist in order to be consumed, and their value is destroyed in the using of them. That single fact rearranges everything downstream of it.
Demand for them can also be engineered away. Nobody is going to invent a substitute for gold's monetary role in a laboratory, but a change of cathode chemistry can delete an entire metal from a supply chain inside one product cycle.
You could build this same instrument for any of them. It would need different drivers, different weights and its own backtest, which would make it a different instrument that happened to look like this one. Recalibrating this gauge and relabelling the dial would produce a number that looked authoritative and meant nothing.
So this page describes them. It does not measure them. If you want a reading on anything other than gold, the honest answer is that this gauge does not give you one.
UNDER THE HOOD
Methodology
Every rule, weight, and threshold — in full
WEIGHTED BY:
The composite score
Eight macro forces are scored on three axes — direction (up / flat / down), strength (1–3), and structural weight (how much that force matters in the current regime). The composite score is a weighted sum, normalized to −30 to +30, with 0 as neutral. Above +5 leans up. Below −5 leans down. The gauge to the right is the live read.
The regime selector
The regime selector reweights the eight forces — it does not change any of the underlying data. The default, Structural / haven driven, reflects the world since 2022, with central bank buying and geopolitics carrying most of the weight. Real-rate driven (classic) puts the old textbook rule back in charge, letting real yields and the dollar do most of the work. Crisis / risk-off assumes a bad week — war, a rout, a failure — and hands the score over to geopolitics and fund flows. The gap between the three tells you how much the reading depends on which theory you believe.
The tug-of-war
The tug-of-war separates each force into pulls-up and pulls-down. Arrow length equals strength × weight. The ball's offset is net pull, capped at ±85% of the track.
Time horizons · Now / Soon / Structural
Catalysts are tagged by horizon: Now (days), Soon (weeks), Structural(months to years). This separation is the single most important thing this model does — it keeps imminent volatility from being confused with the slow tectonic shifts that actually set the trend.
On Bitcoin & "digital gold"
A reasonable question is whether crypto allocation is structurally pulling capital away from gold. The short answer: at the margin, yes — among certain retail cohorts. But the bid that matters most for gold's price floor in this cycle is central bank reserve diversification, and central banks are not buying Bitcoin. Crypto and gold both benefit from the same underlying impulse (faith in fiat eroding), but the buyers are different people with different mandates. Treated here as adjacent, not substitutive.
What this is not. Not a price forecast. Not a trading signal. Not financial advice. A working model for thinking, maintained by hand, updated periodically.
Official gold reserves by country, monthly. The primary record behind the central bank numbers on this page.
STAY IN THE LOOP
Get major gold-signal updates by email
Substantive updates only — the daily readings when they meaningfully shift, weekly positioning changes, and the quarterly demand data when it lands. No every-tick price noise, no commentary, no calls. Double opt-in, one-click unsubscribe.
