XAUUSD trades nearly around the clock, but its behaviour changes completely across the day. Trading gold in the wrong hour is one of the most expensive avoidable mistakes, because gold punishes it harder than any major pair.
Real movement begins. Often sweeps the Asian range first
13:00 – 17:00 (Overlap)
Highest volume. Tightest spreads. Cleanest moves
13:30 / 15:00
US data, then COMEX futures activity — the sharpest moves of the day
17:00 – 22:00
Volume fades. Trends can extend, but thinly
22:00 – 00:00
Rollover. Spreads widen sharply. Avoid
The London–New York overlap is where gold should be traded. For most people it is also the evening in Europe or the morning in the Americas, which is convenient.
The Asian range trap
Gold's Asian session frequently produces a tight, orderly range — 8 to 15 points.
Two things follow, and both matter:
It looks tradeable and mostly is not. Spreads are at their widest, often three to five times the overlap level. A 10-point range with a 0.60 spread means costs are 6% of the entire available move.
The Asian high and low become the day's reference levels. London routinely takes out one of them at the open before the real move. This is the single most reliable intraday pattern in gold, and it is why so many Asian-session breakout trades fail — the breakout *is* the sweep.
Volatility, concretely
Gold moves 1–2% in a normal day. EURUSD moves about 0.4%.
At 3,400, 1% is 34 points. A quiet day for gold is a large day for a currency pair.
Three consequences:
Stops must be wider in price terms. A 15-point stop is inside gold's ordinary noise during the overlap.
Position size must be much smaller. Same money risk, four to five times the price distance.
The same lot size is not the same risk. A trader moving from EURUSD to gold at their usual size has multiplied their risk several times without changing anything visible on the chart.
Use ATR(14) on your trading timeframe. If ATR is 34 points, a stop inside 34 is inside noise. This one measurement prevents most gold stop-outs.
The two fixes, and what they are for
Twice each London day, at 10:30 and 15:00 UTC, a group of banks runs an auction that sets a single benchmark price for gold — the LBMA Gold Price, still widely called the London fix. Central banks, refiners, jewellers and funds settle contracts against it.
Two practical consequences, and neither is a strategy:
Volume concentrates in the minutes around each fix, because participants with fix-referenced business want to trade near that price. Spreads are usually good and moves can be abrupt.
The 15:00 fix sits inside the London–New York overlap, which already carries the day's heaviest flow. When a fix and a US data release fall close together, the combination produces some of gold's sharpest intraday moves.
What to do with this: know the times so an abrupt move at 15:00 is a scheduled event rather than a surprise, and treat the two minutes either side the way you treat any liquidity event — not as an entry.
Anyone selling a "fix trading strategy" is selling a schedule. The schedule is public and the direction is not in it.
Volume profile — what it can and cannot tell you
A volume profile plots how much traded at each price rather than at each time. The price with the most volume is the point of control; the band holding most of the session's volume is the value area.
On COMEX gold futures this is real data, because futures trade on a regulated exchange that publishes volume. The reading is genuinely useful: prices where a lot traded tend to attract price back, and prices where little traded tend to be crossed quickly.
On spot XAUUSD it is not. There is no central venue, so your platform's "volume" is your own broker's tick count. A profile built from it describes one broker's flow, which may be a reasonable proxy or may not — and you cannot tell which from the chart.
The honest use: read volume profile on the futures contract for context, then trade your spot instrument with that context in mind. Do not build a spot profile and treat it as market participation.
The event calendar dominates
Gold is a rates instrument. Three events move it more than any technical level:
FOMC decisions and press conferences (19:00 and 19:30 UTC)
US CPI (13:30 UTC)
NFP (13:30 UTC, first Friday)
Around these, gold can move 40 points in ninety seconds with spreads at five to eight times normal. A stop is a request, not a guarantee.
The rule from the news lesson applies with extra force here: never hold gold through a tier-1 US release at a size you did not choose for slippage.
Worked example
Two traders, identical analysis: long gold from the 3,404 area, stop 3,374, target 3,489.
Trader A enters at 03:00 UTC. Spread 0.55. Gold is in a 3,398–3,412 Asian range. For four hours nothing happens.
At 08:05 the London open drives price to 3,371 — sweeping the Asian low and the swing low — stopping them out by 3 points. Price then reverses and reaches 3,489 by the New York session.
Trader B waits for 13:30, after the London open has resolved and the sweep has happened. Spread 0.18. Price is at 3,408, having reclaimed the swept low.
Same entry area, same stop level, but the sweep is now behind them rather than ahead. The trade runs to target.
Identical analysis. One paid three times the spread to have their stop taken by the most predictable event of the gold day.
Practical rules
Trade the 13:00–17:00 overlap unless you have a specific reason not to.
Mark the Asian high and low before London. Expect one to be swept.
Avoid 22:00–00:00 entirely.
Size from ATR, never from a pip habit carried over from FX.
Check the US calendar first. Gold is a rates instrument with a chart attached.
What to take away
Gold's day has two or three hours that matter, and the overlap is where they are.
The Asian range is a trap: widest spreads, and its extremes are London's target.
Gold moves several times a major pair's range. Wider stops, much smaller size.
ATR tells you what is inside noise. Use it before choosing any stop.
FOMC, CPI and NFP override every technical level on the chart.
Next: putting the session, the levels and the sweep into one method.
