New York Session Gold Trading (XAUUSD): The Overlap and 08:30
The New York session on gold (XAUUSD), from zero to the pro level, on real gold charts. You will learn the New York session hours (08:00 to 17:00 New York time, which is 12:00 to 21:00 UTC in summer and 13:00 to 22:00 UTC in winter), the London-New York overlap (08:00 to noon New York) and the 2026 clock changes that stretched it for a few weeks; the traditional COMEX open at 08:20 and the 08:30 US data candle, measured against every other candle; what New York does to London's high and low (counts on 561 days); the 10:00 checkpoint and the London afternoon auction; the lunch lull and the quiet afternoon after the 13:30 COMEX close (measured); the London close reversal idea, checked; a real failure; an honest test of the New York sweep of London's range on more than two years of gold data, shown as counts and R only; the filters that did and did not help; news risk, the pro routine and a flowchart.
Chapters
- 0:00A real 08:30 New York candle on gold
- 0:59The New York session hours: New York time and UTC
- 1:52The overlap, the COMEX open and the 08:30 candle (measured)
- 3:12What New York does to London's high and low
- 4:1410:00, lunch and the afternoon (measured)
- 5:14The London close reversal idea, checked
- 5:50A real failure
- 6:42An honest test of the New York sweep
- 7:55Filters: what helped and what didn't
- 8:50News risk and the pro routine
- 9:36Should I trade this New York move? (flowchart)
- 10:13Recap
Key points
New York trades 08:00 to 17:00 New York time: 12:00 to 21:00 UTC in summer and 13:00 to 22:00 UTC in winter. The overlap with London (08:00 to noon) is the busiest part of the gold day.
The 08:30 US data candle is often big (typically 12.1 points on 15-minute candles), but on a normal day the 09:30 candle is bigger still.
Over 561 days New York took only London’s high 234 times, only the low 198, both 100 and neither 29.
Lunch and the afternoon after the 13:30 COMEX close are much quieter; the 10:00 turn and the London close reversal were close to coin flips.
Our honest test of fading a New York sweep of London’s range lost (331 trades, -35.4R after costs, every year negative); following the break was close to zero. Use the session map for context and risk, not as a promise of profit.
Chart data: COMEX gold futures OHLC from Yahoo Finance (GC=F), times in UTC. Education only, not a signal.
Read more: New York Session Gold Trading: US Data, the Overlap and Volatility · London Session Gold Trading: How XAUUSD Behaves at the London Open · ICT Kill Zones and the Silver Bullet Model on Gold (XAUUSD): Times, Rules and Honest Limits · New york session · London new york overlap · Comex
Transcript
Gold, fifteen minute chart, 14 October 2026. From 02:00 to 08:00 New York, all through the London morning, price moved inside a range of just 28.5 points. Then at 08:30 New York, the usual time for US data, one candle fell 64.5 points and closed far below London's low. Its full range was 72.4 points, on more than twenty times the volume of a normal London candle. The bounce stopped at 4,129.6, still under London's low. Soon after, it stopped at exactly the same price again. Around noon, gold was 84.6 points below where that candle opened. This is the New York session. You'll learn its hours in New York time and UTC, why the overlap is the busiest part of the day, how big the 08:30 candle really is, what New York does to London's high and low, a real failure, and an honest test.
Chapter one. The New York session hours. The New York session for gold runs from 08:00 to 17:00 New York time. Then comes a one hour break, and the new trading day starts at 18:00. In UTC, it moves with US daylight saving time. In summer, New York is UTC-4, so the session is 12:00 to 21:00 UTC. In winter, New York is UTC-5, so it's 13:00 to 22:00 UTC. London closes at 17:00 London time, which is 12:00 in New York. So from 08:00 to 12:00 New York, both big gold centres are open. This is the overlap. One trap. In 2026, the US changed its clocks on 8 March and 1 November, and Britain on 29 March and 25 October. In those weeks, London closed at 13:00 New York, so the overlap ran an hour longer.
Chapter two. The overlap, the COMEX open, and the 08:30 candle. We measured every one hour gold candle from June 2024 to November 2026, by New York hour. Each bar is the median, the middle value. In the London-only hours before New York, a typical hour moved 10.8 points. In the overlap, 18.4. That's about 1.7 times the range, on 2.7 times the volume. The 09:00 hour was the biggest of the whole day. Inside the overlap, two clock times matter. 08:20 is the traditional COMEX open, from the old trading floor days. On five minute candles, the 08:20 candle was typically 6.2 points, against 5.1 for the one before. A small step. The big step is 08:30. On fifteen minute candles, the 08:30 candle was typically 12.1 points, about 1.8 times the last candle before New York. On average, it was the biggest candle of the day, 18 points, because news days pull it up. But on a normal day, the 09:30 candle, when US stocks open, was bigger: 15.4 points. And of the 20 biggest fifteen minute candles in our data, 7 came at 08:30, and 13 inside the overlap.
Chapter three. What New York does to London's high and low. Mark London's range: the high and low from 02:00 to 08:00 New York. Then watch what New York does with it between 08:00 and noon. We checked 561 days. New York took only the high on 234 days, and only the low on 198. It took both on 100 days, and left London's range alone on just 29. So New York almost always runs the stops on at least one side. On the 432 one-sided days, by the 17:00 close, price stayed beyond the broken edge 254 times. It came back inside 158 times, and closed beyond the other edge only 20 times. The direction is a coin flip. The New York morning went the same way as the London morning 288 times, and the other way 270 times. But it matters: on 289 days, the high or the low of the whole day was set in those four hours.
Chapter four. 10:00, lunch, and the afternoon. 10:00 New York is a checkpoint. More US data often lands then, and the LBMA afternoon gold auction is at 15:00 London, which is 10:00 New York. Many traders say the morning move turns at 10:00. We checked. From 10:00 to noon, price went the same way as from 08:00 to 10:00 on 278 days, and the other way on 280. A coin flip. So reassess at 10:00, but don't treat it as a signal. Then comes lunch. From noon to 13:30, a typical fifteen minute candle moved 6.4 points, against 11.2 in the morning. After 13:30, the COMEX close, it drops to 5.2, and hourly volume nearly halves. So the afternoon is mostly quiet. The big exception is a Fed decision day. The decision usually comes at 14:00 New York, and then the quiet afternoon rule is off.
Chapter five. The London close reversal idea. London closes at noon New York. A popular idea says that when London traders close their positions, the morning move reverses. We compared the New York morning, 08:00 to noon, with the afternoon, noon to 17:00. The afternoon went the other way on 281 days, and the same way on 274. Only a tiny tilt. And the afternoon move was typically 7.6 points, half the morning's 15.2. So it's not a rule. Most afternoons simply go quiet.
Chapter six. A real failure. 10 November 2026. London's range was 4,076.2 to 4,135.1. At 08:15 New York, a candle dipped below London's low and closed back inside. It looks exactly like a sweep and reverse. So the fade trader buys at 4,079.8, with the stop under the sweep low, just 4.2 points of risk, and the target at London's high. Fifteen minutes later came the 08:30 candle. It moved 44.1 points, straight through the stop. By 11:45 New York, price was 54.8 points below the stop. The lesson: a sweep just before 08:30 is not a reversal yet. The data candle decides, and a tiny stop has no chance.
Chapter seven. An honest test. We tested the New York sweep exactly as traded. Between 08:00 and 11:00 New York, price trades beyond one London edge. The first candle that closes back inside is the entry. Stop at the sweep extreme, target the other edge, and close at 13:30. One hour candles, 561 days, and costs of 0.30 per ounce. On paper, the target was typically 1.8 times the risk. The result: 331 trades. 68 hit the target, 177 hit the stop, and 86 closed at 13:30. In total, minus 35.4R. That's about minus 0.11R a trade, and every year lost. On fifteen minute candles: 36 trades, and minus 0.1R, close to zero. But two trades made 10.9R together. Without them, minus 11R. And following the break instead? Stop at London's middle, target one range: minus 8.8R on 359 trades. With a two range target, minus 0.2R. Close to zero, but no edge.
Chapter eight. Filters: what helped, and what didn't. Wider stops. Adding 25, 50 or 100% of the range cut the loss to 24, 14.5 and 14.9R. Still losses. A closer target, London's middle, lost more: 38.9R. The bigger picture helped a little. Fading in the direction of the 20 day average lost 7.1R. Fading against it lost 28.4R. Entry time helped most. Entries in the 08:00 hour, around the data, lost 29.9R. The 09:00 hour lost 9.2R. And the 10:00 hour was the only group that made money: plus 3.6R on 65 trades. That's small, and one split of past data can be luck. But it agrees with the pro advice: let the 08:30 dust settle before you trade.
Chapter nine. News risk and the pro routine. Before 08:15, write the plan. Mark London's high, low and middle, the Asian range, and yesterday's high and low. Check the calendar for 08:30, 10:00 and Fed days. Write your news policy: flat through the release, smaller size, or wait fifteen to thirty minutes for structure. A spike and a whipsaw are normal. Then write three scenarios. New York breaks a London edge and holds, sweeps it and turns, or stays inside. Each needs a trigger, and a point where it's wrong. One attempt per side. Reassess at 10:00. Few new trades after 11:00, as London winds down, and be flat or planned by 13:30.
Chapter ten. Should you trade this New York move? Any no, and you wait. Is London's range marked, and have you checked the news calendar? If not, wait. Is big US data due in the next fifteen minutes? Then no new trade. Let the candle print. Did a fifteen minute candle close outside London's range, or back inside it? A wick alone is not a trade. Is this the first attempt on that side, before 11:00? If not, skip. If all is clear, take the plan, with the stop where the idea is wrong, and be flat by 13:30.
Let's recap. The New York session runs from 08:00 to 17:00 New York. That's 12:00 to 21:00 UTC in summer, and 13:00 to 22:00 in winter. The overlap, 08:00 to noon, is the busiest part of the day, and the 08:30 candle is the biggest on news days. New York almost always takes one of London's edges, and the break held more often than it reversed. Fading the sweep lost in our test, and waiting until after the data helped a little. The full written guide is on xauusdglobal.com. Subscribe for the next lesson, and come trade with us in the XAUUSD Global community. It's free during early access.
Other short videos

London Session Gold Trading (XAUUSD): Open Drive and Judas Swing

Asian Range Breakout on Gold (XAUUSD): London Breakout and Fakeout

Trading Maths for Gold: Why a 70% Win Rate Can Still Lose
