Day Trading Gold: A Complete Day Plan for XAUUSD

Short answer: day trading gold means every trade opens and closes inside one session, so you never hold through the reopen gap. A workable day plan has six parts: a pre-market routine, a fixed trading window (the London open or the New York morning), a short list of allowed setups, risk sized from the stop, daily limits (a maximum number of trades and a loss limit of about two full losses), and a written review. The limits do not make a strategy profitable. They keep a bad hour from turning into a bad month.

In plain words: a day trader is like a shopkeeper who opens at a set time, follows the same routine every morning, sells only the items on the shelf, closes the till at a fixed time and counts the money every evening. The shop does not open at 3 a.m. because the shopkeeper feels lucky. Gold moves enough in a day to tempt you to trade all of it. The plan’s job is to pick the few hours that matter and protect you from yourself in the rest.

This lesson pulls together pieces taught elsewhere in the Academy. If you are still deciding between styles, read scalping vs swing trading first. For the sessions themselves, see the London and New York lessons. This page is about how the day fits together.

What day trading gold means

There are three common styles of short-term gold trading, and the line between them is the holding time:

A day trader reads the higher timeframes (daily and 4-hour) for direction and levels, then trades on the 5-minute to 1-hour charts. The day trader is flat at the end. That one rule removes the weekend gap, overnight swap on CFDs, and the risk of waking up to a different market.

Why gold suits day trading, and why it bites

Gold has three features that make it popular with day traders. It trades nearly around the clock, it has deep liquidity in the London and New York hours, and it moves a lot: the median full-session range in our futures data was $31.6 in 2024, $52.9 in 2025 and $101.3 in 2026. The same features make it dangerous. A $100 day means a stop sized for 2024 is far too tight now, and a position sized for a $30 day can lose three times what was planned.

Data note. All our numbers come from the site’s COMEX gold futures (GC=F) hourly file via Yahoo Finance, 12 June 2024 to 3 November 2026, with every contract-roll session removed. Spot XAUUSD trades a few dollars below the front futures month (the basis), but its intraday shape is almost the same.

The trading day on the clock

The gold trading day starts at 18:00 New York time (the reopen after a one-hour daily break) and runs to 17:00 the next afternoon. We use New York time because the futures day is built on it. The gold market hours tool shows every block below in your own zone.

The hour-by-hour picture is clear in both years. In 2026 the median 1-hour range was $25.4 at 08:00 NY, $31.1 at 09:00 and $28.2 at 10:00, against $12.9 at midnight and $11.0 at 16:00. Everything was smaller in 2024 (the 09:00 hour was $11.5), but the shape was the same. That is why most gold day traders pick the New York morning as their main window, with the London open as a second choice.

Choosing a window that fits your life

The best window is the busy one you can actually watch, every day, without rushing. While the US is on standard time (from 1 November 2026 to 14 March 2027), the New York morning block of 08:00 to 11:59 falls at:

When the US moves to daylight time, every row moves one hour earlier for places that do not change their clocks. For a trader in India with a day job, the New York block lands in the evening, which is why so many Indian gold day traders trade 18:30 to 22:30. For a trader in Singapore it runs past midnight, so the London block in the afternoon may fit better even though it moves less. A window you trade tired is worse than a quieter window you trade fresh. The gold market hours tool converts the blocks for your city on any date.

Where the day’s high and low are set

A useful question for a day trader is: if I take a trade in London, how often will the day’s extreme come later? We assigned each session’s high and low to the block where it printed.

Out of 527 sessions, the high came in Asia 181 times, London 66, the New York morning 150 and the New York afternoon 130. For the low it was 217, 72, 144 and 94. New York set 49% of all the day’s extremes, and London only 13%. Two practical lessons follow:

1. A London trade often meets the day’s extreme later. If you hold a London trade into New York, the biggest move against you may still be ahead. Many day traders take London profits before 08:00 NY or move the stop.
2. Asia often holds an extreme. Asia includes the reopen and the full Asian night, so it often sets the high or low that London and New York then test. That is why the Asian range is on every pre-market sheet.

How many trades a day?

The honest answer is: as many as your rule produces, and no more. A rule that gives you ten trades a day has usually been loosened until it trades on noise. We ran one simple, clearly defined rule over every hour from 02:00 to 10:59 New York time: the hourly round-number sweep from our scalping lesson. A candle pokes through a $50 round number by up to half an hourly ATR and closes back on the other side; entry at the close, stop beyond the wick plus $0.50, target 2R, out by 17:00 NY. One trade open at a time.

The rule produced 443 signals on 308 of 527 days. On 219 days it gave nothing. Of the days with a signal, 199 had one, 88 had two and only 21 had three or more. And the later trades were worse: the first trade of the day averaged -0.23R after costs, the 135 trades after it -0.27R. More trades did not mean more profit. Note that this rule loses money overall (100 of 402 decided trades reached 2R, 25%, below the 33% needed to break even at 2R before costs). It is used here because it is mechanical and gives several trades a day, not because it works.

Most experienced gold day traders settle on one to three trades a day. Our blog post on how many trades a day is normal covers how to find your own number.

Daily risk and loss limits

Risk works at three levels in a day plan:

• Per trade (1R): a fixed share of the account, commonly 0.25% to 1%. Position size comes from the stop distance, never the other way round. The position size calculator does the arithmetic.
• Per day: a hard stop for the day, usually written as a number of R (for example, stop after two losses or at -2R) or a share of the account (1% to 2%).
• Per week: a larger limit (for example, -5R) that sends you back to review or demo.

The figure uses the three real trades from the example day below. With a hypothetical $10,000 account and 0.5% risk ($50 per trade), the stops of $5.5, $4.9 and $14.7 give 0.09, 0.10 and 0.03 lots. The losses would have been $49.50, $49.00 and $44.10. The point: a wider stop means a smaller size, so every loss costs about the same.

What daily limits really do: our replay

We replayed the same 443 signals under six daily rules, with a $0.30 round-trip cost.

Three lessons stand out:

1. Limits cap the worst day. Without a rule, the worst day lost -4.1R, and 56 days lost 2R or more. With a two-loss stop, no day lost more than -2.3R.
2. The tightest limits shrank the drawdown most. One trade a day cut the deepest drawdown from 108R to 72R, mostly because fewer losing trades were taken.
3. Limits did not create an edge. Every version of this rule lost money: -105.8R in total with no rule, -69.9R with one trade a day. If your setup has a negative expectancy, a loss limit only makes it lose more slowly. The limit protects your account while you find out whether the setup works; it does not decide that question.

The other half of the daily-limit story is the profit side: what to do on a good day so you don’t give it back. Our Trader Journey post on daily profit target vs daily loss limit covers give-back rules with a worked example. The risk of ruin calculator shows how per-trade risk and streaks interact.

The pre-market routine (15 minutes)

Do it at the same time every day, before your window opens, and write the answers down.

1. Calendar. Any CPI, NFP, FOMC or other tier-1 US release today? If yes, at what time in your zone, and will you stand aside for it? The gold economic calendar shows them in your time.
2. Yesterday’s levels. Previous day’s high, low and close.
3. The Asian range. High and low from the 18:00 reopen to 02:00 New York time.
4. Pivots. Today’s classic pivot, R1 and S1. The pivot points masterclass explains them, and our daily analysis hub publishes them with the round numbers and moving averages every day.
5. Volatility. The 14-day ATR, and whether yesterday was a wide or narrow day.
6. Bias sentence. One line: “Above X I look for longs at Y; below Z I stand aside.” Plus what would prove it wrong.
7. Risk. Today’s 1R in dollars, the daily stop, and how much of this week’s limit is used.

That is all. If you can’t finish the routine before your window opens, you skip the first hour, not the routine.

The day plan on one page

A full plan for one window could look like this:

• Window: 08:00 to 11:00 New York time, plus the London open from 03:00 to 05:00 on days I’m awake.
• Allowed setups: two, written out in full, for example a liquidity sweep reversal at a marked level and a breakout and retest of the Asian range. Everything else is “no trade”.
• Risk: 0.5% per trade, sized from the stop. No adding to losers.
• Limits: maximum three trades, stop after two losses or at -2R, stop for the week at -5R.
• News: flat two minutes before a tier-1 release; no new trade until 15 minutes after.
• Stop time: flat by 11:30 New York time, or move the stop to break-even and set a final exit time.
• Review: every trade journaled in R with a screenshot. Our trading journal templates lesson has the fields.

A real day played out: Thursday 8 October 2026

Here is one real session from our data, run through the plan with the round-number sweep rule as the setup. It is a range day, and range days are where day traders lose most of their money, so it is worth studying.

Pre-market (written before 02:00 New York time):

• Calendar: no tier-1 US release that day.
• Yesterday (7 October): high 4,197.8, low 4,091.2, close 4,140.7.
• Classic pivots for 8 October: P 4,143.2, R1 4,195.3, S1 4,088.7.
• Asian range (18:00 to 01:59): 4,128.1 to 4,166.8, and price sitting near the pivot.
• 14-day ATR: $93.4.
• Bias sentence: “Price is between the pivot and the Asian high with no news. Expect rotation inside the Asian range unless it breaks with a strong close. Trade the plan’s setups; stop after two losses.”

What happened, trade by trade (New York time, prices from the futures file):

Each loss was about 1R after costs. With no daily rule the day lost -3.1R. With the plan’s “stop after two losses” rule, trade 3 is never taken and the day ends at -2.1R. In dollars, on the sizing above, that is $98.50 instead of $142.60.

The day’s full range was only $42.6 (4,128.1 to 4,170.7), 46% of the 14-day ATR, and the close was 4,157.0, back inside the Asian range. Every short at 4,150 was a fade of a market that was not trending, inside a range that held all day apart from a small poke above the Asian high at 13:00. Three things in the pre-market sheet warned of this: no news, price on the pivot, and a narrow Asian range. The next day, 9 October, gold broke out and reached 4,233.7, which is a reminder that a range day is often followed by an expansion day.

What a trader learns from this day:

1. The plan worked even though every trade lost: the loss was capped at two trades, and the trader stopped with energy and money left.
2. The same level failing twice is information. After the second failure at 4,150, the market was saying that 4,150 was not resistance today.
3. A range day calls for smaller targets or no trades. A 2R target needs room to run, and inside a $42.6 range it had none.

Base rates to keep in mind

• Hour sizes have tripled. Median 1-hour ranges in 2026 were about three times those of 2024 (09:00 NY: $31.1 vs $11.5). Re-measure before you reuse old stop sizes.
• New York sets most extremes: 49% of all daily highs and lows in our sample, London 13%.
• One rule, few trades: 199 of 308 days with a signal had just one.
• Later trades were worse: -0.23R for the first trade of the day vs -0.27R after it.
• Limits cap losses, not expectancy: worst day -4.1R without a rule, -2.3R with a two-loss stop; average still negative.

None of these is a promise about tomorrow. They describe about two and a half years of one futures contract.

Pro nuance

The day’s character changes the plan

Experienced day traders sort the day early into one of three types and adjust:

• Trend day: price leaves the Asian range in London and keeps going in New York. Pullback entries in the trend direction work; fades get run over.
• Range day: price rotates around the pivot inside the Asian range, like 8 October. Fades at the edges with small targets, or no trade.
• News day: a scheduled release decides the direction. Flat before it, trade the structure that forms after it, as our news trading lesson explains.

Clues in the first hours: the width of the Asian range compared with the ATR, whether London closes outside the range, and whether the first New York hour extends or reverses the London move.

Process beats prediction

You can’t control whether a trade wins. You can control whether you followed the plan. Many day traders score each day on rule-keeping (setup on the list, size correct, limits respected) as well as on P&L. A day like 8 October scores full marks on process even though it lost money.

Costs grow with the number of trades

At a $0.30 round trip, a trade with a $5 stop pays 0.06R in costs every time. Three such trades a day, five days a week, is close to 1R a week before a single loss. Wider stops in the busy hours, fewer trades and no trading in the thin hours keep costs small. The scalping vs swing lesson has the cost table.

Worked example (hypothetical)

A trader has $5,000 and risks 0.5% ($25) per trade with a daily stop of -2R ($50).

1. Pre-market: Asian range 4,020 to 4,045, previous-day high 4,060, pivot 4,035, ATR $80, no news.
2. London, 03:30 NY: price breaks above 4,045 and closes at 4,050. The plan’s breakout-retest setup waits for a retest.
3. 04:15: retest of 4,045 holds with a bullish 15-minute candle. Entry 4,047, stop 4,039 (below the retest low), $8 risk. Size = $25 / ($8 × 100) = 0.031, rounded down to 0.03 lots, risking $24.
4. Target: the previous-day high at 4,060 (1.6R), with the rest of the position aimed at 4,071 (3R).
5. 05:10: 4,060 is hit; 0.02 lots are closed for +$26 (0.02 × $13 × 100), and the stop on the last 0.01 lot moves to break-even.
6. 08:40: after US data, price spikes to 4,066 and reverses to the break-even stop. The last 0.01 lot is out at 4,047 for $0.
7. Result: +$26, about +1.1R. One trade, both limits respected, flat by 09:00. The trader stops for the day because the plan’s window is used.

The numbers are hypothetical, not a real trade.

Common mistakes

• Trading every hour. The quiet hours give small moves and the same spread. Pick a window.
• Using last year’s stop size. Ranges in 2026 were about three times 2024’s. A stop inside the normal hourly noise is a donation.
• No daily stop, or moving it. The worst day in our replay without a rule was -4.1R. A limit only works if it is fixed before the session.
• Holding a day trade overnight “because it will come back”. That turns a day trade into an unplanned swing trade with the wrong size.
• Revenge trades after two losses. The trades after the first of the day did worse (-0.27R) even in a mechanical rule; tired, angry trades are usually worse still.
• Trading through tier-1 news by accident. Check the calendar first, every day.
• Judging the plan by one day. 8 October lost money and was well traded. Judge the setup over 50 to 100 trades in your journal.

Quick checklist

1. Did I finish the 15-minute pre-market routine and write the bias sentence?
2. Is there a tier-1 release in my window, and what is my news rule?
3. Is this setup on my written list, and do I know the entry, stop and target before I click?
4. Is my size calculated from the stop at my fixed risk?
5. How many trades have I taken, and how far am I from the daily stop?
6. What type of day is this (trend, range, news), and does my target fit it?
7. Will I be flat by my stop time?
8. Did I journal every trade in R, with a screenshot and one lesson?

Sources

• CME Group: Gold futures contract specifications and trading hours (the futures day from 18:00 to 17:00 New York time)
• US Bureau of Labor Statistics: release schedule (CPI and jobs reports at 08:30 Eastern time)
• Federal Reserve: FOMC meeting calendar
• Our own counts (hour sizes, where the day’s high and low formed, signals per day, the daily-limit replay and the real day): the site’s COMEX gold futures (GC=F) hourly and daily files via Yahoo Finance, 12 June 2024 to 3 November 2026. Every contract-roll session was removed, because the continuous futures series jumps when it switches delivery month. The rule tested is the hourly round-number sweep from our scalping lesson, before slippage, with a $0.30 round-trip cost.

Educational content only, not financial advice. Trading gold with leverage carries a high risk of loss.

Frequently asked questions