Gold 15-Minute Intraday Strategy: A Session Momentum Breakout Tested on XAUUSD

The short answer

We tested a 15-minute gold intraday strategy that buys a break of the previous two hours’ high (or sells a break of the low) in the direction of the 200 EMA, only during London and New York hours, and is flat by 20:00 UTC. On 50 trading days of 15-minute COMEX gold data (21 August to 30 October 2026) it made 119 trades, won 31.1% and lost $263 per ounce after a $0.30 cost.

The loss was not a cost problem: before costs it still lost $228. A version without the trend filter lost $208, and a two-year hourly version of the same idea lost $362. This page shows the rules, the numbers, and what we would change before testing again.

The strategy: session momentum breakout

Momentum breakouts are the simplest day-trading idea there is. If gold breaks the high of the last two hours during an active session, buy and hold for a larger move, protected by a stop and a time limit.

Exact rules

The 200 EMA on a 15-minute chart is roughly the 50 EMA on the hourly chart, which is how many traders combine timeframes. Our multi-timeframe analysis lesson explains top-down bias in more depth, and the London session lesson covers the hours this rule trades.

How we tested it

The results

15-minute results (50 trading days)

Two-year hourly check

Costs

What the results say, in plain English

• Every version lost, in both halves. That consistency is more convincing than the size of any single number. On this data, buying two-hour highs in London and New York did not pay.
• The trend filter made it worse, not better. Without the 200 EMA filter the rule took more trades and lost less. The filter mostly kept it on the long side, which was the losing side in a falling market (longs lost $231 of the $263).
• A longer breakout lost less. The 4-hour version traded less and lost about half as much as the 2-hour version, which suggests the shortest breakouts were the weakest.
• The hourly version was near break-even before costs. At zero cost it lost $161 over 667 trades; at $0.30 it lost $362. On two years of data the idea was roughly random, and costs decided the sign.
• Low win rate, as designed. With a 2R target, a win rate near 31% is below the 33% needed to break even before costs.

Why intraday breakouts on gold are tricky

Gold has a habit of running just past an obvious level, collecting the stops resting there, and turning. Our tested pages on the ICT Judas swing and Turtle Soup look at the opposite trade, fading those false breaks. Neither side of the trade was reliably profitable in its simplest form, which tells you something: the edge, if there is one, is in context (which level, which session, which news day) rather than in the trigger.

Compare this with our Donchian channel test, where hourly breakouts with a trailing channel exit were positive over two years. The difference is the exit: Donchian lets winners run for days, while an intraday rule that must be flat by 20:00 UTC caps the winners and keeps every loser.

When this strategy fails

• Fake London breakouts. The first push after 07:00 UTC often reverses within an hour.
• Post-data reversals. A break on US data can retrace fully within 30 minutes.
• Falling markets with a lagging filter. The 200 EMA kept the rule buying weak bounces for days.
• Late signals. Breakouts after 14:00 UTC have little time to reach a 2R target before the 20:00 exit.

Risk sizing on the 15-minute chart

The average stop in the main test was about $13.22 per ounce. Risking $100 per trade:

• $100 / ($13.22 + $0.30) is about 7.4 ounces, roughly 0.07 lots.
• The worst losing streak was 10 trades, about $1,000 at that size.

Use the position size calculator to size from your own stop, and check the day’s volatility on the gold volatility tool before you trade: when ATR doubles, your position should roughly halve.

What we would test next

1. Retest entries instead of breakouts. Wait for the first pullback to the broken level, as in our breakout and retest lesson.
2. A trailing exit. Let winners run past 20:00 UTC with an ATR trail, accepting overnight risk.
3. Session levels. Use the Asian range or the previous day’s high and low rather than a rolling two-hour window. Our best timeframe for gold guide discusses how timeframe choice changes these trade-offs.

Checklist before you try a 15-minute breakout

• Define the session. Write down the hours you will take signals and the time you will be flat, in your own time zone, using the market hours tool.
• Test more than one breakout length. Our 2-hour breakout lost more than the 4-hour one.
• Decide what the trend filter is for. In our test it kept the rule on the losing side for weeks. A filter should be justified by results, not by habit.
• Plan the exit first. The biggest difference between these losing intraday breakouts and the profitable hourly Donchian test was the exit rule.
• Count the news days. Mark the days with US data in your results; if most losses cluster there, a calendar filter is cheaper than a new indicator.
• Keep a cost buffer. Our hourly check went from −$161 to −$362 when costs were added. Small edges vanish quickly.

Test it yourself

Code the rules, run them on your platform’s M15 XAUUSD history with your real cost, and then change one thing at a time: the breakout length, the filter, the exit. Keep a record of every variant you try, because the more variants you test, the more likely one looks good by chance. The strategies tested hub lists every named strategy we have tested with the same method.

Practise it with other traders

Day-trading rules improve fastest when other traders look at your entries. In the XAUUSD Global community, members share their M15 setups and journals, and challenge each other’s tests. The first month is free, then $19.99 a month.

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

Frequently asked questions