XAUUSD strategies for EAs: what each one looks like on a live account

Why gold is hard for an EA

XAUUSD moves a lot in dollar terms. One standard lot is 100 ounces, so a $10 move in the gold price is $1,000 on one lot and $10 on 0.01 lot. A normal day can easily cover $30 to $50, and a busy one much more.

Gold also reacts hard to US data and to the Fed: CPI, payrolls (NFP) and rate decisions can move it several dollars in seconds. Spreads widen around those moments and at the daily rollover, and the market can gap on Monday open.

All of this means an EA that works on a calm currency pair can fail on gold, and that the size of each trade matters more than the entry.

1. Trend following

Idea: wait until gold is clearly moving one way, then follow it, usually with a trailing stop.

On a live record: a win rate often below 50%, a few large wins paying for many small losses, and flat or slowly falling stretches while gold goes sideways.

What it fears: long ranges with false starts.

2. Breakout

Idea: mark a range, for example the Asian session high and low, and trade when the price leaves it.

On a live record: trades cluster at the same times of day; results depend heavily on spread and slippage at the moment of the breakout.

What it fears: breakouts that reverse straight back into the range.

3. Scalping and short-term mean reversion

Idea: take small, quick moves, often against a short stretch of the price, and get out fast.

On a live record: many trades, short holding times, a high win rate and small average wins. Costs matter a lot: a Raw account, a low-latency VPS and a broker with good execution make a real difference.

What it fears: wider spreads and slippage, and a few large losses that wipe out many small wins.

4. Grid

Idea: when the price moves against the first trade, open more trades at set distances and close them all together on a small bounce.

On a live record: a smooth, rising balance line, a very high win rate, several trades open at once, and an equity line that drops well below the balance from time to time. The drops are the real risk.

What it fears: a long one-way move. Gold can run $100 or more in one direction without a pause, which is why grids on gold need a hard stop.

5. Martingale and recovery trades

Idea: after a loss, trade bigger so that one win gets the loss back.

On a live record: lot sizes that jump after losing trades, a near-perfect win rate, and then, often, one very large drawdown.

What it fears: a losing streak longer than the account can pay for. On a market as fast as gold that streak comes sooner than most people expect.

6. News trading

Idea: trade the seconds or minutes around scheduled releases such as CPI or NFP.

On a live record: few trades, on release days only, and results that vary a lot from broker to broker because spreads and fills are at their worst exactly then.

What it fears: slippage, requotes, and brokers that widen spreads sharply during news.

How to check a gold EA before you buy it

A live account, not a demo. Demo fills are perfect; live fills on gold are not.
Equity, not just balance. On myfxbook, compare the equity line with the balance line. A grid or martingale hides its risk in open trades.
Lot sizes after a loss. If they go up, it is a martingale or recovery system, whatever the description says.
Drawdown and time. Ask whether you could sit through the largest drawdown so far, and expect a deeper one eventually. A few weeks of history says little.
A stop loss on every trade. Look at the trade history: are stop losses set, and how big are the losses when they are hit?

There is a longer checklist in how to verify a forex EA’s track record.

Where Amber Algo GOLD fits

I don’t publish the entry rules of my robots, but I do publish how they control risk, because that is what you need to judge them.

Amber Algo GOLD is not a grid and not a martingale. It sends every trade with its stop loss attached, and with the default settings a trade that hits its stop loss loses about 1% of the free margin. It holds at most one buy and one sell at a time, and it never trades bigger after a loss.

Amber Algo GOLD has been live since Oct 16, 2026: a gain of +3.86%, a maximum drawdown of 1.78% and a profit factor of 2.08 (myfxbook, as of Nov 6, 2026). Every trade is public. Check it on myfxbook.

Risk. Trading with leverage can lose all the money in an account. Past results do not guarantee future results. This article explains common approaches; it is not financial advice.

More guides