Gold trading strategy: what actually tests well on XAUUSD

The strategy families, and what gold does to each

Nearly every gold trading strategy belongs to one of a few families, and XAUUSD's character — violent expansions, long ranging stretches, and a cost structure that punishes short targets — treats them very differently.

The pattern across families: on gold, the entry idea matters less than whether the strategy knows which market condition it is for — and refuses to trade outside it.

What survived our testing

We run 34 strategy implementations through replayed MT5 bar data — the same engine that trades live, fed historical bars, so nothing is reimplemented for the test. Most of the 34 are disabled in the shipped configuration because they did not earn their place on the data. The survivors are documented, with the reasoning, on the strategies page.

What the survivors share is not an indicator. Each trades a defined session window rather than around the clock, waits for a structural reason — an order block, a liquidity sweep, a session-open range — rather than an indicator crossing, sizes from a fixed risk percentage with a fixed reward ratio, and refuses entries when the spread is wide relative to the target.

The disabled majority is the more instructive half: most failures were not bad ideas but good ideas trading in the wrong conditions — breakout logic firing in ranges, reversion logic fighting trends. The methodology behind those decisions is in the backtesting guide.

The regime matters more than the strategy

The single most transferable finding: whether gold is trending or ranging on the day decides more of the outcome than which entry signal fired. The same strategy flips between profitable and unprofitable as the regime flips, which is why averaging a backtest across all conditions hides what is actually happening — see curve fitting for how that mistake gets made.

The practical consequence for a manual trader is to classify the day before picking the playbook. Ours does the same thing mechanically: it classifies the regime and runs different strategy pools in each — the mechanics are on trading modes.

The parts that are not the strategy

Three decisions outside the entry rule move results as much as the rule does. When you trade: the sessions differ enough to be different markets, and the daily rollover is a standing toll — gold trading hours has the measured version. How much you risk: a fixed percentage per trade with a stop derived from structure, not the other way round — risk management. What you refuse: wide spreads, scheduled news, thin holiday sessions. Refusing trades is the cheapest edge available on gold.

This is also the honest answer to why strategy articles disappoint: the entry recipe is the marketable part, and the unmarketable parts are where the results come from.

Testing one yourself

Whatever the source of a strategy — an article, a purchase, your own idea — the sequence that protects you is unchanged: backtest against real spread costs rather than zero-cost fills (how), forward-run it on a demo account long enough to survive both regimes, and only then let it near a balance. The step most people skip is the middle one, and it is the cheap one.

And treat any strategy sold on a win rate with suspicion on principle: gold's cost structure makes very high win rates the signature of martingale sizing, not of a good entry.