Breakout, pullback or scalping? How to tell an XAUUSD EA's style from its test report
You are looking at an XAUUSD EA. The page shows results, but which day trading style is behind them? The breakout at the London open, the pullback, the reversal at a strong level, the news, or minute-long scalps?
It matters: each style has its own good days and bad days. And you don't have to guess. Each one leaves marks in the Strategy Tester report, and paid EAs on the Market come with a free demo that runs there.
The rule to keep in mind: on gold, the strategy is not the entry signal. It is the stop, the lot and the timing that come with it.
The five setups, in one line each
- Overnight range breakout: gold drifts sideways overnight; the entry is a bar closing outside the range when London opens, and the simplest target is the range size.
- Pullback with the trend: gold trends in waves; the entry is the return to the moving average or the broken level, on a bar that resumes the trend.
- Reversal at a strong level: the previous day's high or low, or a round number like 4,000; the entry needs a rejection bar, never the touch.
- The news: Fed rate decisions, payroll and US inflation move gold in seconds; the entry waits for the first move to settle, or it doesn't happen.
- Scalping: trades of minutes, with short targets and stops, in the hours with the most volume.
What each style usually looks like in the report
Read it as "usually", not as a rule: an EA can mix styles.
| Style | Entries by hours | Average position holding time | Trades per month | Average win vs. average loss | Losing stretches |
|---|---|---|---|---|---|
| Breakout | Around the London open | Intraday | Up to one a day | Win at or above loss: the target is the range size | Sideways days, false breakouts |
| Pullback | Through the day | Longer, while the trend runs | Depends on the trend days | Win above loss: the target is the previous high or beyond | A "pullback" that becomes a reversal |
| Reversal | Wherever price meets the levels | Until the middle of the range | Depends on the sideways days | Closer together: the target is inside the day's range | Trend days, when the level gives way |
| News | At the release times | Short | Few: calendar days only | Wider stop and smaller lot, modest target | Back-and-forth spikes |
| Scalping | The hours with the most volume | Minutes | Many | Small wins; spread and commission take a visible share | Cost too large for the target |
Read it from the free demo in four steps
- Install the free demo from the EA's Market page and open the Strategy Tester with Ctrl+R.
- Run 2025 and 2026 with "Every tick based on real ticks" and your broker's data. On gold, only real ticks show whether a short stop or target was hit first.
- In the report, divide the total trades by the months tested. Put the share of profit trades next to the average profit trade and the average loss trade. Then check the average position holding time and the entries by hours.
- Turn on Visualization and run a Fed or payroll day: you see whether the EA stays out, waits or trades the spike.
The full guide on our site has the five setups with entry, stop and target, the comparison table, figures and FAQ: XAUUSD day trading strategy, 5 setups and when to use them.
What every setup on gold needs, and where to check it
The entry changes with the style. These pieces don't, and the test shows them.
- A stop set from the first second. The safest is a stop on the broker's server: filled in on the order, it stays there even if your computer shuts down. An EA that executes its own stop needs MetaTrader running all the time (a VPS is the simple way); its page should say so. In Visualization, check that every losing trade is closed by a stop, never left to run.
- A stop sized by the day's move, not fixed pips. With a 20 USD daily ATR (the average bar size, example value), a 5 USD stop is a quarter of the move; at 60 USD, less than a tenth. In Visualization, compare a calm day with a busy one.
- A lot set by the risk in money. First what you accept to lose on the trade, then the lot that fits.
- The cost in the math. On a 2 USD target, a 0.30 USD spread already takes 15% of what the trade is after. Check that the report charged the commission your account pays.
- No recovery through the lot. In the deals list at the end of the report, the volume should never double after a loss. Doubling is martingale, the shortest road from a bad day to a broken account.
Our three EAs: run the same test on them
Each of our three EAs trades gold on MetaTrader 5 in its own way and has a free demo. One number from each, in the test with real ticks:
- Collie Gold on the Market: short baskets, groups of entries with the same lot, with a stop on every basket, executed by the EA itself, so MetaTrader needs to stay on (a VPS is the simple way); Fed and payroll days are left out. +586% in 2025 on the Moderate profile. Largest continuous drop: 37%.
- Pitbull Gold on the Market: waits for the move and enters with target and stop on the server. +884% in 2025 on the Balanced profile, and 21 of 22 months in profit in 2025 and 2026. Largest continuous drop: 49%, in 2024.
- Beagle Gold on the Market: you choose 1% to 5% of the balance per trade, with the stop on the server and at most two trades open. +184% in 2025 on the Moderate profile, and 21 of 22 months in profit in 2025 and 2026. Largest continuous drop: 27%.
In all three, the lot grows with the balance, with no martingale. Download each demo, run the four steps and let the report tell you the style.
Lucky Dogs Trading Systems


