Gold Pullback Compass
- Uzmanlar
- Sürüm: 1.1
- Etkinleştirmeler: 5
Gold trend-following EA for XAUUSD H1. $10,000 grew to $159,000 in a 2006–2026 backtest at 1% risk per trade (about 14% a year), with every multi-year period profitable. No martingale, no grid, one trade at a time.
Gold Pullback Compass trades XAUUSD on the H1 chart. It waits for a clear trend, lets price pull back, and enters when the trend resumes. Every trade has a stop-loss from the start, a trailing stop, and a size based on a fixed % of your equity, so the account grows with its profits and shrinks its risk after losses.
20 years of gold, one rule set (backtest 2006–2026, $10,000 start, 1,610 trades)
| Risk per trade | Profit | Final balance | Average per year | Worst drop from a peak |
|---|---|---|---|---|
| 0.5% (conservative) | +$31,351 (+314%) | $41,351 | 7% | 12% |
| 1% (standard) | +$149,030 (+1,490%) | $159,030 | 14% | 23% |
| 2% (aggressive) | +$1,850,340 (+18,503%) | $1,860,340 | 29% | 40% |
Higher risk means bigger profits and deeper drops in the same proportion. The 2% row is shown for comparison: a 40% drop is more than most traders can sit through, so the EA defaults to 1%.
Why it holds up
- Profitable in every multi-year period: 2006–2010 (including the 2008 crisis), 2010–2015, 2015–2020, 2020–2025 and 2025–2026. Its oldest years were among its best, so it is not a system fitted to recent markets.
- Not a bet on gold going up: long and short trades win equally often (53.7% and 53.5%), and more than half of all profit came in the years when gold fell.
- Confirmed on a second, independent price source (2016–2026): same profit factor to within 0.1%.
- Stable settings: neighbouring values give similar results, so it does not hang on one lucky setting.
- No martingale, no grid, no averaging down, no hedging. One position at a time.
How it decides
- Trend: moving-average structure on H1 must point clearly up (for buys) or down (for sells).
- Pullback: price must come back toward the trend, but not too deep.
- Market quality: an efficiency filter keeps it out when gold moves sideways and noisy.
- Exit: ATR stop-loss, ATR trailing stop, and break-even once the trade is well in profit. Profits come from letting winners run.
Before you buy
- Choose risk with the table in mind: at 1% the deepest historical drop was 23%; future drops can be deeper.
- Not every year is a winner: at 1% risk, 18 of 21 calendar years were profitable. The losing years were 2014 (−17%) and 2022 (−7%); 2018 was flat.
- Give it time: in the backtest a 3-month stretch ended in a loss about 1 time in 3, a 12-month stretch about 1 time in 5. That is why rentals start at 3 months; 12 months is the fairest test.
- It is patient: losing streaks of up to 12 trades and flat months happen. Closing winners early by hand hurts results.
- Backtest assumptions: 1-minute price model, today's swap rates for all years, no commission (on a commission account expect slightly less). Some filter settings were chosen on 2023–2026 data.
- Past results, including backtests, do not guarantee future results. There is no public live record yet.
Recommended use
- XAUUSD (any broker suffix, e.g. XAUUSD, XAUUSD.m, GOLD) on the H1 chart, one chart only.
- Risk 0.5–1% per trade. A low-spread account and a VPS running around the clock.
- Run the free demo in the Strategy Tester first.
Inputs
- Lot sizing: Risk % of equity per trade (default) or Fixed lot.
- Risk per trade, % of equity (default 1.0); Fixed lot (0.01).
- Skip a trade if the minimum lot would risk more than the chosen % (default on).
- Daily loss limit, % (default 10; blocks new entries until the next day).
- Maximum spread in points (0 = off).
- Move stop to break-even when well in profit (default on).
- Magic number, order comment, maximum slippage.
The strategy settings are fixed at the tested values, so they cannot be over-tuned by accident.
