Donchian Breakout (Gold)
XAUUSD · 4H chart · strategy s13_donchian · 2 variations
Buys gold when a 4-hour candle closes at the highest level of the last 55 candles and rides the trend with a wide trailing stop.
What it is about
This is classic trend following. A 'Donchian channel' is simply the highest high and the lowest low of the last N candles. When price closes above the top of that channel, it has broken out of the recent range, which may be the start of a new trend leg. The famous 'Turtle' traders used this kind of rule.
Most breakouts fizzle and end in a small loss. Now and then a breakout turns into a big, lasting trend, and because the stop trails far behind price, the strategy stays in for weeks and makes many times its risk. Those few big trends pay for everything else.
Gold trends well on the 4-hour chart, and this rule held up across most nearby settings in the research.
How it trades
Gold, at any 4-hour close. About 1-1.5 trades a month. A typical trade lasts about 4-5 days, the average about 9 days, and the big winners run for several weeks.
- Low win rate: about 37% (A) and 34% (B). The average winner is about 3.5-3.9R against an average loss of about 1R.
- Only about 1-1.5 trades a month.
- Losing streaks of up to 7 (A) and 9 (B) trades in a row in the backtest.
- Largest drop from a peak: 7.2R (A) and 11.1R (B). Backtest 2018-today at zero cost: +97.9R (A) and +103.4R (B).
- 7 of 9 years positive. 2018 and 2021 were flat to slightly negative.
- It lives on big trends. Gold's strong rally in 2024-25 gave a large share of the profit (2025 alone +30.9R for A). Years without a clean trend tend to be flat or red.
- Buys only. Sells lost in the research, so a long falling gold market would mostly mean sitting out or small losses.
- Trades last days to weeks, and overnight financing (swap) is not included. The research charged about 4% a year.
- Few trades means a single year says little. The same rule on the Nasdaq faded badly in 2024-26, a reminder that trend edges can fade.
Where it comes from
A classic trend-following rule (Donchian channel breakout, as used by the 'Turtle' traders), tested in the research of October 2026 as part of a trend-following batch. Gold on the 4-hour chart was the robust case. It was tested on 2018-2023 (the 'train' period) and checked on 2024-2026 (the 'test' period). Its daily results are only loosely related to the previous-day high breakout strategy (+0.17 to +0.20 correlation).
Main variation: s13-A ★ Main
The variations
Each variation is the same strategy with a different approach, trading its own paper account. Click one to see exactly what it changes.
The base rules: a close above the 55-candle high. The long lookback means it only fires on fairly significant breakouts, so it trades less and the trades it takes are more often real trend starts.
s13-B20-candle breakoutUses a 20-candle lookback (about 3-4 trading days), so it reacts to smaller breakouts and gets into trends earlier, at the cost of more false starts.
Variations side by side
Live result of all 2 together: 0.0R (zero cost).
| Live paper account (R, zero cost) | Backtest 2018 – today (R, zero cost) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| s13-A★ MainDonchian breakout + chandelier trail | XAUUSD | 4H | – | 0 | – | – | +97.9R | +0.86R | 8/9 | 1.1 | 37% | −7.2R |
| s13-BDonchian breakout + chandelier trail | XAUUSD | 4H | – | 0 | – | – | +102.4R | +0.63R | 7/9 | 1.5 | 34% | −11.2R |
Main = the variation with the most backtest R per R of max drawdown (2018 – today, zero cost). It is picked on that history, so the live accounts are the real test.