Previous-Day High Breakout
XAUUSD · 1H chart · strategy s05_pdh_breakout_trail · 2 variations
Buys gold when an hourly candle closes above yesterday's high during an uptrend, then follows the move with a wide trailing stop.
What it is about
Yesterday's high is a level every trader can see. When gold is already in an uptrend and an hourly candle closes above that level, it often keeps rising. Breakout traders buy, and sellers who had stop-loss orders above the high are forced to buy back.
This is trend following. Most breakouts fail and end in a small loss. The ones that work can run for many hours or days, and the strategy does not cap them with a target. A stop that follows price from a distance lets the big moves pay for the many small losses.
It was the first robust edge found on gold. Many pattern and win-rate ideas were tested first and failed. This one stayed profitable across most nearby settings and in the later test period.
How it trades
Gold, at any hour of the trading day except the 17:00 New York close. About 6-7 trades a month. A typical trade lasts about 7 hours; the average is about 20 hours, and the best trends run for several days.
- Low win rate: about 31% of trades win in A (40% in B). The average winner in A is about 3.5R against an average loss of 1R.
- About 6-7 trades a month.
- Losing streaks are long: up to 17 losses in a row in A, 10 in B.
- Largest drop from a peak: about 22.6R (A) and 18.0R (B). Backtest 2018-today at zero cost: +309.2R (A) and +276.5R (B), all 9 years positive (2023 only just).
- Profits come in bursts. 2025 alone made +109.4R of A's total, while 2018, 2021 and 2023 were close to flat.
- Buys only. Gold rose strongly in 2024-26, which helps a breakout-buying strategy. The short version (selling breaks of yesterday's low) lost money.
- Overnight financing (swap) is not included. Research put it at about 0.04R per trade on gold.
- A large share of the profit came from a few strong trend periods. A long sideways market would mean many small losses in a row.
- It is a gold strategy. On the Nasdaq 100 the same rules were weak (+0.12R per trade) and are not used.
- Checked against TradingView (1-hour chart, 2025-2026): 159 trades there vs 153 in Python, with similar totals. Different data feeds give slightly different breakouts.
Where it comes from
Our own research (October 2026): a broad scan of gold patterns found this as the first edge that held up. It was tested on 2018-2023 (the 'train' period, used to choose settings) and checked on 2024-2026 (the 'test' period): +155.4R and +153.8R for variation A.
Main variation: s05-B ★ 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.
Same as A, but once the trade has been 2R in profit, the stop jumps to at least +1R from the next candle. A trade that went well can then no longer end as a loss. The normal trailing stop keeps working on top. The aim is a smoother ride for a little less profit.
s05-APure trailing stopThe base strategy unchanged: a 1 ATR stop that becomes a 3 ATR trailing stop, and nothing else. Trades that go well can still give back all their profit before the trail catches up. That is the price of giving trends room.
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) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| s05-B★ MainPrevious-day high breakout + ATR trail | XAUUSD | 1H | – | 0 | – | – | +276.5R | +0.39R | 9/9 | 6.7 | 40% | −18.0R |
| s05-APrevious-day high breakout + ATR trail | XAUUSD | 1H | – | 0 | – | – | +309.1R | +0.47R | 9/9 | 6.2 | 31% | −22.6R |
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.