Part 4: Breakouts, system and test · Chapter 13 of 14 · 11 min
A system with patterns, on trial
A system written beforehand (patterns with confirmation, stop at the extreme, target at one height, one position at a time) gives +25.1 R over 89 trades, but 16.09 R come from a single one. Without it +9.0 R remain (+0.10 R per trade), inside the range of random series.
The system
The rules are written before looking at the results and use only information available at the time of entry.
- Patterns allowed: the 13 types of chapters 2-8, without the V spikes (chapter 9: no edge over chance).
- Entry with confirmation: after the breakout you enter at the first close, within 6 candles, that sits at least 0.5 ATR beyond the break level, with no close having come back before. Without confirmation the signal lapses.
- Stop: beyond the opposite extreme of the pattern plus 0.25 ATR.
- Target: one height of the pattern from the entry.
- Time exit: after 48 candles, at the close.
- One position at a time: signals that arrive with a position open are skipped.
- Cost: 0.30 $ per ounce per trade. Risk is measured in R (the distance from entry to stop).
Data: H1 gold, December 2025 - October 2026, 13 pattern types.
The result
There are 149 signals with confirmation; with one position at a time the trades are 89 (the other 60 arrive with a position open). Sum +25.1 R, mean R +0.28, 56% of trades positive, median R +0.11, maximum drawdown (the largest fall from the previous peak) 6.2 R, longest losing streak 6. If each trade risked 1% of capital, 6.2 R would be 6.2% of capital: a drawdown you see and either tolerate or not.
Random series with the same rules (50 series shuffled in blocks of 6 candles) give a median sum of +1.3 R, from −12.7 to +25.8 (5th and 95th percentile), 47% of trades positive (from 38% to 56%), median R −0.26 (from −1.00 to +0.35), maximum drawdown 10.0 R (from 5.0 to 18.5) and longest losing streak 6. Gold sits at the upper edge of the range of chance: +25.1 against a 95th percentile of +25.8.
| Variant (one position at a time) | Trades | Sum R | Mean R | Max drawdown | Random: sum (5th/95th) |
|---|---|---|---|---|---|
| Immediate entry, ½ height | 136 | +11.3 | +0.08 | 15.4 | +0.9 (−15.9 / +23.6) |
| Immediate entry, 1 height | 126 | +20.1 | +0.16 | 10.4 | +4.9 (−17.2 / +21.0) |
| Immediate entry, 2 heights | 114 | +31.0 | +0.27 | 6.9 | +6.1 (−18.8 / +34.7) |
| With confirmation, ½ height | 93 | +20.2 | +0.22 | 5.9 | +2.2 (−9.8 / +17.1) |
| With confirmation, 1 height | 89 | +25.1 | +0.28 | 6.2 | +1.3 (−12.7 / +25.8) |
| With confirmation, 2 heights | 88 | +22.0 | +0.25 | 6.4 | +0.6 (−17.4 / +28.3) |
The six variants are positive and all sit inside the range of chance. They are not six independent confirmations: they share the same patterns and, as shown below, the same enormous trade.
One trade alone is worth 64%
Of the +25.1 R, 16.09 R come from a single trade: 64% of the sum. On 28-30 August 2026 a symmetrical triangle breaks down, the system enters at 4,474.53 $ with the stop at 4,482.67 $ (risk of 8.14 $, 1.8 ATR) and price falls 131 $ in 48 candles. The target (one height of the pattern, 184.19 $, so 4,290.34 $) is not touched: the exit is by time on 1 September at 21:00 at 4,343.28 $.
Without it +9.0 R over 88 trades remain (+0.10 R per trade). Removing the three best (16.09 R, a flag of 4.18 R and another symmetrical triangle of 2.49 R) leaves +2.3 R. In the random series, without the best trade, the median sum is −2.7 R (from −15.9 to +18.3) and without the three best −8.7 R (from −21.4 to +8.3): gold stays inside the range of chance. The best trade of a random series is worth a median 4.4 R (from 2.2 to 7.9): 16.09 R is off the scale, and for that reason it has to be reported as such and not as a typical result.
By half of the period: the first gives +17.2 R (47 trades), the second +8.0 R (42 trades), but without the 16.09 R trade the second half is −8.1 R over 41 trades.
Three trades and the curve
Figure 1 shows a winning trade: a flag of 12-13 April 2026 (swings 4,647.18, 4,738.00 and 4,703.76 $) with an upward breakout on 13 April at 19:00 above 4,738.00 $. The confirmation arrives at 22:00 and the entry is at 4,758.63 $; the stop is at 4,698.93 $ (risk 59.70 $, 3.1 ATR) and the target, with a height of 90.82 $, at 4,849.45 $. The target is reached on 14 April at 22:00: +1.52 R.
Figure 2 shows a losing trade: a falling wedge of 12-13 March (from 5,136.48 to 5,065.05 $) with an upward breakout on 13 March at 11:00 above 5,092.08 $. Entry at the confirmation of 12:00 at 5,116.69 $, stop at 5,052.88 $ (risk 63.81 $, 3.8 ATR), target at 5,188.02 $. The stop is hit at 14:00: −1.00 R.
Figure 3 shows the 16.09 R trade: the symmetrical triangle of 28-30 August (from 4,633.57 to 4,489.99 $) with a downward breakout on 30 August at 20:00 below 4,477.94 $ and entry at 21:00. Figure 4 shows the cumulative R on D1: the steps are small for almost the whole period and one jumps by 16 R on 1 September.
Read the chart in words
H1 gold chart from 12 to 15 April 2026 with the pattern, the breakout, the confirmation candle and the lines of entry, stop and target.
| When | What | Price |
|---|---|---|
| 12 Apr 2026 22:00 UTC | Point 1 | 4,647.18 |
| 13 Apr 2026 07:00 UTC | Point 2 | 4,738.00 |
| 13 Apr 2026 11:00 UTC | Point 3 | 4,703.76 |
| 13 Apr 2026 19:00 UTC | breakout up | 4,746.01 |
| 13 Apr 2026 22:00 UTC | confirmation and entry | 4,743.14 |
| 13 Apr 2026 22:00 UTC | Entry | 4,758.63 |
| 13 Apr 2026 22:00 UTC | Stop | 4,698.93 |
| 13 Apr 2026 22:00 UTC | Target | 4,849.45 |
| 14 Apr 2026 22:00 UTC | target reached | 4,849.97 |
Times in UTC.
Read the chart in words
H1 gold chart from 11 to 13 March 2026 with the pattern, the breakout, the confirmation candle and the lines of entry, stop and target.
| When | What | Price |
|---|---|---|
| 12 Mar 2026 05:00 UTC | Point 1 | 5,136.48 |
| 12 Mar 2026 10:00 UTC | Point 2 | 5,188.52 |
| 12 Mar 2026 20:00 UTC | Point 3 | 5,058.33 |
| 13 Mar 2026 02:00 UTC | Point 4 | 5,126.80 |
| 13 Mar 2026 08:00 UTC | Point 5 | 5,065.05 |
| 13 Mar 2026 11:00 UTC | breakout up | 5,108.42 |
| 13 Mar 2026 12:00 UTC | confirmation and entry | 5,104.01 |
| 13 Mar 2026 12:00 UTC | Entry | 5,116.69 |
| 13 Mar 2026 12:00 UTC | Stop | 5,052.88 |
| 13 Mar 2026 12:00 UTC | Target | 5,188.02 |
| 13 Mar 2026 14:00 UTC | stop hit | 5,038.50 |
Times in UTC.
Read the chart in words
H1 gold chart from 28 August to 2 September 2026 with the pattern, the breakout, the confirmation candle and the lines of entry, stop and target.
| When | What | Price |
|---|---|---|
| 28 Aug 2026 14:00 UTC | Point 1 | 4,633.57 |
| 28 Aug 2026 18:00 UTC | Point 2 | 4,451.50 |
| 29 Aug 2026 23:00 UTC | Point 3 | 4,475.35 |
| 30 Aug 2026 05:00 UTC | Point 4 | 4,470.01 |
| 30 Aug 2026 17:00 UTC | Point 5 | 4,489.99 |
| 30 Aug 2026 20:00 UTC | breakout down | 4,477.12 |
| 30 Aug 2026 21:00 UTC | confirmation and entry | 4,477.50 |
| 30 Aug 2026 21:00 UTC | Entry | 4,474.53 |
| 30 Aug 2026 21:00 UTC | Stop | 4,482.67 |
| 30 Aug 2026 21:00 UTC | Target | 4,290.34 |
| 1 Sept 2026 21:00 UTC | time exit | 4,335.51 |
Times in UTC.
Read the chart in words
Daily gold chart from December 2025 to October 2026 with, below, the curve of the cumulative R of the 89 trades.
| When | What | Price |
|---|---|---|
| 1 Sept 2026 | +16.09 R in one trade | 4,328.00 |
Times in UTC.
Choosing the best variant
A system with many variants invites choosing the best one. A test: for each pattern type with at least 8 trades you choose the variant with the highest mean R (among the six). With hindsight, over the whole period, the 184 chosen trades give +0.29 R. The same procedure applied to the 50 random series gives +0.24 R (from +0.07 to +0.44): choosing the best of six produces a positive result on its own even on series with no structure.
The honest version chooses on the first half and measures on the second. The 88 trades of the second half give +0.12 R; random series with the same procedure +0.02 R (from −0.25 to +0.37). Inside the range of chance.
What the data say
The system cannot be told apart from chance. The result is positive and sits at the upper edge, but a single trade is worth 64% of the sum, the second half of the period without it is negative and choosing the best variant produces a similar result on its own. More data are needed: with a standard deviation of 1 R per trade, an error of ±0.10 R (1.96 × 1 / √n) requires about 384 trades. This book has 89 with a single instrument. The right way to go on is a forward test: same written rules, no changes, at least 100 trades on a demo account or with minimal risk, and the comparison with a random series.
Procedure in five steps
- Write all the rules before looking at the results and use only what is known in real time.
- Measure sum, mean R, maximum drawdown and longest losing streak, with the cost.
- Compare with a random series that has the same rules.
- Remove the best trade (and the three best) and look at the two halves of the period.
- Go on with a forward test before risking money.
Exercises
Chapter exercises
1. On the chart the letters A-D are: the last swing of the wedge, the close of the breakout, the confirmation close and the candle of the stop hit. Which letter is the entry of the system?
Read the chart in words
| When | What | Price |
|---|---|---|
| 13 Mar 2026 08:00 UTC | Point A | 5,065.05 |
| 13 Mar 2026 11:00 UTC | Point B | 5,108.42 |
| 13 Mar 2026 12:00 UTC | Point C | 5,120.46 |
| 13 Mar 2026 14:00 UTC | Point D | 5,038.50 |
Times in UTC.
2. A system sums +25.1 R over 89 trades and one of them is worth +16.09 R. How much remains without that trade?
3. Choosing with hindsight the best variant for each type gives +0.29 R, and the same procedure on random series +0.24 R. What does it mean?
4. With a standard deviation of 1 R per trade, how many trades are needed for an error of ±0.10 R (error = 1.96 × deviation / √n)?
Have a question or want to share your exercise?
Post in the community, or join the free signals room on Telegram.
Educational content, not financial advice. Trading involves risk.