Part 4: Breakouts, system and test · Chapter 10 of 14 · 8 min
True and false breakouts
After the break of a pattern, price comes back with a close within 12 candles in 63% of cases (168 out of 267), as in a random series (62%). Waiting for a confirmation (a close 0.5 ATR beyond the level, with no return) leaves 149 trades out of 268 and gives +0.17 ± 0.27 R: inside the range of chance.
What a false breakout is
The breakout of a pattern is the first close beyond its edge. The book calls a breakout false if, within 12 candles of the close of the breakout, another close goes back to the opposite side of the break level. It is a wide definition (one close is enough) and it is known only after the fact: it serves to measure how often an edge is crossed and then taken back.
How many there are
Of the 268 breakouts of all the patterns of chapters 2-8 (without the V spikes), the false ones are 168 out of 267 (63%, ±6 points; one case has no 12 candles of data after it). In 100 random series with the same candles the share is 62% (5th-95th percentile 57%-69%): the return is the rule, in these data and in chance. By pattern:
| Pattern | False on gold | False in random series (median) |
|---|---|---|
| Head and shoulders | 6 out of 11 (55%) | 67% |
| Double | 38 out of 58 (66%) | 62% |
| Triple | 3 out of 3 | 50% |
| Ascending triangle | 9 out of 11 (82%) | 63% |
| Descending triangle | 4 out of 11 (36%) | 67% |
| Symmetrical triangle | 23 out of 33 (70%) | 65% |
| Rectangle | 2 out of 7 (29%) | 67% |
| Channel | 14 out of 25 (56%) | 65% |
| Rising wedge | 22 out of 31 (71%) | 60% |
| Falling wedge | 11 out of 17 (65%) | 61% |
| Flag | 21 out of 42 (50%) | 65% |
| Pennant | 8 out of 8 | 65% |
| Broadening | 7 out of 10 (70%) | 60% |
The differences between patterns and chance are small and the samples from 3 to 58: the flag comes back less (50% against 65%) and the ascending triangle more (82% against 63%), but with 42 and 11 cases they are inside the margin.
Hindsight and confirmation
If you enter right at the close of the breakout (as in the previous chapters), the 242 trades (26 patterns are skipped for risk) give +0.10 ± 0.26 R (random series +0.01, from −0.13 to +0.21). With hindsight the difference between the breakouts that turn out false and the true ones is enormous: the 154 false ones give −0.41 R, the 88 true ones +0.98 R. But "true or false" is known after 12 candles, not at the close.
The confirmation is an attempt to know it beforehand: you enter only if, within 6 candles of the breakout, a close sits at least 0.5 ATR beyond the level with no close having come back before, and the entry is at that close (stop beyond the opposite extreme, target equal to the height of the pattern from the entry, exit after 48 candles). Out of 268 breakouts the confirmation arrives or the trade is valid in 149 cases (119 with no confirmation or skipped for risk). Mean R +0.17 ± 0.27: 57 targets, 52 stops, 40 time exits. Random series give +0.02 (from −0.19 to +0.25). In the first half of the period +0.17 R (73 trades) and in the second +0.17 (76): the two halves agree, but the error (±0.27 R) is larger than the mean.
Figure 1 shows a breakout that comes back: a symmetrical triangle from 23 to 25 February 2026, upward breakout on 25 February at 16:00 above 5,196.08 $, entry at 5,213.39 $, but at 20:00 a close comes back below the level and the stop (5,165.89 $) is hit: −1.01 R. Figure 2 shows a double bottom (15-16 March, 4,966.22, 5,036.62, 4,971.81 $) with an upward breakout on 17 March at 04:00 above the neckline (5,036.62 $), entry at 5,036.92 $, stop at 4,961.85 $ (4.3 ATR) hit on 18 March at 11:00: −1.00 R. Figure 3 shows a breakout that holds: the rectangle of 15-16 June, upward breakout on 16 June at 07:00 above 4,335.27 $ with no close back, target at 4,370.58 $ reached on 17 June at 17:00: +1.13 R.
Read the chart in words
H1 gold chart from 23 to 26 February 2026 with the five swings numbered, the two lines, the breakout, the close back inside and the entry, stop and target lines.
| When | What | Price |
|---|---|---|
| 23 Feb 2026 23:00 UTC | Point 1 | 5,253.40 |
| 24 Feb 2026 14:00 UTC | Point 2 | 5,101.20 |
| 24 Feb 2026 20:00 UTC | Point 3 | 5,175.24 |
| 24 Feb 2026 23:00 UTC | Point 4 | 5,125.26 |
| 25 Feb 2026 05:00 UTC | Point 5 | 5,211.46 |
| 25 Feb 2026 16:00 UTC | breakout up | 5,215.69 |
| 25 Feb 2026 20:00 UTC | close back inside | 5,147.40 |
| 25 Feb 2026 16:00 UTC | Entry | 5,213.39 |
| 25 Feb 2026 16:00 UTC | Stop | 5,165.89 |
| 25 Feb 2026 16:00 UTC | Target | 5,405.69 |
| 25 Feb 2026 20:00 UTC | stop hit | 5,147.40 |
Times in UTC.
Read the chart in words
H1 gold chart from 15 to 18 March 2026 with the three swings numbered, the neckline, the breakout, the close back inside and the entry, stop and target lines.
| When | What | Price |
|---|---|---|
| 15 Mar 2026 23:00 UTC | Neckline | 5,036.62 |
| 15 Mar 2026 23:00 UTC | Point 1 | 4,966.22 |
| 16 Mar 2026 06:00 UTC | Point 2 | 5,036.62 |
| 16 Mar 2026 10:00 UTC | Point 3 | 4,971.81 |
| 17 Mar 2026 04:00 UTC | breakout up | 5,040.74 |
| 17 Mar 2026 05:00 UTC | close back inside | 5,020.00 |
| 17 Mar 2026 04:00 UTC | Entry | 5,036.92 |
| 17 Mar 2026 04:00 UTC | Stop | 4,961.85 |
| 17 Mar 2026 04:00 UTC | Target | 5,107.32 |
| 18 Mar 2026 11:00 UTC | stop hit | 4,928.84 |
Times in UTC.
Read the chart in words
H1 gold chart from 14 to 18 June 2026 with the five swings numbered, the two lines, the breakout and the entry, stop and target lines.
| When | What | Price |
|---|---|---|
| 15 Jun 2026 02:00 UTC | Point 1 | 4,336.15 |
| 15 Jun 2026 06:00 UTC | Point 2 | 4,303.23 |
| 15 Jun 2026 14:00 UTC | Point 3 | 4,371.13 |
| 15 Jun 2026 22:00 UTC | Point 4 | 4,307.46 |
| 16 Jun 2026 03:00 UTC | Point 5 | 4,335.39 |
| 16 Jun 2026 07:00 UTC | breakout up | 4,337.06 |
| 16 Jun 2026 13:00 UTC | close back inside | 4,333.08 |
| 16 Jun 2026 07:00 UTC | Entry | 4,336.60 |
| 16 Jun 2026 07:00 UTC | Stop | 4,306.87 |
| 16 Jun 2026 07:00 UTC | Target | 4,370.58 |
| 17 Jun 2026 17:00 UTC | target reached | 4,384.94 |
Times in UTC.
Procedure in four steps
- Define "false" beforehand, with a window and a condition on the close.
- Count all the breakouts and compare the share with a random series.
- If you use a confirmation, write it down and count how many trades it leaves.
- Always distinguish what is known at entry from what is known afterwards.
Exercises
Chapter exercises
1. On the chart the four candles are marked with the letters A-D: the last swing of the triangle, the breakout, and two candles after. Which letter is the first close that comes back below the break level?
Read the chart in words
| When | What | Price |
|---|---|---|
| 25 Feb 2026 05:00 UTC | Point A | 5,211.46 |
| 25 Feb 2026 16:00 UTC | Point B | 5,215.69 |
| 25 Feb 2026 20:00 UTC | Point C | 5,147.40 |
| 25 Feb 2026 23:00 UTC | Point D | 5,163.39 |
Times in UTC.
2. Of the 267 pattern breakouts 63% come back within 12 candles, and in a random series 62%. What does it mean?
3. With hindsight the true breakouts give +0.98 R and the false ones −0.41 R. Why can it not be used?
4. Entry with confirmation gives +0.17 ± 0.27 R over 149 trades, and random series +0.02 (−0.19 / +0.25). What do you conclude?
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.