Part 3: Triangles, channels, wedges, flags and the rest · Chapter 7 of 14 · 8 min

Flags and pennants

The flag after a pole of at least 4 ATR: on H1 gold it shows up 42 times and the 41 trades give +0.48 ± 0.45 R, the highest result in the book so far (only 2% of random series do as well), but it is all in the first half of the period (+1.09 against −0.04) and comes from one pattern out of fourteen tested. The pennants are 8 and the 5 valid trades all go to the stop.

The rule

A flag (or a pennant) is a pause after a strong move. The book defines it like this:

  • Pole: a leg between two consecutive swings of at least 4 ATR (H1), completed in at most 12 candles.
  • Consolidation: the next swing retraces at most 50% of the pole, within 30 candles.
  • Flag or pennant: if the consolidation lasts at least 6 candles and in the second half the range is less than 60% of the first, it is a pennant (price narrows); otherwise it is a flag.
  • Breakout: the first close beyond the end of the pole (continuation) or beyond the extreme of the consolidation (failure), within 30 candles of the confirmation of the consolidation.

Trade: entry at the close of the breakout, stop beyond the opposite extreme of the consolidation plus 0.25 ATR (risk 0.5-6 ATR), target equal to the height of the pole, exit after 48 candles, cost 0.30 $.

How many show up

On H1 goldRandom series: median (5th-95th)Breakout as continuationAfter the break, +1.5 ATR first (10 candles)
Flag4249 (42-60)32 out of 42 (76%; random 64%)26 (62%, ±15 points)
Pennant816 (10-21)5 out of 8 (63%; random 50%)5 (63%, ±34 points)

The flags are 42 (4 on H4, 5 on the daily chart), the lowest number of chance (42-60); they last a median of 12 candles and the pole is a median of 85.1 $ high (5.5 ATR). Continuation after the break is 76% against 64% for a random series (the difference sits inside the error: 42 cases, ±13 points), and price travels 1.5 ATR in the direction of the break first in 62% of cases (±15 points), against 37% for any candle. The pennants are 8 (1 on H4): they last 24 candles and the breakout comes a median of 13 candles after the last swing. Their "height" is 115.9 $ but the ATR at the breakout is very small (the consolidation reduces volatility), so in ATR it looks enormous (38.8 ATR).

The trades

TradesMean RTarget, stop, timeMedian target/risk (break-even)Random: median (5th-95th), at leastFirst / second half
Flag41+0.48 ± 0.4514, 14, 131.94 (34%)+0.02 (−0.31 / +0.37), 2%+1.09 (19) / −0.04 (22)
Pennant5−1.02 ± 0.010, 5, 02.16 (32%)−0.06 (−0.65 / +0.77), 100%−1.02 (3) / −1.01 (2)

The flag is the most noticeable result of the book: 14 targets, 14 stops, 13 time exits; 22 winners out of 41; with a target worth a median of 1.94 times the risk break-even would be at 34% of targets, and the data have 34% (14 out of 41). Only 2% of random series reach a mean R this high. But it should be read with three cautions. First, the result is all in the first half of the period: +1.09 R with 19 trades and −0.04 R with 22 in the second. Second, it is one pattern out of fourteen tested: if fourteen patterns are tested, the probability that at least one looks as rare as 2% by pure chance is not small (about 25%). Third, without the three best (+4.69, +2.96, +2.33 R) the mean R falls to +0.26. It is an indication to retest on new data, not a result.

The pennants are 8 and 3 are skipped for a risk beyond 6 ATR: the 5 trades all go to the stop (−1.02 R), but with 5 cases nothing can be said.

Figure 1 shows a winning flag: pole from the low of 4,540.03 $ (16 January 2026 at 15:00) to 4,601.72 $, consolidation with a low at 4,591.10 $, upward breakout on 18 January at 01:00, entry at 4,603.37 $, stop at 4,590.29 $ (4.1 ATR), target at 4,665.06 $ (the height of the pole, 61.69 $) reached at 23:00: +4.69 R. Figure 2 shows a flag stopped out: pole from 4,744.78 to 4,814.50 $ (19-20 April), upward breakout on 20 April at 13:00, entry at 4,818.37 $, stop at 4,776.84 $ hit on 21 April at 07:00: −1.01 R. Figure 3 shows a pennant: downward pole from 4,561.89 to 4,469.25 $ (19 May 2026), downward breakout on 20 May at 02:00, entry at 4,468.12 $, stop at 4,517.77 $ hit at 14:00: −1.01 R.

Figure 1. A winning flag: pole from 4,540.03 to 4,601.72 (16-17 January 2026), upward breakout on 18 January, target at 4,665.06 reached (+4.69 R).Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 16 Jan 2026 09:00 UTC to 19 Jan 2026 07:00 UTC.
Read the chart in words

H1 gold chart from 16 to 19 January 2026 with the three swings numbered (start and end of the pole, extreme of the consolidation), the breakout level and the entry, stop and target lines.

WhenWhatPrice
16 Jan 2026 15:00 UTCPoint 14,540.03
17 Jan 2026 03:00 UTCPoint 24,601.72
17 Jan 2026 18:00 UTCPoint 34,591.10
18 Jan 2026 01:00 UTCbreakout up4,604.97
18 Jan 2026 01:00 UTCEntry4,603.37
18 Jan 2026 01:00 UTCStop4,590.29
18 Jan 2026 01:00 UTCTarget4,665.06
18 Jan 2026 23:00 UTCtarget reached4,690.99

Times in UTC.

Figure 2. A flag stopped out: pole from 4,744.78 to 4,814.50, upward breakout on 20 April 2026, stop at 4,776.84 hit on 21 April (−1.01 R).Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 19 Apr 2026 17:00 UTC to 21 Apr 2026 15:00 UTC.
Read the chart in words

H1 gold chart from 19 to 21 April 2026 with the three swings numbered, the breakout level and the entry, stop and target lines.

WhenWhatPrice
19 Apr 2026 23:00 UTCPoint 14,744.78
20 Apr 2026 01:00 UTCPoint 24,814.50
20 Apr 2026 05:00 UTCPoint 34,781.28
20 Apr 2026 13:00 UTCbreakout up4,826.09
20 Apr 2026 13:00 UTCEntry4,818.37
20 Apr 2026 13:00 UTCStop4,776.84
20 Apr 2026 13:00 UTCTarget4,888.09
21 Apr 2026 07:00 UTCstop hit4,776.00

Times in UTC.

Figure 3. A pennant stopped out: downward pole from 4,561.89 to 4,469.25, downward breakout on 20 May 2026, stop at 4,517.77 hit at 14:00 (−1.01 R).Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 19 May 2026 02:00 UTC to 20 May 2026 22:00 UTC.
Read the chart in words

H1 gold chart from 19 to 20 May 2026 with the three swings numbered, the breakout level and the entry, stop and target lines.

WhenWhatPrice
19 May 2026 08:00 UTCPoint 14,561.89
19 May 2026 13:00 UTCPoint 24,469.25
20 May 2026 00:00 UTCPoint 34,512.00
20 May 2026 02:00 UTCbreakout down4,462.32
20 May 2026 02:00 UTCEntry4,468.12
20 May 2026 02:00 UTCStop4,517.77
20 May 2026 02:00 UTCTarget4,375.48
20 May 2026 14:00 UTCstop hit4,538.53

Times in UTC.

Procedure in five steps

  1. Write the rule: pole (size and duration), consolidation (maximum retracement), breakout.
  2. Define the stop on the consolidation and a target equal to the pole.
  3. Count all the flags and the trades skipped.
  4. Compare with a random series, split the period and look without the three best.
  5. Count how many patterns you tested and keep the highest result as a hypothesis to verify.

Exercises

Chapter exercises

1. On the chart the three swings of a flag are marked with the letters A, B and C. Which letter is the end of the pole (the level price must exceed for the continuation)?

Chart exercise. Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 16 Jan 2026 09:00 UTC to 18 Jan 2026 09:00 UTC.
Read the chart in words
WhenWhatPrice
16 Jan 2026 15:00 UTCPoint A4,540.03
17 Jan 2026 03:00 UTCPoint B4,601.72
17 Jan 2026 18:00 UTCPoint C4,591.10
18 Jan 2026 01:00 UTCbreakout up4,604.97

Times in UTC.

2. The pole of a flag is 60 high and the stop sits 15 from the entry (hypothetical numbers). What is the target worth in R and at what percentage of targets do you break even?

3. The flag gives +0.48 ± 0.45 R and only 2% of random series do as well, but the first half of the period gives +1.09 and the second −0.04. What do you conclude?

4. Fourteen patterns are tested and one has a result that chance obtains in 2% of trials. Why is it not a proof?

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.