Double Tops and Bottoms, Beyond the Textbook

[Fairman]

Double Tops and Bottoms, Beyond the Textbook

A double top has two peaks at similar highs, and a double bottom has two troughs at similar lows. Textbooks call them reversal patterns. In real trading, they need more careful handling.

diagram.png (38.29 KiB) Viewed 48 times

The standard story: price hits resistance twice, fails, and then breaks the neckline (the low between the two peaks), signaling a downtrend.

The problems traders face:

- Entering too early, before the neckline breaks
- Ignoring that the second peak often sweeps the first one, creating a liquidity trap
- Setting stops too tight, right where they're likely to be hunted

A more careful approach:

1. Identify the pattern on a meaningful timeframe, such as the 1H or higher.
2. Note whether the second peak slightly exceeds the first. That could be a sweep, not a failure of the pattern.
3. Wait for a decisive break of the neckline, preferably with displacement.
4. Look for a retest of the neckline as an entry.
5. Place the stop above the second peak, with a buffer.
6. Target a distance equal to the height of the pattern, or the next liquidity level.

Cautions:

- Many double tops become triple tops or breakouts
- Context and trend matter

Patterns are stories, but not every story has the ending you expect.

Use the neckline as your trigger, and stay patient until it breaks.

[PropScalpDesk]

The textbook combination in steps 4 to 6 has a reward-to-risk problem that's easy to miss until you put numbers on it. Say the peaks are at 1.1050, the neckline is at 1.0990, so the pattern is 60 pips tall and the measured target is 1.0930. You short the retest of the neckline at 1.0990 with the stop above the second peak plus a buffer, around 1.1060. That's 70 pips of risk for 60 pips of target, less than 1:1, and the retest you're waiting for doesn't come at all on many breaks.

So something has to give. Either the stop goes above the retest swing rather than above the second peak, which needs a smaller timeframe to define, or the target goes to the next liquidity beyond the measured move. I usually do the first and accept more stop-outs.

Point 2 is the most useful line in the post. A second peak that pokes a few pips above the first and closes back below is often a better signal than a clean, equal double top.

How often does the neckline retest actually come on the pairs you trade?

[FXS Prop Desk]

Ignoring that the second peak often sweeps the first one, creating a liquidity trap

This is the detail that separates the textbook version from what traders actually see. Two practical consequences. First, a stop placed just above the first peak sits exactly where the sweep is likely to reach, so for a double-top short it belongs above the second peak with some buffer. Second, waiting for a candle close below the neckline, rather than selling at the second peak, filters out many patterns that turn into breakouts.

The neckline retest after the break also offers a cleaner entry with a tighter stop, though it doesn't come every time.

[Shadow Trader]

To answer PropScalpDesk's question from my own log: of 40 H1 double tops on EURUSD and GBPUSD where the neckline broke on a close, 17 came back to retest it within six hours. That is a bit over 40 percent, so waiting for the retest means skipping most of the breaks.

The stop above the retest swing is what makes those 17 worth having. The chart shows a typical one with round numbers, peaks at 1.1050 and 1.1054, the neckline at 1.0990 and the measured target at 1.0930.

img_12719_r1.png (62.88 KiB) Viewed 2 times

The break closed at 1.0978, and the retest topped at 1.0992 before rolling over. Shorting 1.0988 with the stop at 1.1000 risks 12 pips against 58 to the target, instead of 70 against 60 in the textbook version.

The tighter stop gets hit more often, which PropScalpDesk already accepts. Across my 17 retests it was taken out 6 times. Usually that was a second push back to the neckline that then failed anyway.

The 4 pip overshoot on the second peak is the detail from point 2, and on this one it was the best clue. Peaks that are equal to the pip have been less reliable for me than a small sweep with a close back under.