Strategy Detail
Harmonic Patterns
Trade a named harmonic pattern. Each one is a five-point price structure defined by Fibonacci ratios; the strategy enters at the completion point, targets a retracement, and stops just beyond the structure.
Harmonic patterns are price structures defined by Fibonacci ratios between swing points. A handful of swing highs and lows, in the right proportions, mark a spot where price has historically turned. The classic names, Gartley, Bat, Butterfly, Crab, Cypher, all describe the same idea with different ratio rules: find where a move is likely exhausted, then trade the reversal.
This strategy trades one named pattern at a time. On every bar it looks for the pattern’s geometry in the recent swing points. When the final point completes a valid structure, it enters in the reversal direction with a fixed plan: a target at a Fibonacci retracement of the move, and a stop just beyond the completion point.
How it finds a pattern
Every harmonic pattern is built from swing pivots, the local highs and lows that anchor a structure. The strategy detects those pivots as they confirm, then checks whether the most recent five (four for AB=CD) hold the pattern’s ratio rules within a tolerance.
A pivot only counts once enough later bars confirm it is a genuine turn, so the strategy never acts on a swing that has not formed yet. Entry happens on the bar the final pivot confirms, which is the earliest point a trader could actually act. That makes the backtest honest: it never enters at a price it could only know in hindsight.
Strictness
Strictness is a single control that bundles two things: how significant a swing has to be before it counts as a pivot, and how close the ratios have to be to the textbook numbers. Tighter settings produce fewer, cleaner structures; looser settings produce more setups with more noise.
| Strictness | Swing size | Ratio tolerance | Character |
|---|---|---|---|
| Strict | 4 bars each side | 3% | Fewest setups, closest to the textbook shape |
| Balanced | 3 bars each side | 5% | The default middle ground |
| Loose | 2 bars each side | 9% | Most setups, looser shapes, more noise |
“Swing size” is how many bars on each side a high or low must exceed to count as a pivot. “Ratio tolerance” is how far each Fibonacci ratio may stray from its ideal value and still qualify. These are the exact values the engine uses, so a run is fully reproducible from the pattern, strictness, contract, and timeframe.
Timeframe
Harmonic structures are rare on daily bars. A strict Gartley turns up only about once a decade on a daily chart, far too seldom to judge. They are far more common intraday, so 15-minute is the default: cleaner swings and fewer but more significant structures. 5-minute produces many more setups at the cost of more noise. You can test any timeframe; the slower ones simply fire less often, which the trade count makes plain.
How it trades the completion
When a pattern completes, the strategy enters at the close of the confirming bar. It places a target at a 0.618 retracement of the structure’s anchor leg and a stop just beyond the completion point, both sent as a bracket so the trade exits on its own when either is reached. If neither is hit within a fixed window, the trade closes at the market.
The sample backtests below are real results, with commissions and slippage, not a count of how often the textbook shape “worked”. Some patterns reverse cleanly and some do not; the numbers tell you which.
Patterns
The nine patterns below share this same entry, target, and stop logic. They differ only in geometry: which Fibonacci ratios define a valid structure and how far the completion point sits from the origin. Each links to its full definition in the pattern catalog, with the ratio rules and the idea behind it, alongside sample backtests across NQ, ES, and gold.
FAQ
Do harmonic patterns work?
Harmonic patterns like the Gartley, Bat, Butterfly, Crab, and Cypher are described everywhere but rarely tested. This family scores each completed pattern by the trade its geometry implies: enter at the completion point, target a 0.618 retracement, and stop just beyond the structure, with commissions and slippage included. The win rate on each pattern page is a real backtest result rather than a count of how often the textbook shape “worked”, so open a pattern and run it to see what the data shows.
Are harmonic patterns reliable, and do they repaint?
The detection is built to avoid hindsight: a pivot only counts once enough later bars confirm it is a genuine turn, and entry happens on the bar the final pivot confirms, the earliest point a trader could actually act. That makes the backtest honest about what was knowable in real time. Reliability still varies by pattern, since some reverse cleanly and some do not, so the per-pattern numbers are the way to tell them apart.
What timeframe and strictness should I test harmonic patterns on?
Harmonic structures are rare on daily bars, where a strict Gartley turns up only about once a decade, so this strategy runs on 15-minute and 5-minute bars, with 15-minute as the default for cleaner swings. Strictness bundles how significant a swing must be and how close the ratios must match the textbook, with tighter settings giving fewer, cleaner structures and looser settings giving more setups with more noise. The catalog lets you compare patterns, and each page reports the backtested result for the settings used.
Pattern catalog (9)
Each pattern shows its shape and a short definition, then its sample backtests run at identical defaults. The shape and description link to the full pattern page.
Patterns
- B = 0.618 of XA
- C = 0.382-0.886 of AB
- D = 1.272-1.618 of BC
- D = 0.786 of XA
Gartley
View pattern →The original harmonic pattern: a 0.786 retracement completing an XABCD structure, read as an exhausted pullback turning back with the larger trend.
ES 15m · Gartley · 2020-2024
Gartley · balanced
- Win rate
- 62.50%
- Profit factor
- 1.40
- Max drawdown
- 1.18%
ES 5m · Gartley · 2020-2024
Gartley · balanced
- Win rate
- 45.28%
- Profit factor
- 0.49
- Max drawdown
- 4.93%
GC 15m · Gartley · 2020-2024
Gartley · balanced
- Win rate
- 40.00%
- Profit factor
- 0.41
- Max drawdown
- 2.74%
GC 5m · Gartley · 2020-2024
Gartley · balanced
- Win rate
- 59.68%
- Profit factor
- 1.35
- Max drawdown
- 1.66%
NQ 15m · Gartley · 2020-2024
Gartley · balanced
- Win rate
- 56.52%
- Profit factor
- 0.94
- Max drawdown
- 1.77%
NQ 5m · Gartley · 2020-2024
Gartley · balanced
- Win rate
- 46.25%
- Profit factor
- 0.69
- Max drawdown
- 7.55%
- B = 0.382-0.5 of XA
- C = 0.382-0.886 of AB
- D = 1.618-2.618 of BC
- D = 0.886 of XA
Bat
View pattern →A deep 0.886 completion that lands close to the structure's extreme, letting a trade enter near it with a tight stop just beyond.
ES 15m · Bat · 2020-2024
Bat · balanced
- Win rate
- 64.71%
- Profit factor
- 1.05
- Max drawdown
- 1.30%
ES 5m · Bat · 2020-2024
Bat · balanced
- Win rate
- 52.17%
- Profit factor
- 0.50
- Max drawdown
- 5.25%
GC 15m · Bat · 2020-2024
Bat · balanced
- Win rate
- 40.00%
- Profit factor
- 0.80
- Max drawdown
- 1.89%
GC 5m · Bat · 2020-2024
Bat · balanced
- Win rate
- 50.88%
- Profit factor
- 0.85
- Max drawdown
- 3.05%
NQ 15m · Bat · 2020-2024
Bat · balanced
- Win rate
- 60.00%
- Profit factor
- 0.58
- Max drawdown
- 4.85%
NQ 5m · Bat · 2020-2024
Bat · balanced
- Win rate
- 46.15%
- Profit factor
- 0.47
- Max drawdown
- 10.63%
- B = 0.786 of XA
- C = 0.382-0.886 of AB
- D = 1.618-2.24 of BC
- D = 1.272-1.618 of XA
Butterfly
View pattern →An extension pattern that completes beyond the origin, fading exhaustion as price prints a fresh extreme.
ES 15m · Butterfly · 2020-2024
Butterfly · balanced
- Win rate
- 58.62%
- Profit factor
- 1.37
- Max drawdown
- 1.79%
ES 5m · Butterfly · 2020-2024
Butterfly · balanced
- Win rate
- 52.86%
- Profit factor
- 0.80
- Max drawdown
- 2.90%
GC 15m · Butterfly · 2020-2024
Butterfly · balanced
- Win rate
- 64.71%
- Profit factor
- 0.64
- Max drawdown
- 4.88%
GC 5m · Butterfly · 2020-2024
Butterfly · balanced
- Win rate
- 58.89%
- Profit factor
- 1.10
- Max drawdown
- 3.13%
NQ 15m · Butterfly · 2020-2024
Butterfly · balanced
- Win rate
- 60.53%
- Profit factor
- 1.14
- Max drawdown
- 6.60%
NQ 5m · Butterfly · 2020-2024
Butterfly · balanced
- Win rate
- 59.52%
- Profit factor
- 0.91
- Max drawdown
- 4.59%
- B = 0.382-0.618 of XA
- C = 0.382-0.886 of AB
- D = 2.618-3.618 of BC
- D = 1.618 of XA
Crab
View pattern →The deepest extension in the family, completing at a 1.618 projection, built for sharp reversals from overshoots.
ES 15m · Crab · 2020-2024
Crab · balanced
- Win rate
- 42.86%
- Profit factor
- 0.43
- Max drawdown
- 1.54%
ES 5m · Crab · 2020-2024
Crab · balanced
- Win rate
- 55.56%
- Profit factor
- 0.74
- Max drawdown
- 2.47%
GC 15m · Crab · 2020-2024
Crab · balanced
- Win rate
- 25.00%
- Profit factor
- 0.37
- Max drawdown
- 2.11%
GC 5m · Crab · 2020-2024
Crab · balanced
- Win rate
- 33.33%
- Profit factor
- 0.27
- Max drawdown
- 3.35%
NQ 15m · Crab · 2020-2024
Crab · balanced
- Win rate
- 56.25%
- Profit factor
- 0.92
- Max drawdown
- 3.39%
NQ 5m · Crab · 2020-2024
Crab · balanced
- Win rate
- 46.43%
- Profit factor
- 0.89
- Max drawdown
- 3.65%
- B = 0.886 of XA
- C = 0.382-0.886 of AB
- D = 2.0-3.618 of BC
- D = 1.618 of XA
Deep Crab
View pattern →A Crab variant with a deeper 0.886 B point, completing at the same 1.618 extension.
ES 15m · Deep Crab · 2020-2024
Deep Crab · balanced
- Win rate
- 76.92%
- Profit factor
- 3.09
- Max drawdown
- 1.10%
ES 5m · Deep Crab · 2020-2024
Deep Crab · balanced
- Win rate
- 46.81%
- Profit factor
- 1.02
- Max drawdown
- 4.27%
GC 15m · Deep Crab · 2020-2024
Deep Crab · balanced
- Win rate
- 35.00%
- Profit factor
- 0.30
- Max drawdown
- 4.71%
GC 5m · Deep Crab · 2020-2024
Deep Crab · balanced
- Win rate
- 48.78%
- Profit factor
- 1.20
- Max drawdown
- 1.75%
NQ 15m · Deep Crab · 2020-2024
Deep Crab · balanced
- Win rate
- 50.00%
- Profit factor
- 1.46
- Max drawdown
- 3.78%
NQ 5m · Deep Crab · 2020-2024
Deep Crab · balanced
- Win rate
- 47.62%
- Profit factor
- 0.91
- Max drawdown
- 4.71%
- B = 0.382-0.618 of XA
- C = 1.272-1.414 of XA
- D = 0.786 of XC
Cypher
View pattern →A distinct geometry where C pushes past A, then price retraces 0.786 of that whole move to complete at D.
ES 15m · Cypher · 2020-2024
Cypher · balanced
- Win rate
- 50.00%
- Profit factor
- 0.26
- Max drawdown
- 3.01%
ES 5m · Cypher · 2020-2024
Cypher · balanced
- Win rate
- 73.08%
- Profit factor
- 1.48
- Max drawdown
- 0.58%
GC 15m · Cypher · 2020-2024
Cypher · balanced
- Win rate
- 66.67%
- Profit factor
- 1.26
- Max drawdown
- 1.17%
GC 5m · Cypher · 2020-2024
Cypher · balanced
- Win rate
- 61.29%
- Profit factor
- 0.92
- Max drawdown
- 2.18%
NQ 15m · Cypher · 2020-2024
Cypher · balanced
- Win rate
- 57.89%
- Profit factor
- 0.65
- Max drawdown
- 5.49%
NQ 5m · Cypher · 2020-2024
Cypher · balanced
- Win rate
- 62.07%
- Profit factor
- 0.69
- Max drawdown
- 2.41%
- XA = 0.382-0.886 of OX
- AB = 1.13-1.618 of XA
- BC = 1.618-2.24 of AB
Shark
View pattern →A five-point structure of two strong extensions into a reaction point where the move is expected to stall.
ES 15m · Shark · 2020-2024
Shark · balanced
- Win rate
- 52.66%
- Profit factor
- 0.64
- Max drawdown
- 18.18%
ES 5m · Shark · 2020-2024
Shark · balanced
- Win rate
- 56.51%
- Profit factor
- 0.81
- Max drawdown
- 17.34%
GC 15m · Shark · 2020-2024
Shark · balanced
- Win rate
- 56.76%
- Profit factor
- 0.93
- Max drawdown
- 7.95%
GC 5m · Shark · 2020-2024
Shark · balanced
- Win rate
- 57.74%
- Profit factor
- 0.89
- Max drawdown
- 8.69%
NQ 15m · Shark · 2020-2024
Shark · balanced
- Win rate
- 51.43%
- Profit factor
- 0.63
- Max drawdown
- 28.48%
NQ 5m · Shark · 2020-2024
Shark · balanced
- Win rate
- 56.51%
- Profit factor
- 0.85
- Max drawdown
- 24.56%
- Pullbacks ~0.5-0.786 of the prior drive
- Drives ~1.13-1.618 of the prior pullback
- Drive 3 ~ Drive 2 in size
Three Drives
View pattern →Three symmetric pushes to a high or low, each separated by a partial pullback, read as a climax.
ES 15m · Three Drives · 2020-2024
Three Drives · balanced
- Win rate
- 55.56%
- Profit factor
- 1.31
- Max drawdown
- 3.27%
ES 5m · Three Drives · 2020-2024
Three Drives · balanced
- Win rate
- 53.82%
- Profit factor
- 0.97
- Max drawdown
- 7.90%
GC 15m · Three Drives · 2020-2024
Three Drives · balanced
- Win rate
- 57.53%
- Profit factor
- 1.18
- Max drawdown
- 6.10%
GC 5m · Three Drives · 2020-2024
Three Drives · balanced
- Win rate
- 59.10%
- Profit factor
- 1.21
- Max drawdown
- 5.15%
NQ 15m · Three Drives · 2020-2024
Three Drives · balanced
- Win rate
- 58.15%
- Profit factor
- 1.21
- Max drawdown
- 5.59%
NQ 5m · Three Drives · 2020-2024
Three Drives · balanced
- Win rate
- 53.38%
- Profit factor
- 0.92
- Max drawdown
- 17.03%
- C = 0.618-0.786 retrace of AB
- CD = AB (equal legs)
AB=CD
View pattern →The building block of every harmonic pattern: two equal-length legs around a partial retracement.
ES 15m · AB=CD · 2020-2024
AB=CD · balanced
- Win rate
- 56.28%
- Profit factor
- 1.03
- Max drawdown
- 5.92%
ES 5m · AB=CD · 2020-2024
AB=CD · balanced
- Win rate
- 53.93%
- Profit factor
- 0.95
- Max drawdown
- 12.78%
GC 15m · AB=CD · 2020-2024
AB=CD · balanced
- Win rate
- 57.24%
- Profit factor
- 1.05
- Max drawdown
- 8.70%
GC 5m · AB=CD · 2020-2024
AB=CD · balanced
- Win rate
- 53.89%
- Profit factor
- 0.87
- Max drawdown
- 16.53%
NQ 15m · AB=CD · 2020-2024
AB=CD · balanced
- Win rate
- 54.69%
- Profit factor
- 1.04
- Max drawdown
- 11.47%
NQ 5m · AB=CD · 2020-2024
AB=CD · balanced
- Win rate
- 59.12%
- Profit factor
- 1.07
- Max drawdown
- 10.96%