Gartley

The classic five-point (X-A-B-C-D) harmonic pattern with a 0.618 B retracement and a 0.786 D completion. Bullish (D a swing low) → +1, bearish → -1.

Quick reference

Item	Value
Family	Harmonic Patterns
Input type	Candle (uses high, low)
Output type	f64 (+1 / -1 / 0)
Output range	{-1.0, 0.0, +1.0}
Default parameters	none (swing threshold 5%, baked)
Warmup period	6
Interpretation	Reversal at the D completion

Formula

swing pivots confirmed by a 5% non-repainting zig-zag (pattern_swing)
last five pivots X-A-B-C-D, legs measured as absolute price differences:
  AB / XA ∈ [0.55, 0.70]   (≈ 0.618)
  BC / AB ∈ [0.382, 0.886]
  CD / BC ∈ [1.13, 1.618]
  AD / XA ∈ [0.74, 0.84]   (≈ 0.786 — the defining D completion)
direction: D a swing low → +1 (bullish), D a swing high → -1 (bearish)

See crates/wickra-core/src/indicators/gartley.rs.

Parameters

None. The swing threshold (SWING_THRESHOLD = 0.05) is a baked-in family constant (pattern_swing.rs) and the Fibonacci windows are documented constants in the detector. Gartley::new is infallible.

Inputs / Outputs

const _: fn(&mut wickra::Gartley, wickra::Candle) -> Option<f64> =
    <wickra::Gartley as wickra::Indicator>::update;

• Python. update((o,h,l,c,v,ts)) → float (never None); batch(open, high, low, close) → 1-D ndarray.
• Node. update(open, high, low, close) → number; batch(open, high, low, close) → number[].
• WASM. update(open, high, low, close) → number.

Warmup

warmup_period() == 6. Five confirmed pivots are required; the earliest bar that can confirm a fifth pivot is the sixth. Pinned by test accessors_and_metadata.

Edge cases

• Bullish Gartley reports +1 (test bullish_gartley_is_plus_one).
• Bearish Gartley reports -1 (test bearish_gartley_is_minus_one).
• Legs outside the windows report 0.0 (test out_of_ratio_does_not_trigger).
• reset clears state (test reset_clears_state).
• Streaming equals batch (test batch_equals_streaming).

Examples

Rust

use wickra::{Candle, Gartley, Indicator};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    // X=100 A=140 B=115.3 C=127.65 D=108.56 → AB/XA≈0.618, AD/XA≈0.786.
    let bars = [
        (149.85, 150.0, 149.85, 149.85),
        (100.0, 148.5, 100.0, 100.0),
        (101.0, 140.0, 101.0, 101.0),
        (115.3, 138.6, 115.3, 115.3),
        (116.453, 127.65, 116.453, 116.453),
        (108.56, 126.3735, 108.56, 108.56),
        (109.6456, 119.416, 109.6456, 109.6456), // D confirms → bullish
    ];
    let mut pat = Gartley::new();
    let mut last = 0.0;
    for (ts, (o, h, l, c)) in bars.iter().enumerate() {
        last = pat.update(Candle::new(*o, *h, *l, *c, 1.0, ts as i64)?).unwrap();
    }
    println!("{last}"); // 1
    Ok(())
}

Python

import wickra as ta

bars = [
    (149.85, 150.0, 149.85, 149.85, 1.0, 0),
    (100.0, 148.5, 100.0, 100.0, 1.0, 1),
    (101.0, 140.0, 101.0, 101.0, 1.0, 2),
    (115.3, 138.6, 115.3, 115.3, 1.0, 3),
    (116.453, 127.65, 116.453, 116.453, 1.0, 4),
    (108.56, 126.3735, 108.56, 108.56, 1.0, 5),
    (109.6456, 119.416, 109.6456, 109.6456, 1.0, 6),
]
pat = ta.Gartley()
print([pat.update(b) for b in bars][-1])  # 1.0

Node

const wickra = require('wickra');
const bars = [
  [149.85, 150.0, 149.85, 149.85], [100.0, 148.5, 100.0, 100.0],
  [101.0, 140.0, 101.0, 101.0], [115.3, 138.6, 115.3, 115.3],
  [116.453, 127.65, 116.453, 116.453], [108.56, 126.3735, 108.56, 108.56],
  [109.6456, 119.416, 109.6456, 109.6456],
];
const pat = new wickra.Gartley();
let last = 0;
for (const [o, h, l, c] of bars) last = pat.update(o, h, l, c);
console.log(last); // 1

Streaming

pat = ta.Gartley()
for o, h, l, c, v, ts in candle_feed:
    signal = pat.update((o, h, l, c, v, ts))
    if signal > 0:
        pass  # bullish Gartley completed at D — reversal long setup
    elif signal < 0:
        pass  # bearish Gartley — reversal short setup

Interpretation

1. Reversal at D. The signal fires when price completes the D leg inside the Gartley windows; the potential reversal zone is around D. Combine with the structure's stop (just beyond X) for risk control.
2. Tightest of the four. Gartley's 0.786 D is shallower than Bat (0.886), Butterfly/Crab (extensions) — the family differs only in these ratio windows.

Common pitfalls

• Window tolerance. The windows use standard harmonic ranges with a small tolerance; real-world patterns that miss a ratio by more than the band will not register — by design, to avoid false positives.
• Confirmation lag. Non-repainting but lags the visual D by the threshold move that confirms the pivot.

References

• Gartley, H. M. Profits in the Stock Market (1935); Carney, S. Harmonic Trading (2010) for the modern Fibonacci ratios.

See also

• Bat, Butterfly, Crab, Shark, Cypher.
• Indicators-Overview — full taxonomy.