Auto-Fibonacci

Fibonacci retracement anchored automatically on the dominant (largest-magnitude) leg among the last six confirmed pivots, rather than the immediate last leg.

Quick reference

Item	Value
Family	Fibonacci
Input type	Candle (uses high, low)
Output type	AutoFibOutput (7 prices)
Output range	bracketed by the dominant leg's start and end
Default parameters	none (swing threshold 5%, 6-pivot history, baked)
Warmup period	2 (two confirmed pivots)
Interpretation	Retracement of the swing the market most respects

Formula

among the last six pivots, pick the leg with the largest |start - end|;
  level(r) = end + r * (start - end)
for r in {0, 0.236, 0.382, 0.5, 0.618, 0.786, 1.0}

Identical level maths to FibRetracement, but the leg is chosen by magnitude across the recent pivot window instead of always the last one. See crates/wickra-core/src/indicators/auto_fib.rs.

Parameters

None. The swing threshold 0.05 and the six-pivot history are baked-in family constants. AutoFib::new is infallible.

Inputs / Outputs

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

• Python. update((o,h,l,c,v,ts)) → (level_0, level_236, level_382, level_500, level_618, level_786, level_1000) or None; batch(high, low) → (n, 7) ndarray (NaN warmup).
• Node. update(high, low) → { level0, level236, level382, level500, level618, level786, level1000 } or null; batch(high, low) → flat number[] length n*7.
• WASM. update(high, low) → object (same camelCase keys) or null.

Warmup

warmup_period() == 2. At least one leg (two pivots) must exist; before that update returns None. Pinned by tests accessors_and_metadata and no_output_before_two_pivots.

Edge cases

• Anchors on the largest leg among several recent swings, not the last one (test anchors_on_the_largest_leg).
• reset clears all state (test reset_clears_state).
• Streaming equals batch (test batch_equals_streaming).

Examples

Rust

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

fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Pivots 130 -> 120 (10) -> 220 (100, dominant) -> 200 (20).
    let bars = [
        (129.87, 130.0, 129.87, 129.87),
        (120.0, 128.7, 120.0, 120.0),  // confirm high @130
        (121.2, 220.0, 121.2, 121.2),  // confirm low  @120
        (200.0, 217.8, 200.0, 200.0),  // confirm high @220
        (202.0, 220.0, 202.0, 202.0),  // confirm low  @200
    ];
    let mut fib = AutoFib::new();
    let mut last = None;
    for (ts, (o, h, l, c)) in bars.iter().enumerate() {
        last = fib.update(Candle::new(*o, *h, *l, *c, 1.0, ts as i64)?);
    }
    let v = last.unwrap();
    // Dominant leg 120 -> 220: 0% on 220, 100% on 120.
    println!("{} {} {}", v.level_0, v.level_500, v.level_1000); // 220 170 120
    Ok(())
}

Python

import wickra as ta

bars = [
    (129.87, 130.0, 129.87, 129.87, 1.0, 0),
    (120.0, 128.7, 120.0, 120.0, 1.0, 1),
    (121.2, 220.0, 121.2, 121.2, 1.0, 2),
    (200.0, 217.8, 200.0, 200.0, 1.0, 3),
    (202.0, 220.0, 202.0, 202.0, 1.0, 4),
]
fib = ta.AutoFib()
print([fib.update(b) for b in bars][-1])
# (220.0, 196.4, 181.8, 170.0, 158.2, 141.4, 120.0)

Node

const wickra = require('wickra');
const fib = new wickra.AutoFib();
const bars = [[130.0, 129.87], [128.7, 120.0], [220.0, 121.2],
              [217.8, 200.0], [220.0, 202.0]];
let last = null;
for (const [h, l] of bars) last = fib.update(h, l);
console.log(last.level0, last.level1000); // 220 120

Streaming

fib = ta.AutoFib()
for o, h, l, c, v, ts in candle_feed:
    levels = fib.update((o, h, l, c, v, ts))
    if levels is not None:
        golden = levels[4]  # 61.8% of the dominant swing

Interpretation

1. Significance over recency. Markets often respect the largest recent swing, not the most recent micro-leg; Auto-Fib tracks that dominant structure.
2. Drop-in for discretionary fibs. Mirrors how a trader picks the "obvious" big swing to draw retracements from.

Common pitfalls

• Window-bounded. Only the last six pivots are considered; a historically huge leg eventually drops out of the window.
• Switches anchor. When a new, larger leg confirms, the levels jump to it — expected behaviour, but the output is not monotone across time.

References

• Fischer, R. Fibonacci Applications and Strategies for Traders (1993).

See also

• FibRetracement, GoldenPocket, FibConfluence, ZigZag.
• Indicators-Overview — full taxonomy.