AdaptiveCci

Lambert's CCI with an efficiency-ratio-adaptive centre line — it leads in trends and stays calm in chop.

Quick reference

Field	Value
Family	Ehlers / Cycle (DSP)
Input type	Candle (high / low / close)
Output type	f64
Output range	unbounded around 0 (≈ ±100 bands)
Default parameters	(period = 20) (Python)
Warmup period	period
Interpretation	> +100 overbought / strong up; < −100 oversold / strong down.

Formula

TP   = (high + low + close) / 3
ER   = |TP_t − TP_oldest| / Σ |ΔTP| over the window      (0..1)
sc   = ( ER·(2/3 − 2/31) + 2/31 )²
mean += sc·(TP_t − mean)                                  (adaptive centre, seeded with SMA)
MD   = mean(|TP_i − mean|) over the window
CCI  = (TP_t − mean) / (0.015 · MD)

The classic Cci centres typical price on an SMA; that SMA's lag delays the oscillator in fast moves. Driving the centre line with a KAMA-style adaptive average (via Kaufman's efficiency ratio) lets it accelerate toward price in a clean trend and slow in noise, so the CCI reaches its bands sooner on real moves and pokes them less on chop. The 0.015 constant preserves Lambert's ±100 convention. Source: crates/wickra-core/src/indicators/adaptive_cci.rs.

Parameters

Name	Type	Default	Valid range	Source	Description
period	usize	20 (Python)	>= 2	adaptive_cci.rs:60	Window for the efficiency ratio, centre line and mean deviation. 0 errors with Error::PeriodZero; < 2 with Error::InvalidPeriod.

The period getter returns the window; value returns the current output if ready.

Inputs / Outputs

From crates/wickra-core/src/indicators/adaptive_cci.rs:

use wickra::{Candle, Indicator, AdaptiveCci};
// AdaptiveCci: Input = Candle, Output = f64
const _: fn(&mut AdaptiveCci, Candle) -> Option<f64> = <AdaptiveCci as Indicator>::update;

A Candle in, an Option<f64> out. The Python binding takes a candle for update and three numpy columns (high, low, close) for batch; Node takes update(high, low, close) and batch(high[], low[], close[]) (NaN warmup).

Warmup

warmup_period() == period. The first value lands once the window is full (first_emission_at_warmup_period pins this).

Edge cases

• Uptrend → positive. A rising typical price gives a positive CCI (uptrend_is_positive pins this).
• Downtrend → negative. A falling typical price gives a negative CCI (downtrend_is_negative pins this).
• Flat window → 0. Zero mean deviation returns 0 rather than dividing by zero (flat_window_is_zero pins this).
• Finiteness. Candle::new rejects non-finite fields, so no in-method guard is needed.
• Reset. c.reset() clears the window, the adaptive mean and the last value (reset_clears_state).

Examples

Rust

use wickra::{BatchExt, Candle, Indicator, AdaptiveCci};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut c = AdaptiveCci::new(10)?;
    let candles: Vec<Candle> = (0..40)
        .map(|i| { let b = 100.0 + f64::from(i); Candle::new(b, b, b, b, 1_000.0, 0).unwrap() })
        .collect();
    println!("last > 0: {}", c.batch(&candles).last().unwrap().unwrap() > 0.0);
    Ok(())
}

Output:

last > 0: true

Python

import numpy as np
import wickra as ta

c = ta.AdaptiveCci(20)
n = 60
base = 100 + np.sin(np.arange(n) * 0.3) * 5
print(c.batch(base + 1, base - 1, base)[-1])

Node

const ta = require('wickra');

const c = new ta.AdaptiveCci(20);
console.log('warmupPeriod:', c.warmupPeriod()); // 20

Streaming

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

let mut c = AdaptiveCci::new(20).unwrap();
let mut last = None;
for i in 0..60 {
    let base = 100.0 + (f64::from(i) * 0.3).sin() * 5.0;
    let candle = Candle::new(base, base + 1.0, base - 1.0, base, 1_000.0, 0).unwrap();
    last = c.update(candle);
}
println!("{last:?}");

Streaming update and batch are equivalent tick-for-tick (batch_equals_streaming pins this).

Interpretation

1. Earlier band touches. In a genuine trend the adaptive centre accelerates, so the CCI reaches ±100 faster than a fixed-SMA CCI.
2. Fewer false pokes. In a range the centre line slows, damping spurious band crosses.
3. Same playbook. Read the ±100 (and ±200) levels and zero-line crosses as for the standard CCI.

Common pitfalls

• Not bounded. Like all CCIs it can exceed ±100/±200 in strong moves.
• Adapts the centre, not the band. The 0.015·MD scale is unchanged; only the centre line adapts.
• Period role. period is the window for ER, centre and deviation, not a smoothing constant.

References

Lambert, D. (1980), Commodity Channel Index (Commodities magazine); Kaufman, P. J. (1995), Smarter Trading (efficiency ratio); Ehlers, J. F. (2013), Cycle Analytics for Traders (cycle-adaptive oscillators).

See also

• Indicator-Cci — the fixed-SMA original.
• Indicator-AdaptiveRsi — the adaptive RSI counterpart.
• Indicator-Kama — the adaptive average it borrows.
• Indicators-Overview — the full taxonomy.