VolumeWeightedMacd

MACD built on volume-weighted moving averages — heavy-volume bars dominate the trend estimate, so crossovers reflect where real participation occurred.

Quick reference

Field	Value
Family	Volume
Input type	Candle (close / volume)
Output type	VolumeWeightedMacdOutput { macd, signal, histogram }
Output range	(−∞, +∞) for each field
Default parameters	(fast = 12, slow = 26, signal = 9)
Warmup period	slow + signal − 1
Interpretation	macd > signal (histogram > 0) = bullish; crossover = trade trigger.

Formula

macd      = VWMA(close, fast) − VWMA(close, slow)
signal    = EMA(macd, signal_period)
histogram = macd − signal

The classic MacdIndicator smooths price with EMAs that ignore volume. The volume-weighted MACD replaces both with a Vwma, so each average is Σ(close·volume) / Σ(volume) over its window — heavy bars pull the line harder. The signal line stays a plain EMA of the MACD, preserving the standard histogram. Source: crates/wickra-core/src/indicators/volume_weighted_macd.rs.

Parameters

Name	Type	Default	Valid range	Source	Description
fast	usize	12	>= 1, < slow	volume_weighted_macd.rs:77	Fast VWMA window.
slow	usize	26	> fast	volume_weighted_macd.rs:77	Slow VWMA window.
signal	usize	9	>= 1	volume_weighted_macd.rs:77	Signal EMA window over the MACD line.

fast >= slow errors with Error::InvalidPeriod; any zero period errors with Error::PeriodZero. periods() returns (fast, slow, signal); value returns the current output if ready.

Inputs / Outputs

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

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

A Candle in, an Option<VolumeWeightedMacdOutput> out. The Python binding returns a (macd, signal, histogram) tuple from update and an (n, 3) array from batch(close, volume); Node returns an object { macd, signal, histogram } and a flat Float64Array of length n*3; WASM mirrors the object with camelCase keys.

Warmup

warmup_period() == slow + signal − 1. The slow VWMA seeds after slow candles; the signal EMA needs signal − 1 further MACD values (first_emission_at_warmup_period pins this).

Edge cases

• Uptrend → positive MACD. A steady advance keeps the fast VWMA above the slow one (uptrend_has_positive_macd pins this).
• Histogram identity. histogram == macd − signal holds every bar (histogram_is_macd_minus_signal pins this).
• Constant volume. With flat volume each VWMA reduces to an SMA; the series stays finite and well-defined (equal_volume_matches_plain_macd pins this).
• Finiteness. Candle::new rejects non-finite fields, so no in-method guard is needed.
• Reset. vwmacd.reset() clears both VWMAs, the signal EMA and the last value (reset_clears_state).

Examples

Rust

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

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut m = VolumeWeightedMacd::new(3, 6, 3)?;
    let candles: Vec<Candle> = (0..60)
        .map(|i| Candle::new(100.0 + f64::from(i), 100.0 + f64::from(i), 100.0 + f64::from(i),
                             100.0 + f64::from(i), 1_000.0, 0).unwrap())
        .collect();
    let out = m.batch(&candles);
    println!("warmup_period = {}", m.warmup_period());
    println!("macd > 0: {}", out.last().unwrap().unwrap().macd > 0.0);
    Ok(())
}

Output:

warmup_period = 8
macd > 0: true

Python

import numpy as np
import wickra as ta

m = ta.VolumeWeightedMacd(12, 26, 9)
close  = np.cumsum(np.ones(80)) + 100.0
volume = np.full(80, 1000.0)
macd, signal, hist = m.batch(close, volume).T
print(macd[-1], signal[-1], hist[-1])

Node

const ta = require('wickra');

const m = new ta.VolumeWeightedMacd(12, 26, 9);
console.log('warmupPeriod:', m.warmupPeriod()); // 34

Streaming

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

let mut m = VolumeWeightedMacd::new(12, 26, 9).unwrap();
let mut last = None;
for i in 0..80 {
    let base = 100.0 + f64::from(i);
    let c = Candle::new(base, base + 1.0, base - 1.0, base + 0.5, 1_000.0, 0).unwrap();
    last = m.update(c);
}
println!("{last:?}");

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

Interpretation

1. Signal crossovers. macd crossing above signal (histogram turning positive) is the bullish trigger; the reverse is bearish — read exactly like classic MACD, but with volume conviction baked in.
2. Zero-line. macd > 0 means the volume-weighted fast trend leads the slow trend; sustained positive readings confirm an uptrend.
3. Volume edge. Compared with plain MACD, the volume-weighted version turns earlier when a move is driven by genuine participation and lags when a move is thin — useful as a confirmation filter.

Common pitfalls

• Needs real volume. On feeds with synthetic or missing volume the VWMAs degenerate toward SMAs; prefer plain MACD there.
• Not identical to MACD. Even with similar shape, the lines differ from EMA-MACD; do not expect overlapping crossovers.
• Spiky volume. A single enormous-volume bar can jerk the fast VWMA; smooth or cap volume if your feed has outliers.

References

Dormeier, B. P. (2011), Investing with Volume Analysis. The volume-weighted MACD is a common construction substituting VWMA for EMA in Appel's MACD (Appel, G., 1979).

See also

• Indicator-MacdIndicator — the EMA-based original.
• Indicator-Vwma — the volume-weighted moving average building block.
• Indicator-MacdHistogram — the standalone histogram.
• Indicators-Overview — the full taxonomy.