BetterVolume

Barry Taylor's (emini-watch) Better Volume — classifies each bar by its effort (volume) against its result (range) over a lookback: low volume, climax up/down, high churn, or climax + churn.

Quick reference

Field	Value
Family	Volume
Input type	Candle (open / high / low / close / volume)
Output type	f64 (a category code)
Output range	one of 0, 1, 2, 3, −3, 4
Default parameters	(period = 14) (Python)
Warmup period	period
Interpretation	1 low volume · ±3 climax up / down · 2 high churn · 4 climax + churn · 0 neutral.

Formula

range      = high − low
low volume : volume          is a new low  of the lookback   -> 1
climax     : volume · range  is a new high of the lookback   -> +3 up bar (close ≥ open), −3 down bar
churn      : volume / range  is a new high of the lookback   -> 2
both       : climax and churn on the same bar                -> 4
otherwise                                                    -> 0

The rules are applied in that order and a later match overrides an earlier one, as in Taylor's original. "New high / low" means strictly beyond every one of the previous period − 1 bars (the current bar is compared against the bars before it), so a run of identical bars stays neutral. A zero-range bar has no defined volume per range and never counts as churn.

Volume-Spread Analysis reads price action through effort versus result. A climax is the most effort meeting the widest result — often the start or end of a move; churn is heavy volume going nowhere — professional absorption near turning points; a low-volume bar shows the absence of interest that marks pullbacks and tests. Source: crates/wickra-core/src/indicators/better_volume.rs.

Parameters

Name	Type	Default	Valid range	Source	Description
period	usize	14 (Python)	>= 1	better_volume.rs:76	Lookback length: the current bar is compared against the previous period − 1 bars. 0 errors with Error::PeriodZero.

The period getter returns the lookback; value returns the current code if ready.

Inputs / Outputs

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

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

A Candle in, an Option<f64> code out. The open matters: it decides the sign of a climax bar. The Python binding takes a candle for update and five numpy columns (open, high, low, close, volume) for batch; Node and WASM take update(open, high, low, close, volume) and batch(open[], high[], low[], close[], volume[]) (NaN warmup).

Warmup

warmup_period() == period. The lookback must hold period bars before the first classification (first_emission_at_warmup_period pins this for period = 3).

Edge cases

• Steady bars → 0. Identical bars are never a strict new high or low, so they classify as neutral (steady_bars_are_neutral pins this).
• Low volume → 1. A bar whose volume undercuts every previous bar in the lookback (low_volume_bar pins this).
• Climax → ±3. The largest volume · range of the lookback; +3 when close ≥ open, −3 otherwise (climax_bars_carry_direction pins both).
• Churn → 2. The largest volume / range of the lookback without being a climax (churn_bar pins this).
• Climax + churn → 4. Both on the same bar (climax_and_churn_together pins this).
• Zero-range bars. volume / range is undefined, so they never churn (zero_range_bars_never_churn pins a run of flat, zero-volume bars to 0).
• Finiteness. Candle::new rejects non-finite fields, so no in-method guard is needed.
• Reset. bv.reset() clears the lookback window and the last value (reset_clears_state).

Examples

Rust

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

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut bv = BetterVolume::new(3)?;
    let steady = Candle::new(100.0, 102.0, 100.0, 101.0, 1_000.0, 0)?;
    let candles = vec![
        steady,
        steady,
        // Same range, a bit more volume on a narrower bar: high churn.
        Candle::new(100.0, 101.0, 100.0, 100.5, 1_500.0, 0)?,
    ];
    println!("{:?}", bv.batch(&candles));
    Ok(())
}

Output:

[None, None, Some(2.0)]

Python

import numpy as np
import wickra as ta

S = (100, 102, 100, 101, 1000)                  # steady bar: open, high, low, close, volume
bars = [S, S, (100, 102, 100, 101, 200),        # low volume
        S, S, (100, 110, 100, 109, 4000),       # climax up
        S, S, (109, 110, 100, 100.5, 4000),     # climax down
        S, S, (100, 101, 100, 100.5, 1500),     # high churn
        S, S, (100, 103, 100, 102, 10000)]      # climax + churn
o, h, l, c, v = (np.array(col, dtype=float) for col in zip(*bars))
print(list(ta.BetterVolume(3).batch(o, h, l, c, v)))
# [nan, nan, 1.0, 0.0, 0.0, 3.0, 0.0, 0.0, -3.0, 0.0, 0.0, 2.0, 0.0, 0.0, 4.0]

Node

const ta = require('wickra');

const bv = new ta.BetterVolume(3);
bv.update(100, 102, 100, 101, 1000);
bv.update(100, 102, 100, 101, 1000);
console.log(bv.update(100, 101, 100, 100.5, 1500)); // 2 (high churn)
console.log('warmupPeriod:', bv.warmupPeriod());     // 3

Streaming

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

let mut bv = BetterVolume::new(20).unwrap();
let mut last = None;
for i in 0..60 {
    let base = 100.0 + f64::from(i);
    let c = Candle::new(base, base + 2.0, base - 2.0, base + 0.5, 1_000.0, 0).unwrap();
    last = bv.update(c);
}
println!("{last:?}"); // Some(0.0): constant volume and range never set a new extreme

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

Interpretation

1. Climax (±3). Maximum effort meeting maximum result. At the end of an extended move it often marks exhaustion (a buying climax up, a selling climax down); at the start of a range break it marks strong participation.
2. High churn (2). Heavy volume relative to the range — effort without result. Professionals are absorbing supply or demand; look for a turn, especially near support / resistance.
3. Climax + churn (4). Both signatures on one bar — the strongest turning-point warning.
4. Low volume (1). No interest: in a pullback it suggests the counter move is weak ("no supply" / "no demand") and the trend may resume.

Common pitfalls

• Reading the codes as a scale. The output is a categorical code, not an oscillator — 4 is not "twice" 2, and averaging or smoothing the codes is meaningless. Branch on the value instead.
• Override order. A bar that is both low-volume and churn (narrow range on little volume) reports 2, not 1, because later rules override earlier ones.
• Needs real volume. On feeds without genuine volume the classification is noise.
• Period choice. A short period makes many bars "new" extremes; a long one flags only the rare outliers.

References

Barry Taylor, Better Volume indicator (emini-watch.com); Williams, T., & Brooks, G. (2005), Master the Markets (Volume Spread Analysis); Wyckoff, R. D. — the original effort-versus-result principle.

See also

• Indicator-EaseOfMovement — Arms' range/volume mover.
• Indicator-ForceIndex — price change times volume.
• Indicator-MarketFacilitationIndex — range per unit volume.
• Indicators-Overview — the full taxonomy.