ElderRay

Elder Ray — Alexander Elder's Bull Power / Bear Power: how far each bar's high and low stretch from an EMA of the close.

Quick reference

Field	Value
Family	Momentum Oscillators
Input type	Candle (uses high, low, close)
Output type	ElderRayOutput { bull_power, bear_power }
Output range	unbounded; bull normally > 0, bear normally < 0
Default parameters	period (Elder uses 13)
Warmup period (warmup_period())	period
Interpretation	Strength of buyers (bull) and sellers (bear) vs. the trend mean.

Formula

ema       = EMA(close, period)
BullPower = high - ema
BearPower = low  - ema

The EMA represents the market's consensus value; Bull Power measures how far buyers lifted the high above it, Bear Power how far sellers drove the low below it. In a healthy uptrend (rising EMA) Bull Power is positive and Bear Power negative-but-recovering; Elder's textbook long setup is a rising EMA with Bear Power negative and ticking up (sellers losing ground).

Source: crates/wickra-core/src/indicators/elder_ray.rs.

Parameters

Name	Type	Default	Valid range	Source	Description
period	usize	13	>= 1	elder_ray.rs:60	EMA period for the consensus mean. 0 errors with Error::PeriodZero.

(Python class wickra.ElderRay(period); Node new ta.ElderRay(period).)

Inputs / Outputs

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

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

Two outputs per bar. Node: update(open, high, low, close) returns { bullPower, bearPower } | null; batch(o[], h[], l[], c[]) returns a flat Array<number> of length n*2 ([bull, bear, bull, bear, …], NaN for warmup). Python: update(candle) returns (bull, bear) | None; batch(open, high, low, close) returns an (n, 2) ndarray.

Warmup

warmup_period() returns period: the first reading lands once the inner EMA is seeded (bar period, index period − 1). Pinned by accessors_and_metadata (warmup_period() == 13) and warmup_then_known_value.

Edge cases

• Known value. warmup_then_known_value pins an ElderRay::new(3): EMA of closes [10, 12, 14] seeds at 12; bar high 16, low 13 → bull 4, bear 1.
• Equivalence to a manual EMA. matches_manual_ema runs a standalone EMA on the closes and reproduces both powers bar-for-bar.
• Reset. reset_clears_state clears the inner EMA.
• Streaming/batch: batch_equals_streaming.

Examples

Rust

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

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut er = ElderRay::new(3)?;
    let bars = [
        Candle::new(10.0, 11.0, 9.0, 10.0, 1.0, 0)?,
        Candle::new(12.0, 13.0, 11.0, 12.0, 1.0, 1)?,
        Candle::new(14.0, 16.0, 13.0, 14.0, 1.0, 2)?,
    ];
    for v in er.batch(&bars).into_iter().flatten() {
        println!("bull={} bear={}", v.bull_power, v.bear_power);
    }
    Ok(())
}

Output:

bull=4 bear=1

The EMA(3) seeds at the third bar with the mean close 12; that bar's high 16 gives Bull Power 4 and its low 13 gives Bear Power 1.

Python

import numpy as np
import wickra as ta

er = ta.ElderRay(3)
op = np.array([10.0, 12.0, 14.0])
hi = np.array([11.0, 13.0, 16.0])
lo = np.array([9.0, 11.0, 13.0])
cl = np.array([10.0, 12.0, 14.0])
print(er.batch(op, hi, lo, cl))   # (3, 2): [bull, bear] rows, NaN for warmup

Output:

[[nan nan]
 [nan nan]
 [ 4.  1.]]

Node

const ta = require('wickra');
const er = new ta.ElderRay(3);
const open = [10, 12, 14];
const high = [11, 13, 16];
const low = [9, 11, 13];
const close = [10, 12, 14];
console.log(er.batch(open, high, low, close)); // flat [bull, bear, …], NaN warmup
const last = er.update(14, 16, 13, 14);
console.log(last); // { bullPower: 4, bearPower: 1 }

Output:

[ NaN, NaN, NaN, NaN, 4, 1 ]
{ bullPower: 4, bearPower: 1 }

Streaming

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

let mut er = ElderRay::new(13)?;
let mut last = None;
for i in 0..40 {
    let base = 100.0 + f64::from(i);
    last = er.update(Candle::new(base, base + 2.0, base - 2.0, base + 0.5, 1.0, i64::from(i))?);
}
println!("{last:?}");
# Ok::<(), Box<dyn std::error::Error>>(())

Interpretation

Elder Ray decomposes momentum into the two camps explicitly: Bull Power is the buyers' reach above value, Bear Power the sellers' reach below it. Read against the EMA's slope:

1. Trend filter + oscillator. Use the EMA slope for trend direction and the two powers for timing within it.
2. Long setup. Rising EMA, Bear Power negative but rising (a higher low in Bear Power) — sellers exhausting in an uptrend.
3. Short setup. Falling EMA, Bull Power positive but falling — buyers exhausting in a downtrend.
4. Divergence. A new price high with a lower Bull Power high warns of fading buying strength.

Common pitfalls

• Ignoring the EMA slope. The powers are only meaningful alongside the trend direction; Bull/Bear Power signals are filtered by the EMA's slope, not taken in isolation.
• Comparing magnitudes across instruments. The powers are in price units; normalise (or compare to ATR) before cross-instrument comparison.

References

Alexander Elder, Trading for a Living, 1993 — the Elder Ray (Bull Power / Bear Power) system.

See also

• Indicator-Ema — the consensus mean the powers are measured against.
• Indicator-ElderImpulse — Elder's impulse system (EMA slope + MACD-hist).
• Indicators-Overview — the full taxonomy.