Expectancy

The expected return per trade in units of average loss (the "R-multiple" expectancy) over a rolling window of returns.

Quick reference

Field	Value
Family	Risk / Performance
Input type	f64 (per-trade or per-bar return)
Output type	f64
Output range	unbounded; signed (> 0 profitable edge)
Default parameters	period is required
Warmup period	period
Interpretation	E = 0.3 nets 0.3R per trade, where R is the average loss.

Formula

mean    = average of the `period` returns
avgLoss = average of the absolute losing returns (rᵢ < 0)
E       = mean / avgLoss          (0 when there are no losing returns)

Dividing the mean return by the average loss makes the figure comparable across systems with different bet sizes — unlike the raw mean return (which is just an SMA of the series). A positive E is a profitable edge, a negative E a losing one. When the window contains no losing returns there is no risk reference, so the indicator returns 0 (undefined R-multiple).

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

Parameters

Name	Type	Default	Valid range	Source	Description
period	usize	none	>= 1	expectancy.rs:54	Window of returns. 0 errors with Error::PeriodZero.

Inputs / Outputs

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

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

Python streams as float | None, batches as a 1-D numpy.ndarray. Node streams as number | null, batches as Array<number>.

Warmup

warmup_period() == period. The unit test accessors_and_metadata pins warmup_period() == 20 for period = 20.

Edge cases

• Positive edge. +2, −1, +2, −1: mean 0.5, avgLoss 1 → 0.5; pinned by positive_edge.
• Negative edge. +1, −2, +1, −2: mean −0.5, avgLoss 2 → −0.25; pinned by negative_edge.
• No losses. An all-winning window returns 0 (no risk reference); pinned by no_losses_returns_zero.
• Flat returns. Zeros are not losses; pinned by flat_returns_are_not_losses.
• Rolling eviction. Old losses slide out, flipping the window to "no losses"; pinned by rolling_window_evicts_old_losses.

Examples

Rust

use wickra::{BatchExt, Indicator, Expectancy};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut e = Expectancy::new(4)?;
    println!("{:?}", e.batch(&[2.0, -1.0, 2.0, -1.0]));
    Ok(())
}

Output:

[None, None, None, Some(0.5)]

Python

import wickra as ta

e = ta.Expectancy(4)
for r in [2.0, -1.0, 2.0, -1.0]:
    print(r, '->', e.update(r))

Output:

2.0 -> None
-1.0 -> None
2.0 -> None
-1.0 -> 0.5

Node

const ta = require('wickra');
const e = new ta.Expectancy(4);
for (const r of [2, -1, 2, -1]) console.log(r, '->', e.update(r));

Output:

2 -> null
-1 -> null
2 -> null
-1 -> 0.5

Interpretation

Expectancy answers "how much do I make per trade for every unit I typically risk?". Because it normalises by the average loss, an expectancy of 0.3 means the same thing whether you bet $10 or $10,000 — a positive, comparable edge. Read it alongside WinRate (hit rate) and the payoff side: a modest win rate with a large expectancy is a strong system.

Common pitfalls

• 0 means "no losses in window", not "no edge". An all-winning window has no risk reference and returns 0; that is a degenerate case, not a flat edge.
• Not the same as mean return. Plain mean(return) is an SMA of the series; Expectancy divides it by the average loss to express the edge in risk (R) units.

References

Van Tharp's expectancy (E ≈ mean(R), with R = return / initial risk), Trade Your Way to Financial Freedom (1998).

See also

• Indicator-WinRate — the hit-rate companion.
• Indicator-KellyCriterion — edge-based bet sizing.
• Indicator-ProfitFactor — gross profit / gross loss.
• Indicators-Overview — the full taxonomy.