Rolling Alpha Pass Node
Rolling Alpha (Jensen's) — Series Input
Overview
The Rolling Alpha Pass Node computes Jensen's Alpha over a rolling window — the excess return of a portfolio above what would be predicted by its beta exposure to a benchmark.
Positive alpha means the strategy generates return beyond what systematic market exposure would explain. It is the key measure of true skill in a market-neutral or long/short context.
Formula
Parameters
| Parameter | Default | Description |
|---|---|---|
| period | 20 | Rolling window in bars |
| riskFreeRate | 0 | Per-bar risk-free rate (e.g. 0.00004 ≈ 1% annual at daily bars) |
Inputs & Outputs
| Slot | Direction | Type | Description |
|---|---|---|---|
| input | Input | { values, timestamps } | Portfolio returns series |
| values | Output | (number | null)[] | Rolling Jensen's Alpha per bar; nulls during warm-up |
| timestamps | Output | number[] | Unix timestamps aligned to input |
Use Cases
Strategy Skill Detection
Positive rolling alpha means the strategy is generating returns independent of market beta — evidence of genuine edge rather than passive exposure.
Regime Filtering
Enable a strategy only when its rolling alpha is positive — switch off during periods when the strategy fails to add value above benchmark exposure.
Factor Attribution
Decompose returns into alpha (skill) and beta (market exposure) components for performance attribution analysis.
Tips & Best Practices
Use Matching Benchmark
Alpha is only meaningful when compared against the correct benchmark. A tech portfolio measured against a bond index will show misleading alpha.
Short Windows Are Noisy
Rolling alpha with period < 60 bars can be very noisy. Use longer periods for stable signal and shorter for responsive but volatile estimates.
Combine with Treynor Ratio
Alpha tells you excess return; Treynor tells you return per unit of beta. Together they give a full picture of market-relative performance.