Alpha

Alpha measures what a symbol's returns deliver beyond what its market exposure predicts, bar by bar. The build regresses the chart symbol's excess returns on a benchmark's over a rolling window and plots the intercept — alpha — as a percentage, sign-colored around a zero line with an optional gradient fill. A dashboard pins the context that keeps the number honest: benchmark, window, current alpha and beta, and the R-squared fit quality.

How to Trade the Alpha?

• Above zero: returns beat what beta-scaled benchmark exposure predicts over the window; below zero, they trail it.
• Zero crosses: outperformance starting or fading — both directions carry alerts on the rolling intercept.
• Weigh by R-squared first: when the benchmark explains little return variance, slope and intercept are loosely pinned down and the print merits little weight.

Alpha is always relative — a stock can post a large reading against a broad index that vanishes against its own sector — so quote readings with their benchmark and window attached, and trust a value that survives window changes over a single print.

Alpha Settings

• Benchmark (default SP:SPX): reference market the symbol is regressed against.
• Window (default 60): returns in each rolling regression; shorter reads faster and noisier.
• Risk-Free Rate (%) (default 0): annual rate subtracted from both return series per the Jensen formulation; 0 reports the raw intercept.
• Alpha Display (default Annualized (Scale)): scaling multiplies the intercept by periods per year; Annualized (Compound) compounds it; Per-Period reports it raw.
• Periods Per Year (default 0): annualization factor, auto-detected from the timeframe at 0.
• Show Dashboard (on) with Location (Top Right) and Size (Small); Gradient Fill (on).

Frequently Asked Questions

How do Alpha and Beta relate?

They are the two outputs of one regression: Beta is the slope of that fit — sensitivity to benchmark moves — while alpha is the intercept left once that exposure is removed. A low-beta symbol with positive alpha earns largely independently of the index, a different profile from a high-beta name with the same headline gain.

Why does my reading differ from a published alpha figure?

Usually convention: sources differ on scaling versus compounding the intercept, on benchmark, window, and whether a risk-free rate is subtracted. Match the Alpha Display and inputs before comparing numbers.

Which window length makes sense?

Common practice runs 60 to 250 bars on daily charts and 36 to 60 on monthly ones. Short windows are noisy and long ones lag regime changes — the useful check is whether a reading persists as the window varies.