InsidersTradesSigma
Performance & Risk Metrics
A time-series measure of price dispersion recalculated sequentially over fixed-length overlapping periods to detect regime shifts and risk changes in real time.
Rolling window volatility divides price history into successive, overlapping windows (commonly 20, 60, or 252 trading days) and computes the standard deviation of returns for each one. As new data arrives, the oldest observation drops out and the window shifts forward, giving a continuous stream of volatility estimates. This captures volatility clustering and stress phases that a single static estimate would average away.
The window length is a trade-off. A short window (20 days) reacts quickly to a new regime but is jumpy; a long window (252 days) is stable but slow to register a shift. Returns-based ratios such as Sharpe and Sortino use a rolling volatility estimate in their denominator, so the choice of window directly shapes the risk-adjusted score.
Formula