N-Sigma Price-Change Outlier Detection with Rolling Volatility

Summary

The indicator measures one-bar changes in the selected price series, then divides each change by an estimated standard deviation based on a rolling sample. It plots the absolute standardized change as a bar, colored by whether the price move was positive or negative. Threshold lines default to two and three standard deviations, which the author associates approximately with 95% and 99% extremeness under a normal model.

The explanation stresses that these levels rely on assumptions: changes should be independent, centered near zero, and have stable variance. Financial returns may violate these conditions as volatility shifts, and standardized changes are not automatically reliable probability statements. The author discusses shorter samples for tracking recent volatility and a longer sample for a long-run estimate, but offers no empirical evaluation of predictive value. The tool flags unusually large changes relative to its estimate; it does not specify a trading rule or establish that outliers reverse or continue.

Key ideas

• The method first differences price and standardizes each change by an estimated rolling standard deviation.
• The display uses absolute standardized values while color indicates the direction of the move.
• Two- and three-sigma thresholds are presented as approximate extremeness levels under a normality assumption.
• Independence and constant variance may fail in markets with changing volatility.
• The indicator identifies unusual moves but provides no entry rule or performance evidence.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.