Standard Deviation

Standard deviation describes dispersion. A small number means recent closes clustered near their average. A large number means they did not. In technical platforms the input is usually the close, the window matches a moving average, and the result is in price points. It is the engine inside Bollinger Bands. Plotted alone, it is a volatility line with no upper bound and no direction.

Because deviations are squared, one far close dominates the window. That is a feature when you want to notice a shock and a distortion when one bad print is not the regime. The number falls only as that print ages out or as new closes cluster. It will not warn you before the shock. It records it.

Use standard deviation to see whether the current distance from a moving average is ordinary for this market. A close two deviations away is the Bollinger tag. Use a collapse in the deviation as a squeeze context. Use the point value to scale a stop: a stop inside one recent deviation is inside ordinary scatter.

Assuming returns are normal, so that 95 percent of closes "should" sit inside two deviations. Market returns have fat tails. Closes outside the bands happen more often than the textbook normal curve says, and they cluster. Also mixing sample and population settings across platforms and thinking the indicator broke.

A fixed window treats every close inside it as one regime. After a crash the deviation stays high and makes later, calmer moves look small. And standard deviation has no sense of path. Two markets with the same dispersion can have opposite trends.