The correlation coefficient has been described in textbooks (see reference below) as a standardized version of the covariance of two variables, and as such, we believe to is a better way to compare two data streams than the covariance, especially if the data streams have different numeric ranges. The correlation coefficient is found in the Regression category of indicators, and is called:
Linear XY Regression: Coefficient of Correlation (r)
However, for those who feel they need the covariance, we have programmed (for NeuroShell release 6) an indicator called Covar, which takes two data streams and produces the covariance between them in a lookback window. (The lookback window must be at least 2). To download the Covar indicator click here. Store the downloaded Covar.dll and Covar.tpl files in the NeuroShell Template folder. The file Covar.bas is the PowerBasic source code for the Covar indicator.
From the reference below, “the variance indicates the variation of a set of scores from their means, whereas the covariance indicates the covariation of two sets of scores from their means.”
Reference: “Multiple Regression in Behavioral Research”, Second Edition, by Elazar J. Pedhazur