I use log returns because I'm looking for a symmetric distribution... price alone is unstable and meaningless for correlation (which depends on standard deviation). The calculation for log returns is simple: r = log10(close1 / close2) You can then apply this to individual securities or an entire system, and compare statistical measures (e.g., mean, std dev, skewness, kurtosis, variance). Excess kurtosis means a fat tail distribution, and greater risk. Positive skew means a tail that extends to positive values, meaning a greater chance of large returns. and so forth... To calculate a correlation between systems/securities, you can put the daily log returns in Excel and use PEARSON(logReturnsSystemA, logReturnsSystemB).