Glossary
R-squared
The share of the variance of one variable explained by a regression on another. In a single-variable regression it is the squared correlation.
R-squared lies between 0 and 1. A value of 0.49 means the benchmark explains 49% of the variance of the asset's returns, and the other 51% is residual.
It does not measure how good an asset is, and a high value in a regression of price levels on price levels is often spurious. High R-squared with trending data is a warning sign.
Example: a correlation of 0.7 gives R-squared of 0.7^2, or 49%. With asset volatility of 24%, the residual volatility is 24% x sqrt(1 - 0.49), or 17.14%.