R-squared

The share of the variation in a dependent variable that a regression explains, between 0 and 1. In a regression with one explanatory variable it equals the square of the correlation between the two.

A stock whose returns have a 0.6 correlation with the index has an R-squared of 0.36 against it: 36% of the variation in its returns moves with the market and 64% is its own. For a large diversified fund the figure is often above 0.9, which means its returns are mostly the market's and its fees should be judged against an index fund.

A high R-squared doesn't make a model right, and a low one doesn't make a slope useless. Adding variables always raises it, so the adjusted version penalises extra variables. Trending series such as revenue and GDP can show a high R-squared with no real relationship at all.

Explained variation, not proof.