Transparency is a feature. This page explains what the market model measures and why we compute it — in plain terms. The precise formulas and numeric conventions are documented in our internal methodology; the summary here is deliberate, not an omission.
The market model relates an asset's returns to a single benchmark (market) series over the same set of periods, fitting the best straight line through the paired returns by ordinary least squares. One pass produces three companion statistics — beta, alpha, and R² — that together describe how the asset moves with, and apart from, the market. The two return series are aligned strictly period for period and use simple returns expressed as decimal fractions; a length mismatch is rejected rather than guessed at.
Beta is the sensitivity of the asset to the market: the slope of the fitted line. A beta of one means the asset tends to move roughly in step with the benchmark, above one means it tends to swing more than the market, below one means it moves less, and a negative beta means it tends to move the opposite way. Because beta is a ratio of co-movement to market movement, the choice of how the underlying dispersion is normalized cancels out, so it does not depend on that convention.
Alpha is the part of the asset's average return that the market does not explain — the intercept of the fitted line, the return you would expect when the market return is flat. This is a per-period figure (sometimes called Jensen's alpha), not an annualized one; any annualization happens later, downstream, with a declared number of periods per year. Positive alpha means the asset did better than its market sensitivity alone would predict.
R² measures how much of the asset's variation the market explains — the goodness of fit of the line. It is the square of the Pearson correlation between the two series and always sits between zero and one: near one means the benchmark accounts for almost all of the asset's movement, near zero means the market explains little of it. As a sanity check, regressing a series on itself gives a beta of one, an alpha of zero, and an R² of one.
