Planar projection of an isotropic random vector (source code)

= Planar projection of an isotropic random vector
{title2=$\mathbb E R_\perp=(\pi/4)\mathbb E R,\quad\mathbb E R_\perp^2=(2/3)\mathbb E R^2$}

Here isotropic means \b[rotationally invariant] in three <dimensions>, a stronger condition than merely having identity <covariance matrix>. Such a <random vector> has uniform direction on the <sphere>, independent of its nonzero length $R$. Its projection onto a fixed plane has length $R_\perp=R\sin\theta$. The solid-angle averages are
$$
\mathbb E\sin\theta=\tfrac12\int_0^\pi\sin^2\theta\,d\theta=\frac\pi4,\qquad
\mathbb E\sin^2\theta=\tfrac12\int_0^\pi\sin^3\theta\,d\theta=\frac23.
$$
Thus $\mathbb ER_\perp=(\pi/4)\mathbb ER$ and $\mathbb ER_\perp^2=(2/3)\mathbb ER^2$ whenever the corresponding moments exist. A mass at zero contributes zero to both lengths and needs no direction. The factors are independent of the radial law; the mean-length factor differs from the root-mean-square factor.