= Solution
Put $K=V^{-1}$, the <precision matrix>, with $X_1$ scalar and $X_2$ containing all remaining coordinates. Holding $X_2=x_2$ in the <multivariate normal density>, the terms involving $x_1$ in its exponent are
$$
-\frac12\left[K_{11}(x_1-\mu_1)^2+2(x_1-\mu_1)K_{12}(x_2-\mu_2)\right].
$$
Completing the square gives a <normal distribution> with mean $\mu_1-K_{11}^{-1}K_{12}(x_2-\mu_2)$ and
$$
\boxed{\operatorname{Var}(X_1\mid X_2=x_2)=\frac1{(V^{-1})_{11}}.}
$$
Equivalently, the block <matrix inverse> identity gives $K_{11}=(V_{11}-V_{12}V_{22}^{-1}V_{21})^{-1}$. The <Gaussian conditional variance from precision> is a reciprocal diagonal entry of the <precision matrix>, rather than the marginal <variance> $V_{11}$.
Back to article page