Simulating One Correlated Normal Given the Other

EasyProbability

Reported in interviews at BlackRock

Problem statement

$X_1$ and $X_2$ are jointly normal with means $\mu_1, \mu_2$, standard deviations $\sigma_1, \sigma_2$, and correlation $\rho$. You have already drawn (or observed) a value $x_1$ of $X_1$. How do you simulate a draw of $X_2$ consistent with that $x_1$?

Give the conditional distribution you would sample from, and the concrete simulation recipe.

Show hint

Can you write the second variable as a linear function of the first plus a piece that is independent of it?

Relevant roles: Research Scientist, Applied Scientist, ML Engineer and Data Scientist.

Related free problems