Yes, I’ve run a few Monte Carlo models for portfolio stress testing. I always start with a clean set of assumptions, discretize the returns carefully, and then run enough scenarios—tens of thousands—so the tail behavior is stable. Once I have the output I’ll run a sensitivity check on volatility and correlations, just to be sure I haven’t overlooked a hidden risk. What’s tripping you up?

Try a quick diagonal‑dominance test first: if your correlation matrix has off‑diagonals that are too high compared to the diagonal variances, the matrix can become ill‑conditioned. Compute the eigenvalues; if any are negative or close to zero, that’s a red flag. Also, before adding the new asset, run the model with just the old set to confirm the baseline matrix is fine. If the baseline is stable, then the new correlations are likely the culprit. You could also shrink the new correlation estimates toward the sample mean or use a Ledoit–Wolf estimator to regularize them. That usually keeps the matrix positive definite without blowing up the discretization.

Sounds good—just keep an eye on the eigenvalue spread after the shrinkage; a well‑conditioned matrix should have all positive eigenvalues with a reasonable gap. Good luck with the run.