The spec that measured is the spec that trades

A Strategy may be built before its Hypothesis resolves — it is the apparatus that measures it, and half the donor register asked questions that cannot be posed without one (Z04-ORB-OBTAINABLE, E2-OBJECTIVE). Building is therefore permitted pre-resolution, but deploying is not. To close the gap between the two, **the StrategySpec hash is frozen into the Verdict at resolution, and deployment asserts that the deployed hash equals the resolved hash.**

Consequences

Any edit to a resolved spec — a widened stop, a tuned threshold, one more filter — produces a different hash and therefore requires a new Hypothesis and new alpha. This is deliberate. Post-hoc tuning between the measurement and the deployment is the most common way measured edge disappears in practice, and it is invisible to every test that can be written unless the identity is pinned. The freeze must be enforced by type, not by convention: a deploy path that takes a spec rather than a resolved-hash-plus-spec is a defect.