---
name: declining-a-credit-specifies-the-requirement
description: Stating precisely WHY your evidence does not discharge a requirement often specifies that requirement better than its own statement did — the refusal carries information the claim would have destroyed.
metadata:
  type: feedback
---

When you could take credit for partially covering something and instead say exactly which conditions
your evidence did NOT present, the enumeration of what is missing IS a specification. Taking the
credit would have destroyed that information, because a satisfied requirement stops being examined.

**Why:** measured 2026-08-21. doyle's never-executed-cells list for #212 carried a row reading "one
real composite roll on a real box landing new-core AND new-adapter together — no lane gates it, no rig
simulates it." I had such a roll (KITSUBITO, 2026-08-04) and could have offered it. Instead I gave the
three reasons it does not discharge the leg, the decisive one being that neither of my runs ever
presented the REFUSED outcome — the adapters updated cleanly or were already current, so nothing was
near a floor. doyle's reply: reason (c) "does not just decline the credit — it TELLS ME WHAT THE LEG
ACTUALLY IS", which his own row had under-specified as "one real composite roll". It is not one
condition, it is TWO OPPOSITES:

  (i) **#159** — an adapter whose floor the RUNNING core does not satisfy but the INCOMING core does:
      pre-fix refused in-cycle; post-fix it must be ADMITTED and its record must register at the new
      version.
  (ii) **#213** — an adapter whose floor NEITHER core satisfies: it must still refuse, and the refusal
      must render REFUSED rather than FAILED and exit 3 rather than 1.

He rewrote the row to name both conditions and the adapter-manifest floor each needs, so field
verification now has a recipe instead of an aspiration. That improvement came out of declining a
credit that was available.

**How to apply:**
- When your evidence partially covers a requirement, report the UNCOVERED conditions by name, not the
  coverage percentage. "This exercises the happy path and never presents a refusal" is actionable;
  "partial coverage" is not.
- Treat an easy credit as a question: what would this have had to show to actually count? Answering
  that is usually the missing spec.
- A requirement stated as one condition is worth suspecting when your evidence satisfies it "except".
  The "except" is often a second, opposite condition hiding inside the first.
- Prior art is still worth citing, scoped: my roll IS cited, as the happy path at a v0.53.0-era
  baseline with post-state captured independently of an exit code known to be invalid.

Related: [[identical-readings-across-opposite-outcomes-indict-the-meter]] (same session; both are
about what evidence does NOT establish), [[dont-take-a-diagnosis-as-measured]].