docs(ms-ai-architect): cond2 ratifisert som hele-fila-lesning; verifisert skår 1 av 46 [skip-docs]

docs/r11-pilot-results.md

@@ -382,11 +382,14 @@ remainder must not be misleading. **Its scope was never fixed:**
   file is a **separate ungrounded claim**, to be flagged on its own, not a
   defeater of this subtraction.
 
-The analysis below takes the **broad** reading. It is not ratified, and it is
-the operator's to settle — the same operator who ratified conditions 2 and 3 as
-human-judged. Under the narrow reading the classifier did exactly what cond 2
-asks and idx 8's cond 2 holds; the finding then becomes "lines 263-265 are a
-second ungrounded claim" rather than "the classifier was wrong".
+**The operator ratified the broad reading, 2026-08-03.** Cond 2 is measured
+against the whole file: a remainder that contradicts the file it sits in is
+misleading, whatever the source says. The stated ground is that these files are
+publicly distributed and read as wholes — a self-contradicting file is a trust
+defect regardless of which half is wrong. The narrow reading (cond 2 scoped to
+the edited passage, with contradictions elsewhere handled as separate ungrounded
+claims) was considered and rejected. Recorded here because the fork was real and
+a later run must not silently re-open it.
 
 The classifier justified cond 2 on one of the two sub-deletions and never
 checked the other:
@@ -396,20 +399,16 @@ checked the other:
   (`**Azure Monitor + Application Insights** (Verified)` / "Drift metrics
   emitteres til Application Insights"), so removing the prerequisite bullet
   leaves no false implication.
-- Sub-deletion (a), `eller managed compute cluster`: **fails under the broad
-  reading, holds under the narrow one.** The same file documents
+- Sub-deletion (a), `eller managed compute cluster`: **fails.** The same file
+  documents
   `**Managed Compute Cluster** (for store volumer)` as a real compute option at
   lines 263-265, with pricing and a usage recommendation. Deleting the
   alternative from the prerequisites leaves the file asserting serverless Spark
   as the only compute requirement seventy lines above a cost section that prices
-  the alternative. Broad: that is a misleading remainder. Narrow: the remainder
-  states what the source grounds, and 263-265 is a separate claim needing its
-  own verdict.
+  the alternative. That is a misleading remainder under the ratified reading.
 
-Either way the candidate does not go in as written — under the broad reading
-because (a) must be dropped from the subtraction, under the narrow reading
-because lines 263-265 must be adjudicated first and may themselves need an edit.
-The remedies are different and the operator picks which applies.
+The candidate therefore does not go in as written: **(a) must be dropped from
+the subtraction**, leaving (b) alone.
 
 There is a further limit the O2 envelope cannot resolve: cond 3 passes for (a)
 only because the source does not *mention* managed compute — not-mentioned
@@ -424,20 +423,19 @@ string** than the one V2b attested, so it is not machine-clean until
 amendment, including dropping idx 17's trailing colon: **amended remainder →
 re-run the check before it counts as verified.**
 
-**The generalisable finding — conditional on the broad reading.** The classifier
-judged cond 2 against the *source* and against citations it chose itself. It did
-not systematically judge it against **the rest of the same file**. Under the
-broad reading that is a defect: internal consistency is a cond-2 dimension the
-wave prompts never assigned, and idx 8 — the single `high` confidence record in
-the set — is the proof that it bites. Under the narrow reading it is not a
-defect at all; the same sweep is still worth running, but its output is a list
-of *additional* ungrounded claims rather than a list of corrections to the
-existing verdicts.
+**The generalisable finding.** The classifier judged cond 2 against the *source*
+and against citations it chose itself. It did not systematically judge it against
+**the rest of the same file**. Under the ratified reading that is a defect:
+internal consistency is a cond-2 dimension the wave prompts never assigned, and
+idx 8 — the single `high` confidence record in the set — is the proof that it
+bites. The remaining 15 have not had the check. It must be run per candidate
+before any of them reaches a ratifier, and its output corrects the existing
+cond-2 verdicts rather than merely adding to them; expect it to move some.
 
-Either way the remaining 15 have not had the check, and it should be run per
-candidate before any of them reaches a ratifier. What changes with the reading
-is what the results *mean*, not whether to gather them — which is why the sweep
-can proceed before the scope question is settled, and the verdicts cannot.
+**Score after hand-verification: 1 of 46 clears both conditions, not 2.** Only
+idx 17 survives. That is the number §10 measurement #2 should be read with — the
+classifier's own "2 of 46" counted idx 8 on a cond-2 justification that was
+half-unchecked.
 
 ## Appendix A — the 15 admitted proposals, hand-verified
 
@@ -537,10 +535,11 @@ deletes (§9.1, V2b), so it is a text change and must be reviewed as one.
 Hand-checking the two affirmative rows (§9.2) showed that half of idx 8's cond-2
 justification was never checked, and that the same file documents the deleted
 alternative in its cost section. Read every `yes` in this column as "the
-classifier believed this". Only **idx 17** has been hand-verified so far, and
-idx 8 is not ratifiable as written under either reading of cond 2 — but **which
-reading applies is an open operator decision** (§9.2), and it governs both the
-count and what a sweep of the remaining 15 would mean.
+classifier believed this". Only **idx 17** has been hand-verified. Idx 8 is not
+ratifiable as written: cond 2 is measured against the whole file (operator
+ratification, 2026-08-03, §9.2), and the file contradicts the remainder at lines
+263-265. **The verified score is 1 of 46, not 2.** The other 15 rows have not had
+the whole-file check.
 
 Reproduce the tally and the machine checks:
 `node scripts/kb-eval/check-o2-returns.mjs`.