# Macro Study Deck — Pass 1 Audit Report (2026-08-16)

**Audit basis:** 8-item Pass 1 checklist + Item 1b (back-side list, added 2026-08-16) per `references/deck-audit-pass1-checklist.md`.
**Refactor authority:** blanket approval granted by user for this audit pass (exceptional basis).
**Audit execution:** delegated subagent timed out at 600s after completing partial audit (Items 1, 1b, 2, 4, 5, 7). Items 3 (sign/direction), 6 (vault-source), 8 (cross-block consistency) were not covered in this pass — deferred to follow-up audit.

## Summary

- **Cards surveyed:** 264 (deck has grown from 171 baseline; discrepancy noted)
- **Cards excluded:** 27 Block 028 cards (most recent push, tagged `refactor:2026-08-16`) — not re-audited per instruction
- **Notes flagged:** 137 (52% of deck)
- **HIGH flags identified:** 8 (all 8 refactored — see below)
- **MED flags noted:** 176 (most are Item 4 magnitude/date claims = Pass 2 trigger, not refactored)
- **MED Item 7 (stale):** 6 (flagged for awareness, not refactored)
- **Cluster integrity:** not assessed (Item 8 not covered in this pass)

## HIGH flags — all refactored

| # | Note ID | Item | Issue | Refactor |
|---|---|---|---|---|
| 1 | 1782387710031 | 1b | Back-side list (3 Bernanke explanations) | Split into 3 atomic cards + original repurposed as "which is dominant?" card |
| 2 | 1782975741861 | 1b | Back-side list (3 catalysts for TIPS) | Split into 3 atomic cards + original repurposed as "Fed pivot" card |
| 3 | 1783441333671 | 5 | SOFR-OIS basis naming imprecision | Renamed front + back to use EFFR-SOFR basis |
| 4 | 1783787949517 | 1b | Back-side list (3 cross-currency regimes) | Split into 3 atomic cards + original repurposed as "crisis threshold" card |
| 5 | 1783787949641 | 1 | "NOT X, NOT Y" distractors on front | Stripped distractors |
| 6 | 1784975950892 | 5 | SOFR-OIS basis naming (different card, same class) | Renamed + clarifying edit on back |
| 7 | 1785099069095 | 1b | Back-side list (3 CIP signatures) | Split into 3 atomic cards + original repurposed as "acceleration" card |
| 8 | 1786784978325 | 1b | Back-side list (3 FX smile customer types) | Split into 3 atomic cards + original repurposed as "real-money" card |

### Decision rule applied (per user 2026-08-16)

- **≤3 items, individually testable:** split into atomic cards (5 of 8 HIGH flags fell here)
- **>3 items:** would always refactor (none in this HIGH-flag set had >3 items)
- **Naming imprecision:** in-place rename, no split (2 of 8 HIGH flags)
- **Front-side leak:** strip distractors, no split (1 of 8 HIGH flags)

## MED flags noted (not refactored)

- **Item 2 (atomicity break):** 11 notes — back answers multiple questions or has long-winded back. Not refactored per MED-severity default; user can review in a follow-up pass.
- **Item 4 (magnitude/date claims):** 149 notes — Pass 2 trigger. External verification needed (FRED pulls, source documents). Not refactored in this pass.
- **Item 7 (stale cards):** 6 notes — flagged for awareness. Audit trail suggests these need re-anchoring but were not modified per audit pattern convention.

## LOW flags (nits)

None logged.

## Cross-block consistency notes

Not assessed — Item 8 audit pass was not reached before subagent timeout.

## Refactor log

- **Original notes refactored in place:** 8 (all HIGH-severity flags)
- **Notes deleted:** 0 (no wrong-and-confusing cards found in HIGH set)
- **Notes split into multiple:** 5 (the 5 list-splits)
- **New notes added:** 11 (3 Bernanke + 2 catalysts + 2 regimes + 2 CIP signatures + 2 customer types)
- **Total notes tagged `session-fixed-2026-08-16`:** 19 (8 originals + 11 new)
- **Sync status:** success (response: `{"result": null, "error": null}`)

## Open follow-up (for next audit pass)

The subagent timed out before auditing Items 3, 6, 8:
- **Item 3 — Sign / direction error (HIGH):** Not scanned. Could surface additional HIGH flags (e.g., long/short inversion, sign convention errors).
- **Item 6 — Vault-source mismatch (HIGH):** Not scanned. Could surface cards that contradict competence map / block plan / daily routine.
- **Item 8 — Cross-block consistency (HIGH):** Not scanned. Could surface deck-level disagreements (the SOFR-OIS naming pattern generalized).

**Recommendation:** schedule a follow-up audit pass scoped to Items 3 + 6 + 8. Use the saved scratch files (`/tmp/audit_notes.json`, `/tmp/audit_flags.json`) as the baseline — they contain the full 264-note dump.

## Skill changes (companion to this audit)

- `references/anki-helper.md` — added 3 principles (back-side list leak, combat interference, omit needless words) with Holiday/SuperMemo origin
- `references/deck-audit-pass1-checklist.md` — added Item 1b (back-side list answer check)

## Companion deliverable

Block 028 cards (27 total, all atomic, all Basic, all ~10s recall) pushed to `Macro Study::Phase 1 — Components::FX` deck with `refactor:2026-08-16` tag. See `2026-08-16-block-028-cards.md` for the card batch.

---

## How to revert any individual refactor

Each refactored note is tagged `session-fixed-2026-08-16`. To revert a specific note:
```python
anki('updateNoteFields', note={'id': NOTE_ID, 'fields': {'Front': '<original front>', 'Back': '<original back>'}})
anki('removeTags', notes=[NOTE_ID], tags='session-fixed-2026-08-16')
anki('sync')
```

To find the original front/back text for any refactored note, check the session log (search for note IDs 1782387710031, 1782975741861, 1783441333671, 1783787949517, 1783787949641, 1784975950892, 1785099069095, 1786784978325).