[MaxGhenis]

What changes

• Slide 13: "We fix CPS errors, so our child poverty rates differ from Census's by design. We're tracing how calibration, sampling noise and rent imputation move each state's rate." Max asked for this: we differ from Census on purpose, so the slide no longer says "diagnosis" or "miss".
• Slides 3 and 4: "the newest version of seven models", since the deck drops a version released after one it shows.
• Slide 8: cites Section 22 and Section 45, as the pinned amount.yaml does.
• Slide 10: "We estimate that they pass SNAP's tests", since those rates rest on our sampled households.
• Notes: new answers on Philadelphia's wage tax, Nevada and South Dakota, PN3's Roadmap figures for Ohio, Oklahoma, Maryland and Indiana, and timing that gives slides 10 and 12 more room with slide 13 still at 10:45.
• compute_family.py: a partial rerun now refuses a file computed on another version or commit.

No chart or data file changes.

Invariants

• Every printed amount keeps its space before the next word (test from Rebuild the APPAM talk around one family's question #101).
• A partial rerun keeps only rows computed on the same pin (test_a_partial_rerun_refuses_another_model_version).
• In Indiana at $39,500, the posted limit pays exactly January through April (test_indianas_posted_limit_stops_at_its_april_limit).
• Slide 6 hard-codes Maryland as a cliff state, and a test now requires MD in the data's list.

Reviews

• ~/reviews/appam-2026/rebuild/deck-review/review-r1-opus-deck.md (Opus 5.5, REQUEST CHANGES at 05812f0): this PR fixes findings 1, 2, 3, 5, 6, 7, 9, 10, 11 and 12, and I read the live TAXSIM page for finding 4, which needs no change.
• ~/reviews/appam-2026/rebuild/deck-review/review-r1-opus.md (Opus 5.5, APPROVE at 05812f0): this PR applies O1 to O5, F2 and F3.
• ~/reviews/appam-2026/rebuild/verify/review-head-roadmap-opus.md (APPROVE): this PR applies items 1 to 6.

Checks

• Vitest: 179 passed. tsc and ESLint report nothing.
• Python: the two new tests passed before the guard also checked the commit. I stopped a wider subset run after 33 passes to free the machine, so the full suite still needs a run after the talk.

Follow-up

Rerun Pennsylvania with the wage tax set, after the talk. The Questions answer sizes it by arithmetic at the 3.74% resident rate.

🤖 Generated with Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
policyengine-slides	Ready	Preview	Nov 5, 2026 3:29pm UTC

[MaxGhenis]

Merged on Max's go ("deploy to policyengine.org/slides whatever we have", and the slide 13 wording he asked for), relayed by the coordinating session. Head b1b4a7e: gh pr checks exit 0, MERGEABLE and CLEAN, and the preview swept clean at 1440×810 and 1280×720. Reviews: review-r1-opus-deck.md (Opus 5.5, REQUEST CHANGES at 05812f0, findings fixed here), review-r1-opus.md (Opus 5.5, APPROVE at 05812f0) and review-head-roadmap-opus.md (APPROVE). The delta itself had no separate review, because the changes come from those reviews' own wording.