Skip to content

Release v1.9.0 — the tower-tip epoch (tip_mass 0.25 + MAP-containment law, ADR-019 Am. 3)#202

Merged
Polichinel merged 1 commit into
mainfrom
development
Jul 24, 2026
Merged

Release v1.9.0 — the tower-tip epoch (tip_mass 0.25 + MAP-containment law, ADR-019 Am. 3)#202
Polichinel merged 1 commit into
mainfrom
development

Conversation

@Polichinel

Copy link
Copy Markdown
Contributor

Promote developmentmain for v1.9.0. Trial-merge clean; merged tree identical to
development. Merge with a MERGE COMMIT, not squash (dev/main-divergence rule).

What ships (MINOR; CONFORMANCE_FLOOR stays 1.0.0)

  • tip_mass 0.5 → 0.25 (ADR-019 Amendment 3): the tower-tip MAP is the median of the top
    floor of the published tower
    . Evidence in research/map_hdi/tip_mass_study.py (bias halved,
    RMSE win at pooled S, exact zero-inflation reads, real-cell C-44 gate passed with margin).
  • MAP-containment law in assert_summarizer_contract (exact draw-count condition;
    asymptotically every HDI of mass > 12.5% provably contains the MAP).
  • research/figures/ tower-figure toolkit + provenance note; reports/plots/ gitignored.
  • Consumer note: published MAPs shift toward the mode on skewed cells (the intended C-32
    direction) — downstream repos (faoapi/reporting) will see numerically different MAPs.

After merge + green CI: gh release create v1.9.0 (version-guard: 1.9.0 > 1.8.1 on PyPI).

🤖 Generated with Claude Code

…containment law (ADR-019 Amendment 3) → v1.9.0 (#201)

The tower-tip MAP is now the median of the TOP floor of the published tower
(tip_mass default 0.5 → 0.25), matching the design intent the name always
promised. Evidence (research/map_hdi/tip_mass_study.py: 1000-replicate
accuracy/stability battery, duplicate-capture frontier, the real-cell C-44
gate): the 0.5 shorth carried a structural rightward bias that does not shrink
with S; 0.25 roughly halves it, beats the shorth on RMSE at pooled S, reads
zero-inflated cells exactly, and passes the real-cell zero-stack gate with
margin (masses <= 0.15 resurrect the C-44 signal loss; 0.20 has zero margin).

New MAP-containment law in assert_summarizer_contract: every floor holding more
than half the tip floor's draws (a mass-m floor spans floor(m*S)+1 draws;
asymptotically m > tip_mass/2 = 12.5%) provably contains the tip — wider floors
by nesting, narrower qualifying floors by the sub-window trim argument. All
published bands (50/90/95/99) qualify; the unguaranteed region shrinks vs 0.5.

Also: research/figures/ — permanent seeded generators for the PRN06 tower
figures (upright tower: 99% base at the bottom narrowing to the red 25% top
floor with the tip on it; degenerate zero-width bands rendered as open
diamonds), reports/plots/ as the gitignored output home, and the provenance/
placement research note (plotting lives under research/, never src/ — the
sibling charters exclude it six times over).

Governance: ADR-019 Amendment 3; Summarize CIC §3 + header amendment;
CHANGELOG [1.9.0]; README banner + chronicle (banner guard green). Consumer
note: published MAPs shift toward the mode on skewed cells — the intended C-32
direction. CONFORMANCE_FLOOR stays 1.0.0. Gate: ruff / mypy (3.10+3.11) /
pytest 100% line+branch / validate_docs all green.

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
@Polichinel
Polichinel merged commit 125905e into main Jul 24, 2026
14 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant