Flip ratified gate-2 amendment 1 live: mean-over-draws estimator (prospective only) by MaxGhenis · Pull Request #97 · PolicyEngine/microcosm-dynamics

MaxGhenis · 2026-07-08T16:34:11Z

Flip ratified gate-2 amendment 1 live: mean-over-draws estimator

Authorized follow-up flip for gate-2 amendment 1, ratified by merge of #96
(merge commit fec27eb51) after the full ceremony: adversarial round AMEND
BEFORE RATIFYING (comment 4915412987)
-> fixes 5b70840 (comment 4916048161)
-> verification RATIFY AS-IS (comment 4916419901).
Mirrors the gate-1 amendment-2 flip (#69) and the gate-2 lock flip (#81).

The estimator moves per-cell scoring from one frozen simulation draw to
the mean over K=20 pre-registered draws (numpy default_rng(5200 + k),
k=0..19) of the cell rate, scored |ln(rbar_candidate,s / rate_a,s)|.
Tolerances, the power cap, the 46-cell membership, and the 4-of-5-seed
conjunction are byte-identical to origin/master — only the candidate
statistic changes. A structured diff of the thresholds subtree confirms the
only changes are the five prose fields below; all 46 tolerances are unchanged.

The five locked-text edits (`flip_on_ratification.flip_edits`, applied exactly)

edit	locked path	old -> new
candidate	`protocol.candidate`	"one replicate, simulation seed s" -> K=20 draws at `default_rng(5200 + k)`, 20-draw mean cell rate; records the committed `4200 + s` registration history and the new `5200 + k` convention explicitly (`registration_divergence` resolved)
statistic	`statistic`	`r_candidate,s` -> `rbar_candidate,s` (mean over K=20 draws of the cell rate)
pass_rule	`protocol.pass_rule`	`r_candidate,s` -> `rbar_candidate,s`; mean of the cell RATE across draws, then `\|ln(rbar / rate_a)\|`
faithful_candidate_oc	`protocol.faithful_candidate_oc`	basis note added: the draw-noise-free half-normal basis now MATCHES the estimator; `p_seed_pass 0.9404` / `p_gate_pass_4_of_5 0.9685` stand unchanged (unachievable under the single-draw estimator, finding 1)
draft_label	`protocol.pass_rule`	stray "DRAFT seed-level conjunction" -> "Seed-level conjunction"

Also carries the ratified fresh_run_artifact_schema (candidate-10 run
contract: per_draw_per_cell_rates shape [20, 46, 5]; undefined-draw
INVALIDATES the run; report-only per-draw dispersion) into the live
protocol, and removes amendment_proposed (preserved in git history at the
ratification merge fec27eb51).

`amendment_history` entry

gate_2 gets its first entry: id: 2026-07-08-mean-over-draws-estimator,
flipped_live: this pull request, with all four ceremony pointers
(review 4915412987 / fixes 5b70840 + comment 4916048161 / verification
4916419901 / ratifying merge PR 96 fec27eb51). PROSPECTIVE ONLY
(no_self_rescue): no committed run's verdict changes; candidates 1-9 stand
FAIL. The single-draw estimator contradicted the ratified OC basis; K=20
restores ~0.939 / 0.967 ~= the ratified 0.9404 / 0.9685. Evidence chain: #94
(chronic-cell forensics) and the candidate 1-9 ladder (issue #42).

Locked-hot test rewiring (`tests/test_gates_derivations.py`)

The gate-1 amendment-2 lesson: ratified bindings must stay hot, not go
dormant with the proposal object. The gate-2 amendment block keyed off
amendment_proposed and would go silently dormant on removal. This PR:

Adds a "Gate-2 amendment 1 RATIFIED bindings" section binding the
load-bearing guarantees to their LIVE post-flip locations: the K=20 / 5200+k
convention, the artifact schema, the OC basis, the history entry with all
four ceremony pointers, and prospective-no-verdict-change vs the nine
committed artifacts (runs/gate2_hazard_v1..v9.json, all gate_2_pass
false).
Keeps the proposal-object tests dormant-safe for future proposals:
_gate2_amendment() now guards on proposal_number == 1 (the gate-1
_amendment_2 pattern), so a future gate-2 proposal (number != 1) never
reactivates these proposal-1-specific bindings.
Updates the two tests that asserted the old locked text
(flip_edits' locked_text_now presence, "one replicate, simulation
seed s") — the live-text verification now asserts the new amended text
in the ratified section.

Verification

black -l 79 clean; ruff clean.
Full suite green (serial, repo .venv).

🤖 Generated with Claude Code

…spective only) Moves gate-2's ratified amendment 1 from the amendment_proposed object into the locked protocol. Per-cell scoring goes from ONE frozen simulation draw to the MEAN over K=20 pre-registered draws (numpy default_rng(5200 + k), k=0..19) of the cell RATE, scored |ln(rbar_candidate,s / rate_a,s)| at BYTE-IDENTICAL tolerances and the same 46-cell, 4-of-5-seed conjunction. Only the candidate statistic changes; the error budget does not (all 46 tolerances, the power cap, governance, scope, and external anchors are byte-identical to origin/master). Ceremony (mirrors the gate-1 amendment-2 flip #69 and the gate-2 lock flip #81): adversarial round AMEND (PR 96 comment 4915412987) -> fixes 5b70840 (comment 4916048161) -> verification RATIFY AS-IS (comment 4916419901) -> ratified by merge of PR 96 (merge commit fec27eb) under the maintainer's standing campaign directive of 2026-07-07, full ceremony. The five locked-text edits (flip_on_ratification.flip_edits, applied exactly): - protocol.candidate: one frozen replicate -> K=20 draws at default_rng(5200 + k); records the committed 4200+s registration history and the new 5200+k convention explicitly (registration_divergence resolved). - statistic: r_candidate,s -> rbar_candidate,s (the mean over K=20 draws of the cell rate). - protocol.pass_rule: r_candidate,s -> rbar_candidate,s (mean of the cell RATE across draws, then |ln(rbar / rate_a)|). - protocol.faithful_candidate_oc: basis note added -- the draw-noise-free half-normal basis now MATCHES the estimator; p_seed_pass 0.9404 / p_gate_pass_4_of_5 0.9685 stand unchanged (unachievable under the single-draw estimator, finding 1). - protocol.pass_rule: the stray "DRAFT" label removed. Also carries the ratified fresh_run_artifact_schema (the candidate-10 run contract: per_draw_per_cell_rates [20,46,5]; undefined-draw INVALIDATES the run; report-only per-draw dispersion) into the live protocol, adds gate_2's first amendment_history entry (id 2026-07-08-mean-over-draws-estimator, flipped_live: this pull request), and removes amendment_proposed (preserved in git history at the ratification merge fec27eb). PROSPECTIVE ONLY (no_self_rescue): no committed run's verdict changes; candidates 1-9 stand FAIL. The single-draw estimator contradicted the locked block's own ratified OC basis; K=20 mean-over-draws restores ~0.939 / 0.967 ~= the ratified 0.9404 / 0.9685. Tests rewired locked-hot (the gate-1 amendment-2 lesson): the load-bearing guarantees -- the K=20 / 5200+k convention, the artifact schema, the OC basis, the history entry with all four ceremony pointers, and prospective-no-verdict-change vs the nine committed artifacts -- rebind to their LIVE post-flip locations in a new "Gate-2 amendment 1 RATIFIED bindings" section. Proposal-object tests stay dormant-safe (guarded on proposal_number == 1, so a future proposal != 1 never reactivates them); the two tests that asserted the OLD locked text are updated to the new live text. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

vercel · 2026-07-08T16:34:13Z

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

Project	Deployment	Actions	Updated (UTC)
social-security-model	Ready	Preview, Comment	Jul 8, 2026 4:34pm

…r the amended mean-over-20-draws estimator; observed-residual count + age-band remarriage deltas; designed count-cell cancellation asymmetric (male net +0.025 ln 4/5, female net +0.044 ln 3/5), share_widowed.75+|female holds the gate 4/5 (#42 comment 4917059482) (#98) The first fresh gate-2 registration scored under the amended mean-over-K=20-draws estimator (gates.yaml gate_2 amendment 1, ratified 2026-07-08, flipped live #97). Frozen spec: issue #42 comment 4917059482. One-shot; published REGARDLESS of verdict. VERDICT: gate_2_pass=False, 1/5 seeds pass (only seed 1). Runtime 168s (K=20 draws x 5 seeds); no undefined draw. Spec = candidate 8 + two deltas, under the amended estimator: * Delta 1 (candidate 9's, unchanged): observed undatable-marriage lifetime-count initial state R = n_marriages - datable in-exposure marriage events (the forensics Q1 residual), RNG-neutral; observed_residual_counts / _delta1_reconciliation imported from candidate 9. Reconciled to remainder 0.0 train-side before the one-shot (per-person identity max 0.0; aggregate max remainder 2.9e-16). * Delta 2 (new): age-band-conditioned remarriage by ego age band (18-34/35-49/50+) x years-since-dissolution x origin x sex, same add-one smoothing (wbar_diss), candidate 8's order-bit stratum and aggregate-preservation rescale REMOVED. Fertility = candidate 8's kernel, unchanged. * Precheck bit-exact (ref-moment dev 0.0, rate_a dev 0.0, holdout sha256 all match). Amended estimator + fresh_run_artifact_schema: rbar_candidate,s = mean over K=20 draws (default_rng(5200+k), k=0..19) of the cell rate; score |ln(rbar/rate_a,s)| scored once. The artifact commits the [20,46,5] per-draw per-cell rate cube (rbar recomputable cell-by-cell), the undefined-draw check (run_invalidated=False, 0 undefined), and report-only per-draw dispersion (per-cell sd + worst-draw |ln|). Failing gated cells (rbar-score/tol): * share_widowed.75+|female seeds 0,2,3,4 (0.231/0.196/0.245/0.192, tol 0.185) -- an elderly-widow-stock LEVEL cell neither delta addresses; it independently holds the gate on 4/5 seeds (decider: broader than the registered modal + delta-targeted cells). * mean_lifetime_marriages|female seeds 2,4 (0.064/0.067, tol 0.051). * mean_lifetime_marriages|male seed 2 (0.052, tol 0.047). * completed_fertility.c1970s seed 2 (0.182, tol 0.171). All 4 gated remarriage cells pass 5/5 (the flagged thin-50+ risk did not bite; only one <20-py cell, age18-34|ysd2|widowed|male, no gated failure). Count-cell tilt vs design (residual +0.046, over-production -0.046, net ~0) was ASYMMETRIC (designed_cancellation_succeeded=False): male net +0.025 ln (4/5, delta 2 removed ~half the over-production) but female net +0.044 ln (3/5, delta 2 removed almost none). The over-production age conditioning targets is male-heavier; delta 1's residual is present for both sexes, so on the low-rate_a seeds (2,4) where c8 was near-perfect the female residual pushes the cell out -- the registration's flagged modal failure. Per-draw dispersion (report-only): 33 cell-seed instances where the worst single draw would clip the tolerance but the 20-draw mean passes -- the draw noise the amended estimator averages out, vindicating amendment 1 on this evidence. Adds scripts/run_gate2_candidate10.py, runs/gate2_hazard_v10.json, tests/test_gate2_candidate10.py (consistency + schema conformance + rbar recompute from per-draw rates + live seed-0 single-draw pin + faithful-copy attestation). Updates test_gates_derivations.py's prospective-amendment no_self_rescue check to include candidate 10 (10 committed gate-2 runs, all still gate_2_pass false). Pre-registered forecast was P(pass) 0.65-0.75. Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

…ro threshold movement) (#112) Moves gate-2's ratified amendment 2 (the tranche split) from the amendment_proposed object into the locked block. STRUCTURAL, zero threshold movement: the locked gate_2.thresholds scored surface is byte-identical to origin/master -- all 46 gated cells, every tolerance and derivation, the option-(a) protocol, the ln(1.5) power cap, and the 4-of-5 seed conjunction unchanged (verified at full-subtree granularity). Only naming, scope prose, the certification map, the unlocked sibling stubs, one governance rule, and the history changed. Ceremony (mirrors flips #69, #81, #97): adversarial round AMEND BEFORE RATIFYING (PR 111 comment 4930448912) -> fixes 81ea41e (comment 4930615774) -> verification RATIFY AS-IS (comment 4930753295) -> ratified by merge of PR 111 (merge commit 8a4a240) under the maintainer's standing campaign directive of 2026-07-07, full ceremony. The six locked-text edits (flip_on_ratification.flip_edits, applied exactly): - description: first clause re-scoped so tranche 2a claims only mh85_23 + cah85_23 (+ ind2023er deaths); household-composition/MX23REL -> gate_2b. - description_formula_clause: the "computes exactly" statutory-formula oracle clause (#78/#80) RETAINED verbatim (fix C1). - holdout_basis: [mh85_23, cah85_23, MX23REL] -> [mh85_23, cah85_23]; MX23REL -> gate_2b.holdout_basis. - scope_note: provision_class_coverage prose gains explicit tranche ids (2a/2b/2c); the certification_scope map is promoted into the locked scope. - scope_note_draft_label: "This DRAFT tranche covers..." relabeled to "tranche 2a_marital_fertility covers...". - scope_note_holdout_xref: the caregiver "named in holdout_basis but not scored in this draft" cross-reference re-pointed to the unlocked gate_2b. The five flip_additions: tranche_id 2a_marital_fertility inside thresholds; gate_2b (relationship_household) and gate_2c (marriage_earnings_joint) UNLOCKED sibling tranches (locked: false, no floor, no scored run, each with its own future lock ceremony naming the verification round); the certification_scope map promoted into the locked scope; and the standing governance rule description_claims_exactly_the_scored_surface promoted into gate_2.governance.amendment_rules (attributed to review #106 finding 4 / PR #111 finding 8), binding 2b/2c/gate_3. gate_2.name ("family and benefit outputs") retained per its C3 disposition (the tranche container). Also applies the two verification-round flip-time nits (comment 4930753295): the 2b bullet-3 legend quote spacing normalized to "poverty/household" to match the derivation_basis quote; and COLA qualified in the A2 own-record example list (its levels ride the Phase-B claiming/mortality components -- #74 marks CA, M, DI; #78 covers CA, M/DI uncertified -- so it is not a pure gate-1 + oracle surface, though still outside the 2a/2b/2c tranches). Adds gate_2's second amendment_history entry (id 2026-07-09-tranche-split; all four ceremony pointers; content: ZERO THRESHOLD MOVEMENT, verdicts v1-v15 FAIL and v16 PASS stand as tranche-2a verdicts, citing #106 finding 4 and the candidate-16 pass #109), and removes amendment_proposed (preserved in git history at the ratification merge 8a4a240). PROSPECTIVE / STRUCTURAL ONLY (no_self_rescue): a naming + certification-scope amendment has no verdict to rescue and re-scores nothing; candidate 16 stays PASS (4/5), candidates 1-15 stay FAIL, as tranche-2a verdicts. Tests rewired locked-hot: the re-scoped description/holdout_basis, tranche_id, the promoted certification_scope map (per-class requires_tranche + coverage tokens bound EXACTLY, both nits asserted), the gate_2b/gate_2c stubs, the promoted standing rule, and the second history entry with all four pointers rebind at their LIVE post-flip locations in a new "Gate-2 amendment 2 RATIFIED bindings" section. The proposal-object tests stay dormant-safe (guarded on proposal_number == 2, so they skip once amendment_proposed is gone), and the D1 merge-safe guard naturally skips post-flip. Full suite: 1319 passed, 68 skipped, 0 failed (-n 8); black -l 79 + ruff clean. Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

vercel Bot deployed to Preview July 8, 2026 16:34 View deployment

MaxGhenis merged commit f6b571a into master Jul 8, 2026
7 checks passed

MaxGhenis mentioned this pull request Jul 10, 2026

Flip ratified gate-2 amendment 2 live: explicit tranche structure (zero threshold movement) #112

Merged

MaxGhenis mentioned this pull request Jul 12, 2026

gate_w1 amendment 1 flip (ceremony record inside) #165

Merged

Navigation Menu

Uh oh!

Flip ratified gate-2 amendment 1 live: mean-over-draws estimator (prospective only) - #97

Flip ratified gate-2 amendment 1 live: mean-over-draws estimator (prospective only)#97
MaxGhenis merged 1 commit into
masterfrom
gate2-amendment-flip

MaxGhenis commented Jul 8, 2026

Uh oh!

vercel Bot commented Jul 8, 2026 •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Conversation

MaxGhenis commented Jul 8, 2026