Registration 18 (A-COLA) build: stage 1 on master, stage 2 and close-out (blind replay) by MaxGhenis · Pull Request #585 · PolicyEngine/microcosm-dynamics

MaxGhenis · 2026-10-25T03:06:01Z

Do not merge. The coordinator lands this after independent review and the tier-count slot.

What this PR is

This is a blind replay of the stage-2 builder's tree for Registration 18 (A-COLA), milestone F0 of the mortality, groups and financing design. The builder's sandbox could not write .git, so it left an uncommitted tree. That tree is preserved as preserve/acola-f0-builder-tree-20261024 (05567445). This branch commits it from a lane that has opened nothing on the restricted list and none of the exposed session's text. The lane also ran the dd222b9 worktree equivalence itself and wrote the provenance record, the commit messages and this body itself.

It would supersede #571, which an exposed session committed. The coordinator chooses which one lands.

Specification: Registration 18, Candidate 2 design: latent-permanent conditioned chained QRF (research memo) #42 (comment), authorized by decision d658.
Design: Design: mortality, groups and financing lane (priced milestones, approved under d1922) #558 at 2c19f5b6 (decision d1922), milestone F0.
Instructions, by path (EV = /Users/maxghenis/microcosm-launch-evidence/dynasim-parity-20260909):
- the cleared builder brief, EV/acola-builder-brief-cleared-20260930.md;
- the stage-2 addendum, EV/acola-build-20260930/stage2-brief-20261024.md, and its erratum, EV/acola-build-20260930/stage2-brief-20261024.erratum.md;
- this replay's brief, EV/acola-build-20260930/f0-blind-replay-brief-20261024.md.

PSID use follows the project's data-access record. Every test uses invented data. Nothing here runs the registered population; scripts/run_acola_registered.py ran only through its own tests on invented fixtures.

axiom: n/a: microsim-only registered test build

What was built

The base is master 2195258c.

Commit	What it does
`4bbed4fd`	Replays stage 1 (`9de858d0`) with `git cherry-pick`. It keeps master's `tests/tier_counts.json`, `tests/README-tiers.md` and exclusion tuple, and its other 24 files are byte-identical to `9de858d0`. Stage 1 holds the separate config and row registry (C0, C1, M0–M4 runnable; C0-D registered and deferred), the runner, the statistics, the nulls with the TR 2026 Table V.C1 capture, the entry script and its guards, and the verbatim specification file (SHA-256 `02287d54…`).
`10c02653`	Fix: two missing spaces in the nulls record's December note. The record was regenerated by its script, and every null value, rational, rate and input hash is unchanged.
`4ff1fec0`	`scripts/compare_acola.py`, the comparison lane's standard-library script, with its tests on invented comparator records.
`a3806e56`	The mechanism-L equivalence script, the golden output for `dd222b9`, and its test.
`5ffb0eee`	The cross-check between stage 1's Table V.C1 capture and master's `data.tr2026` accessor for the COLAs effective December 2027–2029.
`b0e01614`	The delayed-credit disclosure test and the design note `docs/design/acola_trust_fund_implementation.md`.
`f02181e9`	An added test showing that a registered run carries the 48-month cap. It refuses parameters that differ from the committed statutory capture, and `statutory_value_checks` compares `max_delayed_months`.
`dae71963`	The design note: replay citations, the worktree equivalence, and a more precise item 17 (the claimant's own worker record and the aged widow(er) path).
`7cd45ee7`	Housekeeping: the five new `src/` files go into `POST_REVIEW_SOURCE_EXCLUSIONS`, its exact-tuple test and the reachability test, and the tier manifest is recounted.
`6abb6726`	The equivalence provenance record, and the design note's build-commit sentence.

Compared with the builder's tree, four files differ: the two nulls files, the disclosure test and the design note, each from a commit above. The provenance record and the tier files changed as the brief requires. No shared benefit module changes. This branch edits none of cola_track_a/, scenario_benefits.py, estimates/cola_age_profile.py, track_a_v2/, fra68_track/, ss/, claiming.py, engine/, gates.yaml, track_b/ or runs/*.json.

Invariants and the tests that check them

Property tests use Hypothesis; every input is invented. Test files are under tests/acola_track/.

Invariant	Tests
Accounting. Component totals sum to T_s; T_base = T_base,OASI + T_base,DI; P = s_OASI·P_OASI + s_DI·P_DI where both denominators are positive.	`test_statistics.py::test_accounting_identity_and_cut_bound_on_invented_components` (property), `::test_component_accounting_and_spouse_three_classes`
Identity. The ratio of means equals P when S_base = S_reform. Under mechanism L the sets are equal by construction, so equal counts are not evidence; the artifact says so.	`test_statistics.py::test_accounting_identity_and_cut_bound_on_invented_components` (property; every reform benefit positive, so the sets are equal)
Zero reform. With `annual_reduction = 0`, every change statistic, falling share and null is exactly 0, while median benefit levels stay equal and positive.	`test_nulls.py::test_zero_reform_is_exactly_zero` (property), `::test_build_zero_reform_nulls_are_all_exactly_zero`, `test_statistics.py::test_zero_reform_has_exact_zero_changes_share_and_floors`, `test_timing_properties.py::test_cut_bounds_monotonicity_and_zero_reform_on_rounded_paths` (property)
Bounds. Under a cut, worker-path individual changes lie in [N_full, 0], with a tested allowance for propagated dime flooring. Dual entitlement is an intended, labelled exception: a cut to the claimant's own PIA can raise the spouse excess by more than it lowers the own benefit.	`test_timing_properties.py::test_cut_bounds_monotonicity_and_zero_reform_on_rounded_paths` (property), `test_statistics.py::test_accounting_identity_and_cut_bound_on_invented_components` (property), `::test_mint_order_falling_share_and_individual_cut_bounds` (property; band medians of individual changes in [−100, 0]), `::test_valid_dual_entitlement_can_raise_benefits_under_registered_cut`, `::test_dual_entitlement_bound_violation_is_structural_under_dime_floors` (property)
Monotonicity (intended). A larger cut, or an earlier first reduced year, weakly lowers P and every individual benefit, outside the dual-entitlement exception.	`test_statistics.py::test_larger_cut_weakly_lowers_p_and_every_individual_change` (property), `test_timing_properties.py::test_cut_bounds_monotonicity_and_zero_reform_on_rounded_paths` (property)
Determinism. The same seed gives the same draw, and person order changes no statistic or floor.	`test_runner.py::test_same_draw_seed_is_deterministic`, `test_timing_properties.py::test_registered_draw_root_and_module_stream_are_deterministic` (property), `test_statistics.py::test_cost_and_mint_order_independent_including_every_floor`
Weighted median. Permutation and weight-scale invariance; the ordinary median under equal weights; bounded by min and max; the exact-W/2 midpoint in exact arithmetic; zero weights excluded and counted.	`test_statistics.py::test_median_bounds_permutation_and_weight_scale` (property), `::test_equal_positive_weights_match_ordinary_median` (property), `::test_registered_exact_weighted_median`, `::test_near_half_weight_uses_exact_binary_fraction`, `::test_zero_weights_are_excluded_and_counted_for_each_median`
Nulls. N = 100·(Π f_k − 1) with each f_k < 1; more reduced increases lower the null; a larger cut lowers it and a larger COLA raises it; the committed record reproduces from its pinned inputs, with 3, 3, 19 and 20 factors.	`test_nulls.py::test_null_is_exact_product_with_factors_less_than_one`, `::test_more_reduced_increases_strictly_lowers_the_null`, `::test_larger_reduction_lowers_null_and_larger_cola_raises_it` (all properties), `::test_committed_nulls_reproduce_exactly_from_pinned_inputs`
compare.py. Each four-corner interval contains the point; C is monotone in d and, for fixed d, in p while p − d > 0; the precision gate and the PWBM and MINT branches follow the registration; interval subtraction is [a, b] − [c, d] = [a − d, b − c]; null fields are reported, never guessed.	`test_compare.py::test_oact_corners_contain_point_and_monotonicity` (property), `::test_pwbm_corner_interval_contains_point` (property), `::test_oact_precision_gate_boundary`, `::test_mint_interval_branches_and_formats`, `::test_interval_subtraction_and_null_propagation`, `::test_null_fields_are_reported_never_guessed_and_empty_band_serializes`
Differential. P against v2's all-alive weighted-total cell; the ratio of means against A7; the floors against v2's and A7's conventions.	`test_statistics.py::test_ratio_means_and_union_differential_against_existing_functions`, `::test_cost_floor_matches_v2_and_ratio_means_floor_matches_a7`
Guards. The entry script refuses a wrong pointer, a HEAD other than `--registered-commit`, a dirty tree, an existing artifact or sidecar, and a changed specification file.	`test_registered_run_script.py::test_preflight_refusal_matrix`, `::test_changed_specification_is_refused`, `::test_existing_artifact_or_sidecar_is_one_shot_refusal`
Unchanged registries. `cola_track_a`'s `REGISTERED_ROWS`, the exercise-1 and exercise-3 constants and their consumers are byte-unchanged.	`test_unchanged.py` (2 tests)
Start year. Determination years 2010, 2011 and 2027 reach the legacy emitter. On C0, eligibility in 2027, 2028, 2029 and 2030 carries 3, 2, 1 and 0 reduced increases; C1 carries one more.	`test_timing_properties.py::test_future_start_year_reaches_the_actual_legacy_emitter`
No NaN. Undefined floors, empty bands and refused draws serialize as `null` with a reason under `allow_nan=False`.	`test_statistics.py::test_empty_band_and_undefined_floor_serialize_with_reasons`, `test_registered_run_script.py::test_undefined_values_and_reasons_serialize`, `::test_nonfinite_value_is_rejected_before_either_file_is_written`
Mechanism L unchanged since `dd222b9`. Byte-identical canonical output at `dd222b9` and at HEAD.	`test_mechanism_l_equivalence.py` (4 tests)
Delayed-credit disclosure. L applies the 48-month cap to invented people born 1917–1942 who claim after FRA, and a registered run cannot carry another cap.	`test_delayed_credit_disclosure.py` (27 items)
TR 2026 COLA cross-check. The Table V.C1 capture and master's `data.tr2026` accessor agree for the COLAs effective December 2027, 2028 and 2029.	`test_tr2026_crosscheck.py`

Tests run

Every run used .venv/bin/python -m pytest -p no:cacheprovider with POPULACE_DYNAMICS_PSID_DIR set to an empty directory, serially (pytest-xdist is not installed).

Selection	Tree	Result
`tests/acola_track`	`dae71963` plus the housekeeping diff later committed as `7cd45ee7`	289 passed (1,398 s)
`tests/acola_track/test_mechanism_l_equivalence.py`, rerun after the provenance commit	`6abb6726`	4 passed
Cleared brief §7 legacy list plus dry-run tests: `tests/cola_track_a`, `tests/fra68_track/test_registered_run_script.py`, `tests/track_a_v2/test_matrix_estimands.py`, `tests/estimates/test_cola_age_profile.py`, `tests/test_scenario_benefits.py`, `tests/fra68_track/test_dry_run_membership_text.py`, `tests/track_a_v2/test_dry_run_protocol.py`, `tests/baselines/test_track_a_inputs.py`	same as the first row	500 passed, 2 skipped
`tests/estimates/test_birth_evidence_artifact.py -k post_review`	`6abb6726`	2 passed
`-k tiers_match_policy_manifest`, full collection	`7cd45ee7`	1 passed, 24,907 deselected, 2 skipped

The 2 legacy skips. tests/cola_track_a/test_statutory_capture_oracle.py skips because the local policyengine-us checkout is not at the captured revision.
The 2 collection skips. Two Track B modules skip at collection because populace.fit is not installed in the repo venv.
Formatting. Black (79 columns), ruff and git diff --check are clean on all 25 Python files the branch adds or changes.

Tier counts

Tier	Base `2195258c`	This branch
unit	18,088	18,352
artifact	4,253	4,278
integration_psid	1,450	1,450
reproduction_legacy	520	520
oracle_policyengine	308	308
Total	24,619	24,908

The recount is correct at this branch's base only. Master has moved, so the coordinator's tier-count slot recounts at merge.

Mechanism-L equivalence and provenance

Registered side. git worktree add --detach .equiv-dd222b9 dd222b9af926c773a0f92732e2d08be507793378 inside the clone. PYTHONPATH=$PWD/.equiv-dd222b9/src .venv/bin/python scripts/acola_mechanism_l_equivalence.py $PWD/.equiv-dd222b9 --output <scratch>.
- An in-process rerun listed the 117 imported populace_dynamics files. All 117 resolve under the worktree's src and equal dd222b9's blobs.
Build side. The same script at 7cd45ee7 with PYTHONPATH=$PWD/src. It imports the same 117 files, byte-identical.
Result. Both outputs, and both in-process reruns, are byte-identical to tests/acola_track/mechanism_l_dd222b9.json:
- output SHA-256 07c59d69544503f87b43e8be5df57f01d773f597aa52b8f4027f40f7f9cc87de;
- 7 rows × 69 invented persons, 483 records;
- script SHA-256 5d7eef14e1cd4556e7a6d4d28f3b43f3cb4c442a68ffa6569724b5a831ea7a27.
- No REG18_MECHANISM_DRIFT. The golden file was not regenerated.
Worktree. Removed after the runs.
Provenance file. tests/acola_track/mechanism_l_dd222b9.provenance.json (SHA-256 504e60b5…) records the method, both commands, the registered and build commits, the output and script hashes, and the 117 imported files. It also keeps this replay's first-attempt runs (build side at a3806e56, the same output) and the stage-2 builder's earlier git archive export record verbatim (builder file SHA-256 b48fc549…).
Mechanism paths since dd222b9. git diff --name-status dd222b9 <master> over cola_track_a/, scenario_benefits.py, estimates/cola_age_profile.py, track_a_v2/, fra68_track/, ss/, claiming.py, engine/ and scripts/run_track_a_registered.py. At 2195258c, f8ffc7d9 and current master fe5ab5ab, it shows only two added files: engine/forward_earnings_s1_v1.py and ss/statutory_delayed_credit.py. No existing file changed.

Coordination (ruling 14)

No open PR touches cola_track_a/, track_a_v2/, scenario_benefits.py, estimates/cola_age_profile.py or acola_track/.
These touch the exclusion-tuple files (scripts/first_estimates_birth_evidence.py, tests/estimates/test_birth_evidence_artifact.py), so merge order will need a rebase of the tuple: Coverage-rule registry (CE1) #583, SIPP income, assets, pension and SSI readers (D2) #581, PSID income and wealth readers and field registry (D1a) #575, Aggregate-table captures and readers, part 1 (D4) #574, Federal SSI interface and source-to-input contract (SSI0) #573, Retirement-plan rule layer (WP2) #572, SM1: differential mortality by lifetime earnings, education and race (normalized within age-sex-year cells) #570, OP1: open-population harness around the sealed loop #566, Track B insured status: quarters of coverage; fully, currently and disability insured #562, Track B P2: history-based participation gate, developed on invented data (d1803, reserve) #561, IC3 employer gate block: thresholds, partitions, and rulings (draft for referee) #230 and Pre-IC3 noise floors: the Workstream B aggregate-side battery, promoted to v1 (E1, E2, E6/E7, E11; E11-detail/E12 gaps recorded) #223.

Against #571, by hash only

Of the 37 paths this branch changes from 2195258c, 30 are byte-identical at #571's head f78efbc6. Seven differ. For six, #571 holds the builder's version (checked against the builder's hash list) and this branch holds a commit above:

docs/design/acola_spec_only_nulls.json and src/populace_dynamics/acola_track/nulls.py (10c02653);
tests/acola_track/test_delayed_credit_disclosure.py (f02181e9);
docs/design/acola_trust_fund_implementation.md (dae71963, 6abb6726);
tests/tier_counts.json and tests/README-tiers.md (7cd45ee7).

The seventh is tests/acola_track/mechanism_l_dd222b9.provenance.json, which each branch wrote for itself.

🤖 Generated with Claude Code

This commit replays 9de858d0eaf14df7affd5d4e810568a6d41584e9, the stage-1 build of Registration 18 (A-COLA, decision d658), with `git cherry-pick 9de858d0` onto origin/master 2195258. Stage 1's parent, 75cd352, is 510 commits behind master. Conflicts and how each was resolved: - tests/tier_counts.json: kept master's version. Stage 1's counts were taken on its old base; the close-out recounts the tiers. - tests/README-tiers.md: kept master's version, for the same reason. - The exclusion tuple (POST_REVIEW_SOURCE_EXCLUSIONS) did not conflict: stage 1 did not edit it. The close-out appends the new src/ files to it. The other 24 files of 9de858d0 are byte-identical to it (checked blob by blob against 9de858d0's tree). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Stage 1's December note read "include December2030" and "determination year2029". The note is prose in the committed nulls record, which the second registration comment will cite, so it should read correctly. - src/populace_dynamics/acola_track/nulls.py: the note now reads "December 2030" and "determination year 2029". - docs/design/acola_spec_only_nulls.json: regenerated with scripts/build_acola_spec_only_nulls.py. Only the december_note line changes; every null value, exact rational, rate and input hash is unchanged. tests/acola_track/test_nulls.py passes (17), including the check that the committed file reproduces exactly from its pinned inputs. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

scripts/compare_acola.py is the standard-library script the validation-only lane runs after the run artifact merges and the seal opens. It fetches nothing. This commit adds the stage-2 builder's file unchanged (SHA-256 checked against the builder's hash list). What it does, as read in the code: - verify_seal hashes every file the seal lists (SHA-256 and byte count), refuses on any mismatch, missing file, path escape or an unlisted comparator record, and returns the seal's own SHA-256; compare_files then checks the artifact's SHA-256 against its .env.json sidecar before it parses the artifact. - OACT: CL = p - d and C = 100 d / (p - d), with p and d each +/- half their last printed unit (from the record's decimal counts) and the interval over the four corners; p - d > 0 is asserted at every corner. The 0.10 precision gate is applied mechanically and its branch recorded. - PWBM: C = 100 (R - B) / B; a point when both stored values carry six or more decimals, otherwise four corners. - MINT5: reading S (shortest decimal of the stored double, shifted two places for a percent format) with half-up and half-even checks against the HTML's decimals; reading L as a sensitivity. The headline statistic follows the record's median_type; "unsettled" compares both. - Gaps P - C, N - C, P - N and P - N_model per row (N_MINT, or M1's null, on the MINT5 rows) by [a,b] - [c,d] = [a-d, b-c], with the model's mean, draw SD and floor beside each. Decimal arithmetic at 50 digits; decimals are written as JSON strings. No pass band, no verdict. tests/acola_track/test_compare.py tests it on invented comparator records in the acola-comparator-v1 schema and on the stage-1 invented dry-run artifact layout. No comparator value is used anywhere. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

The stage-2 addendum (section 3, item 3) requires showing that mechanism L, the cola_track_a benefit path rows C0, C1 and M0-M4 use, gives identical outputs on invented data at the build and at the registration commit dd222b9. These four files are the stage-2 builder's, unchanged (SHA-256 checked against the builder's hash list): - scripts/acola_mechanism_l_equivalence.py (SHA-256 5d7eef14...7a27) builds one fixed invented population (seed 20261024) and projection state, then runs the legacy reference_benefit_rows for every runnable A-COLA row and writes canonical JSON of each person's base, reform and component amounts. It imports only legacy interfaces that exist at both commits, and no acola_track code, so it runs against a dd222b9 checkout. The population includes every birth cohort 1917-1942 claiming at 70, dually entitled spouses, DI and converted DI workers, a projected widow and each opening-stock component. - tests/acola_track/mechanism_l_dd222b9.json is the golden output (SHA-256 07c59d69...87de): 69 persons on each of 7 rows. - tests/acola_track/test_mechanism_l_equivalence.py reruns the script at HEAD and asserts byte equality, raising REG18_MECHANISM_DRIFT with the differing row/person ids otherwise; it also pins the script and golden hashes in the provenance record and checks that the script's rows equal the A-COLA registry's track_row() values. - tests/acola_track/mechanism_l_dd222b9.provenance.json is the builder's record (SHA-256 b48fc549...9ed8). The builder's sandbox could not create the dd222b9 worktree, so it records a git-archive export instead. The next commit replaces it with the worktree run. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

The stage-2 addendum (section 3, item 4) asks for a test that stage 1's Table V.C1 capture and master's data/tr2026 accessor give the same intermediate COLA increases effective December 2027, 2028 and 2029. tests/acola_track/test_tr2026_crosscheck.py is the stage-2 builder's file, unchanged. It reads the three captured cells through acola_track.nulls.tr2026_inputs(), which verifies the snapshot's pinned hashes, and the same three years through populace_dynamics.data.tr2026.cola_path(2027, 2029). It asserts the determination years, December timing, the V.C1 source class and equal percents as Decimals. It prints no other accessor value. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Both files are the stage-2 builder's, unchanged; later commits in this branch amend them. tests/acola_track/test_delayed_credit_disclosure.py covers the addendum's section 3, item 5. Mechanism L computes a retirement claimant's factor with claiming.benefit_factor (cola_track_a/benefits.py:345 at master 2195258 and at dd222b9). The test takes the invented late claimants of the equivalence population, one per birth year 1917-1942, claiming at 70, where 70 lies more than 48 months past FRA. For each it shows that worker_record's factor is the 48-month factor, that the same people get a larger factor and larger base and reform benefits when only the invented bundle's max_delayed_months is raised to 60, and that 48 is SSAParameters' default. docs/design/acola_trust_fund_implementation.md maps every Registration 18 element to its code and tests, states the invariants, and holds the section "Disclosures for the second registration comment": the interpretations (cleared brief section 1 item 2 and rulings 4-16), the delayed-credit cap with the erratum and the d1408 diagnostic note, C0-D's deferral, and the TR 2026 capture's URL, file, SHA-256, table, rows and column. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

The builder's disclosure test shows the 48-month cap on an invented parameter bundle and as SSAParameters' default. The disclosure is about the registered run, whose parameters come from a policyengine-us checkout, so the test did not yet show that the registered run carries the same cap. A registered run compares its runtime parameters with the committed statutory capture and refuses any mismatch: run_acola raises when cola_track_a.runner._parameter_consistency, which calls statutory.statutory_value_checks, is inconsistent, and that check covers max_delayed_months. The new test, test_registered_parameter_binding_carries_the_48_month_cap, shows: - the committed capture data/external/track_a_statutory_parameters.json and captured_ssa_parameters() both carry 48 months; - the delayed-credit check passes for that bundle and fails for one with 60 months, so a registered run cannot carry another cap; - under the committed FRA schedule, a claim at 70 by anyone born 1917-1942 is more than 48 months past FRA, and benefit_factor at 70 equals the factor at FRA + 48 months. Test only; no source change. The disclosure file now passes 27 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Edits to the stage-2 builder's design note: - Header: cite the stage-2 brief's erratum and this blind replay's brief beside the cleared brief and the stage-2 addendum. - Mechanism preservation: the builder could not create the dd222b9 worktree and ran the registered side on a git-archive export, so its note listed the worktree run as outstanding. That run is now done: the dd222b9 worktree inside the clone, PYTHONPATH on its src, all 117 imported modules resolving there, and output byte-identical to the golden file; the build side at a3806e5 also matches. The paragraph now says so, and keeps the builder's export as history. It also records that the dd222b9..origin/master path check gives the same result at f8ffc7d. - Disclosure 17: "including late-entitled linked worker records" was imprecise. The text now names the two routes read in the code: the claimant's own worker record (benefits.py:345, the same line at dd222b9) and, on the aged widow(er) path, the deceased worker's factor, which widow_benefit_path receives as deceased_claim_age_factor and which passes delayed credits through (ss/benefits.py widow_benefit). It adds that a registered run carries the cap because it refuses parameters that differ from the committed capture (48 months; pinned by the previous commit's test), and gives the statutory window, 840 - FRA months: 60 at FRA 65, falling to 50 for 1942 births, as the erratum states. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…iers Housekeeping for the five new src files of Registration 18's build, as the cleared builder brief's ruling 3 and the stage-2 addendum's item 6 require. - scripts/first_estimates_birth_evidence.py: append the five acola_track files (__init__, config, nulls, runner, statistics) to POST_REVIEW_SOURCE_EXCLUSIONS, so the historical reducer's source identity does not change. - tests/estimates/test_birth_evidence_artifact.py: the exact-tuple test lists the same five paths, and the reachability test asserts that the acola_track modules are collected but unreachable from the birth-evidence reducer. These two files are the stage-2 builder's versions, byte for byte. - tests/tier_counts.json and tests/README-tiers.md: recounted with EV/merge-watchers/recount-tiers.py on this tree (empty PSID root). Base 2195258: unit 18,088, artifact 4,253, integration_psid 1,450, reproduction_legacy 520, oracle_policyengine 308, total 24,619. Now: unit 18,352, artifact 4,278, others unchanged, total 24,908. That is one unit item more than the builder's recount (24,907), from the delayed-credit parameter-path test added on this branch. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

The stage-2 builder's sandbox could not create a git worktree, so its provenance record described a git-archive export of dd222b9's imported sources and left the worktree run and the build-commit pin open. This commit records those runs, made in this replay. - Registered side: `git worktree add --detach .equiv-dd222b9 dd222b9` inside the clone; scripts/acola_mechanism_l_equivalence.py run with the clone's venv and PYTHONPATH on the worktree's src. An in-process rerun listed the 117 imported populace_dynamics files, all under the worktree's src, equal to dd222b9's blobs and to the builder's list. - Build side: the same script at 7cd45ee with PYTHONPATH on the clone's src; the same 117 files, byte-identical. - Both outputs, and both in-process reruns, are byte-identical to tests/acola_track/mechanism_l_dd222b9.json (SHA-256 07c59d69..., 7 rows x 69 invented persons). Script SHA-256 5d7eef14... No REG18_MECHANISM_DRIFT; the golden file was not regenerated. - The record keeps the builder's export record verbatim (file SHA-256 b48fc549...) and notes the replay's first attempt, which made the same runs with the build side at a3806e5 before it stopped. - The worktree is removed. The design note's mechanism-preservation paragraph now names the build commit 7cd45ee beside a3806e5. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

vercel · 2026-10-25T03:06:07Z

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

Project	Deployment	Actions	Updated
social-security-model	Ready	Preview	Oct 25, 2026 3:06am UTC

MaxGhenis and others added 10 commits October 24, 2026 11:44

Navigation Menu

Repository navigation

Registration 18 (A-COLA) build: stage 1 on master, stage 2 and close-out (blind replay) - #585

Registration 18 (A-COLA) build: stage 1 on master, stage 2 and close-out (blind replay)#585
MaxGhenis wants to merge 10 commits into
masterfrom
acola-reg18-build-blind-20261024

MaxGhenis commented Oct 25, 2026

Uh oh!

vercel Bot commented Oct 25, 2026

Uh oh!

This branch was successfully deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

MaxGhenis commented Oct 25, 2026

What this PR is

What was built

Invariants and the tests that check them

Tests run

Tier counts

Mechanism-L equivalence and provenance

Coordination (ruling 14)

Against #571, by hash only

Uh oh!

vercel Bot commented Oct 25, 2026

Uh oh!

This branch was successfully deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant