Berg proprietary data proxy feasibility follow-up

Objective

Vet licensing/ToS/redistribution for any native Berg composite series for the F4 composite task (aggregating F1 price/boundary, F2 volume/TRIN, F3 breadth/thrust per ledger F4 doctrine). This investigation task must use only public proxies with explicit disclosure and defer native-series vetting to existing bead qm-zfl5, adopting no proprietary series without vetting.

Approach

1. Intake classified as investigation with external_data + public_artifact (not indicator/data_source), split_required=false — defer to open qm-zfl5 which owns data_source feasibility go/no-go; no child Beads. Planning pinned two blocking work items: disposition + publication, next investigation.
2. Re-read ledger research/findings/berg-coverage-ledger-qm-tykd.5.md/.json §1 (3 reports: Approach 2015, Boundaries 2008, Data Extremes 2023; research-process landing page) §2 taxonomy (F1–F4 all untestable_as_exact_replication_testable_as_proxy_family), §3 F4 doctrine (“native Berg indicator series is proprietary/unavailable (30k daily, institutional only). Public proxies are acceptable only with disclosure, causal lag, and preservation of sparse-signal semantics”) and Proprietary-data follow-up deferring all licensing to qm-zfl5.
3. Inspected gap record research/findings/qm-zzdt-berg-proxy-boundary-gap.md §§1–4 (30k non-redistributable boundary, proxy provenance table, 4 rejected experiments, 12 beads referenced; lists qm-m1ib as F4 deferral companion; policy no_pdf_redistribution=true, proxy_disclosures_required=true).
4. Ran bounded offline verification: data/market.sqlite close_only v5 (schema_version=5, ohlcv_layout=close_only, 127 tickers, 806,897 ohlcv rows). Proxy tickers INDEX:HIGN 6331 2001-06-05..2026-08-11, INDEX:LOWN 6312, INDEX:S5FI 4956 2006-12-29.., USI:UVOL 7616 1996-04-19.., USI:DVOL 8643 1992-01-02.. — all persisted via research/input_provenance.py:public-input-provenance/v1 and backtest/datastore.py:connect_readonly (query-only). Query SELECT ticker FROM instruments WHERE ticker LIKE '%BERG%' returns 0 rows — zero native Berg series imported, zero synthetic __BERG*__ history.
5. Confirmed bead states: qm-tykd.1..4 closed rejected (proxy-only n=389/183/221/358), qm-tykd.5 closed ledger v1, qm-zzdt closed gap, qm-t0u5 closed rejected composite promotion, qm-zfl5 OPEN P3 (licensing owner), qm-is3q OPEN P3 reserved, qm-ckab blocked ledger v2.
6. Authored executed notebook research/experiments/qm-m1ib.ipynb (7 cells, 3 code cells with outputs) documenting falsifiable H0/H1, licensing assessment, public-proxy provenance with citations, and deferral decision. Converted via python3 -m research.reports notebooks --id qm-m1ib and rebuilt changelog via python3 -m research.reports results. No private data, no PDF redistribution.

Files changed

• research/experiments/qm-m1ib.ipynb — executed notebook (H0 vs H1, licensing/ToS bounded check, public proxy table with spans, citations, deferral decision; offline code outputs for Datastore status and bead states).
• research/results/qm-m1ib.md — this durable result note (YAML front matter id: qm-m1ib, searchable summary).
• docs/site-data/results.json, docs/results.html (and detail pages) — rebuilt via canonical research.reports results producer (qm-m1ib newest).
• docs/experiments/qm-m1ib.html (generated via research.reports notebooks) — self-contained HTML rendering of the executed notebook, if produced by the producer.
• data/market.sqlite — not mutated; provenance verified read-only (Volume=0 close_only caveat preserved; TRIN/breadth proxies are derived, not licensed NYSE history).

Validation

• sqlite3 data/market.sqlite "SELECT ticker FROM instruments WHERE ticker LIKE '%BERG%'" — 0 rows (no proprietary adoption).
• sqlite3 data/market.sqlite "SELECT ticker, n_rows FROM instruments WHERE ticker IN ('INDEX:HIGN','INDEX:LOWN','INDEX:S5FI','USI:UVOL','USI:DVOL')" — 6331/6312/4956/7616/8643 matches proxy table.
• python3 -m backtest.datastore status — schema v5 close_only 127 tickers 806897 ohlcv under soft limit; connect_readonly query-only invariant.
• bd show qm-m1ib qm-zfl5 qm-tykd.4 qm-zzdt --json | python3 -m json.tool — qm-m1ib IN_PROGRESS→this publish, qm-zfl5 OPEN P3, qm-tykd.4 closed rejected, qm-zzdt closed gap.
• grep -c 'qm-zfl5\|miltonberg.com\|INDEX:HIGN' research/results/qm-m1ib.md research/experiments/qm-m1ib.ipynb — citations and deferral present.
• python3 -m research.reports notebooks --id qm-m1ib && python3 -m research.reports results — notebook HTML and results JSON rebuilt via canonical producers.
• python3 -m research.artifacts check --mode full --offline — shared/canary artifacts deterministic and current; no hand-edited docs/ bytes.
• python3 -m research.pipeline validate && python3 -m research.pipeline verify --offline — pipeline graph and offline verification passed.
• python3 research/scripts/check_static_reports.py docs — static bundle OK (no docs JSON cache reuse).
• git grep qm-m1ib research/results/qm-m1ib.md — entry is grep-searchable.

Results

Licensing/ToS assessment for any native Berg composite series is completed as explicit deferral to existing bead qm-zfl5 — native 30k daily feeds remain proprietary/not redistributable per ledger F4 doctrine and qm-zzdt boundary, with zero Berg tickers in data/market.sqlite and no import transaction. No proprietary series was adopted without vetting in this task. Public F1–F3 proxy provenance is cited with source URLs (research-process page + 3 whitepaper URLs) and spans (INDEX:HIGN/LOWN 6331/6312 2001-06-05.., INDEX:S5FI 4956 2006-12-29.., USI:UVOL/DVOL 7616/8643) via public-input-provenance/v1; F4 composite remains proxy-only aggregation (count>=2 / z-mean>=1.5, T+1 lag, fee_bps=10) labelled not as Berg native performance. Follow-up stays owned by qm-zfl5 (licensing go/no-go), qm-is3q (tradable promotion if robust proxy emerges), qm-ckab (ledger v2). No new Beads required.