Point-in-Time Alternative Data Ingestion, Entity Resolution & Survivorship-Free Panel Construction

Naive alternative-data backtests routinely overestimate live out-of-sample alpha due to three structural data leaks: (1) aligning corporate filings by fiscal period-end (fp) or calendar filingDate, which introduces lookahead bias whenever disclosures arrive after the 16:00:00 ET market close and silently overwrites historical states upon subsequent restatements; (2) evaluating signals over surviving index constituents rather than historical registrant unions; and (3) passing historical text through commercial large language models whose pre-training cutoffs post-date the evaluation window. We document the deterministic, zero-LLM ingestion architecture of Starling Quantitative (https://starlingquant.com) across 3,607 historical Russell 3000 SEC Central Index Keys (CIKs) from January 2019 through November 2026, explicitly retaining 555 non-surviving and delisted issuers (15.4%) with automatic -30.0% Shumway (1997) performance-delisting haircuts. Every structured XBRL item and primary filing is joined to its exact down-to-the-second SEC EDGAR acceptanceDateTime, mapped to canonical 10-digit CIK and composite point-in-time OpenFIGI identifiers, stamped with One-Document-One-Witness (source_artifact_id) provenance, and quarantined to live forward shadow testing whenever a vendor feed relies on mutable watermarks. Finally, we specify our dual institutional delivery schemas: Feed Tier A (pre-quorum ScoutPitch feature envelopes) and Feed Tier B (post-quorum calibrated MasterDecision consensus). Canonical Web Monograph: https://starlingquant.com/guides/point-in-time-alternative-data-and-survivorship-free-panelsComplete Technical Specification: https://starlingquant.com/llms-full.txt

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-14
DOI
https://doi.org/10.5281/zenodo.24222467
Primary Topic
Financial Markets and Investment Strategies
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Point-in-Time Alternative Data Ingestion, Entity Resolution & Survivorship-Free Panel Construction

Starling Quantitative Research
Zenodo (CERN European Organization for Nuclear Research)
Financial Markets and Investment Strategies
article

Point-in-Time Alternative Data Ingestion, Entity Resolution & Survivorship-Free Panel Construction

Starling Quantitative Research
article en

Abstract

Naive alternative-data backtests routinely overestimate live out-of-sample alpha due to three structural data leaks: (1) aligning corporate filings by fiscal period-end (fp) or calendar filingDate, which introduces lookahead bias whenever disclosures arrive after the 16:00:00 ET market close and silently overwrites historical states upon subsequent restatements; (2) evaluating signals over surviving index constituents rather than historical registrant unions; and (3) passing historical text through commercial large language models whose pre-training cutoffs post-date the evaluation window. We document the deterministic, zero-LLM ingestion architecture of Starling Quantitative (https://starlingquant.com) across 3,607 historical Russell 3000 SEC Central Index Keys (CIKs) from January 2019 through November 2026, explicitly retaining 555 non-surviving and delisted issuers (15.4%) with automatic -30.0% Shumway (1997) performance-delisting haircuts. Every structured XBRL item and primary filing is joined to its exact down-to-the-second SEC EDGAR acceptanceDateTime, mapped to canonical 10-digit CIK and composite point-in-time OpenFIGI identifiers, stamped with One-Document-One-Witness (source_artifact_id) provenance, and quarantined to live forward shadow testing whenever a vendor feed relies on mutable watermarks. Finally, we specify our dual institutional delivery schemas: Feed Tier A (pre-quorum ScoutPitch feature envelopes) and Feed Tier B (post-quorum calibrated MasterDecision consensus). Canonical Web Monograph: https://starlingquant.com/guides/point-in-time-alternative-data-and-survivorship-free-panelsComplete Technical Specification: https://starlingquant.com/llms-full.txt

Zenodo (CERN European Organization for Nuclear Research)
Financial Markets and Investment Strategies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon