Methodology

Stratum's live signals are deterministic calculations from published academic models, computed from regulator-published XBRL facts. No model weights, no black boxes — every result exposes its inputs, and every input traces to a filed document on EDGAR. A signal is a reason to read the filings, not a conclusion about the company.

P-01 · Earnings quality

Beneish M-Score (Beneish, Financial Analysts Journal, 1999): an 8-ratio probit model built to tell apart firms later found to have manipulated earnings; M > −1.78 is that study's classification cut-off. It is a statistical resemblance, not a finding about any one company. Supplemented by the Sloan accrual ratio (Sloan, The Accounting Review, 1996): high accruals relative to assets were associated with later earnings reversals.

P-12 · Restatement indicators

Dechow–Ge–Larson–Sloan F-Score (Contemporary Accounting Research, 2011), Model 1: a logistic model fitted on firms subject to SEC enforcement releases (AAERs). F is the ratio of predicted to unconditional misstatement probability; F > 1 is above the baseline rate and > 1.85 is the study's “substantial” band. RSST accruals are approximated where full components are not tagged; the evidence block records the construction used.

P-13 · Interest coverage

BIS zombie definition (Banerjee & Hofmann, BIS Quarterly Review, 2018): interest coverage below 1.0× for three consecutive fiscal years. Context from Altman Z″ (Altman, 1995 four-variable, book-equity variant) — no market data required.

P-07 · Pension funding

Defined-benefit funded ratio from tagged plan assets and benefit obligations; severity scales the deficit against shareholder equity.

P-05 · Effective tax rate

Persistent effective-tax-rate gap versus the statutory rate across three years. A low ETR is frequently legitimate; the signal triages persistence and magnitude for review of the tax footnote.

P-09 · Language change (on demand)

Loughran–McDonald dictionary densities (Journal of Finance, 2011) — uncertainty, negative, and litigious language per 1,000 words in the latest annual report versus the prior one. Note: the full LM dictionary requires a commercial licence for commercial use; Stratum ships a research subset and loads the licensed CSV if provided.

P-08 · Insider filing timing (on demand)

Clusters of Form 4 filings in the 14 days preceding 8-K events over the trailing year. A timing pattern is a reason to read the filings — not an accusation.

P-02 · Subsidiary jurisdictions (on demand)

Subsidiary jurisdictions parsed from Exhibit 21 of the latest 10-K, screened against a configurable watchlist. Mention-count proxy; ownership and revenue attribution belong to the LLM-enrichment roadmap.

The AI layer (P-03, P-10, P-11)

Three signals use a language model (DeepSeek V4 by default; any OpenAI-compatible endpoint) for what deterministic math cannot do: reading claims, strategic language, and asset descriptions out of prose. The AI layer operates under three hard rules:

1. Quote or it didn't happen. Every extracted claim must include a verbatim passage; Stratum verifies the passage exists in the filing text and silently discards anything unverifiable. The discard count is reported in each signal's evidence. 2. Metered. Every call is logged to a cost ledger (/api/costs: spend, tokens, cache-hit rate, cost per company) and stops at a configurable daily budget. 3. Deterministic before probabilistic. Where a signal can be split (P-10's 13D counts, P-11's PP&E intensity), the deterministic part runs first and stands alone if no LLM key is configured.

The agentic layer

Bounded, budget-aware — not open-ended autonomy. Triage ranks the universe for $0 (pure arithmetic, no model call); investigation runs a hard-coded 4-step sequence capped at 5 steps / 2 LLM calls, enforced by a counter in code, not a prompt; the watch agent attaches a memo to an alert that has fired only when budget allows, never blocking delivery. The full design rationale lives in the repository (stratum/agents/README.md).

Composite

60% weight on the strongest signal, 40% on the mean of available signals; the level is the maximum across signals, and coverage is always reported. One signal at its upper threshold is enough to surface a company in the daily brief — by design.

Honest limits

Tag coverage varies by filer; sub-ratios that cannot be computed default to their neutral value and are listed in the evidence. These are screening models with meaningful false-positive rates — Stratum ranks where to look first; the filings remain the ground truth. Nothing here is investment advice.