Normalize schemas while preserving the difference between missing values and real zeros.
SZL HOLDINGS / PURIQ FINANCE
DOMAIN EXPERIENCE V4 · SOURCE BOUND
FINANCIAL INTELLIGENCE
PURIQ Finance
Provenance-first financial signal and decision console. One governed runtime underneath; a domain-native workbench above it. Observed evidence, modeled analysis, human authority, and audit receipts remain separate states.
INGESTPRICESTRESSDECIDEAUDIT
Signed price observatory
Open observatoryInspect public BTC, ETH and SOL prints. Provider signatures authenticate assertions; they do not establish market accuracy or independent sources. Keys come from the provider over HTTPS.
NOT CHECKED
No accepted price
Full-record signature, freshness and source policy are required.
Review only. Local evidence digests are unsigned. No orders or capital movement.
Reference-price research
MODELED / SYNTHETIC. Compare equal-venue and declared-group medians under outliers and duplicated sources. Group independence is an assumption; this is not an accuracy or performance benchmark.
NOT RUN
| Scenario | Venue median | Group median |
|---|
Verification evidence and source revision
No observation.
PURIQ / RESEARCH AUDIT
What survives the data checks?
Audit an annual panel before using it in a financial thesis. Inspect exclusions, release timing, coverage, and supplied robustness estimates. No model is fitted here.
SYNTHETIC EXAMPLE / NOT AUDITED
The example uses fictional entities and estimates. Suppression is never converted into an observed zero. Unknown release dates are excluded from the as-of sample.
Run the audit to inspect the sample.
The result is a stateless calculation over caller-supplied records. Export includes your inputs and an unsigned integrity receipt. It does not establish source authenticity or causal validity.
Sample attrition
No audit yet.
Robustness comparison
Reported estimates only. A sign change or fewer than 30 declared clusters prompts review; the threshold is a screening heuristic, not a validity test.
No specifications evaluated.
Entity-year coverage
Supply expected_entities, start_year, and end_year to define the intended sample. A sample inferred only from submitted records can hide absent entities.
Full audit and integrity receipt
No audit yet.
Research references
Selected public primary sources reviewed on 7 October 2026. These inform the design; the authors and projects do not endorse this product.
Use the date information became available to guard against look-ahead bias.
Make data preparation, evaluation, and data health checks explicit parts of research.
Historical vintages describe what was known at the time. This audit checks supplied release dates; it does not fetch vintages.
Suppressed employment and wage fields can be zero-filled. Interpret disclosure flags before classifying a value.
Retain period-specific documentation when comparing tax-return migration across years.
Keep filing dates, units, and reporting periods with financial observations.
Stanford economist whose work connects economics and machine learning. Reading focus: identification and research design.
Duke financial econometrician. Reading focus: forecast evaluation and volatility.
Applied econometrics: causality, policy evaluation, sensitivity, and robustness.
PURIQ / SOURCE WORKBENCH
Inspect the evidence before the thesis.
NOT PROBED
Configuration is not connectivity. Status describes the last request for an adapter, not every market or instrument. Source revision: UNAVAILABLE.
Inspect the 16-source access register
| Source | Last-request state | Access |
|---|
One bounded observation
Public observations only. No account, wallet, order, credential entry, model inference, or trading authorization. Polymarket and Kalshi contract equivalence is not established. Filings retain period and filing-date context.
Source instrument
UNAVAILABLE
No chart observation.
Returned records
Source envelope and provenance
Digests establish internal consistency, not signatures, investment performance, or source authenticity. Display is limited to 250,000 characters; export preserves the accepted public response.
No accepted observation.
PURIQ / ADVISORY ANALYTICS V2
From source evidence to measured inputs.
Daily-close indicators are modeled analysis, not investment advice or an execution quote. Paper use only. Live history comes from the canonical Coinbase adapter; the fixture is synthetic. Missing daily intervals block calculation.
NOT COMPUTED
No inputs accepted.
Receipts are unsigned and caller-held. Integrity does not prove authenticity, profitability, or trading authority. No shared portfolio state or ledger is written.
Bounded portfolio statistics
Enter 3–300 positive closes per asset, at most 16 assets. These are unverified caller inputs. The output contains per-asset statistics and simple means, not a weighted portfolio or backtest.
Full source binding and computation evidence
No evidence yet.
Loading governed evidence…
Canonical product source: https://github.com/szl-holdings/a11oy/tree/main/verticals/puriq-markets · Portfolio: a-11-oy.com/products/