Legal

Data, Rankings & AI Methodology Disclosure

Effective date: August 31, 2026

Swoopr Investment surfaces and ranks market information. This page explains, in plain English, where that data comes from, generally how rankings and scores are built, and what their outputs do and don't represent.

1. Where the Data Comes From

2. How Rankings and Scores Work

Tools like the market-mover, IPO, earnings, and breakout screens rank or score assets by comparing them against a defined, disclosed set of criteria, for example, price momentum, relative volume, valuation figures, or event proximity. In general:

3. AI-Generated Content

4. How Swoopr Separates Data, Calculation, Model Output, and Generated Explanation

A financial interface can combine several different types of information. They should not receive the same level of trust simply because they appear in the same card. Swoopr separates them into four layers.

Layer 1: Source Data

Source data is information obtained from an identified external or internal record: market data, company filings, fund documents, regulator publications, economic series, or other documented sources.

Where the distinction matters, Swoopr exposes the provider or authority; the field or series; the data-as-of time; whether the feed is real-time, delayed, end-of-day, estimated, or otherwise limited; material adjustment rules; and known coverage limitations.

A label such as "live" describes interface behavior only when the underlying timeliness is also stated accurately. Swoopr does not imply exchange-real-time data when a feed is delayed.

Layer 2: Deterministic Calculation

A deterministic calculation applies a documented formula to known inputs. Examples include percentage change, position-size math, fee compounding, moving averages, valuation ratios, and portfolio weights.

When a calculation is reused across the site, the implementation relies on one tested source of truth rather than duplicating the formula in multiple components.

A deterministic formula can be correct even when an input is stale or inappropriate. Calculation validation and data validation are therefore separate controls.

Layer 3: Model or Score

A score, classification, projection, backtest, simulation, or ranking is a model output. It can be useful without being a fact.

Every material model documents: inputs; formula or algorithm; weighting; range; assumptions; missing-data behavior; limitations; version; and what the output is not intended to mean.

Swoopr avoids labels that can reasonably be read as personalized buy, sell, or invest recommendations unless that product behavior has been explicitly designed, reviewed, and authorized for that purpose.

Layer 4: Generated Explanation

A generated explanation turns structured information into natural-language context. Generated text does not become the source of a new factual number or regulatory rule.

For higher-risk financial surfaces, Swoopr's target control is:

  1. Structured inputs are collected from approved sources.
  2. Deterministic numeric values are validated before generation.
  3. The model receives those values and approved context.
  4. The model is instructed not to invent missing numbers, current prices, legal requirements, tax rules, or recommendation claims.
  5. Generated output is checked for unsupported numeric substitutions and prohibited claims.
  6. Source links and data timestamps remain visible alongside the explanation.
  7. High-risk outputs can be routed to additional review or disabled when evidence is insufficient.

If a generated statement cannot be traced to approved context, it is removed, qualified, or withheld rather than presented with false precision.

5. Missing Data Policy

Swoopr distinguishes between: zero; not available; not meaningful; not supported by the current source; and not calculated. A missing value does not silently become zero.

For example, a valuation ratio may be not meaningful when its denominator is negative. A missing filing fact may be unavailable for a period. An unsupported provider field may be absent. Those conditions mean different things and are represented separately.

6. Current-Value Claims

Prices, yields, rates, contribution limits, regulatory thresholds, and similar values can become stale. Every reusable current-value component has a defined freshness rule and, where appropriate: a source authority; a data-as-of timestamp; a review and refresh trigger; and fallback behavior when the source is unavailable.

A page build date is not a substitute for a data-as-of date.

7. Review Tiers

Swoopr applies stricter review controls when content or tooling can materially affect interpretation of: taxes; retirement contribution or withdrawal rules; securities regulation; margin and leverage; options loss mechanics; crypto regulation; market-data timeliness; recommendation-like scores; and current regulatory thresholds.

Lower-risk educational definitions can use a longer review cadence, but they still require accurate sourcing.

8. Correction Standard

When Swoopr finds an error, the response matches the error:

9. Simulated and Historical Results

Any simulated, backtested, or historical figures shown on the Service do not reproduce actual order fills, real-time liquidity, real bid/ask spreads, or real market impact. Actual trading results will differ from simulated ones. See our Investment & Trading Risk Disclosure for more on backtest and model limitations.

10. Related Pages

Read this alongside the Financial Disclaimer, the Investment & Trading Risk Disclosure, the Conflicts of Interest Disclosure, and the Swoopr Source Verification Ladder.

11. Contact

Questions about how a specific score or ranking is built can be sent through our Contact Us page.