Algorithmic Score Disclosure

1. Purpose

FilingIQ is committed to transparency about how we generate scores, ratings, and indicators. This disclosure explains the types of algorithmic assessments we produce, our general methodology, what these scores represent, and their inherent limitations. We believe users have the right to understand the tools they use for research.

2. Types of Scores and Ratings

FilingIQ produces the following categories of algorithmic assessments:

2.1 Insider Scores

Insider Scores are composite assessments of individual Form 4 transactions filed with the SEC. Each score is derived from multiple analytical dimensions, including:

• Insider track record: How the filing insider's previous purchases performed afterwards, measured as the share of them followed by a positive return and the consistency of that excess return.
• Conviction: The dollar value of the trade, the share of the insider's holdings it represents, and its size compared with the insider's own typical trade.
• Role and information access: The insider's position at the company, from chief executive to director or large shareholder, as a proxy for how much they know about its business.
• Timing: How close the trade is to the company's earnings release, and how far the stock had fallen from its recent high when the insider bought.
• Confirmation: Whether other insiders, in the same company or sector, are buying or selling at the same time.

2.2 Alpha Scores

Alpha Scores provide an algorithmic assessment of trade quality by combining multiple scoring dimensions into a single composite value on a 0-100 scale. Alpha Scores are designed to help users quickly identify potentially noteworthy transactions by synthesizing multiple factors into a single, comparable metric.

2.3 Transaction Indicators

Transaction Indicators are contextual flags that highlight notable characteristics of individual trades. These are not scores but binary or categorical markers that call attention to specific trade attributes, such as:

• First purchase by an insider at a company
• Significant position changes (large percentage increases or decreases)
• Cluster buying or selling (multiple insiders transacting in a short period)
• Transactions by officers versus directors
• Open market purchases versus other transaction types

2.4 Transactions We Do Not Score

Some reported transactions say little about what an insider expects from the company, or their reported price cannot be compared with the stock's price. We show them, but give them no score and do not count them as signals:

• Derivative transactions, such as option grants and exercises.
• Trades under a pre-arranged trading plan (Rule 10b5-1).
• Amendments to an earlier filing; the original filing is scored.
• Transactions in preferred shares.
• Sales made only to cover the tax on vesting shares (sell-to-cover), and purchases through an employee stock purchase plan (ESPP).
• Transactions whose reported price is not a price of the listed ticker, for example ordinary shares of a foreign company traded abroad or prices in another currency. They are not scored, but the insider still counts as a buyer or seller.

Purchases that are part of an initial public offering or a financing round are scored but not highlighted as signals, because their price is set by the offering, not by the insider.

3. Methodology Overview

Our scoring methodology follows these general principles:

• Data source: Scores are derived from statistical analysis of publicly available filings, primarily Form 4 (insider transactions) and Form 13F (institutional holdings) from SEC EDGAR and, for congressional trades, the financial disclosures of Members of the U.S. House and Senate, combined with market price data from third-party providers (for example the price before and after a trade).
• Multi-dimensional analysis: Multiple analytical dimensions are weighted and combined algorithmically to produce composite scores.
• Model versioning: Every score carries the model version and the timestamp it was computed. Weight definitions are held in source control.
• Proprietary formulas: The exact weightings, formulas, and algorithmic details are proprietary and constitute the intellectual property of Sprint Zero UG (haftungsbeschränkt). We disclose the general approach and dimensions but not the specific implementation.

4. What Scores Are

FilingIQ scores and ratings are:

• Statistical interpretations of filing data patterns based on historical analysis.
• Relative rankings that compare individual trades or patterns against historical distributions of similar transactions.
• Research tools designed to help users filter, sort, and prioritize SEC filing data more efficiently.
• Quantitative summaries that distill multiple data points into accessible metrics.

5. What Scores Are NOT

FilingIQ scores and ratings are expressly not any of the following:

• Investment recommendations or financial advice. No score, rating, or indicator produced by FilingIQ should be interpreted as a recommendation to buy, sell, or hold any security.
• Predictions of future price movements. Scores reflect historical patterns and statistical analysis, not forecasts of future market behavior.
• Guarantees of insider intent or conviction. While our models analyze patterns that may suggest conviction, we cannot determine an insider's actual intent or motivation.
• Substitutes for professional financial analysis. Our scores are research aids, not replacements for the judgment of qualified financial professionals.

6. Limitations

Users should be aware of the following limitations of our algorithmic scoring:

• Data completeness: Scores are based on publicly filed data, which may be incomplete, contain errors, or be subject to amendments. Not all relevant information may be captured in SEC filings.
• Historical bias: Historical patterns do not predict future outcomes. Market conditions, regulatory environments, and corporate governance practices evolve over time.
• Contextual gaps: Scores may not capture all relevant context surrounding a transaction, such as private information known to insiders, corporate events, or industry-specific factors.
• Model updates: When scoring models are updated or recalibrated, historical scores may be recalculated, potentially changing previously assigned values.
• Sector and size bias: Model performance may vary across different market capitalization ranges, industry sectors, and market conditions.

7. Methodology Evolution

FilingIQ reserves the right to update, modify, or replace its scoring models at any time. We are committed to continuous improvement of our analytical tools. When significant methodology changes occur:

• We will communicate major changes through the Service and, where appropriate, via email to registered users.
• We may provide transition periods during which both old and new scores are available.
• Historical scores may be recalculated under updated methodologies for consistency.

8. Questions

If you have questions about our scoring methodology, how scores are calculated, or how to interpret the data presented on FilingIQ, please contact us at support@filingiq.io. We welcome feedback and are committed to providing clarity about our analytical approach.

---

For general legal inquiries, please contact legal@filingiq.io.