Data & methodology
Understand the source, date, and limits of a result before using it in your research.
Where the data comes from
Financial statements are derived from public filings in SEC EDGAR, including annual 10-K, 20-F, and 40-F reports. Quarterly coverage primarily uses domestic 10-Q filings; fourth-quarter flows may be derived from annual and nine-month reports when compatible evidence exists. Financial tables expose source filing links where available.
Market prices come from Yahoo Finance through the yfinance data pipeline. Price refreshes are scheduled, not a real-time quote feed. Ownership research uses 13F, Form 4, and Schedule 13D/13G filings, each with different reporting delays and coverage.
Which companies are covered?
Coverage depends on companies and filings ingested into the database. Not every exchange, security type, company, or historical period is supported. A missing value means evidence is unavailable or unsuitable for that calculation; it should not be interpreted as zero. Some ticker aliases resolve to the company’s primary stored share class.
How are framework scores calculated?
The app displays 17 scored frameworks. Each applies a defined set of tests to the available evidence. The individual framework pages identify published rules, app-defined adaptations, qualitative proxies, source material, and verdict logic. An investor’s name describes the inspiration for a screen, not an endorsement or a prediction of their trades.
BUY, PASS, NEUTRAL, and FAIL have framework-specific meanings. PASS may mean a company clears a checklist without satisfying a purchase valuation. Ranked strategies depend on the comparison universe. Consensus summarizes those screens; it is not a probability of profit.
Annual statements, quarterly data, and TTM
Current scores use trailing 12-month flows and the latest balance sheet for supported criteria, with annual history or annual fallbacks where appropriate. TTM requires four consecutive compatible quarters. Quarterly EPS is not simply added to create TTM EPS. Intrinsic-value models, custom frameworks, raw financial screener filters, and historical backtests retain annual inputs.
A recent score-computation date does not mean every input is from that date. Check the financial basis, latest source filing, and market-data date shown beside the results. Different frameworks can legitimately use different periods.
Refresh timing and historical limits
Scheduled jobs refresh market data, process new filings, and publish score updates. Updates can lag or fail; the app shows freshness information where available. Historical coverage varies by issuer and period. Flagged share-basis mismatches can affect per-share ratios and valuations.
Backtests depend on the available universe, timing conventions, assumptions, and historical data quality. Treat them as research simulations rather than achievable returns or forecasts. Review the limitations displayed with each result.
Are the scores AI-generated?
The quantitative screens use defined calculations and criteria. Some frameworks include separately labeled qualitative assessments. Optional AI narratives and debates are distinct generated interpretations; they can make mistakes and are not statements by the named investors.
AI generation is currently a private preview for approved accounts. Signing in does not automatically enable it. Guests can still inspect stock scores, underlying criteria, financial statements, and the public framework guides.
Report a data concern
Contact Studio Pinball with the ticker, metric, period, and source filing you are comparing. This helps distinguish a data issue from a difference in definitions or dates.