Normalize two reporting periods, then compute every movement — absolute, percentage, and percentage-point.
Compare two reporting periods, rank what actually moved, and put your thesis through a test built to disprove it.
Free cash flow diverged from revenue
Free cash flow changed -56.0% while revenue changed +17.9%. The 73.9-point divergence merits follow-up.
From reported numbers to a tested conclusion.
Rank each change by financial magnitude and by how much it matters for this particular company.
State a thesis. The system plans the checks, runs them, and hunts for evidence that would prove you wrong.
Open any finding to see the arithmetic behind it and where interpretation begins.
Most research tools agree with you.
Ask one to evaluate a thesis and it will find support for it, because that is what retrieval optimizes for. Earnings Delta plans checks that could falsify your claim, shows challenging evidence at the same weight as supporting evidence, and returns Challenged when the numbers say so.
Materiality and confidence are different questions.
Materiality asks how big a movement is and whether it matters for this company. Confidence asks how well the claim is supported. Collapsing them into one score destroys the most useful finding of all — a large movement whose cause is not yet established.
- Filed
- Read straight from a fact filed with the SEC.
- Derived
- Computed from filed values — a ratio, or a quarter reconstructed by subtraction.
- Interpretation
- A real pattern whose cause the filings do not establish.
Every number traces to a filing.
Every figure on screen is read from XBRL data filed with the SEC. Open any finding and you get the arithmetic, the us-gaap concept it was tagged under, and a link to the filing on sec.gov.
Filers make that harder than it sounds. A fourth quarter is never reported on its own, cash-flow statements run year to date, and companies disagree on which concept a number belongs under. Each figure records how it was obtained, and anything reconstructed by subtraction says so rather than passing as reported.