Evidence first. Assumptions made visible.

Data sources and reporting periods

Cerebro’s analysis pipeline uses EOD Historical Data (EODHD) for company fundamentals and market inputs, with supplementary filing evidence where supported. Source availability and field definitions vary across companies.

A financial statement’s period end is not the date it was fetched. A quote timestamp is not the financial reporting period. An analysis snapshot records when the analysis was produced. Refreshing a calculation does not create a newer company filing.

Checking the basis of a valuation

Per-share estimates require a consistent currency, financial period and share basis. Multiple share classes, depositary receipts, splits and unusual listing structures can complicate the conversion from company value to value per traded share.

The calculation checks available evidence and may exclude a method when essential inputs cannot be supported. Missing or unreliable data should not be interpreted as a zero valuation. Coverage is not guaranteed for every stock.

Choosing and combining methods

The available methods include discounted cash flow, earnings-based and relative-valuation approaches. Specialist business types may require different inputs and models. The applicable methods depend on the business and the evidence available.

Where a headline estimate combines applicable models, the weights influence the result. The models may share inputs, so agreement is not independent validation. Wide disagreement makes a blended estimate less useful as a single conclusion.

Forecasts, comparisons and analyst targets

An intrinsic valuation projects future business cash flows and discounts them to today. Revenue growth, margins, reinvestment and the discount rate affect the output. Scenario ranges reflect assumptions; they are not statistical confidence intervals.

A forward market comparison uses eligible forecast earnings and peer multiples when sufficient evidence is available. Peer selection and accounting differences matter, and the market can price a whole peer group richly or cheaply. Analyst price targets remain a separate reference.

Scanner plans and recorded outcomes

Scanner ideas describe a possible trade plan and its entry conditions. The tracker records outcomes using its status rules and available price data. The history includes unsuccessful ideas as well as successful ones.

Those records are not audited brokerage returns. Execution timing, spreads, fees, slippage and price gaps can make real outcomes different. A proposed stop is not a guarantee of a particular exit price.

Ownership reports and refreshes

Reported holdings and insider transactions reflect disclosures. Publication can lag the underlying transaction or reporting period. A change in reported holdings does not reveal every trade made between filings.

Refresh processes update stored research on a schedule and as data becomes available. Check the displayed dates and original filings for decisions that depend on recency. A scheduled refresh does not mean every input updates in real time.

How to challenge a result

Start by checking the ticker and listing, currency, share basis, financial period and analysis date. Then compare the underlying figures with the company’s filings. Consider whether the model fits the type of business.

If a result looks inconsistent, bring the ticker, analysis date and the relevant source to the member community. Treat a large difference between a model and the market as a question to investigate rather than proof that either must be correct.

Source references

Check original disclosures and provider definitions when interpreting a figure.

• EODHD data documentation
• SEC EDGAR company filings