The method, in full
This page is the whole method written out, so you can judge it before you trust it. It is a long-term, value-investing framework: understand the business, insist on a durable advantage, check that management compounds capital, and only then argue about price.
Why four questions and not a score
A single score hides its own assumptions. Four questions keep the disagreement visible: a company can have a wonderful business and an impossible price, or a fair price and a business nobody can explain. Both are a no, for different reasons — and you should be able to see which.
The four questions
- Business Model Can you explain how it makes money without hand-waving?
- Moat Has the advantage held for ten years, or just one good one?
- Management Does capital go in and come back out larger?
- Margin of Safety Is there room in the price to be wrong?
The three valuation models
Each model answers the same question — what is this business worth today? — from a different direction. We run all three and show all three, including when they disagree. When they do, the disagreement is the finding.
- Cap Rate What the owner earnings are worth at a 10% capitalisation rate, counting maintenance capex only.
- Margin of Safety Future earnings growth discounted back to today — then cut in half.
- Payback Time The price at which free cash flow pays the whole purchase back inside eight years.
How the reference entry price is set
Of the three model prices, we take the highest as the reference entry price, and compare it with the current price. At or below it, the stock is below the reference entry price; within 15% above, it is close; further above, it is not. The word "reference" is doing real work in that sentence: it is a yardstick for your own research, not a recommendation, and a stock can sit below it and still be a company you should never own.
The four thresholds
- ROIC ≥ 10% Below that, an index fund is the better home for the money.
- Debt ≤ 3 years of earnings Debt a downturn can absorb.
- Four growth rates ≥ 10% Revenue, EPS, free cash flow, book value.
- Price ≤ reference entry price The discount has to already be there.
Where the numbers come from
Fundamentals come from companies' own filings with the US Securities and Exchange Commission — annual reports (10-K) and quarterly reports (10-Q), via SEC EDGAR — together with daily market prices from established market-data providers. Nothing in the app is a private estimate you cannot check: every figure on a report links back to the filing it came from. Coverage today is US-listed stocks.
Where the AI sits — and where it does not
The AI reads the filings and writes the structure: a summary per question, the evidence it leaned on, and what it is worried about. It does not compute the valuations — those are arithmetic on filing data, run the same way every time — and it does not decide anything. Treat its text as a well-organised first draft by a diligent assistant who can still be wrong, and check it against the filing underneath.
What this method will not do
- It will not tell you what a stock does next. No price targets, no timing, no predictions.
- It will not replace reading the filing. It gets you to the right pages much faster.
- It will not catch a fraud, a regulatory shock, or a business that is about to be disrupted. Clean numbers are a starting point, not a guarantee.
- It cannot know your situation — your horizon, your tax position, how much of this you can afford to be wrong about.
The honest summary
Three models nodding and four thresholds clearing means one thing: this company has earned a few hours of your attention. Everything after that — the reading, the judgement, the decision, the consequences — is yours.
See it on a real company →