How it works
01
Every two weeks we sweep about six thousand US filers. A quick first cut keeps about a thousand worth loading in full, and Buffett's and Graham's documented screens run on each of those. A first read of each survivor's annual report weeds out what the numbers can't see: one-off profits, a business too hard to predict, a problem hiding in the risk factors. The best of the rest get the full treatment below, and become the Desk.
02
Ten to fifteen years of financial statements from the SEC's XBRL data, normalised so a company that changed how it tags its debt still lines up year to year. Then the business, risk and management discussion sections of the latest 10-K, split into passages we can cite.
03
Returns on equity and capital, margins and their stability, owner earnings, maintenance capex, debt payback, the $1 retained test and share-count trends. Then a valuation range under bear, base and bull assumptions with a sensitivity grid. Same code, same answer, every time.
04
Separate passes look at predictability, the moat, management and capital allocation, then a dedicated pass argues against owning the business. Each claim must quote the filing; quotes are checked word for word. The model never writes a number.
05
The Desk is the same for everyone. Your lens decides how you see it: the margin of safety you need, the smallest company worth your time and which names come first all adapt to your capital, horizon, temperament and time. Then we keep watching: prices are refreshed nightly, and new filings and price moves past your margin trigger alerts.
Drag the amount and watch which Buffett you'd be learning from. This is the same function the app runs on your profile.
Read the reasoningIf you're investing
£100k
Partnership Buffett · 1956–1969
Bargains in places big money can't go, bought with a wide margin of safety.
Small capital, a wide hunting ground. Graham-style bargains in small and overlooked companies, special situations, and deep margins of safety. Concentrated when conviction was high.
Illustrative research settings for a profile type. They aren't advice. Thresholds are our reading of Buffett's record.
Hot tips, price targets, buy buttons or a claim that we know what Buffett is thinking. You'll get careful research and a plain account of what it can't tell you.
Start free