This page answers the questions a skeptical person asks before paying for anything. I have tried to answer them the way I would answer them out loud, and to be equally clear about the things I will not tell you and why.
It reads the published accounts of about 26,500 companies and turns each one into six colored squares. You set what counts as a pass. You look at the list. That is the whole product.
Anyone who owns a few shares, or wants to, and would like to know if the company behind them is sound and fairly priced. No finance background needed. Six squares, six colors, and you decide what passes.
Someone who wants to be told what to buy. It never will.
Gap (cash) and Gap (growth): the distance between what a business is worth and what it costs, measured two different ways. Debt: how much it owes against what it owns. Profit: how well it turns capital into profit. Cash and Sales: whether each is growing, and the shape of the trend.
Green passes, orange is borderline, red fails. Against your thresholds, not mine.
Because the company did not publish the number that square needs. Some tools fill that hole with an estimate. I leave it gray. A guess with a color on it is still a guess, and it is more dangerous than a hole, because it looks like an answer.
You do. Six sliders, two handles each, and the count updates as you drag. I calculate and display. The thresholds are yours and so is the decision.
The exact calculations are private. Every choice behind them is not, and this is the part worth judging me on.
The median of the year-over-year changes, the middle value, not the rate connecting the first year to the last.
Take a real free cash flow series: 32.9 → 32.2 → 7.7. Measured first to last that is −21.6% a year, and a model fed that number will project the company shrinking for a decade. It was Amazon, and the cash fell because Amazon spent it on warehouses. The median of the same three numbers is +19.7%. Same data, different question.
One freak year drags an average. A company that grew 12%, 14%, 11% and then 400% because it bought a competitor averages out to a number that never happened. The median does not care how big the freak year is, only where it sits in the order. That is what lets it run across thirty thousand companies with nobody checking them one at a time.
Cash. Specifically the cash left over after paying for everything the business needed. Earnings can be shaped by accounting choices; cash is harder to dress up.
The ten-year median of what the market has paid for a dollar of that company’s cash. Never today’s number. Today’s contains today’s price, so using it means the answer is built out of the question, an expensive stock scores as cheap and a beaten-down one scores as expensive.
The government bond yield of the currency the company reports its accounts in. Not the American one. In 2021 the Japanese ten-year paid about 0.05% and the American one 1.52%, and applied once for every year of a forecast that gap decides whether Toyota looks cheap. I use 23 government curves for this, and only yields published within the last 18 months. A frozen rate from three years ago gives a confident, precise, wrong answer, and nothing about the screen tells you it is old.
Yes. Capped at 25% a year and floored at −20%. No company grows 60% a year for a decade, and a model that assumes it will produces valuations that are arithmetic rather than judgment.
Both, and they are separate tests. Three consecutive years of falling cash fails even when the average still looks healthy. The median tells you what a normal year looks like; the shape tells you whether normal still applies. A model that only runs the first test keeps buying companies on the way down.
No. Not the equations, not the weights, not the constants. That is the product, and I would rather say so plainly than pretend the reason is something else.
I will not say, here or anywhere else. Publishing a list of what passes would make this a recommendation service, which it is not and will not be. The list lives in the app, against your thresholds.
No. I am not a broker and not a registered adviser. Bufetico does not tell you what to buy, what to sell, when, or how anything will perform. There is no buy button and there never will be. Investing involves risk.
Three reasons. The data license does not allow the provider’s numbers to leave the app. A card showing a percentage would make me the author of a performance claim, which is regulated in the United States precisely because it works. And nobody shares the week they were down 13%, so the public record would be a rigged shop window.
Official closing prices, updated after each market close. No live prices, no news, no analyst ratings, no price targets. Alerts go out after the close, never during it.
About 26,500 analyzed across 30 exchanges. Around 30,800 are findable by search, because the search box also reaches things the list excludes on purpose. Searching and analyzing are two different jobs.
Because their currency has no government bond curve to work with, and there is no honest substitute. That is 105 companies across 17 currencies, and their valuation squares stay empty. I tried setting a rate by hand for the Argentine peso: at 33%, which is roughly what money costs there, one company came out with a gap of over 2,400%. That is not a judgment about the business, it is arithmetic. An honest hole is worth more than a number wearing the authority of the screen.
Two things worth naming. 4,041 companies that had stopped trading were still being priced, and they sat at the top of the list as 94% bargains. And 505 preferred shares and listed bonds were inheriting their parent company’s accounts, which made them look like 97% bargains. Both are excluded now. A list sorted by how attractive something looks concentrates its own defects at the top, which is exactly where you are most likely to be reading.
You get one email. It names the company and which square changed color, and that is the whole message. No urgency, no link to a broker.
No. Broker connections are read-only. Bufetico cannot trade, cannot move money, and never stores your password. You enter that in your broker’s own window. You can also just type in what you own.
$19 a month. There is no free trial.
The exact calculations stay private. That is deliberate: it is the product. What is public is every choice behind them, so you can judge whether the reasoning is sound before you pay anything.
If you have a question this page does not answer, write to me and I will add it here.
Bufetico
