A stock screen looks like a search box and behaves like an argument. Every threshold you set is a claim about what a good business looks like, and the list of names that falls out is only as good as the claims that produced it.
The clearest way to see that is to take one screen apart. The default screen Finapolis loads, called Finapolis Recommend, cuts a universe of 5,957 listed companies down to 26. It does that with 10 conditions, and each one is arguable.
What this article covers
- A screen's thresholds are the argument. The list of tickers is just the output.
- The default Finapolis screen uses 10 conditions and leaves 26 names out of 5,957.
- 8 of the 10 test the business. Only 1 touches price, and none of them tests value properly.
- A screen is a filter, not a verdict. Everything it surfaces still has to be analysed.
- What a screen excludes is as important as what it includes, and is much easier to forget.
The funnel, in 2 numbers
Load the Screener and it reports 5,957 records before you touch anything. That is the investable universe it tracks. Apply Finapolis Recommend and the count badge on the preset reads 26.
Cutting 5,957 to 26 is a rejection rate of about 99.6%. A filter that aggressive is not looking for the best companies in any general sense. It is looking for companies that clear every one of 10 specific bars at the same time, which is a much narrower and much more opinionated question. Counts move as the data updates, so treat 26 as a reading rather than a constant.
The 10 conditions, and what each one rules out
Two of the conditions are set as global filters at the top of the page, and 8 are metric cards with an explicit floor or ceiling.
| Condition | Threshold | What it is really asking |
|---|---|---|
| Overall Score | A | Only names the platform already grades top of its 4-letter scale |
| Report available | yes | Only names with a published research report behind them |
| Market cap | $5,000M and up | nomicro or small caps: liquidity and disclosure quality |
| Gross margin | 40% and up | Pricing power, or at least a product that is not a commodity |
| Operating margin | 15% and up | The pricing power survives the cost of running the company |
| ROIC | 15% and up | Capital put into the business earns a real return on the way out |
| ROE | 12% and up | The same test seen from the shareholder side |
| FCF yield | 2% and up | The profits show up as cash, and the price is not absurd against it |
| Debt / equity | 1.0 and below | The balance sheet is not carrying the story |
| Interest coverage | 5x and up | Operating profit covers the interest bill several times over |
Grouped by what they test, the shape of the argument becomes obvious. Gross margin and operating margin ask whether the business can charge more than it costs to serve. ROIC and ROE ask whether reinvested money compounds instead of evaporating. Debt to equity and interest coverage ask whether the company survives a bad year without a financing round. Market cap and the report requirement are not quality tests at all: they are coverage and liquidity constraints.
Note the asymmetry. 8 of the 10 conditions describe the business. Exactly 1, free-cash-flow yield, involves the share price, and at a floor of 2% it is a weak test. This screen is looking for quality and is close to indifferent about what you pay for it.
A screen does not find good investments. It finds companies that match a description, and the description is yours.
What this screen deliberately misses
Every threshold that lets something in keeps something else out, and the exclusions are the part people forget.
- Anything cheap and mediocre. A statistically cheap company on a 6 times earnings multiple with thin margins fails at the first hurdle. Deep-value investors would consider this screen close to useless.
- Entire sectors, by construction. A 40% gross-margin floor removes most retailers, distributors, airlines and homebuilders, because that is not how their economics work. It is not a judgement on those businesses, it is a definition that excludes them.
- Banks and insurers. Debt to equity below 1.0 is structurally impossible for a leveraged balance sheet. Financial companies are filtered out by a rule that was not really written about them.
- Early compounders. A company reinvesting everything into growth can have a negative or tiny free-cash-flow yield and fail the 2% test while doing exactly what you would want it to do.
- Anything the platform has not graded A or written up. Those 2 conditions are about coverage, not about the company, and they will exclude good businesses for administrative reasons.
Why the bands are more useful than the sliders
Each metric card carries a row of named bands above its slider, and the vocabulary is specific to the metric rather than generic. P/E runs Cheap, Fair, Expensive, Premium. ROE runs Weak, Fair, Good, Strong. Gross margin runs Thin, Average, Strong, Elite. Interest coverage runs Distressed, Weak, Adequate, Strong, Very strong. Debt to equity runs Low, Moderate, High, Very high, Extreme.
Those labels do the work that a raw number cannot: they tell you where a threshold sits relative to the market before you have any feel for the metric. Clicking a band respects the metric's direction, so a band on a higher-is-better metric opens the range upward from its floor, and a band on a lower-is-better metric opens it downward from its ceiling.
Each card also prints the share of the universe inside the current range, which is the number to watch while you build. It tells you immediately whether a condition is doing any filtering. A card reading 81% is barely a filter; a card reading 39% is carrying real weight. A histogram behind the slider shows the distribution, and an Outliers toggle switches between a trimmed view and the full range, so a handful of extreme values do not compress the scale into uselessness.
Watch the result count after each one. The condition that collapses the count is the one your screen actually rests on, and it is often not the one you thought was doing the work.
Building your own
The metric picker groups its filters into 6 tabs: Valuation, Profitability, Ratios & Growth, Performance, Score and Themes. Above them sit 7 global filters: Sector, Industry, Country, Overall Score, Report, and Include and Exclude lists for naming specific tickers in or out.
Three things make a screen reusable rather than a one-off. Filter combinations can be saved and come back as their own pill with a live count badge, so you can keep a value screen and a quality screen side by side and see how many names each is finding today. The results table has tabs for Overview, Valuation, Profitability, Ratios & Growth and Performance, and you can add your own tab with the column set you care about. And the result set exports, so a screen can be the first step of work that continues elsewhere.
A screen is also worth revisiting as an argument, not just re-running as a query. If a condition has stopped excluding anything, it is decoration. If one condition is doing 90% of the filtering, the other 9 are decoration too.
What a screen cannot tell you
A screen is a filter on reported numbers. It has no view on whether a moat is widening, whether the last 3 years of margins are repeatable, whether an accounting change flattered the figures, or whether a regulator is about to alter the economics. It cannot tell you why ROIC is 15%, and the why is the entire question.
That is the hand-off. The Screener narrows the universe, the Analyzer grades and values each survivor and shows the evidence behind every metric, and the Reporter writes the argument up with each claim traced to a filing. Our methodology page documents the ingestion and validation chain the numbers come through.
FAQ
How many stocks does the Screener cover?
5,957 records with no filters applied. Separate tabs cover ETFs and Themes, and there is an options screen alongside the equity one.
Why does the default screen return so few names?
Because 10 conditions have to be true at once. Each is individually reasonable, and requiring all of them together is what takes 5,957 to 26. Intersecting filters compounds much faster than most people expect.
Is a high ROIC threshold always a good idea?
It selects for businesses that convert capital into profit efficiently, which is a genuine marker of quality. It also selects for asset-light businesses and against capital-intensive ones, and it says nothing about whether that return is durable or already reflected in the price.
Why is there barely any valuation filter?
Because this particular screen is a quality screen. Free-cash-flow yield at a 2% floor rules out only the most extreme prices. If you want cheapness, that is a condition you add yourself, and the Valuation tab has the multiples to do it with.
Can I screen for a sector the default excludes?
Yes, but not by editing one number. Start from Custom, which clears every filter, and build conditions that suit that sector's economics. A margin floor written for software will not describe a distributor.
How current are the numbers?
Fundamentals update as filings and feeds land, so counts move. The figures in this article were read in June 2026 and are used to show how the screen is built, not as a live snapshot.
Is a screen investment advice?
No. A screen is a filter you configured, returning companies that matched it at that moment. Finapolis is a research platform, not a registered investment advisor or a broker-dealer.
