TL;DR: Use a stock screener in layers. Start with a cheapness filter such as earnings yield or a price-to-earnings ceiling, add a quality filter such as return on capital or return on equity so cheap-and-broken stocks fall out, require a balance sheet that can survive a bad year, and treat whatever survives as a research list, never a buy list. The screen's job is to save you from reading ten thousand financial statements. It cannot read the thirty that remain for you.
Key Takeaways
- Cheapness alone finds value traps. Pair every valuation filter with a quality filter.
- Filter in a deliberate order: valuation, quality, balance sheet, then size and liquidity practicalities.
- Loosen one filter at a time when the results are empty. A screen with no output is information about your thresholds, not about the market.
- Every survivor still needs the full research pass. The screener ranks numbers; it does not read footnotes.
What is a stock screener actually doing?
A screener is a database query wearing a friendly interface. Thousands of listed companies each carry the same standard fields: price ratios, profitability ratios, growth rates, debt figures. You set conditions on those fields, and the screener returns every company that satisfies all of them at once. Nothing about the process involves judgment, which is both its power and its limit. It will happily return a superb business and a melting one side by side if both happen to share the same ratios this quarter.
That is why the published mechanical strategies are built the way they are. The best known example, covered in our explainer on Magic Formula investing, refuses to screen on cheapness alone: it ranks on earnings yield and return on capital together, so a stock must be cheap and good before it scores well. Steal that structure even if you never follow the formula itself.
Step 1: Set the cheapness filter
Pick one primary measure and understand what it leaves out:
- Price-to-earnings (P/E) is the familiar one, and the most easily distorted by debt and one-off items.
- Earnings yield (EBIT divided by enterprise value) compares operating earnings with the full cost of owning the business, debt included, which is why mechanical value strategies prefer it.
- Price-to-book suits asset-heavy businesses and misleads on companies whose value sits in intangibles.
Set the threshold generously at first. The point of the first pass is a list in the dozens, not a list of three, because later filters and later research will do the cutting.
Step 2: Add the quality filter
Now require the cheap stocks to be decent businesses. Return on capital or return on equity sustained over several years, not one good quarter, is the standard test. Add a margin floor if your screener supports it. This single layer removes most value traps: companies that screen cheap because earnings are collapsing tend to show poor or deteriorating returns on capital before the price fully adjusts, and companies that are cheap because the market is ignoring them tend to show perfectly respectable ones.
Step 3: Require a survivable balance sheet
Value investing means buying companies other people are worried about, which makes the balance sheet the difference between patience and peril. Filter for debt the business can carry through a bad year: a debt-to-earnings ceiling, or an interest coverage floor, using whatever your screener offers. A cheap stock with a fragile balance sheet is a countdown. A cheap stock with cash and modest debt can wait for the market to change its mind, and so can you.
Step 4: Research the survivors properly
The screen ends here. What remains is a shortlist, and the shortlist goes through the same process as any other candidate: the business in plain sentences, five years of financials, what the price assumes, and the risks written as checkable statements. Our stock research checklist runs that process in a fixed order. A candidate you intend to hold with real money can also be rehearsed first, which is the role of paper trading: the screen graded the numbers, the checklist grades the company, and the rehearsal grades you.
The mistakes that make screens lie
- One-quarter inputs. Ratios built on a single unusual quarter, a one-off gain or a write-down, screen companies in and out for reasons that vanish next quarter. Prefer trailing-twelve-month figures and check anything surprising against the filings.
- Survivorship in reverse. A screen shows today's listings. Companies that failed are absent, which makes any historical test of your filters look better than it was.
- Threshold worship. A stock at a P/E of 10.1 is not meaningfully different from one at 9.9. If a company you like fails by a hair, the correct response is to look at it anyway, not to obey the decimal.
FAQ
How many filters should a value screen have? Three to five. Fewer and the list is unmanageable noise. More and you have described one specific company, usually one that no longer exists at that price. Cheapness, quality, balance sheet, plus at most a size or sector constraint, is the working range.
How often should I re-run the screen? Quarterly suits most investors, because the underlying financials only change quarterly. Running it weekly mostly re-sorts the same data by price noise and invites trades the research has not earned.
Does a low P/E stock count as a value stock? Not by itself. Low P/E is one cheapness reading. A value stock, in the sense this guide uses, pairs a low price relative to earnings or assets with a business good enough and a balance sheet strong enough for that price to be a misjudgment rather than a warning.
Sources
- Greenblatt, Joel, The Little Book That Beats the Market (2005, updated 2010): the published two-ratio screening method this guide's filter order is modeled on.
- Investor.gov (U.S. Securities and Exchange Commission), investor education material on researching investments and reading financial statements, accessed October 2026.
Educational content, not financial advice. Screening is a research starting point. No filter setting in this guide is a recommendation to buy or sell any security.
