Tools & Data · Intermediate · 4 min
Using screeners
You can build a screener filter from clear criteria, read the results as a list to study rather than a list to buy, and recognize survivorship bias.
From thousands to a few
Thousands of shares are listed on the major exchanges, and thousands more funds, currencies and futures trade every day. Nobody can study them all.
A screener does the first sorting for you. In seconds, it reduces a market to the handful of names that match your criteria. Used well, it turns an impossible task into a manageable reading list. Used badly, it becomes a machine for finding confirmation.
Filter by filter
A screener is a tool that filters a large set of assets by criteria such as market value, valuation ratios, dividend yield, performance or technical conditions, and lists those that pass.
A fictional screen, built from criteria of the fundamental analysis rubric:
- start: 5,000 listed shares,
- market value above 2 billion: 1,200 left,
- price-earnings ratio below 15: 310 left,
- net debt below twice EBITDA: 140 left,
- payout ratio below 60%: 85 left.
Good practice:
- Write the criteria first, from your method, before running the screen.
- Use few filters. Every additional filter narrows the list but also excludes companies for arbitrary reasons.
- Check the data behind each filter. Is the P/E trailing or forward? How current is the data?
- Save the screen, so you can run the same criteria again later.
What screens hide
A screener answers one question: which assets meet these numbers today? It cannot say why a P/E is low, whether profits are about to collapse, or whether the company is a value trap. That is the research that starts after the screen.
It also hides history. Screeners list companies that exist today. A backtest that runs today's screen on past data ignores all the companies that went bankrupt, were taken over or were delisted. This survivorship bias makes past results look better than they were. Brown, Goetzmann, Ibbotson and Ross (1992) showed how strongly survivorship can distort performance studies.
Common mistakes
- Treating the result as a buy list. It is a reading list.
- Tuning filters until a favorite appears. That is confirmation bias with a spreadsheet.
- Ignoring data definitions. Different screeners calculate the same ratio differently.
- Backtesting on today's survivors. Include delisted companies, or treat results with great caution.
The takeaway
A screener filters thousands of assets down to a short list that meets your criteria. It saves time, but the list is a starting point for research, not a signal.
Check yourself
Quick quiz
4 questions. Instant feedback, nothing is saved.
-
1Tap the filter that removed the most companies.
Tap the chart.
Requiring a market value above 2 billion cut the list from 5,000 to 1,200, removing 3,800 companies.
-
2A screen returns 85 shares with low P/E ratios. What is the careful next step?
A screen finds candidates. The reasons behind the numbers decide whether any of them is attractive.
-
3What is survivorship bias in a screener backtest?
Companies that went bankrupt or were delisted drop out of today's lists. Results then look better than they were.
-
4Why should screen criteria be written before running the screen?
Criteria come from your method. Tuning them to the result defeats the purpose.
of 4 correct.