Table of contents

• Why most people use Screener wrong

• The query language, properly explained

• Twelve copy-paste queries, one per strategy

• What to do after the screen runs

• Four ways Screener lies to you

• Custom ratios and the Excel trick almost nobody uses

• Free or ₹4,999? An honest answer

• Key takeaways

Why most people use Screener wrong

Screener.in is the most-used stock research tool in India, and most of its users have never written a query.

They open a company page, look at the top row of tiles — market cap, P/E, ROCE, book value — decide the stock is "cheap" or "not cheap," and close the tab. Bas price tag dekh liya. That is not screening.

The actual power of the site sits behind a plain text box at screener.in/screen/new/, where you type conditions in something close to English and get back every listed Indian company that satisfies them. Ten seconds of typing replaces a week of scrolling. Yet the box intimidates people, because Screener ships with almost no documentation and the third-party guides floating around are — bluntly — full of errors. A claim you will meet repeatedly is that Screener can't do OR logic and has no technical indicators. Both are wrong, and we'll show you Screener's own screens that prove it.

This article is the manual we wish existed. By the end you will have twelve working queries, a clear picture of what the tool cannot do, and a list of the specific ways it will quietly hand you a wrong answer without ever showing an error message.

One caution before we start, and it's the whole reason this article exists in this order. A screen is a filter, not a verdict. Look at what a good screen would have told you about ITC.

Figure 1. ITC's compounded sales growth, profit growth and stock price CAGR over 10, 5, 3 years and the latest period. Growth stayed positive for a decade while the stock returned 1% a year and fell 35% in the last year.

Sales and profit compounded at 7–10% for ten years, and average return on equity was 28% over ten years, 32% over five and 35% over three. A quality screen built on return on capital — #6 or #12 later in this article — would have surfaced ITC year after year. The stock returned 1% a year over that decade, then fell 35% in the last twelve months. Seedhi baat: a screen tells you about the business. It tells you nothing about what you are paying for it.

The query language, properly explained

A Screener query is one plain-text expression. There's no drag-and-drop builder, no wizard. You type conditions, join them with AND, and hit Run this query.

Figure 2. Annotated diagram of a Screener.in query showing six syntax rules — crore units, period suffixes, the cosmetic percent sign, manual ANDs, ratio-to-ratio comparison, and null handling with COALESCE.

Six rules cover ninety percent of what you need.

Company-level money totals are in ₹ crore. Market Capitalization > 500 means five hundred crore, and the same goes for Sales, Debt and Net Profit. The most common beginner mistake is typing > 5000000000 and wondering why nothing comes back. Per-share figures are the exception and are quoted in rupees — Current price, Book value, EPS, Graham Number, DMA 50 and DMA 200. That is precisely why Current price < Graham Number is a valid comparison.

You write the AND yourself. Newlines are whitespace and nothing more — Screener does not infer a join between two lines. Every condition except the last one needs an explicit AND after it.

The % sign is decoration. Return on capital employed > 22% and Return on capital employed > 22 are identical. Screener stores percentages as the number 22, not 0.22, so a stray percent sign never changes your result. Handy, but it means a typo like Debt to equity < 0.5% silently reads as < 0.5.

Period suffixes carry most of the meaning. Three families:

Suffix	Means	Example
(none) or 12m	Trailing twelve months	Sales growth
last year / ann	Latest annual report	OPM last year
preceding year	The annual report before that	Gross block preceding year
3Years 5Years 10Years	Compounded growth over that window	Profit growth 5Years
Average … 5Years	Multi-year average of a ratio	Average return on capital employed 5Years
3Years back	The level as it stood three years ago	Debt 3Years back
latest quarter	Most recent reported quarter	Net Profit latest quarter

Note the trap in that table: Sales growth 5Years is a CAGR, while Debt 3Years back is a level. Mix them up and your screen means something entirely different from what you intended.

A ratio can be compared to another ratio. This is where Screener gets genuinely powerful, and where most users never go. All of these are valid:

Price to Earning < Industry PE
Debt < Debt 3Years back
Profit growth 3Years > Sales growth 3Years
Return on equity > Average return on equity 3Years
Gross block > 1.2 * Gross block preceding year

OR works, and so do parentheses. Screener's own official "Capacity expansion" screen nests boolean groups inside brackets. Both AND/OR and the field names are case-insensitive. What's undocumented is operator precedence when you mix AND and OR without brackets — so always bracket your OR blocks explicitly, exactly as Screener does in its own screens.

Beyond the comparison operators there is a small set of functions. COALESCE is the only one Screener documents; others such as least, abs and if are accepted by the query box but undocumented, so lean on them lightly.

What definitely does not exist is RANK or PERCENTILE. Screener does give you per-company medians over time — Sales growth 5years median, Med ROCE 5Yr, Historical PBV 5 Years — but nothing that computes a statistic across the whole universe and places a company inside it. Remember that; it comes back to bite in a moment.

The one that catches everybody

A line with no comparison operator does not throw an error. It filters nothing.

Free cash flow 3years AND
Sales growth 3Years AND
Debt to equity < 0.5

That screen has one working condition. The first two lines are decorative. Screener chup-chaap runs it and returns a list that looks entirely plausible. There are dozens of public screens on the site carrying exactly this bug, saved by people who never realised.

The rule: every line in your query must contain >, <, >=, <= or =. Read your query once, count the operators, count the lines. If they don't match, you have not built the screen you think you built.

And the one that deletes companies behind your back

A loss-making company has no P/E. A company that listed in 2024 has no ten-year growth figure. Both are null, and a null silently fails every comparison — so those companies disappear from your results even though you never asked to exclude them.

COALESCE is how you take that decision back:

COALESCE( Average return on capital employed 10Years, Average return on capital employed 5Years ) > 20

Read as: use the ten-year average; if it doesn't exist, fall back to five-year. Screener documents this idiom itself, and it is the difference between "the best compounders" and "the best compounders that happen to have been listed since 2016."

Twelve copy-paste queries, one per strategy

Below are twelve screens, each mapped to a documented strategy. Paste one into the query box, hit run, and adjust from there.

Two ground rules before you do.

Thresholds are yours to tune, not ours to dictate. Copy karo, chalao, phir apne hisaab se adjust karo. These numbers are sensible starting points for the Indian market in 2026, not sacred. If a screen returns two names, loosen it. If it returns two hundred, tighten it. A screen returning zero results is almost never a syntax problem — it's a threshold problem.

Most of these exclude banks and NBFCs by accident. Any screen containing ROCE, operating margin, current ratio or debt-to-equity does — more on why in the traps section. The exceptions are #5, #9 and #10, which contain no such field and will happily return financials. If you want financials, screen them deliberately, not by leftover.

1. Coffee Can — buy great businesses, seal the can

Market Capitalization > 500 AND
Sales growth 10Years > 10 AND
Average return on capital employed 10Years > 15 AND
Debt to equity < 1

The classic Indian quality screen: a decade of double-digit sales growth alongside a decade of high return on capital. Ten-year fields mean young companies drop out, which is the point — the strategy is explicitly about businesses with a long track record. → Coffee Can Investing

2. Magic Formula — cheap and good, ranked

Market Capitalization > 500 AND
Earnings yield > 8 AND
Return on capital employed > 20 AND
Current price > 20

Greenblatt's method needs you to rank every stock on earnings yield and return on capital, then combine the ranks. Screener has no ranking function — so this filters, and you rank in a spreadsheet afterwards. That two-step is unavoidable. → Magic Formula

3. GARP — growth at a price that isn't silly

PEG Ratio < 1 AND
PEG Ratio > 0 AND
Profit growth 3Years > 15 AND
Debt to equity < 0.5 AND
Market Capitalization > 1000

The PEG Ratio > 0 line is not redundant. Without it you catch companies with negative earnings growth, whose PEG is negative and therefore "less than 1." → Growth at a Reasonable Price

4. Dividend compounders

Dividend yield > 1.5 AND
Average dividend payout 3years > 20 AND
Profit growth 5Years > 5 AND
Debt to equity < 0.5 AND
Market Capitalization > 2000

Be honest about what this can't do: Screener has no "consecutive years of dividend increases" field, so a true Dividend Aristocrats screen is impossible here. This is the closest proxy — a sustainable payout backed by growing profit. You verify the streak by hand on the company page. → Dividend Aristocrats / Kings

5. Golden cross — the trend just turned

DMA 50 > DMA 200 AND
DMA 50 previous day < DMA 200 previous day AND
Market Capitalization > 1000 AND
Volume 1week average > 50000

Yes, on Screener. The previous day comparison is what makes it a crossover rather than a state — it fires only on the day the 50-day moving average crosses above the 200-day, not every day it happens to be above. Run it daily or it's useless. → Moving-Average Crossover

6. High-ROIC compounders

Average return on capital employed 5Years > 20 AND
Sales growth 5Years > 12 AND
Debt to equity < 0.3 AND
Market Capitalization > 1000

The five-year window instead of ten catches businesses that got good more recently. Looser on history, stricter on leverage. → High-ROIC Compounders

7. Piotroski F-Score — cheap but not broken

Piotroski score > 7 AND
Price to book value < 1.5 AND
Market Capitalization > 300

Screener pre-computes the F-Score, which saves you nine separate accounting tests. Pairing a high score with a low price-to-book is the original Piotroski setup: cheap stocks, minus the ones quietly dying. → Piotroski F-Score

8. Graham defensive

Current price < Graham Number AND
Current ratio > 2 AND
Debt to equity < 0.5 AND
Profit growth 10Years > 3 AND
Market Capitalization > 500

Graham Number is a native field — no need to compute √(22.5 × EPS × book value) yourself. In a market as expensive as India's usually is, expect this screen to return very little. That emptiness is information. → Graham's Defensive Investor

9. Cheap relative to its own industry

Price to Earning < Industry PE AND
Price to Earning > 0 AND
Price to Earning < 15 AND
Return on equity > 15 AND
Market Capitalization > 500

Industry PE is the only sector-relative field in the entire query language. It's blunt, but it beats comparing a cement company's P/E to a software company's. → Low P/E

10. Promoters are buying, and nothing is pledged

Change in promoter holding 3Years > 2 AND
Pledged percentage = 0 AND
Profit growth 3Years > 0 AND
Market Capitalization > 300

Insiders buying their own stock over three years is one of the few genuinely informative signals available for free. The zero-pledge condition removes the situations where a promoter stake is really a lender's collateral. → Insider Buying

11. Deleveraging turnaround

Debt < Debt 3Years back AND
Profit growth 3Years > 20 AND
OPM > OPM last year AND
Market Capitalization > 300

Three conditions, three different time bases — a level comparison, a CAGR and a TTM-versus-annual comparison. This is the query that best shows off what the suffix system can actually do. → Turnarounds

12. Quality, well off its high

Down from 52w high > 40 AND
Average return on capital employed 5Years > 12 AND
Debt to equity < 0.75 AND
Piotroski score > 5 AND
Market Capitalization > 500

A falling knife with a track record and a clean balance sheet is a different animal from a falling knife without one. The F-Score floor is what separates the two. → 52-Week Low

The guard rail you should append to almost everything

AND Market Capitalization > 500
AND Current price > 20
AND Volume 1year average > 25000

Three lines that remove the microcaps, the penny stocks and the names you'd never be able to sell in a hurry. Add them to any screen above unless you have a specific reason not to.

What to do after the screen runs

The screen is the cheap part. Bees naam mil gaye — ab kya? Twenty names is not a portfolio, and this is exactly where most retail research quietly stops.

Figure 3. Map of a Screener.in company page showing what each tab contains — Analysis, Profit & Loss, Cash Flow, Ratios, Investors, Peers, Documents, the four summary tables, and Excel export.

A workable order of operations, roughly fifteen minutes per name:

Read the Cons first. The Analysis tab auto-generates a Pros and Cons list. It is blunt and occasionally unfair, and it is the cheapest thirty seconds in the entire workflow. Lines like "The company has delivered a poor sales growth of 9.62% over past five years" will kill a third of your shortlist before you've opened a single statement.

Then the ten-year P&L. Not the last quarter. You are looking for whether operating margin holds up across a full cycle, whether the tax rate looks normal, and whether "other income" is doing suspicious amounts of work. Click any line item to expand its schedule.

Then cash flow. Screener added Free Cash Flow and a CFO/OP row in April 2026, which is the single most useful addition it has made in years. Profit is an opinion; cash from operations tracking well below operating profit for several years running is the tell.

Then the Ratios tab. Debtor days, inventory days, cash conversion cycle. A conversion cycle drifting upward year after year is working capital quietly eating the business, and it never shows up in the P&L.

Then Investors. Promoter holding, pledge, FII and DII stakes, and the number of shareholders quarter by quarter. A rising shareholder count next to a falling promoter stake is retail buying what insiders are selling.

Then Documents — annual reports, credit ratings, concall transcripts with AI summaries. This is where the screen ends and actual research begins.

One warning about the four summary tables at the top of the page. They look like a single grid, but their bottom rows are not the same period. Compounded Sales Growth and Compounded Profit Growth end in a TTM row. Stock Price CAGR ends in a 1 Year row. Return on Equity ends in a Last Year row. Reading across that bottom line as if it were one column is one of the most common errors on the site.

Four ways Screener lies to you

Screener fails permissively. It does not crash on a bad screen — it returns a confident, wrong list. These four are the ones that cost real money.

Figure 4. Four Screener.in traps — ROCE excluding banks, the consolidated-versus-standalone heuristic, restated historical data, and Screener's stricter definition of Debt.

ROCE deletes every bank. Screener computes capital employed as share capital plus reserves plus borrowings. For a bank, borrowings are the raw material — so HDFC Bank shows a return on capital employed of roughly 7%. Any screen containing Return on capital employed > 15 mechanically excludes the entire financial sector, and you will never notice because nothing errors. The same applies to operating margin, current ratio, inventory turnover, cash conversion cycle and the Altman Z-Score — all meaningless or absent for financials. Screen banks separately, on Return on equity, Net Profit and growth in the loan book. Screener's own bank rows, Financing Profit and Financing Margin %, are constructions no other data source uses and are frequently negative — read them on the company page rather than filtering on them.

Consolidated or standalone is decided per company, not by you. Screener defaults to consolidated but silently falls back to standalone when the two sets of numbers are close and standalone covers a longer period — or, separately, whenever standalone is the more recent filing. So a single screen can be comparing one company's consolidated debt-to-equity against another's standalone. There is no query switch to force a basis — the /consolidated/ URL suffix works only on company pages. Check which basis is showing before you trust any margin or leverage number.

"3 years back" is today's restated number. When a company revises its historicals, Screener overwrites its records. Debt 3Years back gives you what the company now says its debt was, not what you would have seen at the time. Add survivorship bias — delisted and merged companies simply vanish from the universe — and any screen that looks backwards is quietly flattering itself. Asked about backtesting on the ValuePickr forum, Screener's own team called it a wishlist item — "very complex and would require heavy resources." Treat Screener as a filter for today. It is not a backtest, and no amount of cleverness makes it one.

Screener's "Debt" is stricter than everyone else's. Current maturities of long-term debt and lease liabilities are both folded in. Sales are net of excise, VAT and sales tax. Director remuneration is pushed into employee cost. Your Debt to equity < 0.5 filter is therefore tighter than the same filter anywhere else, and your numbers will not tie out against a broker terminal or the annual report. Pick one source and stay inside it.

Which leads to the bigger question: which strategies can honestly be run here at all?

Figure 5. Matrix of what Screener.in supports for rules-based strategies — multi-year quality thresholds, composite scores and simple momentum signals work; ranking, backtesting, derivatives and exchange filtering do not.

The short version. Multi-year quality thresholds, composite scores like Piotroski and Altman Z, and simple trend signals all work natively. Sector-relative valuation works partly. Ranking, backtesting, anything involving derivatives, and filtering by exchange do not work at all — and no threshold you type will change that. For a ranked strategy, filter on Screener and rank in a spreadsheet. For a backtest, you need a different tool and a point-in-time dataset.

Custom ratios and the Excel trick almost nobody uses

Two free-tier features that most users don't know exist.

Custom ratios. Under Accounts → Settings → Custom ratio, you can define your own metric from the same field vocabulary — give it a name, a short name and a formula. Once saved, it works everywhere: inside screen queries, as a results column, and in peer comparison. This is how you build a rough DCF value, a distance-from-52-week-low field, or any composite score Screener doesn't ship. It costs nothing.

The Excel template. On any company page, click Export to Excel and you get ten years of financials in a structured workbook. Add your own rows and formulas — a median P/E, a normalised margin, whatever your process needs — then re-upload the modified file at screener.in/excel/. From that point on, every company you export comes back with your calculations already applied.

Think about what that gives you: your own analysis template, applied automatically to any listed Indian company, in about two clicks. It is the most valuable free feature on the site and almost nobody uses it. The one thing to watch is that your formulas point at fixed cell ranges — Screener's own worked example is =MEDIAN(B14:L14) — so a company whose statements are laid out differently can leave your rows pointing at the wrong data. Sanity-check the first bank you export, and keep a separate template for financials if it drifts.

Free or ₹4,999? An honest answer

Figure 6. Comparison of Screener.in's free Hobby Investor tier against the ₹4,999/year Active Investor plan across eight usage limits, plus the four premium-gated features.

Every single query in this article runs on the free tier. Every one of them, every quarter, forever. The free plan is not a crippled trial — it is a genuinely complete screening tool with volume limits.

What premium actually buys is volume plus four gated capabilities: CSV export of screen results, the industry filter, multiple watchlists, and Screener AI credits. Everything else on the list is simply a bigger number — 800 stock alerts instead of 10, 75 screen alerts instead of 2, 60 quick ratios instead of 18. Custom ratios, saved screens and Excel automation are ticked on both tiers.

Tension mat lo about the decision. The honest test is this: if you are copying screen results into a spreadsheet by hand more than once a month, or you want alerts on more than two screens, ₹4,999 a year pays for itself in saved time. If you run four screens a quarter and read the results on the page, stay free. There's a full refund within 48 hours if you get it wrong.

One thing worth knowing either way: Screener's pricing page puts no number on saved screens for either tier. It is screen alerts that are explicitly rationed — two free, seventy-five on premium. And a screen alert is arguably the best feature on the whole site: Screener re-runs your query as new results are filed and emails you what's new, so you never have to remember to check.

Key takeaways

• Every line in a query needs a comparison operator. A bare field name filters nothing and Screener will never tell you. Count your operators against your lines.

• Company totals are in crore, per-share figures are in rupees, percentages are plain numbers, and you write the AND yourself. Those four rules prevent most beginner failures.

• OR, parentheses, ratio-vs-ratio comparisons and moving averages all work — despite what most online guides claim.

• ROCE screens silently delete the entire banking sector. Screen financials separately, on ROE.

• Screener is a filter for today, never a backtest. Restated historicals and survivorship bias contaminate anything backwards-looking.

• The screen is 10% of the work. The ten-year P&L, the cash flow statement and the Documents tab are the other 90%.

• The free tier runs all twelve of these queries. Pay for premium when you're exporting, not before.

What this means for your process

A screen is only as good as the rule set behind it. The twelve queries above are the mechanical layer of twelve different strategies — but the query is the easy part. What matters is knowing why the thresholds are where they are, how long you're meant to hold, what makes you sell, and how many names you're supposed to own.

That's what the playbooks are for. Start with Coffee Can Investing if you want the buy-and-forget end of the spectrum, Magic Formula if you'd rather run something mechanical and rebalance annually, or the Moving-Average Crossover if you want signals rather than statements. All three are free to read, no login.

Everyone has a hot tip. Almost no one has a plan.

Bharatstox publishes NSE and BSE market journalism, corporate filing explanations, live market panels and attributed research calls with visible analyst details and timestamps. Visit Bharatstox for market context, and Style of Picks for the documented strategy playbooks behind these screens.

Disclaimer: This article is for informational and educational purposes only and does not constitute investment advice or a recommendation. Investing in securities markets is subject to market risks. Read all related documents carefully before investing.

Sources

• Screener.in knowledge base — accounting definitions and formulas, null values and COALESCE, custom ratios, standalone vs consolidated, TTM vs last year, export limits, Excel templates, why numbers differ

• Screener.in pricing and changelog, retrieved August 2026

• Screener's own reference screens — Capacity expansion, Golden Crossover, Magic Formula, Debt reduction

• ValuePickr — Screener team on backtesting

• ITC figures from screener.in/company/ITC, retrieved August 2026