▲ NSE Screener
GUIDE

What every number in NSE Screener means, and how to use it

A field-by-field reference for the screener, plus walkthroughs of the stock drawer, time-series lab, factor model and portfolio optimiser. Each indicator lists what it measures, how it is computed here, and how traders and quants typically use it.

01How the app fits together

The app is a funnel: start wide with a screen, then look closer at a few names, then decide how much of each to own.

Data refreshes automatically after the close on weekdays: the NSE equity list, daily prices from Yahoo Finance, then indicators, technical history, factor returns and fundamentals. The status chips in the screener's top bar show how many tickers and bars are loaded.

02Using the screener

A screen is a list of rules that must all be true (AND). Results are sorted by any column and capped at 2,000 rows.

Building a rule

  1. + Add filter, pick a field (grouped as Price, Volume, Trend, Momentum and so on), pick an operator and type a value.
  2. Operators: > >= < <= = !=, between (two values), in (comma list, for text), contains (text), and is null / not null.
  3. Compare to another field with the vs toggle, e.g. close > sma_200 or sma_50 > sma_200 × 1.05. The multiplier lets you demand a margin.
  4. Press Run screen. Click a column header to sort; click again to flip direction.

Other controls

  • Search narrows by symbol or company name.
  • Only symbols with current data (on by default) hides stocks whose last bar is more than 4 days older than the market's latest. Turn it off to find suspended or illiquid names.
  • Columns chooses which fields show. Presets switch in the columns that matter for them.
  • Saved screens stores the current filters, sort and columns in this browser. They are not shared across devices.
  • Export downloads the current result as CSV.

Always add a liquidity floor. Most presets include turnover_cr >= 1 or higher. Without it, extreme values (biggest gainers, lowest P/B) are dominated by thinly traded stocks you could not trade at the shown price.

03Indicator reference

Every screener field, grouped as in the field picker. Field names in mono are what you see in exports and saved screens.

Info

FieldWhat it isHow to use it
Symbol, Companysymbol nameNSE trading symbol and listed company name from NSE's equity list.Search or use in to build a watchlist screen: symbol in TCS,INFY,HCLTECH.
SeriesseriesNSE series. EQ trades normally; BE is trade-for-trade (no intraday netting); BZ flags non-compliant companies.Add series = EQ for anything you plan to trade actively.
Sector (NSE)industryIndustry from NSE index classification. Stocks outside the mapped indices show Unclassified.Use in to stay within sectors you understand, or to compare peers.
PSU flagis_psu1 if the stock is in Nifty PSE, CPSE or PSU Bank, else 0.Government-owned firms often trade at structurally lower multiples; screen them separately when hunting value.
Last bar, Bars of historylast_date barsDate of the latest daily bar and the number of bars stored.Require bars >= 250 before trusting 200-day or 1-year fields; new listings have fewer.

Price & volume

FieldWhat it isHow to use it
Open, High, Low, Closeopen high low closeLatest daily bar, in ₹, at today's price scale.Mostly for display and for field-vs-field rules like close > sma_50.
Change %change_pctToday's close vs the previous close, adjusted for splits and bonuses.Daily movers. Pair with vol_ratio so the move has participation behind it.
Volume, Avg vol (20d)volume avg_vol_20Shares traded today and the 20-session average.Raw volume is not comparable across stocks; use turnover or the ratio instead.
Vol / avg 20dvol_ratiovolume ÷ avg_vol_20How unusual today's volume is for this stock.> 2 is notable, > 3 is a spike. A big up-move on a spike suggests institutional buying; a spike on a down day can mark capitulation or bad news.
Turnover (₹ cr)turnover_crclose × volume ÷ 1e7Value traded today in crore rupees.Your main liquidity filter. As a rough rule, keep any single order under 1–2% of daily turnover to limit market impact.

Trend

FieldWhat it isHow to use it
SMA 20 / 50 / 200sma_20 sma_50 sma_200Simple moving averages of the adjusted close over 20, 50 and 200 sessions.Classic trend filters. close > sma_200 = long-term uptrend; sma_50 > sma_200 is the “golden cross” regime. The 20-day tracks swing trends.
% vs SMA 50 / 200pct_vs_sma_50 pct_vs_sma_200(close ÷ SMA − 1) × 100How far price sits above or below the average.Positive = above trend. Very large values (e.g. +40% over the 200-day) mean an extended stock that often consolidates; small positive values in an uptrend are pullback entries.

Momentum

FieldWhat it isHow to use it
RSI 14rsi_14Relative Strength Index over 14 sessions, 0–100: the share of recent price change that came from up-days.< 30 oversold, > 70 overbought. In strong uptrends RSI can stay above 70 for weeks, so use it for timing within a trend (35–45 dips) rather than as a standalone sell signal.
Return 1W / 1M / 3M / 6M / 1Yret_1w … ret_1yTotal price change over each horizon on adjusted closes, in %.Momentum investing buys 3–12 month winners. Academic momentum skips the latest month because 1-month returns tend to reverse: try ret_6m > 30 with ret_1m not extreme.

Range

FieldWhat it isHow to use it
52W high / lowhigh_52w low_52wHighest and lowest prices over the last year.Reference levels many traders watch for breakouts and support.
% from 52W high / lowpct_from_52w_high pct_from_52w_lowDistance from each extreme, in %. Distance from the high is ≤ 0.Stocks within 3% of a 52-week high with rising volume are breakout candidates; stocks near the low are either value or falling knives, so check fundamentals.

Leading (second-order) signals

These look at changes in the trend and in volatility, on daily log returns of the adjusted close. μ is the average daily log return (in %/day), var is its variance (in basis points²). “15” and “60” are the window lengths in sessions.

FieldWhat it isHow to use it
Edge scoreedge_scoreA 0–100 rank of each liquid stock's odds of beating the average stock over the next 20 trading days. It is built walk-forward from the fields that have predicted returns. See Edge score.Its strongest evidence is on the downside: avoid the bottom 20. Higher scores improve the odds modestly; combine with your own analysis.
μ15, μ60lead_mu15 lead_mu60Average daily log return over the last 15 and 60 sessions: the short- and medium-term drift.μ15 of +0.3%/day ≈ +4.6% over 15 days. Positive and rising drift = momentum building.
Accellead_accelμ15 − μ60Is the short-term trend faster than the medium-term one?> 0 means the stock is accelerating relative to its recent norm.
Accel change 5dlead_accel_chgAccel today minus Accel 5 sessions ago: the change in acceleration.Positive = the acceleration itself is increasing, an early sign before price trends become obvious on a chart.
var15, var60lead_var15 lead_var60Variance of daily log returns over 15 and 60 sessions, in bp². √var15 ≈ daily volatility in bp.var15 = 400 bp² means a typical daily move of about 2%.
RoCV 5dlead_rocvvar15 now ÷ var15 five sessions ago − 1Rate of change of volatility, in %.Rising volatility alongside rising acceleration often marks the start of a move.
Vol regimelead_vol_ratiovar15 ÷ var60Short-term vs medium-term volatility.< 0.6 = volatility compression (a quiet coil, which often precedes expansion). > 1.5 = an active, expanding regime.
Today's move (z)lead_z(today's log return − μ15) ÷ σ15, both as of yesterdayHow many standard deviations today's move is relative to the recent norm.|z| > 2 is unusual. Measured against yesterday's statistics, so today's move cannot distort its own yardstick.
SecondOrderSignallead_signal1 when RoCV > 0 and Accel change > 0: volatility and acceleration are both rising.An early-trend detector. Combine with a trend filter and a confirmation flag; see Recipes.
Short_CONFIRMlead_short_confirm1 when z > 1: today closed more than one recent standard deviation above the 15-day drift.A short-horizon breakout confirmation.
CONFIRMlead_confirm1 when today's log return exceeds the stock's whole-history mean plus one standard deviation.A stricter, long-memory breakout test.
Next-session band (15d)lead_band15_lo lead_band15_hiclose × e^(μ15 ∓ σ15)A one-standard-deviation range for tomorrow's close, if recent drift and volatility persist.Rough stop and target placement: a close below the low band is a larger-than-normal adverse move.

Check the track record. The Signal Lab tests every one of these flags against what happened next. Over the current two-year history, stocks flagged by SecondOrderSignal, Short_CONFIRM and CONFIRM have tended to lag the average stock over the following week, because big moves partly reverse. Use them as “don't chase” warnings unless the Lab shows otherwise. For market-wide early warnings, use Market Pulse.

Technical (15/45 system)

FieldWhat it isHow to use it
SMA 15, SMA 45ta_sma15 ta_sma45Fast and slow moving averages used by the drawer's technical system.Faster than 50/200, suited to multi-week swings.
SMA 15 slopeta_sma15_slope(SMA15 ÷ SMA15 five days ago)^(1/5) − 1Average daily growth of the fast average, in %/day.> 0 = short trend rising. Steepening slope = strengthening trend.
15/45 cross todayta_cross1 if SMA15 crossed above SMA45 today, −1 if below, 0 otherwise.Screen ta_cross = 1 for fresh bullish crossovers; add a turnover floor to avoid whipsaws in illiquid names.
μ45ta_mu45Average daily log return over 45 sessions (%/day): the medium-term drift.The centre of the theoretical bands below.
th_high / th_lowta_th_high ta_th_lowclose × e^(μ45 ± μ15)A drift-based channel: medium-term drift plus or minus the short-term drift.Close above th_high = short-term drift is adding to the medium trend.
Vol band high / lowta_vol_high ta_vol_lowclose × e^(μ45 ± σ15)A volatility channel: medium drift plus or minus one 15-day standard deviation.Expected one-day range; useful for sizing stops in volatility units.
15d return variance, Variance ROCta_ret_var15 ta_var_roc15Variance of daily % returns over 15 sessions, and its change vs 15 sessions ago (%).Variance ROC above 0 = volatility expanding.
Pivot P, R1, S1ta_piv_p ta_piv_r1 ta_piv_s1P = (H + L + C) ÷ 3 · R1 = 2P − L · S1 = 2P − HClassic floor-trader pivots from today's bar, for use next session. The drawer also shows R2/S2.Intraday traders treat P as the bias line, R1 as first resistance and S1 as first support.

Factor (risk & return attribution)

Estimated by regression of each stock's last 252 daily excess returns (needs 120+ days) on the market and on this app's own factor returns (see Factors).

FieldWhat it isHow to use it
Beta vs Nifty 50f_betaSensitivity to the market: beta 1.3 means the stock moves ~1.3% for a 1% Nifty move, on average.< 0.7 = defensive; > 1.3 = aggressive. Portfolio beta is the weighted average.
Blume-adjusted betaf_beta_adj0.67 × beta + 0.33Beta shrunk toward 1, because extreme betas tend to drift back toward the average.Better for forecasting next year's risk than raw beta.
CAPM alpha, t-statf_alpha f_alpha_tAnnualised return not explained by market exposure, and its statistical significance.Only alphas with |t| ≥ 2 are distinguishable from luck. One year of data makes this a noisy estimate.
R² vs marketf_r2Share of the stock's daily variance explained by the market (0–1).Low R² = stock-specific story; beta is less reliable when R² is under 0.1.
Total vol, Idiosyncratic volf_total_vol f_idio_volAnnualised volatility, and the part not explained by the market.High idiosyncratic vol = diversifiable risk; size such positions smaller.
4-factor alpha, t-statf_ff_alpha f_ff_alpha_tAlpha after controlling for market, size, value, quality, momentum and low-volatility factors.The stricter test: a stock that “beat the market” only because it was a small-cap momentum stock shows no factor alpha.
Factor loadingsf_b_smb f_b_hml f_b_rmw f_b_wml f_b_lvolExposure to Size, Value, Quality, Momentum and Low-vol factors.Positive SMB = behaves like a small-cap; positive WML = behaves like a recent winner. Use to check what style you are really holding.

Fundamentals & scores

FieldWhat it isHow to use it
Market cap (₹ cr)mcap_crclose × shares outstanding ÷ 1e7Company size. Roughly: large-cap > ₹1 lakh cr, mid-cap ₹30,000 cr–1 lakh cr, small-cap below.Set a floor (₹1,000–2,000 cr) to avoid microcap data problems.
Price / bookpbMarket cap ÷ shareholders' equity.Most meaningful for banks and asset-heavy firms. Low P/B plus high ROE is the classic value-with-quality combo.
P/E (TTM), P/E ex one-offspe pe_normMarket cap ÷ trailing net income; the normalised version excludes unusual items where reported.Prefer pe_norm: one-off gains can make a P/E look artificially cheap. Blank when earnings are negative.
Earnings yieldearn_yieldNet income ÷ market cap, in % (the inverse of P/E).Works with losses (negative yield) and compares directly to bond yields.
ROE, Debt / equity, GP / assetsroe de gpaReturn on equity, leverage, and gross profit per rupee of assets.ROE > 15% with D/E < 0.5 is a common quality bar. Gross profitability (GP/A) is a robust quality measure in research. Lenders have no gross-profit line.
Value, Quality, V+Q scoresvalue_score quality_score qv_score0–100 percentiles. Value = average rank of book/price and earnings/price. Quality = average rank of ROE and GP/A. V+Q = percentile of their average.90+ = top decile of the market. Ranking beats raw thresholds because it adapts to how expensive the whole market is.
V+Q within sectorqv_sector_scoreThe same composite ranked only against stocks in the same sector.Removes sector bets: banks always look “cheap” on P/B and software “expensive”, so this finds the best of each industry.

04Presets explained

One click loads a ready-made screen. Treat each as a starting point and edit the thresholds.

PresetRulesIdea behind it
Top gainers / loserschange_pct > 0 (or < 0), turnover ≥ ₹1 crToday's movers with some liquidity.
Near 52W highwithin 3% of high, close > SMA200, turnover ≥ ₹2 crBreakout candidates in long-term uptrends.
Near 52W lowwithin 5% of low, turnover ≥ ₹1 crContrarian list; confirm fundamentals before buying.
RSI oversold / overboughtRSI < 30 (or > 70)Mean-reversion candidates.
Volume spikevol_ratio ≥ 2.5, change > 2%Up-moves with unusual participation.
Uptrend (50 > 200)SMA50 > SMA200, close > SMA50Stocks in a confirmed trend regime.
3M momentum3M return ≥ 25%, 1M return > 0Strong winners still moving up.
Pullback in uptrendclose > SMA200, RSI 35–45, 6M return > 15%Buy-the-dip in a strong stock.
SecondOrderSignal onlead_signal = 1Acceleration and volatility both rising.
Signal + breakoutlead_signal = 1, Short_CONFIRM = 1Early signal confirmed by today's move.
Vol compressionvar15/var60 < 0.6, Accel > 0Quiet coil with positive drift; often precedes expansion.
Significant alpha (4-factor)factor alpha t ≥ 2, turnover ≥ ₹5 crOutperformance not explained by style exposures.
Low-beta defensivesbeta 0–0.7, R² ≥ 0.1, turnover ≥ ₹10 crLiquid stocks that fall less in sell-offs.
Top V+Q / Best V+Q within sectorscore ≥ 90, mcap ≥ ₹2,000 crCheap and profitable, market-wide or per sector.
Cheap & profitableP/B 0–1.5, ROE ≥ 15%, mcap ≥ ₹1,000 crSimple value-quality screen.
Large-cap qualitymcap ≥ ₹50,000 cr, ROE ≥ 18%, D/E ≤ 0.5Durable compounders.
High-beta moversbeta ≥ 1.3, turnover ≥ ₹10 crStocks that amplify market moves.

05Stock drawer & TA history

Click any result row to open the drawer on the right.

  • Chart with SMA 50 (blue) and SMA 200 (purple), range buttons, and a crosshair readout.
  • Stat tiles with the key numbers from the screener row.
  • Technical analysis table: the 15/45 system, theoretical and volatility bands, the leading signals with plain-language verdicts, and pivots for the next session.
  • Daily TA history: one row per trading day, each computed only from data available on that day. Download it as CSV to test a rule in a spreadsheet.
  • Links to the full-page TA history and Time-series views for that stock.

Point-in-time history is what makes testing honest. Each TA history row is never revised by later data, so you can check “what did the signal say on that day, and what happened next?” without look-ahead bias.

06Time-series lab

Statistics on one stock's daily log returns. Open it from the drawer, or by clicking a symbol on the Factors page.

BlockWhat it tells youHow to read it
Return & riskCAGR, volatility, Sharpe (vs a 6.5% risk-free rate), max drawdown.Sharpe > 1 over several years is strong. Drawdown tells you the pain you would have sat through.
Tail risk: VaR / CVaRVaR 95 / 99: the daily loss exceeded on 5% / 1% of days. CVaR 95: the average loss on the worst 5% of days.Size positions so a CVaR day does not breach your risk limit.
DistributionSkew, excess kurtosis and the Jarque-Bera normality test.Kurtosis > 0 = fatter tails than a bell curve, so big moves happen more often than “σ” suggests. Negative skew = crash-prone.
Autocorrelation (ACF, PACF, Ljung-Box)Do today's returns predict tomorrow's? On squared returns: does volatility cluster?Significant return ACF = exploitable trend or reversal (rare). Significant squared-return ACF = volatility clustering (common), which is why GARCH helps.
Stationarity (ADF)Augmented Dickey-Fuller test on log price and on returns.Prices are usually non-stationary (they wander); returns should be stationary. A stationary price hints at mean reversion.
Random-walk testsLo-MacKinlay variance ratios and the Hurst exponent.Variance ratio > 1 or Hurst > 0.5 = trending; < 1 or < 0.5 = mean-reverting; ≈ 1 / 0.5 = random walk.
Volatility modelsRolling vol, EWMA and GARCH(1,1).GARCH persistence near 1 means volatility shocks fade slowly. Use its current estimate, not the long-run average, for near-term risk.
TrendLog-linear regression channel and rolling 63-day slope.Price near the lower channel in a rising regression = trend pullback.
SeasonalityAverage returns by weekday and by month.Usually noise with only two years of data; look for consistency, not one big bar.
Forecast coneA 20-session range from the GARCH volatility term structure.Shows where price is likely to be, not where it will go. Use it for option strikes and stop distances.

07Factors page

Shows which investment styles have been paying in Indian equities, built from this app's own data.

Each month-end, the 1,000 most liquid EQ stocks are ranked and split into top 30% and bottom 30%. A factor's daily return is the equal-weighted return of the long leg minus the short leg, held until the next month-end. Fundamentals are point-in-time: a figure is only used after its reporting lag (60–90 days), so there is no look-ahead.

FactorLong minus shortWhat a positive premium means
MKT · MarketNifty 50 minus risk-freeEquities beat cash.
SMB · SizeSmall minus big market capSmall-caps outperformed.
HML · ValueHigh minus low book-to-marketCheap stocks outperformed expensive ones.
RMW · QualityHigh minus low ROEProfitable firms beat unprofitable ones.
WML · Momentum6-1 month winners minus losersRecent winners kept winning.
LVOL · Low volatilityLow minus high 63-day volatilityCalm stocks beat jumpy ones. Carries negative market beta.
V+Q, V+Q within sectorComposite of B/P, E/P, ROE, GP/AReported for tracking; not used in regressions.

For each factor the page shows annualised premium, volatility, Sharpe, t-stat, CAGR, max drawdown and hit rate (share of positive days), plus cumulative return lines and the stocks with the strongest exposures. Use the 6M / 1Y / All buttons to see whether a style is gaining or fading.

Read the t-stat first. Factor history currently starts in mid-2025, so most premia have |t| below 2 and cannot be told apart from noise. Returns are before transaction costs, which hit the small-cap and momentum legs hardest.

08Portfolio lab

Pick 2–40 stocks and the lab shows five ways to weight them, with risk numbers for each.

Settings

Expected return (μ)

Historical: the sample mean, very noisy. CAPM: risk-free + adjusted beta × equity risk premium (default 6%). Blend (default): half of each, a simple shrinkage that stops the optimiser chasing last year's winners.

Covariance (risk model)

Sample: raw, unstable with many stocks. Ledoit-Wolf (default): shrinks correlations toward their average, more stable. Factor: risk from the six factor exposures plus stock-specific noise.

Lookback

How many days of history to estimate from. Only days on which every chosen stock traded are used.

Cap per stock

Maximum weight in any one stock. Long-only and fully invested. A cap of 10–20% stops the optimiser concentrating.

The five portfolios

MethodHow it weightsWhen to use it
Equal weight1/N each.A tough benchmark; beats many optimised portfolios out of sample.
Inverse volatilityWeight ∝ 1 / volatility.Simple risk balancing that ignores correlations.
Minimum varianceLowest possible portfolio volatility.Defensive; ignores expected returns entirely, which makes it robust.
Maximum SharpeBest expected return per unit of risk (the tangency portfolio).Most sensitive to bad μ estimates; use Blend or CAPM μ and a cap.
Risk parity (ERC)Each stock contributes the same share of total risk.Balanced diversification when you have no return view.

Reading the results

  • Risk contribution: the share of portfolio variance from each stock. A 10% weight can be 25% of the risk.
  • Diversification ratio: weighted-average stock volatility ÷ portfolio volatility. Higher = more benefit from diversification.
  • Effective N: 1 ÷ Σw². A 20-stock portfolio with effective N of 6 is really concentrated in about 6 names.
  • Factor exposures: the weighted sum of each stock's loadings, so you can see if you accidentally built a small-cap momentum bet.
  • Efficient frontier: the best return available at each risk level, given your μ and covariance.
  • Correlation heat map and eigenvalues: a large first eigenvalue share means one common driver (usually the market) dominates.
  • Backtest chart: growth of ₹1 for each portfolio against Nifty 50.

The backtest is in-sample. Weights were chosen using the same window they are tested on, so it will flatter the optimised portfolios. Use it to compare behaviour (drawdowns, volatility), not to predict returns.

09Signal Lab: does a signal actually make money?

The Signal Lab replays every preset, flag and screener field on history and tells you how it did after costs. Nobody can promise profits; the Lab tells you the real odds.

  • How it tests. Every week, it records what each signal said using only data available that day. It then measures those stocks over the next 5, 20 and 60 trading days against the average liquid stock.
  • Costs included. Each pick pays STT, stamp duty, exchange charges and an estimated spread (about 0.3–0.6% per round trip).
  • Two halves. Each result is repeated on the earlier and later half of the history. An edge must hold in both to count.
VerdictMeaningWhat to do
Edge after costsBeat the average stock after costs, t ≥ 2, positive in both halves.Worth using, sized sensibly. Re-check as history grows.
Edge eaten by costsA real gross edge that trading costs consume.Combine with other signals, or trade it less often.
Ranks, no long edgeThe field sorts stocks usefully, but its best 10% alone doesn't clear costs.Use it to avoid the bad end rather than to pick winners.
Reliably negativeLagged the average stock, t ≤ −2, in both halves.Think twice before buying what it flags.
No reliable edgeWithin what chance alone could produce.Don't rely on it by itself.

In the screener, each preset button carries a coloured dot for its verdict. When a preset is active, a track-record line appears above the results, and you can switch it between 5, 20 and 60 days.

Reading IC. For fields, the Lab reports the information coefficient: the rank correlation between today's value and the next period's returns. Quant funds treat a steady IC of 0.05 as useful and 0.10 as strong.

10Market Pulse: early warnings for the whole market

Market Pulse scores the market from −100 to +100 for the next 1–3 months. It shifts the odds; it does not predict a date.

  • Inputs. Eight index-level indicators with 18 years of history: Nifty vs its 200-day average, 12-1 month momentum, drawdown from the 1-year high, India VIX level and spikes, the volatility regime, and how Nifty 500 and Bank Nifty are doing relative to Nifty 50.
  • No peeking. Each indicator's direction and weight was set using 2008–2016 only. The score was then tested on 2016–2026, which it never saw. The odds shown on the page come from that test.
  • Reading it. Favourable, Neutral and Below average are the top, middle and bottom thirds of readings. The page shows how often Nifty rose over the next 20 or 60 days in each third, compared with all days.
  • What's driving it. Bars show which indicators push today's score up or down.
  • Breadth. Breadth gauges (% of stocks above their 200-day average, new highs minus new lows, advance-decline, breadth thrust) are shown for monitoring. They only cover the app's two years of stock data, so they are not yet in the score.

How to use it. Lean in when the score is favourable and lean back when it is below average. Adjust position sizes and how much cash you hold, rather than going all-in or all-out. Indian markets have historically rewarded buying after falls and fear, so a weak, fearful market can score as favourable.

11Edge score: a leading signal for each stock

Every liquid stock gets a score from 0 to 100 for the next 20 trading days. It appears as a screener column, in each stock's drawer, and as two presets.

  • What goes in. The fields the Signal Lab found predictive, weighted by how strongly they predicted. Currently that means momentum (12-1 month return, distance from the 52-week high, % above the 200-day average), size and liquidity, low volatility, profitability, valuation and volume. The mix updates itself as evidence changes.
  • Walk-forward, no peeking. On every past date, the weights used only results whose 20-day outcome was already known. That is how a live system would have worked. So its Lab record is an honest out-of-sample test, not a fitted one.
  • What the evidence says. The ranking is real: rank correlation with the next 20 days is about 0.065 (t ≈ 2). The chance of beating the average stock rises steadily from about 38% in the bottom decile to 48% in the top. Most of the value is on the downside: the bottom 20% lagged clearly, while the top deciles were roughly average.
  • In the drawer. You see the score, its decile, how that decile did historically, and which inputs push it up or down.

How to use it. Use it as a filter: skip stocks with an Edge score below 20, and prefer higher scores when choosing between candidates. Don't buy a stock just because its score is high. Check the score's current record in the Signal Lab; it updates every day.

12Backtesting a strategy

The Backtest page turns any rule set into a trading strategy and shows how it would have done after costs.

  1. Start from a starter strategy or any screener preset. Every Signal Lab result also has a “Backtest this” button, and an active preset's track-record line in the screener has a “Backtest →” link.
  2. Rules: “only buy stocks where…” filters, a ranking field, and how many stocks to hold.
  3. Rebalance every week, month or quarter.
  4. Market timing (optional): hold cash when Nifty is below its 200-day average, or when Market Pulse reads below average.
  5. Costs: none, realistic (default) or double as a stress test.

Results show CAGR, Sharpe, max drawdown and turnover, all after costs, against Nifty 50 and the average liquid stock. You also get a plain-English verdict with a t-stat and a two-halves check, a growth-of-₹1 chart, every rebalance with its top picks, and what the strategy would hold today. The link in the address bar saves the exact strategy, so you can bookmark or share it.

Beware of overfitting. Every tweak you try raises the odds that a great-looking result is luck. The page counts your runs in a session and warns you. Trust strategies that are simple, make economic sense, hold up in both halves, and also pass in the Signal Lab.

13Recipes

Combinations that put several indicators to work together. Edit the thresholds to taste.

Early trend with confirmation
lead_signal = 1
lead_short_confirm = 1
close > sma_200
vol_ratio >= 1.5
turnover_cr >= 2

Rising acceleration and volatility, a strong day, a long-term uptrend and above-normal volume. Sort by lead_z.

Quality compounders on a dip
qv_score >= 80
close > sma_200
rsi_14 between 35 and 50
mcap_cr >= 5000

Fundamentally strong companies pulling back within an uptrend.

Squeeze setup
lead_vol_ratio < 0.6
pct_from_52w_high >= -10
sma_50 > sma_200
turnover_cr >= 1

Volatility has compressed near the highs of an uptrend. Watch for a close above lead_band15_hi to trigger.

Defensive core for a portfolio
f_beta between 0.4 and 0.8
quality_score >= 70
de <= 0.5
turnover_cr >= 10

Export the result, then load 10–20 names into the Portfolio lab with Ledoit-Wolf covariance and Minimum variance or Risk parity.

Stock-picker's alpha, not style
f_ff_alpha_t >= 2
f_r2 >= 0.2
bars >= 250
turnover_cr >= 5

Outperformance that survives controls for size, value, quality, momentum and low-vol. Check the stock in the time-series lab before acting.

14Data & limitations

Know what the numbers can and cannot tell you.

  • Sources. The stock list is from NSE. Prices and fundamentals come from Yahoo Finance, which can have gaps or late corporate-action adjustments. Daily returns larger than ±60% are discarded as likely bad prints.
  • History. Prices go back about two years and factor returns about 16 months. Statistics needing long samples (alpha, factor premia, seasonality) are noisy.
  • Survivorship. The universe is today's NSE list; stocks delisted before tracking began are absent, which flatters historical averages.
  • Costs. No figure here includes brokerage, STT (0.1% each way on delivery), stamp duty, spreads or market impact.
  • Sectors. Only stocks in NSE's mapped indices have a sector; the rest are Unclassified, which weakens sector-neutral scores.
  • Update timing. Data refreshes once per weekday after the close. Intraday prices are not shown.
NSE Screener is a research tool, not investment advice. Check any idea against your own analysis and risk limits before trading.