Skip to content
Alpha Nova research guides

Stock screener methodology and data

Documentation updated · Describes the current app; not a market-data timestamp.

Alpha Nova combines public Yahoo Finance price history with NSE end-of-day reports. Its stock screens describe observed price, volume and delivery conditions. Calculations depend on the observations available from each source; they do not provide forecasts, execution prices or probabilities of profit.

Which stocks and sessions are included?

The screener normally uses the official Nifty 500 constituent list. If that list is unavailable, it can fall back to the 500 most-traded EQ symbols in the latest NSE archive session. The workspace labels the universe and reports coverage. A cached constituent list can lag index changes, and missing observations can reduce the usable history.

Volume and delivery

Volume ratio = latest traded quantity ÷ average traded quantity across available observations in the prior 20 sessions. Delivery ratio = latest delivered quantity ÷ average delivered quantity across available prior observations. The latest session is excluded from each baseline. Missing values are omitted rather than replaced with zero. Fewer observations can make a ratio less representative.

Delivery percentage is the share reported in the NSE delivery file. A high delivery ratio does not identify a buyer or establish institutional accumulation. The screener preset and Delivery radar use different classification rules.

Returns and relative strength

Returns use (latest close ÷ earlier close − 1) × 100 over 5, 21, 63 and 126 observations for the week, month, quarter and half-year labels. Three-month relative strength subtracts the Nifty 50 return over 63 observations from the stock return. It is measured in percentage points, not as a percentile or a risk-adjusted performance measure.

Trend, highs and breakouts

Moving averages are simple averages of 50, 150 or 200 daily closes. The screener’s 52-week high and low are the maximum and minimum of up to 252 available closes, not intraday extremes. Its 20-day breakout instead compares the latest NSE close with daily highs from available prior archive sessions. These definitions can differ from other platforms.

The fresh golden-cross flag compares the current 50/200-day ordering with the ordering 10 observations earlier. It is not an event-by-event count of all crosses inside that window. RSI uses Wilder-style smoothing on up to 120 available closes. Volatility annualizes the sample standard deviation of up to 20 log returns using √252.

How is the 0–100 composite score calculated?

Each available factor is ranked within the available screener rows. The percentile is the share of finite observations less than or equal to that value, multiplied by 100. The score takes a weighted mean of those factor percentiles:

Unavailable factors are excluded and the remaining weights are renormalized. The app adds 5 points above the 200-day average, subtracts 5 below it, and adds 3 when the 50-day average exceeds the 200-day average. The result is clamped to 0–100 and rounded. Scores can change with coverage and are not win probabilities; no backtested return is implied.

Are the observations synchronized or real time?

No. Yahoo history, NSE archives and cached responses can have different dates. A latest price-history observation may represent a partial session while delivery belongs to an earlier completed session. Verify dates before comparing them. Failed data requests remain unavailable or may show cached observations; they are not replaced with fabricated prices.

Where can I check the sources?

Explore the six worked screener guides or read about Alpha Nova’s data and storage. These formulas document the current implementation; they are not endorsed by the source providers.