How screen results are produced

This reference explains how the Research terminal screener turns its coverage universe into the ranked table you read. It covers the four-step sequence the screener runs, how multiple filters combine, how sorting and paging work, the saved views and hand-offs that load a ready-made screen, and which outputs are standard finance measures versus GNG Research's proprietary models. For the controls themselves, see Screen and filter. For the columns, see Screening columns and data.

Reference

Conceptually the screener does four things in sequence, then lets you read and refine the output.

Step 1: build the universe

The screener starts from GNG Research's full coverage of listed stocks, ETFs, and preferred shares, with one consolidated, regularly refreshed row of metrics per security. The universe is roughly 5,600 securities. Source data is drawn from AlphaVantage for prices and fundamentals, Nasdaq quotes delivered through AlphaVantage, and public company filings from the U.S. Securities and Exchange Commission sourced through EDGAR, then enriched with GNG Research's own computed metrics and ratings.

Step 2: apply the filters

Every filter you set becomes a condition on that universe.

Numeric ranges apply as greater-than and less-than bounds. Leaving one side blank means no bound on that side.
Multi-selects apply as "is one of these values."
Dates apply as a from-and-to window.
Text applies as "contains."
An optional exclude-N/A condition drops rows missing that metric.

How the conditions combine matters.

Multiple filters across different fields combine with AND. A company must satisfy all of them.
Multiple values inside one multi-select combine with OR. A company matches if it has any of the chosen values.

The search bar layers a partial match on symbol, name, or sector on top of the filters. The portfolio selector adds a symbol filter that restricts the universe to the holdings of one or more of your own portfolios.

Step 3: sort and page

Matching rows are ordered by the column you chose to sort on, with a symbol tie-breaker that keeps paging stable. Rating-style columns sort in their natural good-to-bad order rather than alphabetically. Rows are returned a page at a time so even large result sets stay fast, and more rows stream in as you scroll. The page size follows your tier setting.

Step 4: present

Only the columns you selected are rendered, each formatted to its natural unit. Optional per-column summary statistics (sum, average, median, count) are computed over the rows currently loaded and shown in a summary row under the header.

How multiple filters combine

The combination rule decides how strict a screen is.

Stacking filters across fields narrows the result, because each added field is another AND condition. A screen with a yield floor, a payout-ratio ceiling, and a sector list returns only companies that pass all three.
Adding values inside one multi-select widens that field, because the values are joined with OR. Selecting three sectors returns companies in any of the three.

To loosen a screen, remove a field or blank one side of a range. To tighten it, add a field or raise a floor.

Standard metrics versus proprietary outputs

The screener mixes two kinds of columns, and they are documented differently.

The standard, public columns are conventional finance and technical-analysis measures computed on company financials and price history. Their meaning is what any finance reference would give, and they are defined in full in Data dictionary.

The GNG proprietary outputs are GNG Research's own models: the Safety Score, the Quality Score, the seven-tier Rating, the GNG Quant Score and Rating, the blended Fair Value and its discount, and the buy-zone and trim prices. For those, this documentation tells you what the output measures, the input categories it draws on, its scale, and how to read it, but the exact formulas, weights, coefficients, and thresholds are proprietary and are not published. In broad terms: the Safety and Quality scores express dividend safety and business quality on a 0-to-100 scale where higher is stronger; the Rating places a security into one of seven valuation bands from the strongest buy level down to a trim level; the GNG Quant Score is a 0-to-100 quantitative composite with a matching Strong Buy to Strong Sell rating; the blended Fair Value combines several per-method valuation estimates into a single estimate; and the buy-zone and trim prices are quality-adjusted thresholds derived from that fair value. Use them as a starting filter and a relative ranking, not as a guarantee. For how to read each, see Ratings overview and Fair value and valuation.

Saved views

A saved view stores the full current configuration so you can reload the exact screen in one click. A view captures the search text, the filters and custom filters, the selected columns, the sort, the page size, and the summary toggles.

Free accounts include one saved view. Pro and above include unlimited saved views.
From the Views list you can load, copy, rename, edit, or delete a view.
You can copy one of GNG Research's shared global views into your own to use as a starting point.
While a view is loaded, an indicator shows its name and warns you of unsaved changes with an Update View button. If you open a ticker with unsaved changes, the screener asks whether to save first.

Hand-offs into the screener

The screener can receive a ready-made screen specification from elsewhere in the product and load it as a view. For example, an AI-generated screen or a "research these holdings" hand-off from a portfolio arrives as a prepared filter-and-column layout you can then refine.

Export

On Pro, the result can be exported as CSV, Excel, or PDF, or opened in a connected Google Sheet. Export respects your tier: locked premium columns are stripped from the file, and exports are metered, so Pro includes a fixed number of exports per week while Free and Pro Lite cannot export.

Data and timing

The universe refreshes on a schedule, largely daily for fundamentals, so figures reflect the latest processed snapshot rather than live intraday values. Summary statistics are computed over the rows currently loaded into the table, not necessarily the entire matched universe at once.

Limitations and disclosures

Because data refreshes on a schedule, a screen reflects the most recent processed snapshot, not real-time prices. Summary statistics cover only the loaded rows. Screening surfaces candidates and rankings, it does not make recommendations. The proprietary scores and fair value figures read as guidance, so this surface carries the disclosure below.

GNG Research provides equity research and educational tools, not investment advice. Nothing on the platform is a recommendation to buy or sell any security. Do your own research and consider your circumstances before making any investment decision.