Understand the disclosure.

Ownership persistence and shared holdings

The portfolio subtotal uses the newest exact calendar quarter end with accepted company filings numbering at least 80% of the preceding observed quarter. This coverage threshold keeps a newly opened quarter with only a handful of filings from replacing the prior portfolio subtotal; it does not certify completeness. Newer filings remain visible when their quarter is selected in the map. The portfolio reference period is independent of the map window and company filter. Only companies with one accepted matching named holding at that date enter the company count. For valuation, reported shares are multiplied by a positive, unadjusted official close on or before that date, no more than seven calendar days earlier. The most recent eligible price is selected; NSE takes precedence over BSE only when their dates tie. A dual-listed company is counted once. Arithmetic uses decimal source values, and the CSV retains the exact subtotal, shares, price date, exchange, price-source receipt and filing ID. Displayed rupee values are rounded. Holdings without eligible prices remain in the company count but are excluded from the value. Missing filings, absent names and ambiguous matches are excluded from both; no earlier stake is carried forward. A total with no valued holdings is unavailable, not zero. This retrospective disclosed subtotal is neither a complete portfolio nor a present-day valuation.

The persistence map uses exact calendar quarter ends and one latest accepted whole filing per company and date. Its company scope includes a participant’s accepted reported-name matches at any collected quarter end. Formatting-only candidates do not create a match. A missing filing, no accepted name match, and multiple matching rows are distinct observations. Missing holdings are never filled with earlier values or zeros.

A collected run counts consecutive quarter ends with exactly one matching row. Missing quarters, absent matches and multiple rows break the run. It describes recurring disclosures, not uninterrupted ownership or a holding period. The map’s longest-run figure covers all collected quarters for that company. Cell shading describes reported stake size; arrows compare the same normalized name with the immediately preceding calendar quarter, only when both have one matching row. Share counts remain unadjusted and changes in company capital may affect percentages.

The shared holdings explorer selects one exact quarter end. A company contributes only if the selected participant has one accepted matching row in that filing. Other names are grouped by Unicode, case and whitespace normalization; punctuation variants and reviewed aliases are not merged into peer groups. Repeated peer-name rows within a company and generic category labels retained in older source rows are excluded. Each shared company counts once per peer name, and a participant’s own matched aliases are excluded. The denominator is the selected participant’s comparable companies, not its complete portfolio. Names without a deal-profile match can still appear with their filing evidence.

Co-appearance does not establish coordination, a counterparty, beneficial ownership or any legal relationship. Coverage uses the current collection roster and remains incomplete. Both views use the latest accepted revisions retrospectively; the reporting date is not the publication date. Source filings, original reported names, precise shares and percentages, identity-link basis and checksums accompany CSV exports. Absent observations retain blank numeric fields.

Open the ownership persistence map →   Explore shared holdings →

Independent data checks

The public data quality audit separates retained-source accuracy from historical completeness and reuse rights. It publishes findings, date-level coverage and corrections. A successful ingestion check alone does not establish a complete archive.

Reported values are derived from quantity multiplied by the exchange-reported price, which may be a rounded weighted average. They are not certified trade consideration. Both purchase and sale disclosures contribute to combined reported value.

Reuse and scope

block.deal is an independent explorer, not an exchange service. Exchange reports remain subject to their source terms. Public availability does not itself grant a redistribution licence. Downloaded data includes source provenance so its origin remains visible.

Read the audit and unresolved rights findings →   Inspect actual coverage →   Open sources and downloads →

Daily briefings and pattern screens

Briefings summarize available disclosures for one observed date. Each exchange and deal class has its own report status. An empty export is not proof of a zero-deal day. A newly observed name or company/participant pair means its first appearance in the retained archive, whose earlier coverage may be incomplete.

Screening windows contain 7, 30, 90 or 365 calendar days ending on the latest observed deal date. Repeated purchases count distinct disclosure dates. Several rows on one date count once for this measure. Prior purchases use the immediately preceding window of equal length. The number of company buyers counts distinct reported-name groups in that window.

Participant comparisons use the same dates for every selected name. Common companies indicate disclosed activity in the same security profiles; they do not establish coordination, counterparties, returns or current positions.

Dated shareholding patterns

Named holders come from official NSE XBRL filings and BSE public/promoter tables. The scope covers listed equity-company ISINs, including SME companies. The initial collection uses the latest available filing per company; subsequent accepted quarters and revisions are retained. One complete filing is selected per company, so filings from both exchanges are never added together.

Participant pages show holdings matching the reported name or reviewed aliases. Possible matches based only on formatting appear separately. Neither establishes a verified legal identity. Holdings have individual reporting dates and may be stale. Reporting thresholds, missing filings and differently reported names leave other positions unobserved. An absent match does not mean zero ownership, and the result is not a complete portfolio.

Company shares, ownership percentages, filing dates and original source links accompany each named row. Category aggregates and generic suspense-account labels are excluded. NSE filings are matched using their reported ISIN. BSE table filings are linked to the current listing through the exchange company code; this is an issuer reference, not proof of historical security continuity. Share counts are not adjusted for later corporate actions. The public data does not include PANs or contact details. Failed refreshes retain previous accepted snapshots.

Inspect company filing coverage →   Open the pattern screener →

Unusual activity comparisons

The discovery view covers the last 30 calendar days. Each daily company, exchange, deal-class and reported-side observation is compared with its own preceding 365 calendar days. At least ten prior days with disclosures are required. Missing dates are not inserted as zeros. Current-day and future observations do not enter the baseline.

Reported value, disclosed-side count and distinct reported-name groups each have a prior median, a current-to-median multiple and an empirical percentile. Percentiles use a midrank for tied values. Baseline dates and sample size are shown alongside each result. Incomplete reports can still affect these observations. A high multiple does not establish wrongdoing, statistical significance or an investment signal.

Historical ownership backfill

Available reporting periods are collected from official company indexes, with original source evidence retained. Collection resumes within a bounded request budget. Older unsupported formats, unavailable reports and periods awaiting collection remain gaps; the archive is not a complete historical ownership record. The retained company universe does not reconstruct every formerly listed issuer.

A reporting-date selector displays one accepted filing revision per company and period. Participant pages offer the latest whole filing per company or an exact historical reporting date. Historical selections never carry older positions forward. Coverage lists distinguish a collected filing without a matched name from an uncollected filing, across companies with a name match or formatting candidate at any collected date. This is a retrospective view of collected filings, including later revisions; it does not reconstruct what was publicly known on a past date. Historical shares and percentages are as reported, without adjustment for later corporate actions. Missing named holders cannot be treated as confirmed sales or zero ownership.

Explore unusual activity →   Browse collected reporting dates →

Ownership changes between reporting dates

Comparisons use consecutive available reporting dates for the same company, with one latest accepted revision at each date. The dates, day gap, exchanges and source filings are shown. Intervening periods may be missing. A switch between exchanges can change labels and categories.

Names are compared after Unicode, case and whitespace normalization. Punctuation differences and aliases are not merged across filings. Repeated names within a filing remain unresolved, with all rows retained and no summed or inferred delta. A unique name present in both filings has a reported share-count difference and a percentage-point difference (later percentage minus earlier percentage). Increased, decreased and unchanged labels refer to the percentage stake; share counts may still differ.

A name present on only one date is described as named only in the earlier or later filing. The absent side and change remain unknown, never zero. These are disclosure observations, not confirmed new positions, exits, purchases or sales. Share counts are unadjusted; corporate actions, capital changes, disclosure thresholds and naming differences can affect comparisons.

Explore ownership changes →

Quarterly ownership reviews

Reviews compare exact consecutive calendar quarter ends: 31 March, 30 June, 30 September and 31 December. A company must have a collected filing at both endpoints to enter the change rankings. Earlier available dates and irregular reporting dates are never substituted. Coverage lists distinguish both filings, only the earlier filing, only the later filing, and neither. The denominator is our current collection roster, not a reconstruction of all historically listed companies.

Increases and decreases are ranked by the magnitude of the percentage-point change. Names on only one date are ordered by their reported stake at that date, with the missing side and change left unknown. Repeated names remain unresolved. Counts describe name–company observations, not distinct investors. Percentages across different companies are never added together.

Related bulk/block deals use the same company profile and an unambiguous reported-name group, after the earlier quarter end through the later quarter end inclusive. Purchases and sales remain separate. A reviewed participant alias may link differently reported names; matching does not establish legal identity. Formatting-only candidates and ambiguous filing-name rows are excluded from this attribution. Missing disclosures do not establish zero trading, and activity may repeat across filing names linked to the same group. Do not sum review rows into market totals. These deals provide context and do not explain or reconcile the ownership change.

Reviews use the latest accepted whole filing revision for each date. Historical backfills and later revisions can update these reports. Source links, checksums, exact dates, original named rows and coverage accompany the downloadable JSON and CSV. This is a retrospective review of collected evidence, not a record of everything publicly known at an earlier time.

Browse quarterly ownership reviews →