Coverage & Settlement ledger
the limits, stated in writingBottom lineThe limits of every dataset, stated in writing — what is covered, since when, how fresh, and where the boundary sits — so any number you trust here is trusted with its boundary attached. Counts and dates only; no characterisation.
📖 Glossary & methodology → — every custom metric on this site, defined. 🧪 Detection spec-sheets → — every pre-registered study, failures included.
Page as of 2026-10-01 · built 2026-10-03 04:10:41 · scope: NSE-listed Indian equities (EQ/BE/BZ; SME excluded)
This page documents the coverage and known limits of every dataset behind Patearn. Each figure below is a count or a date reproducible from our own tables as of the stamped date. We publish the boundary of each dataset deliberately: a Patearn read should be trusted only to the extent its inputs are, and those inputs are stated here.
Replay the Tape — scrub to a past date, zero look-ahead →Export coverage & provenance memo →Job runs — did each scheduled job run, and how did it end →Evidence pack (print) →Replay any date (live API) →Move anatomy (what precedes a move) →How to read the charts →
Credibility robust core the only delivery-graded set
117
names with ≥10 graded promises — credibility here means tested delivery, not disclosure
Universe (survivorship-correct)
4,718
securities ever observed · 2,962 active · 1,756 delisted/inactive (retained)
Fundamentals availability
modeled
annual +90d / quarterly +50d synthetic lag — not point-in-time (§6)
Credibility settlement funnel — honest, monotone, denominator-first
2,225
concall symbols touched
→
947
scored
→
475
≥1 promise resolved
→
324
≥3 resolved
→
117
robust core ≥10
The headline is the robust core, not the breadth: “symbols touched” counts every name with a concall on file, most of which have few or no resolved promises yet. Coverage shrinks left-to-right; each step is a stricter, more honest count. 947 carry an LLM credibility snapshot; the resolved-promise distribution and tier matrix below count the 943 with a full point-in-time series (level + momentum), so they total 943, not 947 (the 4-name gap is names scored but without a settled series yet).
Universe construction & survivorship policy
| Metric | Count |
|---|---|
| Total securities ever observed | 4,718 |
| Currently listed & active | 2,962 |
| Delisted / inactive (retained) | 1,756 |
| Left-censored at archive floor | 736 |
| Continuity-break events | 236 |
| Renames confirmed (ISIN handover) | 1,076 |
| Rename candidates (unconfirmed) | 8 |
Archive span: 2004-04-17 → 2026-10-01. As-of mechanism: first_date <= d <= last_date (security_master.universe_on)
- Survivorship is correct only from the archive floor 2004-04-17; delistings BEFORE the archive start are not captured (left-censoring).
- Asymmetric: the price/survivorship spine keeps delisted names, but the fundamentals universe is the ~listed-today set — the two universes differ.
- Universe = NSE mainboard EQ/BE/BZ series; SME (SM/ST) excluded.
- Renames auto-confirmed via shared-ISIN clean handover; ISIN is sparse in the historical bhav feed, so rename resolution is biased toward currently-listed names until the authoritative NSE symbol-change feed is loaded.
Our security universe is built from the raw daily bhav-copy archive, not a current-constituents list. Delisted, suspended and surveillance-series names are kept, so historical analysis can see companies that later disappeared rather than only today's survivors. A company is followed across a symbol rename as one continuous security where an ISIN handover confirms it; demergers, mergers and schemes of arrangement are recorded as continuity-break events so a structural gap is never mistaken for a price move. One asymmetry we state plainly: the price/survivorship spine retains delisted names, while the fundamental history covers the names listed today — the two universes are not identical.
Per-data-class coverage matrix
| Dataset | Source | Coverage | Grain | Basis | Latest | Freshness |
|---|---|---|---|---|---|---|
| As-traded equity OHLC | NSE bhav copy | 6,318 symbols · 2.1× today's active setarchive | daily | as traded | 2026-10-01 | current |
| Exchange delivery qty/% | NSE sec_bhavdata_full | 6,318 symbols · 2.1× today's active setarchive | daily | as traded | 2026-10-01 | current |
| Index OHLC + PE/PB/divyield | NSE ind_close_all | 5,749 days | daily | as traded | 2026-10-01 | current |
| Split/bonus/dividend/rights | NSE corp-action feed | 2,842 symbols · 95.9% of active | event | event | 2026-10-28 | current |
| Stock-futures OI / PCR / basis / max-pain | NSE F&O bhav (UDiFF) | 561 symbols · 18.9% of active | daily | as traded | 2026-10-01 | current |
| FII/DII/Pro/Client long-short | NSE participant-wise OI | 3,653 days | market-level | as traded | 2026-10-01 | current |
| Market net FII/DII flows | NSE fiidii | 72 days | market-level | ingested | 2026-10-01 | current |
| Named bulk/block deals | NSE bulk/block feed | 958 symbols · 32.3% of active | event | ingested | 2026-10-01 | current |
| Index constituents + weight | niftyindices | 826 symbols · 27.9% of active | snapshot | ingested | 2026-10-02 | current |
| NSE EQUITY_L allowlist | NSE EQUITY_L.csv | 2,593 symbols · 87.5% of active | snapshot | ingested | 2026-10-02 | current |
| Market news headline | RSS (MC/Mint/ET/BS) | — symbols | event | ingested | 2026-10-02 | current |
| DVPT baselines / R-P / character / key-price | computed | 4,034 symbols · 1.4× today's active setarchive | derived | derived | 2026-10-01 | current |
| Signed accumulation/distribution tape | computed | 4,214 symbols · 1.4× today's active setarchive | derived | derived | 2026-10-01 | current |
| CPR structure (D/W/M) | computed | 2,635 symbols · 89.0% of active | derived | derived | 2026-10-01 | current |
| Stock RS vs broad & sector + rank | computed | 4,034 symbols · 1.4× today's active setarchive | derived | derived | 2026-10-01 | current |
| Index returns/MA/52w + RS phase | computed | 3,610 days | derived | derived | 2026-10-01 | current |
| Strength/direction, RSI-of-RS, Mansfield | computed | 177 series | derived | derived | 2026-10-01 | current |
| RS band % / regime / break state | computed | 131 series | derived | derived | 2026-10-01 | current |
| Down/up capture, down-excess | computed | 190 series | derived | derived | 2026-10-01 | current |
| Typed state-change events | computed | 3,045 symbols · 1.0× today's active setarchive | event | derived | 2026-10-02 | current |
| Current Screener snapshot ratios | Screener.in scrape | 112 symbols · 3.8% of active | snapshot | ingested | 2026-07-16 | current |
| Historical financial time-series | Screener.in scrape (legacy, source IS NULL) + NSE XBRL results (source=NSE-XBRL-*, 2026-07→) | 2,318 symbols · 78.3% of active | quarterly | modeled | 2026-10-01 | current |
| Quarterly shareholding | Screener.in scrape | 2,114 symbols · 71.4% of active | quarterly | modeled | 2026-10-01 | current |
| Business description corpus | Screener.in | 596 symbols · 20.1% of active | snapshot | ingested | 2026-07-16 | current |
| AI business dossier vers | Gemini flash-lite (grounded) | 3,797 symbols · 1.3× today's active setarchive | snapshot | ingested | 2026-06-25 | current |
| Multi-label theme tags | index-seed / AI / human | 384 symbols · 13.0% of active | snapshot | ingested | 2026-09-27 | current |
| Per-symbol news tags vers | rule gazetteer / classifier | 1,302 symbols · 44.0% of active | event | ingested | 2026-10-02 | current |
| 14-pattern patearn score | computed | 279 symbols · 9.4% of active | derived | derived | 2026-10-02 | current |
| Transcript metadata + path | Screener → BSE PDF | 2,225 symbols · 75.1% of active | event | event | Sep 2026 | Sep 2026 |
| Reported quarterly numbers | Screener quarterly table | 117 symbols · 4.0% of active | event | event | Sep 2025 | Sep 2025 |
| LLM guidance/behavior/redflags vers | Gemini Flash on transcript | 943 symbols · 31.8% of active | event | ingested | Sep 2025 | Sep 2025 |
| Promise MET/MISSED/PARTIAL | deterministic vs results | 943 symbols · 31.8% of active | event | event | Sep 2025 | Sep 2025 |
| Credibility score + rank (snapshot) vers | computed PIT | 947 symbols · 32.0% of active | derived | derived | Sep 2017 | Sep 2017 |
| PIT credibility level + momentum + tape vers | computed PIT | 943 symbols · 31.8% of active | derived | derived | Sep 2026 | Sep 2026 |
| Capital-allocation (C) composite | computed | 2,146 symbols · 72.5% of active | derived | derived | 2026-10-02 | current |
Basis — as traded: a real NSE exchange date · ingested: a real first-seen/fetch time · event: a real event date · derived: computed from a real-dated source · modeled: a synthetic uniform lag (see §6). “days” coverage = distinct trading days for market-level/index classes (which have no per-stock split). Every coverage ratio on this table is measured against today’s active set. Where a dataset covers more symbols than are listed today it is shown as a multiple (e.g. “2.1× today’s active set”) rather than a percentage — the archive retains delisted and renamed series (see Universe), so exceeding the live universe is by design, and a percentage above 100 would be meaningless.
Resolved-promise distribution (the honesty bar)
Concall Credibility measures a management's stated guidance against the results that subsequently landed — a measurement of delivery-versus-guidance, not a judgment of management integrity. A company's credibility level becomes meaningful only once its guidance has resolved. Until then the level is held at a fixed ceiling and reflects only the quantification rate — how specific and falsifiable the guidance was — which is a disclosure metric, not evidence of delivery. Roughly one-third of scored names currently have no resolved promises, and many high bands rest on fewer than three. We therefore define a robust core of names with at least ten graded promises, and report credibility tiers for that core only. Momentum is undefined for the first period and stabilises after about four. Every figure carries its as-of call date and its resolved-promise count; verify each against the original transcript.
Robustness
Unproven-ceiling names (no resolved guidance — level capped, not earned): 468
Momentum-ready (≥2 periods): 942 · stable slope (≥4): 883
Tape: 154 earning-trust · 81 deterioration
Tier × resolved count
| Tier | 0 | 1–2 | 3–9 | ≥10 |
|---|---|---|---|---|
| A+ | 0 | 43 | 25 | 7 |
| A | 0 | 15 | 24 | 12 |
| B | 409 | 21 | 39 | 24 |
| C | 48 | 15 | 47 | 40 |
| D | 11 | 57 | 72 | 34 |
Tier × resolved-promise count. A high tier in a low-resolved column is a thin-sample read, shown here on purpose — a tier is only as strong as the count beside it.
Modeled-vs-filed disclosure
The rule:
report_date = period_end + 90d (annual) / 50d (quarterly) — a uniform synthetic lag applied to every company identically. Two failure modes: late filers leak (treated as knowable earlier than they were); early filers are wrongly greyed (treated as unknown when already public).Fundamental history is assigned an availability date using a uniform modelled lag — 90 days after period-end for annual results, 50 days for quarterly — applied identically to every company. This is a modelled approximation of when results became public, not an observed filing date. A company that filed late will appear knowable earlier than it was; one that filed early will appear unknown when it was already public. Accordingly these surfaces are labelled “modeled-availability” and must not be read as point-in-time. A true first-seen date is captured only going forward (it does not exist for historical periods).
Measured modelled-lag error (real first-seen − modelled date): median Noned, p10 Noned / p90 Noned, over — matched periods. Look-ahead-injecting cases (modelled date earlier than real): —.
Replay the Tape — zero look-ahead receipts
WHAT WE MODEL: a fundamental period's report_date is a SYNTHETIC +90d (annual) / +50d (quarterly) lag off period-end — applied identically to every company.
HOW WE DE-MODEL: where a real exchange (BSE) filing date exists we PREFER it (leak 0 by construction); 84.2% of archived periods now carry a real date, and a conservative p95 lag covers the rest. The blended expected leak is 0.95%.
THE PROOF: the effective PIT read uses the real date, so a backtest can never see a datum before it was public. The receipts below show the leak the naive model WOULD have injected — and that the real dates fix it.
Worst would-have-leaked — ATLASCYCLE 2021-06-30 Q: modelled 2021-08-19 vs real 2023-06-09 (+659d)
Representative conservative — SUMMITSEC 2022-03-31 A: modelled 2024-08-26 vs real 2022-05-25 (-824d)
Receipts — the real BSE filing date vs the modelled report date, per period. Positive = modelled earlier than real (a backtest would have seen it early); negative = conservative (greyed while already public).
Every dataset behind Patearn, its origin, refresh cadence and timestamp basis. MODELED = an availability date we estimate (period-end + a uniform lag) until a real exchange filing date is captured; everything else carries an as-traded or ingested timestamp. 36 data classes.
| Key | Dataset | Source | Cadence | Timestamp basis |
|---|---|---|---|---|
| concall_extraction | LLM guidance/behavior/redflags | Gemini Flash on transcript | per transcript | INGESTED |
| company_profile | AI business dossier | Gemini flash-lite (grounded) | periodic | INGESTED |
| equity_universe | NSE EQUITY_L allowlist | NSE EQUITY_L.csv | periodic | INGESTED |
| fno_oi | Stock-futures OI / PCR / basis / max-pain | NSE F&O bhav (UDiFF) | daily ~19:00 IST | AS_TRADED |
| bhav_eq | As-traded equity OHLC | NSE bhav copy | daily ~19:00 IST | AS_TRADED |
| bulk_block_deals | Named bulk/block deals | NSE bulk/block feed | daily (going-forward) | INGESTED |
| corp_action | Split/bonus/dividend/rights | NSE corp-action feed | as announced | EVENT |
| fii_dii_flows | Market net FII/DII flows | NSE fiidii | daily | INGESTED |
| index_ohlc | Index OHLC + PE/PB/divyield | NSE ind_close_all | daily ~19:00 IST | AS_TRADED |
| participant_oi | FII/DII/Pro/Client long-short | NSE participant-wise OI | daily ~19:00 IST | AS_TRADED |
| bhav_delivery | Exchange delivery qty/% | NSE sec_bhavdata_full | daily ~19:00 IST | AS_TRADED |
| news_headline | Market news headline | RSS (MC/Mint/ET/BS) | continuous | INGESTED |
| concall_results | Reported quarterly numbers | Screener quarterly table | per results | EVENT |
| concall_corpus | Transcript metadata + path | Screener → BSE PDF | per results | EVENT |
| company_about | Business description corpus | Screener.in | periodic | INGESTED |
| fundamentals_live | Current Screener snapshot ratios | Screener.in scrape | periodic | INGESTED |
| shareholding_history | Quarterly shareholding | Screener.in scrape | quarterly (modelled) | MODELED |
| fundamentals_history | Historical financial time-series | Screener.in scrape (legacy, source IS NULL) + NSE XBRL results (source=NSE-XBRL-*, 2026-07→) | quarterly (modelled legacy; XBRL rows carry the real broadcast) | MODELED |
| capital_allocation | Capital-allocation (C) composite | computed | nightly | DERIVED |
| capture | Down/up capture, down-excess | computed | nightly | DERIVED |
| cpr_signals | CPR structure (D/W/M) | computed | nightly | DERIVED |
| dvpt_signals | DVPT baselines / R-P / character / key-price | computed | nightly | DERIVED |
| index_signals | Index returns/MA/52w + RS phase | computed | nightly | DERIVED |
| mep_signals | Signed accumulation/distribution tape | computed | nightly | DERIVED |
| pattern_scores | 14-pattern patearn score | computed | on run | DERIVED |
| rs_extras | Strength/direction, RSI-of-RS, Mansfield | computed | nightly | DERIVED |
| rsband | RS band % / regime / break state | computed | nightly | DERIVED |
| signal_events | Typed state-change events | computed | nightly | DERIVED |
| stock_rs | Stock RS vs broad & sector + rank | computed | nightly | DERIVED |
| cci_credibility | Credibility score + rank (snapshot) | computed PIT | per settle | DERIVED |
| cci_series | PIT credibility level + momentum + tape | computed PIT | per settle | DERIVED |
| concall_settlement | Promise MET/MISSED/PARTIAL | deterministic vs results | per results | EVENT |
| theme_tags | Multi-label theme tags | index-seed / AI / human | weekly reseed | INGESTED |
| index_membership | Index constituents + weight | niftyindices | on reconstitution | INGESTED |
| survivorship | Universe construction policy | raw bhav archive | nightly | DERIVED |
| news_symbol_tags | Per-symbol news tags | rule gazetteer / classifier | continuous | INGESTED |
Strategy validation — we test our strategies and report what fails
We test our own strategies and publish what fails. Every strategy we have backtested is recorded with its results net of realistic cost (tier spread + slippage), walk-forward 2012–26, no look-ahead — the failures kept as visible as the wins. The honest verdict so far: nothing that scales beats a Nifty 500 buy-and-hold net of cost. One low-turnover corner does clear the bar once the cost model accounts for the size of the money — and that result is not yet in the run table below, which is itself a defect we are recording rather than hiding. We state all of that plainly, because a research process is only trustworthy if its negative results are on the record — and only honest if its positive ones are too. No performance claim follows from this surface — it is the rigor evidence behind the product, not a strategy lens or a return promise.
Methodology & limitations
No proven performance
Patearn's rankings express a percentile rank-gap — where a name sits on a given factor relative to its peers — not a forecast of returns. No claim of investment performance is made or implied anywhere in this product. The lead-time study that would test whether any of these signals precede price is not yet built; until it is, every score is a descriptor of present, point-in-time evidence, and should be treated as decision-support, not a prediction.
Coverage is missing-not-at-random
Concall coverage is missing-not-at-random: India's transcript mandate is phased by company size and era, so small-cap and earlier-period absence is systematic. A study restricted to names with concalls is implicitly tilted toward larger, more recent companies — read coverage with that selection in mind.
Data source & licensing
Fundamental and concall data are presently obtained from public web sources (including Screener.in and BSE filings); price and delivery data derive from NSE's published bhav-copy archive. These sources are public and each figure links to its origin, but the present collection method is scraped, not a licensed feed. Migration to owned or licensed data sources is planned as part of a backend rebuild ahead of production distribution.
Infrastructure & SLA
Single-node deployment; no high-availability / disaster-recovery today. Data-delivery SLA, freshness monitoring with alerting, and a SOC 2 / security path are on the procurement roadmap, not yet in place.
Regulatory posture
Patearn is an analytical decision-support tool that supports SEBI Research Analyst Regulations workflows (evidence, as-of dating, source linkage). It is informational only, is not investment advice or a recommendation, and is not a substitute for the registrations or reviews that distribution of research to others may require.
Graceful degradation — what an uncovered ticker shows
When you query a name we do not cover for a given dataset, Patearn shows the coverage status and the sample size — never a fabricated value. Absence is reported as absence. Three states are kept distinct: not covered (outside the dataset), zero (a real measured zero), and not in source (the field does not exist in the upstream feed).
RELIANCE — Prices ✓ (daily) · Fundamentals ✓ modeled-availability · Credibility ✓ robust core (n_resolved ≥10) · Participant flow: market-level only, no per-stock split.
A thin micro-cap — Prices ✓ · Fundamentals: not covered (outside the listed-today set) · Credibility: quantification-rate only, n_resolved = 0 (unproven) · F&O: not an F&O underlying.
A thin micro-cap — Prices ✓ · Fundamentals: not covered (outside the listed-today set) · Credibility: quantification-rate only, n_resolved = 0 (unproven) · F&O: not an F&O underlying.
Trust-design principles — the diligence checklist this page pre-empts
1Survivorship addressed first
Universe built from the raw archive with delisted names retained; a survivorship-correct universe-as-of-date exists.
2Point-in-time honesty is graduated
Fundamentals are labelled modeled-availability with the exact lag disclosed; Credibility is genuinely PIT by construction. We never call modelled data point-in-time.
3Sample size travels with every score
No tier or credibility read appears without its resolved-promise count and as-of date; a robust core is defined and tiers reported only within it.
4No performance claim without a backtest
The product speaks in percentile rank-gaps; the lead-time study is openly marked not-built.
5Grain is never overstated
Market-level data (participant OI, FII/DII flow) is labelled market-level; we do not claim per-name participant flow.
6Absence is a first-class value
Coverage gaps return coverage + n, never a fabricated number; “not covered”, “zero” and “not in source” are distinct states.
7Every figure is sourced and reproducible
Each count maps to a named table/column; each company-level claim links to its origin.
8Characterisation is banned
Copy states stated-vs-actual facts, never judgments of people.
9Data provenance is declared
Scraped-today vs licensed-later is stated as a known item, not buried.
10The as-of clock is global and visible
The whole page renders from one stamped moment; freshness is shown per class.