{"status":"monthly audit","reproduce":{"rule":"ORDER BY md5('\u003cseed\u003e|\u003cstratum\u003e|' || \u003crow id\u003e) within each stratum's population filter; first n rows","seed":"e94c0ff9bc780226f1d53df8931de7a3","window":["2026-09-01T00:00:00Z","2026-09-27T23:55:28Z"],"statement":"Every audited row, its frozen evidence and both labels are in the data-room CSV for this month.","sampler_version":"v1"},"audit_month":"2026-09","limitations":["Completeness is measured only against the candidate companies the evidence generator surfaces (name similarity and shared identifiers). A missed match with no shared name or identifier is invisible to it, so completeness is an upper bound and the duplicate rate a lower bound.","Two different companies with the same name are the dominant precision failure; a name match with no second identifier is the weakest link method and is labelled as such.","WARN notices, permits and trademarks mostly concern employers that are not in the company spine at all; their low link rate is a coverage fact, not a matching defect, and their precision is measured separately.","Investor identity has no registry identifier (no CIK/LEI for most funds), and only a short summary of the source article is stored, so many investor-to-round links cannot be decided from stored evidence and are reported as unknown.","The raters are AI models; agreement between two models is not the same as agreement with a human expert.","Single-lane samples of 30 to 80 rows give intervals roughly 6 to 15 points wide; the cross-lane estimate is tighter, and single-lane figures are indicative. Differences between months within overlapping intervals are not evidence of change.","Prediction calibration is not measurable until each prediction's horizon (90 or 180 days) has closed after snapshots began.","The second rater (claude-sonnet-5, prompt rater-v2) was run through Claude Code sessions with the exact published prompt, not through the metered API call the monthly job uses, so its spend is not in the usage ledger. The model and prompt are the same.","The adjudicator (Claude Opus 5.5) is the same model family as the primary rater and sided with it on most disagreements; agreement between the primary and the adjudicated label is therefore not an independent check.","The acquired-company stratum was re-drawn with the same seed after 524 placeholder links were removed; the pre-cleanup result is published beside it in the sample notes."],"methodology":{"strata":"Precision strata sample event-to-company links per data lane (funding news, Form D, executive hires, contracts, acquisitions by role, product launches, job postings, WARN notices, permits, 8-K filings, trademarks) and investor-to-round links. Completeness strata sample records that were NOT linked and companies that were newly created, and ask whether an existing company was missed. Pair strata sample the current output of each duplicate-candidate generator. A duplicate stratum samples companies uniformly at random and asks whether each has a duplicate row.","sampling":"Every month a stratified random sample is drawn from the links Fundz made in that month. The seed is recorded before any row is labelled, and the order of rows within each stratum is md5(seed | stratum | row id), so anyone with the seed can re-draw exactly the same rows. The evidence each rater sees is frozen at draw time with a SHA-256 checksum, so a label can be re-checked even after the underlying data changes.","labelling":"Each row is labelled by a primary rater from the frozen evidence. A second rater, a language model (disclosed by model name), labels the same rows blind to the primary label. Agreement is reported per stratum as raw agreement and Cohen's kappa. Every disagreement is re-read and adjudicated with a written note; the adjudicated label is final. Labels are append-only.","automation":"An automated analyst runs daily checks over new links and the duplicate backlog. A check may change data without a person only when its precision on a frozen, versioned gold set is at least 0.97 over at least 30 labelled rows, the change is reversible, it touches no company on an active customer's watchlist or lead list, and it is within a per-run cap. Everything else goes to a human review queue.","statistics":"Proportions are published with 95% Wilson score intervals. The cross-lane figure is a population-weighted stratified estimate with a finite-population correction. Rows a rater could not decide are reported as unknown and excluded from the denominator, never counted as correct.","predictions":"Predicted probabilities are frozen daily in an append-only snapshot (a fixed control panel of companies chosen by a hash of their id, independent of any score, plus the top-ranked companies per target and the rows each list actually showed). Once a prediction's horizon has closed, the realised rate is compared with the predicted probability per probability bucket (a reliability table), and the Brier score is reported against the base rate.","publication":"A metric is published only with its sample size, its 95% interval and its measurement month, and only when its sample is at least 30 decided rows. Anything else is shown as not yet measured, with the reason. Limitations are always published beside the numbers.","independence":"The primary rater for the first samples is Claude (an AI model) working in supervised sessions, and the second rater is also a model. No human labeller has labelled these samples. For that reason every audited row, its evidence and both labels are published, and a buyer is invited to draw and label a fresh sample of their own with the published seed procedure."},"predictions":{"ledger":{"note":"ledger grading covers funding only; the v3 partitions begin 2026-08-23 and none has reached its horizon","graded_partitions":[{"model":"fundzscore_v2","target":"funding","horizon":"60.0 days","by_cohort":{"top":{"n":2000,"brier":0.07344,"status":"measured","base_rate":0.08,"brier_skill":0.0022,"reliability":[{"n":0,"bucket":[0.0,0.01],"realised":{"k":0,"n":0,"ci95":null,"status":"not_measured","estimate":null},"mean_predicted":null},{"n":0,"bucket":[0.01,0.02],"realised":{"k":0,"n":0,"ci95":null,"status":"not_measured","estimate":null},"mean_predicted":null},{"n":0,"bucket":[0.02,0.05],"realised":{"k":0,"n":0,"ci95":null,"status":"not_measured","estimate":null},"mean_predicted":null},{"n":1308,"bucket":[0.05,0.1],"realised":{"k":83,"n":1308,"ci95":[0.0515,0.078],"status":"measured","estimate":0.0635},"mean_predicted":0.0861},{"n":574,"bucket":[0.1,0.2],"realised":{"k":54,"n":574,"ci95":[0.0728,0.1207],"status":"measured","estimate":0.0941},"mean_predicted":0.1306},{"n":118,"bucket":[0.2,0.5],"realised":{"k":23,"n":118,"ci95":[0.1335,0.2755],"status":"measured","estimate":0.1949},"mean_predicted":0.21},{"n":0,"bucket":[0.5,1.0],"realised":{"k":0,"n":0,"ci95":null,"status":"not_measured","estimate":null},"mean_predicted":null}],"brier_climatology":0.0736},"random":{"n":2000,"brier":0.0084,"status":"measured","base_rate":0.0085,"brier_skill":0.0032,"reliability":[{"n":1453,"bucket":[0.0,0.01],"realised":{"k":5,"n":1453,"ci95":[0.0015,0.008],"status":"measured","estimate":0.0034},"mean_predicted":0.0022},{"n":308,"bucket":[0.01,0.02],"realised":{"k":6,"n":308,"ci95":[0.009,0.0418],"status":"measured","estimate":0.0195},"mean_predicted":0.0139},{"n":236,"bucket":[0.02,0.05],"realised":{"k":6,"n":236,"ci95":[0.0117,0.0543],"status":"measured","estimate":0.0254},"mean_predicted":0.0361},{"n":3,"bucket":[0.05,0.1],"realised":{"k":0,"n":3,"ci95":[0.0,0.5615],"status":"measured","estimate":0.0},"mean_predicted":0.058},{"n":0,"bucket":[0.1,0.2],"realised":{"k":0,"n":0,"ci95":null,"status":"not_measured","estimate":null},"mean_predicted":null},{"n":0,"bucket":[0.2,0.5],"realised":{"k":0,"n":0,"ci95":null,"status":"not_measured","estimate":null},"mean_predicted":null},{"n":0,"bucket":[0.5,1.0],"realised":{"k":0,"n":0,"ci95":null,"status":"not_measured","estimate":null},"mean_predicted":null}],"brier_climatology":0.00843}},"partition":"fundzscore_v2_holdout","scored_on":"2026-06-01"}],"ungraded_rows_by_model":{"fundzscore_v3":136333,"fundzscore_v1_opco":4000}},"report":{"model":"fundzscore_v3","ledger":null,"targets":{"funding":{"reason":"no snapshot has passed its 180-day horizon + 90-day grace yet","status":"not_yet_measured","target":"funding","horizon":"180 days","surface":"Likely to raise next","definition":"company raises a funding round within 180 days","grace_days":90,"available_from":"2027-06-24"},"acquired":{"reason":"no snapshot has passed its 90-day horizon + 90-day grace yet","status":"not_yet_measured","target":"acquired","horizon":"90 days","surface":"Likely to be acquired","definition":"company is acquired within 90 days","grace_days":90,"available_from":"2027-03-26"},"acquires":{"reason":"no snapshot has passed its 180-day horizon + 90-day grace yet","status":"not_yet_measured","target":"acquires","horizon":"180 days","surface":"M\u0026A brief: Acquires N% in 6 mo","definition":"company acquires a company within 180 days","grace_days":90,"available_from":"2027-06-24"},"exec_hire":{"reason":"no snapshot has passed its 90-day horizon + 90-day grace yet","status":"not_yet_measured","target":"exec_hire","horizon":"90 days","surface":"company page next-event panel","definition":"company hires an executive within 90 days","grace_days":90,"available_from":"2027-03-26"}},"not_measurable":[{"reason":"the retired /prod/predict model output is not stored anywhere; it cannot be backtested and should be retired or snapshotted","surface":"legacy \"Next predicted event\" panel (funding and acquisition pages)"}]}},"audit_sample":"2026-09b","generated_at":"2026-09-28T00:49:27Z","user_reports":{"open":0,"reports":0,"subjects":0,"confirmed_share":{"n":0,"note":"confirmed by an analyst; only as good as that read","unit":"proportion","label":"share of resolved customer reports confirmed as real errors","min_n":10,"reason":"no labelled sample yet","status":"not_yet_measured"}},"schema_version":"fundz-data-quality/1","automated_checks":{"gate":"auto-apply only at gold precision \u003e= 0.97 over \u003e= 30 fired rows","checks":[{"gate":"eligible","gold":{"k":98,"n":98,"ci95":[0.9623,1],"unit":"proportion","as_of":"2026-09-27","label":"C1 gold precision","min_n":30,"status":"published","estimate":1.0},"check":"C1","description":"same parser_id, 2+ canonical orgs","automated_fix":true},{"gate":"blocked","gold":{"k":22,"n":35,"ci95":[0.4634,0.7683],"unit":"proportion","as_of":"2026-09-27","label":"C2 gold precision","min_n":30,"status":"published","estimate":0.6286},"check":"C2","description":"same CIK after trimming zeros","automated_fix":true},{"gate":"review only","gold":{"k":32,"n":38,"ci95":[0.6958,0.9256],"unit":"proportion","as_of":"2026-09-27","label":"C3 gold precision","min_n":30,"status":"published","estimate":0.8421},"check":"C3","description":"same normalised name (needs a 2nd signal)","automated_fix":false},{"gate":"review only","gold":{"k":19,"n":44,"ci95":[0.2968,0.5778],"unit":"proportion","as_of":"2026-09-27","label":"C4 gold precision","min_n":30,"status":"published","estimate":0.4318},"check":"C4","description":"shared domain (never merged on this alone)","automated_fix":false},{"gate":"review only","gold":{"n":4,"unit":"proportion","label":"C5 gold precision","min_n":30,"reason":"sample n=4 is below the publication minimum n=30","status":"not_yet_measured"},"check":"C5","description":"shared LinkedIn company slug","automated_fix":false},{"gate":"review only","gold":{"n":7,"unit":"proportion","label":"C6 gold precision","min_n":30,"reason":"sample n=7 is below the publication minimum n=30","status":"not_yet_measured"},"check":"C6","description":"event text names the linked org (role check for acquisitions)","automated_fix":false},{"gate":"review only","gold":{"n":1,"unit":"proportion","label":"C7 gold precision","min_n":30,"reason":"sample n=1 is below the publication minimum n=30","status":"not_yet_measured"},"check":"C7","description":"8-K filing CIK vs parser company CIK vs org CIK","automated_fix":false},{"gate":"review only","gold":{"unit":"proportion","label":"C8 gold precision","min_n":30,"reason":"no labelled gold rows fire this check","status":"not_yet_measured"},"check":"C8","description":"EDGAR company name for the CIK vs org name","automated_fix":false},{"gate":"review only","gold":{"unit":"proportion","label":"C9 gold precision","min_n":30,"reason":"no labelled gold rows fire this check","status":"not_yet_measured"},"check":"C9","description":"homepage title / og:site_name vs org name; redirects; parked/scanner hosts","automated_fix":false},{"gate":"eligible","gold":{"k":76,"n":76,"ci95":[0.9519,1.0],"unit":"proportion","as_of":"2026-09-27","label":"C10 gold precision","min_n":30,"status":"published","estimate":1.0},"check":"C10","description":"domain accretion (3+ unrelated roots) and junk hosts","automated_fix":true},{"gate":"review only","gold":{"unit":"proportion","label":"C11 gold precision","min_n":30,"reason":"no labelled gold rows fire this check","status":"not_yet_measured"},"check":"C11","description":"LinkedIn slug consistent with name/domain (Apollo accounts[0] fallback)","automated_fix":false},{"gate":"review only","gold":{"n":3,"unit":"proportion","label":"C12 gold precision","min_n":30,"reason":"sample n=3 is below the publication minimum n=30","status":"not_yet_measured"},"check":"C12","description":"WARN notice name vs org, with state agreement","automated_fix":false},{"gate":"review only","gold":{"n":6,"unit":"proportion","label":"C13 gold precision","min_n":30,"reason":"sample n=6 is below the publication minimum n=30","status":"not_yet_measured"},"check":"C13","description":"investor alias clusters / person-as-investor / new row duplicates an old one","automated_fix":false},{"gate":"review only","gold":{"unit":"proportion","label":"C14 gold precision","min_n":30,"reason":"no labelled gold rows fire this check","status":"not_yet_measured"},"check":"C14","description":"investor named in the source article (NOT MEASURABLE: only a ~400-char summary is stored)","automated_fix":false},{"gate":"review only","gold":{"unit":"proportion","label":"C15 gold precision","min_n":30,"reason":"no labelled gold rows fire this check","status":"not_yet_measured"},"check":"C15","description":"ingest churn: a new org whose key matches an older canonical","automated_fix":false},{"gate":"review only","gold":{"unit":"proportion","label":"C16 gold precision","min_n":30,"reason":"no labelled gold rows fire this check","status":"not_yet_measured"},"check":"C16","description":"funding round's own EDGAR link names another org's CIK (review only)","automated_fix":false}],"gold_version":"gold_v1"},"entity_resolution":{"precision":{"lanes":[{"k":72,"n":77,"ci95":[0.8568,0.9719],"note":"primary/model agreement 97% (kappa 0.8207) over 77 rows","unit":"proportion","as_of":"2026-09","label":"funding rounds (news)","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_funding","estimate":0.9351,"population":1706,"company_family":{"k":72,"n":77,"ci95":[0.8568,0.9719],"unit":"proportion","as_of":"2026-09","label":"funding rounds (news) (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":0.9351}},{"k":77,"n":77,"ci95":[0.9525,1],"note":"primary/model agreement 99% (kappa 0.0) over 77 rows","unit":"proportion","as_of":"2026-09","label":"Form D filings","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_form_d","estimate":1.0,"population":1207,"company_family":{"k":77,"n":77,"ci95":[0.9525,1],"unit":"proportion","as_of":"2026-09","label":"Form D filings (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":1.0}},{"k":72,"n":75,"ci95":[0.8889,0.9863],"note":"primary/model agreement 96% (kappa 0.7145) over 77 rows","unit":"proportion","as_of":"2026-09","label":"executive hires","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_exec_hire","estimate":0.96,"population":2132,"company_family":{"k":73,"n":75,"ci95":[0.9079,0.9927],"unit":"proportion","as_of":"2026-09","label":"executive hires (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":0.9733}},{"k":71,"n":76,"ci95":[0.8551,0.9716],"note":"primary/model agreement 91% (kappa 0.4266) over 77 rows","unit":"proportion","as_of":"2026-09","label":"contracts / agreements","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_contract","estimate":0.9342,"population":2418,"company_family":{"k":72,"n":76,"ci95":[0.8723,0.9793],"unit":"proportion","as_of":"2026-09","label":"contracts / agreements (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":0.9474}},{"k":72,"n":77,"ci95":[0.8568,0.9719],"note":"primary/model agreement 94% (kappa 0.3573) over 77 rows","unit":"proportion","as_of":"2026-09","label":"acquisitions: acquirer","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_acq_acquirer","estimate":0.9351,"population":2851,"company_family":{"k":72,"n":77,"ci95":[0.8568,0.9719],"unit":"proportion","as_of":"2026-09","label":"acquisitions: acquirer (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":0.9351}},{"k":62,"n":75,"ci95":[0.7257,0.8958],"note":"primary/model agreement 92% (kappa 0.7416) over 77 rows","unit":"proportion","as_of":"2026-09","label":"acquisitions: acquired company","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_acq_acquiree","estimate":0.8267,"population":2762,"company_family":{"k":63,"n":75,"ci95":[0.7408,0.906],"unit":"proportion","as_of":"2026-09","label":"acquisitions: acquired company (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":0.84}},{"k":66,"n":77,"ci95":[0.762,0.9183],"note":"primary/model agreement 94% (kappa 0.7281) over 77 rows","unit":"proportion","as_of":"2026-09","label":"product launches","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_product_launch","estimate":0.8571,"population":1377,"company_family":{"k":66,"n":77,"ci95":[0.762,0.9183],"unit":"proportion","as_of":"2026-09","label":"product launches (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":0.8571}},{"k":69,"n":76,"ci95":[0.8219,0.9547],"note":"primary/model agreement 95% (kappa 0.7573) over 77 rows","unit":"proportion","as_of":"2026-09","label":"job postings (ATS)","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_job_posting","estimate":0.9079,"population":145948,"company_family":{"k":74,"n":76,"ci95":[0.909,0.9928],"unit":"proportion","as_of":"2026-09","label":"job postings (ATS) (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":0.9737}},{"k":34,"n":39,"ci95":[0.7329,0.944],"note":"primary/model agreement 90% (kappa 0.5939) over 40 rows","unit":"proportion","as_of":"2026-09","label":"WARN layoff notices","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_warn","estimate":0.8718,"population":40,"company_family":{"k":38,"n":39,"ci95":[0.8682,0.9955],"unit":"proportion","as_of":"2026-09","label":"WARN layoff notices (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":0.9744}},{"k":63,"n":72,"ci95":[0.7792,0.9328],"note":"primary/model agreement 92% (kappa 0.7215) over 77 rows","unit":"proportion","as_of":"2026-09","label":"office lease permits","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_permit","estimate":0.875,"population":107,"company_family":{"k":66,"n":72,"ci95":[0.8299,0.9612],"unit":"proportion","as_of":"2026-09","label":"office lease permits (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":0.9167}},{"k":73,"n":76,"ci95":[0.8903,0.9865],"note":"primary/model agreement 95% (kappa 0.314) over 77 rows","unit":"proportion","as_of":"2026-09","label":"SEC 8-K filings","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_8k","estimate":0.9605,"population":846,"company_family":{"k":74,"n":76,"ci95":[0.909,0.9928],"unit":"proportion","as_of":"2026-09","label":"SEC 8-K filings (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":0.9737}},{"k":63,"n":76,"ci95":[0.729,0.8972],"note":"primary/model agreement 91% (kappa 0.6868) over 77 rows","unit":"proportion","as_of":"2026-09","label":"trademarks (USPTO)","min_n":30,"method":"share of sampled links judged correct (strict: exact legal entity)","status":"published","stratum":"p_trademark","estimate":0.8289,"population":5122,"company_family":{"k":70,"n":76,"ci95":[0.8383,0.9633],"unit":"proportion","as_of":"2026-09","label":"trademarks (USPTO) (company family)","min_n":30,"method":"share of sampled links to the right company or its parent, subsidiary or sibling","status":"published","estimate":0.9211}}],"overall":{"k":null,"n":873,"ci95":[0.844,0.9687],"unit":"proportion","as_of":"2026-09","label":"event-to-company links, all lanes (population-weighted, strict)","min_n":30,"method":"stratified, population-weighted, Agresti-Coull variance with FPC","status":"published","estimate":0.9064},"definition":"strict: a link counts as correct only if it is the exact legal entity; company_family: a link to the parent, subsidiary or sibling of the right company also counts as correct","company_family":{"overall":{"k":null,"n":873,"ci95":[0.9245,1],"unit":"proportion","as_of":"2026-09","label":"event-to-company links, all lanes (population-weighted, company family)","min_n":30,"method":"stratified, population-weighted, Agresti-Coull variance with FPC","status":"published","estimate":0.9676},"overall_excluding_job_postings":{"k":null,"n":797,"ci95":[0.8995,0.9492],"unit":"proportion","as_of":"2026-09","label":"event-to-company links excluding job postings (population-weighted, company family)","min_n":30,"method":"stratified, population-weighted, Agresti-Coull variance with FPC","status":"published","estimate":0.9244},"related_entity_links_in_sample":23},"label_set_version":"labels-v2","overall_excluding_job_postings":{"k":null,"n":797,"ci95":[0.867,0.9246],"unit":"proportion","as_of":"2026-09","label":"event-to-company links excluding job postings (population-weighted, strict)","min_n":30,"method":"stratified, population-weighted, Agresti-Coull variance with FPC","status":"published","estimate":0.8958}},"duplicates":{"generators":[{"k":16,"n":19,"ci95":[0.6243,0.9448],"unit":"proportion","as_of":"2026-09","label":"same normalised name","min_n":15,"method":"share of sampled candidate pairs that are the same company","status":"published","stratum":"x_pairs_same_name","estimate":0.8421,"population":13517},{"k":11,"n":17,"ci95":[0.413,0.8269],"unit":"proportion","as_of":"2026-09","label":"shared website domain","min_n":15,"method":"share of sampled candidate pairs that are the same company","status":"published","stratum":"x_pairs_shared_domain","estimate":0.6471,"population":8235},{"k":14,"n":20,"ci95":[0.481,0.8545],"unit":"proportion","as_of":"2026-09","label":"shared LinkedIn company page","min_n":15,"method":"share of sampled candidate pairs that are the same company","status":"published","stratum":"x_pairs_linkedin","estimate":0.7,"population":1467},{"k":20,"n":20,"ci95":[0.8389,1.0],"unit":"proportion","as_of":"2026-09","label":"shared parser company id","min_n":15,"method":"share of sampled candidate pairs that are the same company","status":"published","stratum":"x_pairs_parser_id","estimate":1.0,"population":5549},{"k":12,"n":18,"ci95":[0.4375,0.8372],"unit":"proportion","as_of":"2026-09","label":"same CIK after zero-padding","min_n":15,"method":"share of sampled candidate pairs that are the same company","status":"published","stratum":"x_pairs_cik","estimate":0.6667,"population":683}],"prevalence":{"k":14,"n":74,"ci95":[0.1162,0.2929],"note":"lower bound: only generator-surfaced candidates are checked","unit":"proportion","as_of":"2026-09","label":"canonical companies with a duplicate row","min_n":30,"method":"uniform random sample of canonical non-fund companies","status":"published","estimate":0.1892}},"completeness":[{"k":5,"n":26,"ci95":[0.0851,0.3788],"unit":"proportion","as_of":"2026-09","label":"WARN (no org linked)","min_n":20,"method":"share of unlinked records that match an existing company","status":"published","stratum":"r_unlinked_warn","estimate":0.1923,"population":176},{"k":1,"n":25,"ci95":[0.0071,0.1954],"unit":"proportion","as_of":"2026-09","label":"permits (no org linked)","min_n":20,"method":"share of unlinked records that match an existing company","status":"published","stratum":"r_unlinked_permit","estimate":0.04,"population":1821},{"k":3,"n":22,"ci95":[0.0475,0.3334],"unit":"proportion","as_of":"2026-09","label":"trademarks (no org linked)","min_n":20,"method":"share of unlinked records that match an existing company","status":"published","stratum":"r_unlinked_trademark","estimate":0.1364,"population":54511},{"k":39,"n":40,"ci95":[0.8712,0.9956],"note":"Measured on the parser-company path (sampler v1). Since 2026-09-27 the product resolves 8-K filings by CIK first, which this figure does not measure; it describes the gap that change closes.","unit":"proportion","as_of":"2026-09","label":"8-K (filer resolves to no org)","min_n":20,"method":"share of unlinked records that match an existing company","status":"published","stratum":"r_unresolved_8k","estimate":0.975,"population":1036},{"k":3,"n":42,"ci95":[0.0246,0.1901],"unit":"proportion","as_of":"2026-09","label":"organizations created in the window","min_n":20,"method":"share of newly created companies that duplicate an existing one","status":"published","stratum":"r_new_org","estimate":0.0714,"population":4445}],"investor_links":{"k":27,"n":34,"ci95":[0.632,0.8965],"note":"43 of 77 sampled links could not be decided from stored evidence","unit":"proportion","as_of":"2026-09","label":"investor-to-round links","min_n":30,"status":"published","estimate":0.7941},"rater_agreement":{"all":{"n":1384,"raw":0.9342,"kappa":0.8939,"status":"measured","raw_ci95":[0.92,0.9461],"kappa_ci95":[0.8728,0.915],"interpretation":"almost perfect"},"by_kind":{"pairs":{"n":100,"raw":0.87,"kappa":0.7172,"status":"measured","raw_ci95":[0.7902,0.9224],"kappa_ci95":[0.5738,0.8606],"interpretation":"substantial"},"recall":{"n":220,"raw":0.9773,"kappa":0.9641,"status":"measured","raw_ci95":[0.9479,0.9903],"kappa_ci95":[0.933,0.9952],"interpretation":"almost perfect"},"duplicate":{"n":100,"raw":0.96,"kappa":0.8543,"status":"measured","raw_ci95":[0.9016,0.9843],"kappa_ci95":[0.7145,0.9942],"interpretation":"almost perfect"},"precision":{"n":964,"raw":0.9284,"kappa":0.7134,"status":"measured","raw_ci95":[0.9104,0.9431],"kappa_ci95":[0.6483,0.7786],"interpretation":"substantial"}},"polarity":{"n":1384,"raw":0.9371,"kappa":0.8122,"status":"measured","raw_ci95":[0.9231,0.9488],"kappa_ci95":[0.774,0.8504],"interpretation":"almost perfect"},"second_rater":"language model (named in methodology.labelling of the document); see limitations"},"investor_links_source_verification":{"bias":"Only 160 of the 320 sampled links could be measured. Links whose round source is BusinessWire, FinSMEs or a Google News redirect were excluded (the page could not be fetched or decoded into article text), as were links with no stored source URL or an unreadable page. The measured half therefore over-represents rounds reported by sources that serve plain article text, and the excluded sources may have a different precision. It is a second, independent measurement of investor-link precision, not a replacement for the blind audit's p_investor_link stratum (which could decide only 16 of 30 from stored evidence).","seed":"20260928","label":"investor-to-round links, checked against the round's source article","rater":"Claude model (Opus 5.5) session of the investor-identity build; not a human","sampled":320,"measured":160,"instrument":"Investors::LinkVerifier (app/services/investors/link_verifier.rb), run(n: 320, seed: 20260928); then a hand read of every measured link","measured_on":"2026-09-28","hand_read_strict":{"k":149,"n":159,"ci95":[0.8881,0.9655],"unit":"proportion","as_of":"2026-09-28","label":"hand read, strict","min_n":30,"method":"a reader of the article judged that this investor participated in THIS round (one of the 160 was undecidable)","status":"published","estimate":0.9371},"hand_read_lenient":{"k":153,"n":159,"ci95":[0.9201,0.9826],"unit":"proportion","as_of":"2026-09-28","label":"hand read, lenient","min_n":30,"method":"the same hand read with borderline links counted as correct (the investor-identity build's lenient reading; 4 links differ from strict)","status":"published","estimate":0.9623},"article_mentions_investor":{"k":157,"n":160,"ci95":[0.9463,0.9936],"unit":"proportion","as_of":"2026-09-28","label":"investor named in the article","min_n":30,"method":"the investor's name (or a curated alias) appears in a readable copy of the round's stored source article; proves NAMED, not that it invested in this round","status":"published","estimate":0.9813}}},"publication_rules":{"min_n":30,"statement":"A metric is published only with its sample size, its 95% interval and its measurement month, and only when its sample is at least 30 decided rows. Anything else is shown as not yet measured, with the reason. Limitations are always published beside the numbers."},"identifier_coverage":{"note":"canonical, non-fund organizations; domains are the union of both stores","as_of":"2026-09-28","all_canonical_non_fund":{"fields":{"cik":{"k":60366,"n":233503,"estimate":0.2585},"ein":{"k":48769,"n":233503,"estimate":0.2089},"lei":{"k":5416,"n":233503,"estimate":0.0232},"domain":{"k":184870,"n":233503,"estimate":0.7917},"linkedin":{"k":97283,"n":233503,"estimate":0.4166}},"definition":"all canonical non fund","population":233503},"with_event_last_12_months":{"fields":{"cik":{"k":5158,"n":16460,"estimate":0.3134},"ein":{"k":4031,"n":16460,"estimate":0.2449},"lei":{"k":805,"n":16460,"estimate":0.0489},"domain":{"k":14560,"n":16460,"estimate":0.8846},"linkedin":{"k":12219,"n":16460,"estimate":0.7423}},"definition":"with event last 12 months","population":16460}},"provenance_coverage":{"lanes":[{"lane":"funding rounds + Form D","links":2913,"share":{"estimate":0.0},"reason":"no method column on this lane (the link records WHAT, not HOW)","with_method":0},{"lane":"executive hires","links":2134,"share":{"estimate":0.0},"reason":"no method column on this lane (the link records WHAT, not HOW)","with_method":0},{"lane":"contracts","links":2421,"share":{"estimate":0.0},"reason":"no method column on this lane (the link records WHAT, not HOW)","with_method":0},{"lane":"product launches","links":1379,"share":{"estimate":0.0},"reason":"no method column on this lane (the link records WHAT, not HOW)","with_method":0},{"lane":"acquisitions","links":5613,"share":{"estimate":0.0},"reason":"no method column on this lane (the link records WHAT, not HOW)","with_method":0},{"lane":"job postings (ATS board ownership)","links":145948,"share":{"estimate":0.0},"reason":"no method column on this lane (the link records WHAT, not HOW)","with_method":0},{"lane":"office lease permits","links":107,"share":{"estimate":1.0},"reason":"method recorded in office_lease_permits.match_method","with_method":107},{"lane":"trademarks","links":5122,"share":{"estimate":1.0},"reason":"method recorded in trademark_filings.match_method","with_method":5122}],"window":["2026-09-01T00:00:00Z","2026-09-27T23:55:28Z"],"statement":"cross-lane provenance column NOT MEASURED: the er/identity-core provenance branch was not available when this was computed","registry_entity_links":{"links":1363952,"reason":"registry links (5500, H-1B, UCC, trademarks, 8-K, WARN) always carry method, confidence and run id","with_method":1363952}},"version":1,"payload_sha256":"e77c5e97347aabb30c5d5c6585ffed03db10100c6f4264cdc9f4e4b687282fc4","published_at":"2026-09-28T00:49:27Z","versions":[{"version":1,"published_at":"2026-09-28T00:49:27Z"}]}