Verify a decision

{"actor":"judge","decided_at":"2026-08-25T17:24:35.568Z","decision":"review_revise","investigation_id":"f28d76b9-cdf1-4557-be03-b49d1155a986","new_score":0,"page_slug":"2026-violent-crypto-wrench-attack-wave-h1-chainalysis-report","prev_score":0,"reason":"The disputed-plus-unverifiable share of claims is 19.4% (7 of 36), placing this page in the revise band, but the nature of the errors — not just their count — pushes the penalty to the top of that band. Most seriously, claim_findings[12] attributes a quote to Chainalysis ('the single highest-correlation protective factor') that does not appear in the report and reverses its actual finding: the report says disclosing crypto holdings creates risk, not that it reduces it. The page's own later section on protective steps states the correct, opposite advice, so the page contradicts itself while crediting a real research firm with safety guidance it never gave, on a topic where bad advice carries physical risk. Separately, claim_findings[24] and [25] cite two real, verifiable violent incidents (a ~$1M Paris-suburb robbery and a ~$820K home invasion) to a Decrypt article that contains neither — the underlying events are true but the sourcing is fabricated. claim_findings[31] also mischaracterizes a CoinDesk statistic as an early-2026 comparison when it actually describes full-year 2025 versus 2024, inflating the page's own 'record wave' thesis. Weighed against this, the review found the headline Chainalysis and CertiK statistics (incident counts, dollar totals, YoY changes, France's share) all confirmed precisely, and the page did not over-identify any victim, correctly preserving anonymity even for the legally-protected London torture case. The remaining discrepancies (a rounded '70+' vs. 77 figure, a mislocated arrest, a five-vs-six suspect count) are minor and do not materially affect the verdict.","score_delta":-15,"sequence_num":3,"submission_content_hash":null,"submission_id":null,"submission_kind":null,"submission_valence":null,"v":1}