| Evidence grading reformed (A1/A2 split) |
Methodological |
Distinguishes DOJ investigative products from documents in Epstein's possession |
| CO_APPEARED_WITH methodology caveat |
Methodological |
Clarifies documentary co-mention does not equal real-world co-presence; provides graph density stats |
| "What We Did NOT Find" section |
Structural |
12 null results systematically presented as boundary evidence |
| Geographic distribution table |
Structural |
Maps Montreal/Vancouver/Toronto/National footprint |
| Grand Prix recurring access pattern |
Analytical |
Three data points across 13 years explicitly connected to race calendar |
| Black book density quantification |
Analytical |
Stroll at 99.6th percentile of 1,971 contacts; full distribution provided |
| Full billionaire candidate table |
Analytical |
6 candidates compared honestly, not just Stroll |
| Desmarais/Power Corp (Part 7D) |
Structural |
Own section; honest assessment (institutional, not personal connection) |
| Procurement evidence boundary |
Analytical |
Explicit: strong on network map, thin on criminal extension to Canada |
| "Black money" standalone |
Structural |
Extracted from Pinker section; standalone operational significance |
| Pinker reframing |
Language |
"Pipeline" -> "network"; Grade D interpretation acknowledged |
| Isabel Maxwell thread |
New thread |
20+ Neo4j rels, 16+ EFTA docs, tech/intelligence background |
| Kravis-CIBC board connection |
New finding |
Former CIBC board member connects Trilateral co-member to Montreal bank |
| Stroll initiation pattern |
New finding |
Zero evidence of Stroll->Epstein contact in 60,806 files |
| COUQ Foundation gap |
New finding |
Zero hits in full-text DB; documented only in secondary reporting |
| Pattison gap |
New finding |
Zero hits in full-text DB or Neo4j; Trilateral roster name only |
| 9 language/accuracy fixes |
Language |
"Proves"->"strongly suggests"; flight interpretation; Belzberg caveat; DB alerts; Dershowitz; Lesley; Verchere methodology |