Episode 7: Verification and Synthesis
6 parts
How six agents corroborated the corpus, entity resolution built one namespace from 2.38 million fragments, and what cross-domain synthesis actually found. Each part builds on the last — investigative analysis drawn from court filings, financial records, and released documents. Browse all episodes or the full investigation archive.
- Part 1 External Corroboration: How Six Agents Verified the Corpus Methodology
Six parallel research agents independently verified corpus findings against public registries, FAA records, and court filings (PAPER TRAIL Project, 2026...
- Part 2 519K Splink Clusters: Turning 2.38 Million Fragments Into One Namespace Methodology
Probabilistic entity resolution using a statistical matching method reduced 2.38 million raw Named Entity Recognition entries to 519,000 unified cluster...
- Part 3 Daubert Admissibility: Building a Pipeline for the Courtroom Methodology
Every analytical method in the pipeline was selected to satisfy the legal standard courts use to decide whether scientific evidence is admissible -- est...
- Part 4 9 Leads and 3 ACH Matrices: What Cross-Domain Synthesis Actually Found Methodology
Cross-domain synthesis across 232,000 events produced nine investigative leads and zero confirmed findings -- every lead fell below the 0.75 confidence...
- Part 5 5,000 Iterations: Why Monte Carlo Confirms the Conclusions Methodology
A Monte Carlo simulation ran 5,000 iterations with eight Beta-distributed parameters to test whether the project's conclusions change under parameter un...
- Part 6 Error Disclosure: What Compound Error Ceilings Mean for Every Conclusion Methodology
Eight pipeline stages contribute measured error rates that compound to a worst-case correct rate of 8.4%, meaning any conclusion that passes through eve...