The Distance
Pick any two names in the corpus. The median gap between them is five handshakes. Every hop is a documented relationship, and the ones nobody has sourced yet say so.
The objection to a book like this is that it draws lines between people who merely happen to know each other. Fair. So here is the test, run the other way: take two names out of the corpus at random, and count the handshakes.
The median is five. Not five for people picked because they connect — five for any two of the 1,037 entities this project documents, dealt blind. A Treasury official and a lab founder. A think-tank fellow and a defence contractor. Five.
That is the whole finding, and it is not an argument. It is a number you can generate yourself, repeatedly, and check hop by hop.
Each hop states what kind of evidence it has, because the two kinds are not the same thing. 181 of the 1,400 links carry a receipt on the relationship — a filing, a disclosure, a court record naming both parties. The rest are sourced at their endpoints: every one of the 1,400 has at least one party carrying its own documentation, 1,980 receipts in total, and those are one click away on the hop itself.
That distinction is the honest part. “Both of these people are documented, and this relationship is asserted” is weaker than “this relationship is documented”, and a path that blurs the two is selling you something. Every link without its own receipt is a research task nobody has done yet, visible rather than smoothed over.
Sometimes there is no route at all. That answer is kept too. A corpus that connects everything to everything has stopped being evidence and become a mood.
Shortest documented path, breadth-first, through the entities THE RECORD documents. A dashed hop has no receipt on the link itself — the two ends carry their own sourcing, and the gap is shown rather than smoothed over. “No documented route” is a real answer here, not a failure.