Computational reconstructions of three self-checking historical trade ledgers: the VOC at Batavia, the Sound Toll, and the Spanish-empire royal treasuries

A defensive deposit timestamping three short computational-history results and the code behind them. Each rests on a historical accounting record whose own structure supplies the ground truth, so the data checks the analysis. 1. VOC at Batavia. The Dutch East India Company Bookkeeper-General ledgers, 1701-1801: a commodity-by-route-by-time tensor over 18,321 voyages and 1,693 products. The double-entry books reconcile to 98.6 percent, and route-value concentration roughly halves across the century. 2. Sound Toll Registers. Every ship passage through the Oresund, 1497-1857, 2,152,705 records (STRO 2.0). The dominant trade axis flips from Amsterdam entrepot primacy to a St-Petersburg-to-London trunk, datable to 1807-1815 and confirmed by the Napoleonic closure of the strait that cut annual passages from 7,386 (1806) to 134 (1808). Entity resolution links 59.6 percent of named passages into multi-voyage captain careers. 3. Spanish-empire royal treasuries. The TePaske-Klein Royal Treasuries data, 74 American-mainland treasuries, 11,537 caja-year blocks, 202,585 line items, c. 1540-1825. The finding is a recording asymmetry: 84 percent of situado transfer value resolves to a named place, but under a year-aware test fewer than one percent of sent value is corroborated by a matching receiver record within a year. The periphery itemizes what it sends up, the hub treasuries lump what they receive, so the reconciliation residual is the finding, invisible to aggregate regressions. Large raw inputs are not redistributed here. They are obtained from their original public providers (the KNAW Bookkeeper-General Batavia dataset, the STRO 2.0 figshare deposit DOI 10.6084/m9.figshare.27176202, and the Real Hacienda data from El Colegio de Mexico) via the included code. These are working-paper drafts, not peer reviewed, with limitations stated in each draft. Deposited to establish a timestamped, citable record of priority while the manuscripts are finalized.