Structured records built from named sources, each row citing a place in the one it rests on: a chapter of a chronicle, a page of an edition, a table in a paper. Every record states on its own page what it includes, what it deliberately leaves out and where the scholarship disagrees.
The Caesar dataset is free: the complete table, the CSV and reuse under CC BY 4.0. It is there so you can see exactly how these records are built before you spend anything. The other five are research editions: the page carries the method, the field definitions, the sources and five example rows; the edition carries the complete table as CSV, XLSX and JSON with a full source register, € 19.95 each; the five together are € 79.95 instead of € 99.75, still five separate datasets. See the research editions in the shop →
Every count, duke, earl and banneret lord known to have died between taking the cross and coming home, classified by what the chronicles actually say killed him. The siege camp runs level with the battlefield, and for seven of these men the sources give no cause at all.
The operations Richard commanded in person from Vézelay to Acre, with Sicily and Cyprus counted rather than skipped. He won every engagement he fought and never once laid siege to Jerusalem.
All the operations of the Gallic War across nine campaigning seasons, including those Caesar delegated to his legates and reported in a sentence. Each row references the book and chapter of the Commentarii where it appears.
Excavated deposits in which at least three people carry perimortem injury attributed to violence, from Jebel Sahaba to San Juan ante Portam Latinam. Where the reading is contested — and at the most famous sites it is — both readings are in the row.
Fleet actions in which a squadron of the Republic or of the VOC and WIC met an enemy fleet of at least five ships, from the Strait of Malacca and the Brazilian coast to the Channel and the Mediterranean.
Attacks claimed in a Kommando statement or attributed to the group, with the generation, the target and the legal outcome kept in separate columns — including the eight for which no perpetrator was ever identified.
See the research editions in the shop →
Each research edition is € 19.95 and holds one dataset: the complete table as CSV, as an XLSX workbook and as JSON with a schema, plus a README with the method and the full source register. The bundle is those same five editions together, one folder each, for € 79.95 instead of € 99.75 — it is a discount, not a merged file.
Each dataset states what it includes and what it deliberately leaves out, before you look at a single row. Where a date, a location or a cause of death is contested in the scholarship, the entry says so rather than picking a side quietly. That is the difference between a reference table and a list, and it is why the criteria sit on the page next to the data instead of in a footnote.
Every figure in a dataset's header is counted from the rows themselves rather than written into the page, so a corrected row corrects the summary with it. Corrections are welcome and are credited. If you use one of these in your own work, a link back is all the attribution the licence asks for.
Writesides of History publishes free, in-depth history articles every week. If you enjoy reading, you can support the work: become a member for exclusive deep-dive dossiers and early access, or browse the maps, classroom packs and dossiers in the shop.
Become a Member Donate Visit the Shop