Sign in
Dylanvir
Account
Atlas

← the contents·the plates · time

The centuries, accumulating

Every battle the atlas knows, banded by era. Choose a band and the atlas builds up to it: the band you are reading is inked, every band before it is ghosted in its own colour, and the bands after it are not drawn. Drag to pan, wheel to zoom.

The Modern Agec. 1800 – present

this band · 168 of 190 mapped · 88.4%

plotting 312 of 350 battles · 89.1%denominator · every battle the atlas records from the beginning of the book through The Modern Age (c. 1800 – present)

Through The Modern Age the plate carries 312 of 350 battles. This band on its own is 168 of 190. Earlier bands are drawn faded; later bands are not drawn at all. 312 of 350 battles are plotted. The atlas keeps coordinates in two stores that do not overlap at all: 142 come from the database and 170 from the hand-maintained gazetteer, and no battle carries one in both. 38 battles are on no map here, and 22 of those name a place the atlas has never given a coordinate. Nothing was geocoded to fill the gap.

the bands accumulate

312 marks on the plate — how sure the atlas is of each position

  • 116 drawn as a solid disc · known site — the gazetteer names the ground
  • 142 drawn as a dashed ring · precision not stated — a database coordinate — the row declares no standard, and none was invented for it
  • 31 drawn as a hollow ring · region only — located to a region, not to a field
  • 23 drawn as a ring inside a halo · best guess — the gazetteer's own word for it, carried rather than dropped or promoted

142 of these 312 marks are database coordinates · 142 have no chapter in the book and carry no link

Gazetteer coordinates declare their own precision: high is a known site, med is region-level, low is the gazetteer's own word for a best guess (23 of them). Database coordinates declare no precision at all and are marked undeclared rather than promoted to high.

0 battle(s) fall in no era band. They are drawn but carry no era link, because a link that guesses its chapter is worse than no link.

the whole atlas · 312 of 350 (89.1%) · denominator: every distinct battle id known to the atlas: db rows plus gazetteer-only ids

The Modern Age · 22 of 190 records are on no map, and why
  • Battle of France 1940 — place named (Belgium) but no coordinate recorded
  • Catalonia Offensive 1938 — place named (Catalonia) but no coordinate recorded
  • Battle of Rostov 1941 — place named (Rostov-on-Don) but no coordinate recorded
  • Battle of Przemyśl 1918 — place named (Przemyśl) but no coordinate recorded
  • Battles for Vilnius 1918 — place named (Vilnius) but no coordinate recorded
  • Battle of Našice 1944 — place named (Našice) but no coordinate recorded
  • Battles of Rzhev 1942 — place named (Rzhev) but no coordinate recorded
  • Battle of the Vistula River 1914 — place named (Warsaw) but no coordinate recorded
  • American airborne landings in Normandy 1944 — place named (Cotentin Peninsula) but no coordinate recorded
  • Battle of Pirajá 1822 — place named (Bahia) but no coordinate recorded
  • Battle of Java 1942 — place named (Java) but no coordinate recorded
  • Russian invasion of East Prussia 1914 — place named (East Prussia) but no coordinate recorded
  • Battle of the Dnieper 1943 — place named (Dnieper) but no coordinate recorded
  • Battle of Bautzen 1945 — place named (Bautzen) but no coordinate recorded
  • Operation Nordwind 1944 — place named (Alsace) but no coordinate recorded
  • Battle of Białystok–Minsk 1941 — place named (Poland) but no coordinate recorded
  • Battle of the Ruhr 1943 — place named (Ruhr) but no coordinate recorded
  • Battle of Mullaitivu 2009 — place named (Mullaitivu) but no coordinate recorded
  • Battle for Vlasenica 1943 — no coordinate in either store
  • Battle for Vlasenica 1942 — no coordinate in either store
  • Battle for Glinu 1943 — place named (Glina) but no coordinate recorded
  • Battle of Dervenakia 1822 — place named (Q3560699) but no coordinate recorded
The Modern Age · the 168 marks this band puts on the plate

IWhat the 89 per cent is hiding

312 of 350 battles carry a coordinate, which is 89.1% and reads like a solved problem. It is not distributed. The Late-Medieval Crisis is 20 of 33 — 60.6% — so a reader looking at that band is looking at a third of it. That is why the band figure sits above the plate and in the rail rather than in a note underneath: a percentage for the whole book is not a percentage for the century anyone is actually reading.

IITwo stores, no overlap, one of them not in the book

The coordinates come from two places that share nothing: 170 from the hand-maintained gazetteer and 142 from the database, with no battle carrying one in both. Only the gazetteer’s battles are published, so 142 of the marks on this plate belong to battles with no chapter to open. They are drawn, because the position is real; they carry no link, because a link to a chapter that does not contain them would be a link that lies.

Those 142 database coordinates also declare no precision at all. They are drawn as a dashed ring — not as a solid disc, which would invent a certainty, and not as a hollow ring, which would invent a coarseness the record never claimed.

IIIWhat is not on the plate

38 battles are on no map here, and 22 of them name a place the atlas has never given a coordinate — Warsaw, Java, the Dnieper. Every one is listed by name under its band with the reason it is missing. None of them was geocoded: a location interpolated to fill a hole is the exact failure this atlas exists to avoid, and it would be invisible on the plate once drawn.

basemap · Natural Earth 110m land, public domain · publishes · Natural Earth is public domain and publishes. Caveat carried openly: no script in war-atlas/tools emits this file, so the source string is a self-assertion with no reproducible provenance.

Every plate in this suite draws one coastline through one projection — the original plate runs the same engine.