Idle Sails · about & methods

How this chart is made

Idle Sails is a spectator’s idler: leave the tab open and a procedurally generated, historically grounded age of sail plays itself out. This page documents what stands behind the spectacle — where the numbers come from, how they were interpreted, what the simulation invents, and where it knowingly departs from the record.

Two goals, held together

Every design decision answers to two co-equal aims. The first is simulative: a zero-friction, deterministic world — the same seed always produces the same sea, at any clock speed, which is what makes closing the tab and returning days later exact rather than approximate. The whole control surface is a speed slider and click-to-inspect; there is nothing to build, buy, or win.

The second is historical sensitization. The dataset is built knowing that the archive is not the past (after Michel-Rolph Trouillot’s Silencing the Past): the surviving quantitative record of this era is Euro-centric, institutional, and full of silences that enter at every stage — at the making of sources, of archives, and of narratives. The chart tries to be honest about that rather than hiding it behind a confident-looking map. The rules it follows are set out in the project charter: no silent zeros (a trade known to have existed is never implicitly absent), no fabricated precision (rank only where sources support ranking), basin-local assembly (no forced commensuration between seas with incommensurable records), declared boundaries (every dataset states what it excludes and why), a public silences register, and sober treatment of coerced human movement wherever it appears.

How the world sails itself

The simulation is a seeded, deterministic world running a flowing clock from 1550 to 1850, after which the chart is “redrawn” over a designed ten-year epilogue decade (1850–1860) — the age of sail winding down — and the cycle begins again. Every vessel is generated the moment she puts to sea: her lane is drawn from the era’s trade-flow weights, then her type, flag, tonnage, cargo, name, and master follow from period-specific vocabularies, and her entire fate — every leg, every port call, whether and when she is lost — is rolled at spawn from her own random stream. Nothing about her is decided later, which is why fast-forwarding a saved world in a few big steps lands on exactly the sea that leaving the tab open would have produced.

Ships move along pre-computed route polylines: curved, seasonal tracks baked offline by a least-time router over parametric wind and current fields with per-class sailing polars (square-riggers cannot point closer than ≈60° to the wind; the monsoon reverses; the Roaring Forties carry the eastbound Indiamen). Junks and dhows sail their own polars. Passage durations were calibrated against recorded at-sea sailing times — Middle Passage legs within a fraction of a percent of the SlaveVoyages medians, England to the East Indies against East India Company records — with the known residuals documented in the project’s calibration report rather than smoothed away.

Where the traffic comes from

Since 2026 the sim spawns from an evidence-classed trade-system flow matrix: sixty-odd named trade systems (the Carrera de Indias, the Sound-toll Baltic trades, the VOC spice circuit, the Nanyang junk trade, the English coastal colliers…), each carrying per-decade voyage-count ranges assembled basin by basin from the sources that basin actually has. Each seeded world draws one reading from every range — two seeds may honestly disagree about how busy a trade was, because the historians do. Port prominence is an output of these flows, not an input: the 1560s’ busiest port on this chart is Istanbul, a city the ranked European record could never place there. See prominence, derived from the flows.

The anchor series behind the counted systems include the Sound Toll Registers (Baltic passages, 1497–1857), the Trans-Atlantic Slave Trade Database (SlaveVoyages), Chaunu’s reconstruction of the Carrera de Indias from the Seville records, Dutch-Asiatic Shipping (the VOC’s complete voyage list), the Nagasaki registers kept under the Tokugawa trade restrictions (where a state did count, Chinese hulls outnumber Dutch ten to one — the anchor for the junk-trade reconstruction), the Canton trade records of the European companies, and the Marseille échelles du Levant series. Where no series exists, systems are classed as reconstructions from scholarship or as declared assertions with wide bounds — never as silent zeros.

The evidence classes

Click any vessel and her ledger ends with one italic line saying what kind of evidence stands behind the trade her voyage samples. There are five:

counted

A surviving series records this trade — a toll register, a customs house, a company ledger. Even these are soft ground: the Sound registers are biased by evasion and exemption, and the register says so.

proxied

Inferred from adjacent records — a trade visible only in someone else’s books.

reconstructed

Scholarship’s estimate, not a surviving ledger — strong secondary literature (the Nanyang junk trade after Blussé, Reid, and Sugihara).

asserted

Our own stated estimate where the archive is silent — wide bounds, reasoning on record (Caribbean smuggling, the Chinese coastal grain fleets). An assertion you can see is safer than a zero you can’t.

state

A naval voyage — outside the commercial record entirely.

How the ports were chosen

The world sails 105 ports, chosen in three layers. The backbone is the flow matrix itself: the ports the flow systems need (Hamburg, Danzig, Riga, St Petersburg, Venice, Genoa, Marseille, Boston, New York, Havana, Surat, Calcutta, Amoy…). On top of that sits a curated diversity tranche — minor ports that never troubled a ranking but carry a monopoly, a chokepoint, a single precious cargo, or the flag of an otherwise excluded people (Dejima, Tranquebar, Banda Neira, Naha). Last come declared exceptions kept for what they say about the record itself — Istanbul above all, present without a rank for decades because the sources support presence but not ranking; that state is data, not a defect. The 1815–1850 extension added the ports of the later age of sail (Singapore, Hong Kong, Sydney, Montevideo, New Orleans) alongside a whaling-grounds node off Peru, the Hudson’s Bay supply post at York Factory, and the Persian Gulf pair Basra and Bandar Abbas — each with its founding window and, where it changed hands, the flag of the time.

Ports also live and die: founding and destruction windows are enforced against every lane (no Kingston before 1655; Smeerenburg’s blubber town burns out by 1660; Kaffa closes in 1783), a destroyed harbour keeps a faint ruin mark because the chart remembers, and the dot speaks the dominant port name of its time — Jayakarta before Batavia, Port Royal before Kingston, Hugli before Calcutta, St John’s for the Banks fishery outside Louisbourg’s brief French decades.

Reading the chart

Ship glyphs — shape is class, colour is flag

• Slim dart — merchantman, brig, snow, sloop: the ordinary carriers of the sea.
• Full rounded hull — East Indiaman, fluyt, and the great hulls of the earlier era: galleon, carrack, junk.
• Sharp barbed arrowhead — frigate, sloop-of-war.
• Broad hexagonal hull — ship of the line.
• Boxy hull with a transverse bar — slave ship. The bar is deliberate and sober: these hulls should never read as just another trader.

Colour always means allegiance — the legend lists every flag. Company ships (EIC, VOC, WIC, the Danish and Swedish East India companies…) fly their company’s colour and take their naming culture from their parent nation.

Masters and captains

Every vessel sails under a named shipmaster, drawn from period naming pools per culture. The title travels where the history put it: European naval commanders are captains and merchant shipmasters are masters, while in the Indian Ocean the nakhoda — the shipowner-master of the dhow and Mughal marine — carries the title in the name, Ottoman masters are … Reis (with Greek kapetans alongside, as that empire’s marine truly sailed), and Makassar’s masters carry Daeng. These are plausible names, not attested individuals; any resemblance to a real voyage is the period’s, not the record’s.

The sober treatments

Where the chart knowingly differs from the record

• Vessels are statistical samples, not registered hulls. Each ship stands for the traffic of her lane in her decade; names, masters, and tonnages are plausible vocabulary, not archival entries.
• One seed, one reading. Each world realizes a single value from every trade’s [low, high] evidence range. Uncertainty lives between worlds, not within one.
• Coverage is partial and declared. The folded flow systems account for roughly 85–95% of recorded world voyage volume depending on the decade; the remainder is either unsailable at this port roster or lives in the silences register.
• Steam is a declared boundary. This chart follows sail. The age of steam opens in exactly the decades the world sails into last — the P&O mail to Hong Kong from 1845, the first ocean steamers on the Atlantic — but the wind-and-current engine that shapes every route here cannot produce a steamer’s track, so steam is surfaced (the epilogue era note, this paragraph) and never sailed. A steam layer is deferred, on record.
• Spawn pacing is clamped. Traffic follows the realized totals but is held within 0.5–1.6× the era mean, so the 1550s never feel empty and the 1810s never swamp the renderer — a spectator concession, on record.
• A residual floor keeps authored lanes alive. About 3% of volume is spread across era-active lanes so no authored trade fully vanishes between decades; naval movements are a fixed ~6% pool outside the commercial matrix.
• Weather is climatology, not events. Routes are seasonal averages; the one era-climate signal is a bounded +7% passage-time factor on North Atlantic legs during the Maunder Minimum (1645–1715). Hurricane and wartime risk raise loss odds, not individual storms.
• The loop is a fiction. History does not restart in 1550; the ten-year epilogue “redrawing” decade (1850–1860) exists so an idler can run forever, and the era label says so while it happens.

Sources & further reading

• Michel-Rolph Trouillot, Silencing the Past: Power and the Production of History (Beacon, 1995) — the framework behind the charter and the silences register.
• Sound Toll Registers Online — Danish sound passages 1497–1857. soundtoll.nl
• SlaveVoyages — Trans-Atlantic & Intra-American Slave Trade Databases (voyage counts, durations, mortality). slavevoyages.org
• Pierre & Huguette Chaunu, Séville et l’Atlantique, 1504–1650 — the Carrera de Indias reconstruction.
• Bruijn, Gaastra & Schöffer, Dutch-Asiatic Shipping in the 17th and 18th Centuries — the complete VOC voyage list. huygens.knaw.nl/das
• The Deshima Dagregisters (Japan-Netherlands Institute / Brill) and the Nagasaki trade registers — the East Asian anchor.
• Leonard Blussé, Anthony Reid, and Kaoru Sugihara — scholarship behind the Nanyang junk-trade reconstruction.
• H.B. Morse, The Chronicles of the East India Company Trading to China — the Canton series.
• The Marseille chamber of commerce series on the échelles du Levant — the Ottoman-Mediterranean cross-check.
• CLIWOC — Climatological Database for the World’s Oceans 1750–1854 (wind regimes behind the route baker). García-Herrera et al., Climatic Change (2005).
• ICOADS — wind climatology corroboration. icoads.noaa.gov
• Natural Earth — coastline and land mask. naturalearthdata.com

Full per-page sourcing and caveats live on each research page; the route-engine calibration report (fit and hold-out legs, error table, known residuals) ships with the project repository.

Idle Sails · about & methods · the chart is a claim under revision, and this page is its list of admissions. ⚓ Return to the chart