Ship Counts and Loss Figures: Why Four Traditions Give Four Numbers

Dutch, English, Spanish and Portuguese accounts of the same sea battles disagree on ship counts and losses, and the reasons are structural.

10 November 2026 · Part of Dutch Republic & VOC — a guide to 15 articles →

Ask four historians how many ships fought and how many were lost at a seventeenth-century sea battle and you may get four different answers, each defensible. The disagreement is rarely carelessness. It comes from what each tradition counted, who wrote the count, and what the number was for.

One battle, several totals

Take the Battle of Lowestoft on 13 June 1665 in the new-style calendar. Britannica's account gives the Dutch loss as eight ships destroyed and nine captured, seventeen in all, against a single English ship lost. The Royal Museums Greenwich, describing a painting of the same action, speaks of a Dutch defeat costing about thirty ships. Both are respectable institutions, and the gap between seventeen and thirty is not an error in arithmetic. The larger figure may include vessels burnt, lost in the pursuit or damaged beyond use, but the museum's text does not say, and a reader cannot tell.

The Four Days' Battle of June 1666 is worse. Most accounts agree the Dutch lost four ships. English losses appear in different summaries as ten, seventeen sunk plus six captured, or twenty, depending on whether the compiler counted captures, burnings, ships later written off, or only hulls that physically went down. The Rijksmuseum holds a painting of English ships taken after that fight, a reminder that for the winners a captured hull was a trophy and a number at the same time.

These are not outliers. They are what the evidence looks like once you leave the handful of battles for which a muster roll survives.

Why the numbers drift

Four mechanisms account for most of the disagreement.

What counts as a ship. A fleet of the Republic in 1665 contained great two-deckers, smaller frigates, fire ships, ketches and hired merchantmen carrying a few guns. An English list of "ships of the line" and a Dutch list of "ships present" are not the same set. When Spanish and Dutch sources count at Abrolhos in 1631, the Spanish commander is described with twenty warships, twelve troop caravels and twenty-four merchantmen under his convoy. Which of those fought, and which merely sailed nearby, decides whether the odds were even or two to one. Our account of Abrolhos sets the sixteen Dutch ships against twenty Spanish warships, a comparison that only works if you leave the merchantmen and the transports out.

What counts as a loss. A ship sunk, a ship captured, a ship burnt by its own crew to stop capture, a ship that limped home and was never repaired: each is a loss to somebody and not to somebody else. Navies also differed in whether a recaptured prize was counted twice. The Dutch gained and lost English hulls in the same week in 1666, and a total that nets them off differs from one that adds them.

Who was writing, and when. Dutch admiralty reports went to boards that had to justify expenditure to the States General. English dispatches went to a court that wanted a victory story. Portuguese viceroys wrote to a crown that had little patience for defeats. Frederick Danvers, writing in the 1890s from Portuguese chronicles, repeats the Portuguese habit of claiming the field after the engagements off Goa in 1637, although the Dutch blockade resumed. A figure inherited from such a source carries its author's purpose with it.

Which battle it is. Counts also drift because the battles themselves are defined differently. The two actions at the Schooneveld in June 1673 are counted as two battles by some and one by others. A loss figure for "the battle" then belongs to one or the other, or to both together.

Four traditions, four habits

The table is a generalisation, and every row has exceptions. Its use is to remind a reader which way each source tends to lean before taking its figure at face value.

The calendar and the guns

Two more things disturb comparison. The Republic used the Gregorian calendar and England the Julian until 1752, so one action carries two dates ten days apart, and a careless compiler may merge two reports of the same fight as though they were different days. And guns are counted in different ways: a ship "of seventy-six guns" in one list may be seventy-three in another, as the sources for the Dutch flagship that blew up at Lowestoft show. The ship could be rated by its design, by the guns it carried that day, or by the number of ports.

Casualties are the least stable figure of all. For the Battle of the Sound in 1658, our own account in Why Amsterdam Fought for Denmark gives about forty-one Dutch and forty-five Swedish ships and roughly fourteen hundred and twelve hundred casualties, but flags both as the figures of standard summaries rather than of a careful modern count. Even the sides' strengths are matters of convention.

Chronology of the disputed counts

What a record can honestly do

A dataset has to choose. Our record of Dutch Golden Age naval battles uses a five-ship threshold at inclusion and then deliberately does not tabulate ship counts or casualties. A single figure in a cell would give a disagreement a precision it does not have, and a reader who sorts by "ships lost" would be sorting by whichever tradition the compiler happened to follow. What the record does instead is carry a certainty column and a source for each row, so that the outcome, which is already a judgement, is at least traceable.

That is not a counsel of despair. Where a muster roll exists, as for several of the great fleets of the 1650s and 1660s, the strengths on paper can be compared with the reports. Where one does not, a careful writer gives a range, names who said what, and refuses to let a round number stand for a count. The institutional structure behind the Dutch figures, five admiralty boards each keeping its own books, is explained in our guide to the five Admiralties, and it is one reason that a single "Dutch fleet" total is harder to find than it looks.

Counting as a political act

Numbers were also weapons. After the Raid on Chatham in 1667 the English and Dutch each published accounts that suited the peace negotiations in progress. A year earlier the loss lists had been used the same way at home. In the Mediterranean, Johan van Galen's squadron fought a war in which prizes were wealth, so a captured ship was a financial entry as well as a military one, and its value decided how carefully it was listed. The same was true on the other side of the world, where the Company's fleets fought for a trading empire and wrote their losses into ledgers that served shareholders.

Conclusion

Four traditions give four numbers because they were counting four slightly different things for four different purposes. The honest response is neither to pick the most convenient figure nor to give up, but to treat each number as a claim with an author. A loss figure is a piece of testimony. Read it as one, and the disagreements stop being embarrassing noise and become information about who was counting and why.

Frequently asked questions

Why do sources disagree on how many ships were lost at Lowestoft? Britannica gives seventeen Dutch ships lost, eight destroyed and nine captured, while a Greenwich museum description speaks of about thirty. The difference may reflect vessels burnt, damaged or lost in the pursuit, but the sources do not say.

Were Dutch and English figures for the Four Days' Battle ever reconciled? Not fully. The Dutch loss of four ships is widely agreed, but English losses appear as ten, twenty or seventeen sunk plus six captured, depending on how sinkings, captures and write-offs are counted.

Why does the dataset not list ship counts? For most of these actions the national traditions differ by more than the differences the numbers are meant to show. A single figure would suggest a precision the evidence lacks, so the record gives outcome, certainty and source instead.

Does the calendar difference affect the counts? Indirectly. England's Julian dates ran ten days behind the Gregorian ones used in the Republic, so reports of one action can look like reports of two, and tallies can be double-counted or mistakenly merged.

As an Amazon Associate I earn from qualifying purchases. Affiliate links in Recommended reading — and in Further reading on older articles — may earn Writesides of History a commission at no extra cost to you.