Roman defeats before and after the fall of the West

Iosif Lazaridis has done something unusually useful for anyone interested in Roman history. His new catalogue brings together 1,709 named land and sea battles, from the traditional beginnings of Rome to the fall of Theodoro in 1475. Dates, places, outcomes, sources and doubtful cases sit alongside an interactive map. It is a remarkable piece of historical compilation, and the downloadable records make it possible to ask questions that a map alone cannot answer (Where the Romans Fought).

One choice in particular caught my attention: the medieval empire we usually call Byzantine is included as Roman. That is historically accurate. Its people called themselves Rhomaioi, Romans, and their state continued the Roman imperial tradition. Yet a single name can also obscure profound differences. Language, institutions, military organisation, territory and opponents changed enormously across the centuries. These developments were gradual: the late Western Empire was already Christian, and Greek had long been part of the Roman world (Metropolitan Museum of Art).

I therefore decided to test two contrasts: Eastern against Western imperial courts, and battles before and after 476. Two strong chronological patterns emerge: a long-run rise in defeat odds and a large difference across 476. Defeats make up slightly more than half of the recorded battles after 476, against roughly a quarter in the earlier Imperial period. The direct comparison of the two courts is much less certain.

Why compare the two courts and why choose 476

Constantine inaugurated Constantinople in 330. After Theodosius I died in 395, his sons Arcadius and Honorius ruled the eastern and western halves of the empire. The enduring distinction between their courts gives an East–West comparison a historical basis.

There is a qualification. Sharing imperial power had precedents, and 395 did not create two modern nation-states overnight. As Kaj Sandberg argues, treating that year as a clean constitutional divorce oversimplifies the continuing unity claimed by the emperors (Sandberg 2008).

The other familiar date is 476, when Odoacer displaced Romulus Augustulus from the western throne. This is the conventional fall of the Western Roman Empire, although even the precise timing of the deposition has been reassessed (Cristini 2024).

Roman authority continued in several forms: Julius Nepos retained his imperial title in Dalmatia until his death in 480. I use 476 as a recognisable historical boundary, with the battle recorded in 476 placed on the earlier side. It offers a practical dividing line between periods whose institutions often overlapped (Sandberg 2008).

These two tests address different questions. Comparing the courts concerns overlapping imperial administrations. Comparing periods concerns a much longer transformation, from earlier Rome to the Eastern Roman and successor-state record.

The long view shows a changing battle record

Before dividing the record into periods, it helps to look at the whole chronology. Figure 1 shows the proportion of recorded battles ending in defeat in each century. The upward tendency is evident, but the points also make clear how uneven that history was.

There are 1,469 battles in the main analysis: 916 victories, 512 defeats and 41 draws. I leave out the 236 civil-war battles because Roman forces fought on both sides, along with four non-civil battles whose outcome is unknown. A defeat is counted against all known outcomes, so draws remain in the denominator.

Figure 1. The Roman defeat record across the centuries

The time model uses each battle’s recorded year as a continuous predictor; the century bins only summarise the plotted points in Figure 1. Across 1,469 battles, the model estimates a 12.5% increase in the odds of a recorded defeat per 100 years, with a war-clustered 95% interval of 9.4–15.8%. The fitted long-run tendency coexists with large reversals: the seventh-century peak and subsequent falls are plainly visible in Figure 1.

A catalogue spanning more than two millennia necessarily combines different historical circumstances. This led me to examine how its internal contrasts vary across periods.

Upgrade to paid for the full analysis: the rise in Roman defeats, the East–West comparison, and the tests of whether 476 marked a turning point.