Some thoughts on Sabin’s “Lost Battles”

Philip Sabin, Lost Battles: Reconstructing the Great Clashes of the Ancient World (London: Continuum Books, 2009) Bookfinder link

Big battles are always a popular topic, but even the best-documented ancient battles are difficult to understand. The few ancient accounts which survive never answer every question which modern readers ask, and often disagree with each other or say things which are difficult to believe.  Several plausible interpretations are always possible, and deciding between them is a matter of judgement not proof.  One way to resolve these debates is to apply a new methodology.  Sabin’s book argues that wargaming is just such a methodology and that it has been unfairly neglected as a tool for understanding ancient battles.  To support this, Sabin designs a wargame then presents scenarios for 35 land battles in the ancient Mediterranean world from Marathon (490 BCE) to Pharsalus (48 BCE) with comments on the major uncertainties and how his wargame can help clarify them.  Wargamers have studied these battles many times in the past, but few have Dr. Sabin’s training, or explain their reasoning in such detail.

Perhaps because of its unusual methodology, many aspects of this book are rather conservative.  For example, Sabin tends to give ancient sources the benefit of the doubt, on the grounds that they are probably more reliable than modern speculation. But he doesn’t think the Hellenistic tactical writers like Aesclipidotus are very valuable, repeating the traditional charge that they are bookish and impractical. Phalanxes of spearmen with large shields are treated differently depending on whether or not they are Greek. And Sabin believes the Greek and Roman sources that say that their own armies often beat other people’s armies several times their strength in direct confrontations, losing only a few hundred of their own men out of tens of thousands.

Sabin’s wargame focuses on the grand tactical level. Armies are divided into units, which each represent hundreds or thousands of men. Each unit is rated once for quality (Average, Levy, or Veteran) and once for fighting style (such as Heavy Infantry or Cataphracts). Up to two named generals per side provide benefits to their army as long as they live, but may be killed if they are too bold. The battlefield is divided into 20 squares, which represent both a square of ground several hundred meters across and an abstract part of the battle line such as the left wing. Attacks and movement are from square to square, with troops of only one side allowed to occupy a square at a time. A limited supply of command points and a randomized morale system prevent players from perfectly controlling their troops. Major terrain features such as a river or fortified camp can affect movement and combat. A few special rules allow for bad weather, surprise, tired soldiers, and other special conditions. This abstract approach was a wise decision. We don’t know enough about many aspects of ancient warfare to simulate them in detail. Even within his simple system, there is room for debate about how to classify particular contingents and how to represent an army’s deployment.

The bulk of the book consists of 35 scenarios for land battles from Marathon to Pharsalus.  Each battle piece consists of a few pages discussing the sources, modern scholarship, how to describe the armies and the battlefield under his rules, and any insights he has had from gaming the battle. Sabin argues that wargaming battles reminds us of the dynamic nature of combat, and forces us to consider why generals made the decisions they did. He often makes arguments about plausibility in terms of the frontages and fighting values of his wargame. While sometimes circular (the rules are based on his guesses about historical battles, and his guesses are influenced by the rules), this is less subjective than other ways of deciding between competing theories.

However, unless you agree with Sabin’s assumptions, his simulation has limited value. Although Sabin explains that many years of playtesting, experimentation, and debate are behind his rules, this evidence is unavailable to readers. Because of the lack of quantitative evidence on ancient battles, he can’t test his model mathematically as students of more recent battles can. Sabin is dismissive of the complex calculations behind such modern wargames (p. xvi), but these calculations reflect a complex reality. Writers on ancient battle sometimes use incidents at less famous battles for comparison, but Sabin makes limited and anecdotal use of this evidence. I have grave doubts about two of his assumptions.

The first is that each battle’s outcome was the expected one, in the statistical sense (p. 21). In other words, Cannae was almost certain to be a great Carthaginian victory, but more or less Romans might have broken free, and they could have won by a stroke of luck. This assumption is convenient, and may be necessary, but is certainly difficult to defend. Battles are chaotic events. If both sides did not think they had a reasonable chance of winning, they would not take place at all. Maybe the losing side was consistently mistaken about the strength of their army, but this is hard to prove. This is especially dangerous because the best documented battles tend to be “near-run things” or ones where the outcome was different than expected.  Battles where the side with every advantage won don’t make memorable stories.

The second assumption is that the number of men in two opposing armies was often grossly disparate. Smaller armies could win anyways not just because of luck or better leadership, but because their soldiers were much more effective. Sabin suggests that man for man, veteran troops were nine times as effective as levies, and average troops three times as effective (p. 20).His argument for this is not entirely clear, since he acknowledges that most of the accounts of large armies beating small ones cannot be trusted. He mentions Cannae, Magnesia, Carrhae, and Pharsalus as battles where armies beat opponents with about twice as many soldiers. The question is whether these famous victories of the few over the many should be seen as typical. Greek and Roman authors insisted that their side often beat huge armies of barbarians, but it is usually clear that they had no idea how many soldiers were in the ‘barbarian’ army, and authors were expected to say that barbarian armies were huge. A more subtle argument is woven through the book, where Sabin argues that particular Greek or Roman contingents needed a similar advantage over their enemies to win specific fights. Implicitly, if two thousand Spartans are worth six thousand Thebans, twenty thousand Greeks could be worth sixty thousand Thracians. But is this a valid extrapolation?

To keep the number of units on both sides roughly the same, Sabin decides to make units of better troops represent smaller numbers of men. This has remarkable implications. For example, at Cunaxa we know that 10,000 Greek hoplites made up about half of Cyrus’ infantry line. In Sabin’s reconstruction, the 10,000 Greek hoplites count as 8 units of Veteran Hoplites, while the 20,000 barbarian infantry of Cyrus’ left wing count as only 4 units of Levy Heavy Infantry (each significantly less effective than a Greek hoplite unit). This certainly helps to reproduce the battle pieces in our sources, but not to test whether these descriptions are plausible. Given that he assumes that Greek and Roman soldiers were each worth several ‘barbarians’, it is no surprise that he concludes again and again that the ‘barbarians’ must have had large numbers of men to stand any chance with his model. For example, for the battle of Crimisus in 341 BCE, Sabin assumes that most of the Greeks were Average and most of the Carthaginians Levies. He naturally concludes a few sentences later that as the battle was a “close-run thing” the Carthaginians had about three times as many soldiers as the Greeks did (p. 163). Similar ratios appear in his accounts of Marathon, Cunaxa, and Caesar’s Gallic battles.

Sabin’s model also gives both players knowledge of every factor which can affect the battle. They know the exact type and quality of the troops on both sides, how much the terrain and weather will impede their soldiers, and which officers will flee if threatened rather than standing their ground. Sometimes this is reasonable, particularly when two experienced armies from the same culture fought each other. But ancient generals were often unsure of the number and effectiveness of their own troops, let alone those of the enemy, and battles were full of surprises. Pompeius clearly did not expect that his cavalry would run away at Pharsalus, and there is no sign that Artaxerxes believed that his infantry would run away from Cyrus’ Greeks at Cunaxa. If they had expected this to happen, they would have fought differently, and indeed players are likely to adjust their tactics to reflect this. A model which emphasized commanders’ ignorance and the role of chance, such as the famous wargame Piquet, might give a more realistic experience at the cost of being less likely to reproduce particular battles as described in the sources.

Whether or not you agree with Sabin’s assumptions, this book has many good features. It is clearly written, although it lacks an index. Sabin clearly states his views and provides at least the outline of an argument for each. The book begins with an excellent summary of the problems of studying ancient battles, with each new approach supplementing but not replacing the techniques available in Delbruck’s day. The wargame is fun to play, and must have been great fun to write. The one battle I played (the Granicus, based on Arrian’s version of events) ended in the plausible outcome of the Persians fled but Alexander and many of his men dead. And Sabin acknowledges many of these potential criticisms in the text. Studying so many battles leads to some interesting observations on ancient warfare in general, such as that the infantry line was usually deeper in larger armies. But in the end, each student is still dependent on critical reading, comparative evidence, and their own intuition to reconstruct these lost battles.