Collection: Total War’s Missing Infantry-Type

This week, we’re going to take a break from the more serious topics to look at infantry tactics and compositions in the Total War series, particularly in the light of the recent Total War: Warhammer III, a real-time strategy game set in a late-medieval/early-modern high fantasy setting (the Warhammer setting) and how well (or poorly) some of those tactics correspond to their actual real-world historical counterparts.

In particular, I want to look at how Creative Assembly, the studio behind this series of games, is attempting to create infantry units that occupy a flexible space between ranged and melee combat in an effort to shake up a Total War formula that increasingly is beginning to feel a touch stale. But in the process, as we’ll see, Creative Assembly has essentially given up on modeling how firearm infantry in particular worked on most early modern battlefields and moreover has never actually attempted the most common historical solution to the need for an infantry unit that could engage in both fire (ranged combat) and shock (melee combat), what I’ll call ‘composite’ infantry units.

As always, if you like what you are reading here, please share it; if you really like it, you can support me on Patreon. And if you want updates whenever a new post appears, you can click below for email updates or follow me on twitter (@BretDevereaux) for updates as to new posts as well as my occasional ancient history, foreign policy or military history musings.

(The banner image is a mix of a screenshot from Total War: Warhammer II and, via Wikipedia, a model of a section of pike and shot from the Army Museum in Stockholm.)


What put me in mind of this topic was actually the unique faction quirk for the new human faction in Total War: Warhammer III, Kislev. Kislev’s unique tactical element is that most of their infantry units are ‘hybrid’ units, capable of fighting effectively in both melee and at range. This is unusual in Total War generally: most units that have any kind of ranged attack are lightly armored and poorly armed and thus if forced into a close-combat fight will handle it poorly.

The inspiration here seems to be the historical streltsy, which were, prior to Peter the Great (r. 1682-1725), the standard firearm-infantry of Muscovy/Russia. You can tell this inspiration was on the minds of the developers because the streltsy are just straight up in the game as a unit, called ‘streltsi’ (and I am going to use the spelling difference there to distinguish between historical streltsy and the game’s streltsi).

Via Creative Assembly’s Kislev Roster Reveal, an art reference for the Streltsi. The actual historical streltsy wore far less armor than this, being essentially unarmored because of course they were primarily firearm troops.

Of course the game’s streltsi must have a fantastical element and in this case it is their melee weapon. The historical streltsy were primarily arquebusiers (that is, firearm troops) who used a relatively short (usually at or below five feet) polearm, the bardiche, alongside their firearms. This isn’t as radical as it sounds. Early muskets and arquebuses were long and very heavy and so generally used with a fort-rest, essentially a monopod that supported the end of the barrel. The historical streltsy combined that with their bardiche which lacked a halberd’s spear-point and so had a flat space to rest the arquebus. This of course forced significant compromises in the design of the bardiche – thus its short length.

In-game, the streltsi take this up to eleven by combining their bardiche with their arquebus, replacing the rest of the firearm with an axehead. This looks cool, but is, I should note, utterly impractical – it makes a very heavy firearm even heavier (an arquebus, not including shot and powder, might weigh around 5kg, generally heavier than a full length halberd) without the obvious utility of the arquebus-rest (meaning that in a real world scenario, rather than a game, they’d have to carry a fork-rest too). Also it means firing a weapon with a significant kick while resting an axe-blade underneath your shoulder, so if something goes wrong, well, I hope you didn’t need that arm. Of course the actual solution here to combine the two weapons was the bayonet, which arrived in Russia in the early 18th century under Peter the Great (r. 1682-1725), along with lighter muskets that no longer required a firing rest. The bayonet allowed musket-infantry to be their own pikemen thus at last driving the pike from the battlefields of Europe (on pikes, see below).

Creative Assembly took this ‘hybrid’ concept and then applied it to much of Kislev’s infantry roster. ‘Kossars’ are archers who are also halfway decent in melee, ‘Armored Kossars’ are front-line heavy infantry who also carry pistols which they can volley, and the ‘Ice-guard’ are both high-damage archers and high-damage (but fragile) infantry. For most other factions, a ‘switch-hitting’ melee-ranged unit is typically an elite late-game unit (e.g. Sisters of Avalorn, Shades), but chances are any given Kislev army is going to be mostly composed of such units, which as a faction is what makes them unique.

Via Creative Assembly’s Kislev Roster Reveal, an art reference for the Ice Guard, a hybrid bow-and-polearm (or bow-and-two-swords) unit and Kislev’s elite infantry. I have some problems with the tendency for these units to armor their upper chest but not their belly, but more to the point, polearms like these ice-glaives have to be carried in the hands, so where does the glaive go when you need both hands to fire the bow?

Hybrid Infantry in the Real World

What I find interesting here is that this unit type, ‘hybrid infantry’ – specifically effective, armored hybrid heavy infantry – seems to be increasingly common in Total War games (especially Warhammer), in part because it provides interesting tactical options, but actual historical hybrid infantry of this sort is relatively rare. I should be clear here we are dealing with infantry today – mounted archers (e.g. steppe warriors) often carried contact weapons and shifted between them and bows (carried in bow-cases when not in use). But infantry that was expected in the normal course of battle to both fight at range and also in ‘shock’ (meaning in melee combat), while not unheard of, was relatively rare.

The actual streltsy at best skirt the definition themselves. Streltsy literally means ‘shooter’ in Russian as far as I understand (I do not speak Russian) and the streltsy themselves were actually pretty clearly dedicated firearms troops. They were, for instance, generally unarmored (except for a helmet) in a period where shock cavalry and shock infantry (e.g. pikes) generally still wore at least some armor. And while those bardiches were effective close-combat weapons, the streltsy do not seem to have relied on them to, for instance, ward off cavalry. Instead, streltsy fought from ‘gulyay-gorod‘ (‘walking forts’), which is to say a form of ‘wagon fort’ not too dissimilar to how the Ottomans and Mughals protected there musketeers on the battlefield from fast moving cavalry (but quite dissimilar to the pike-and-shot system of central and western Europe). So while the real streltsy had some shock capability, unlike Kislev’s units in TW: Warhammer, they were unarmored and expected to use field fortifications to keep them out of shock, particularly against enemy cavalry. Attempting to send unarmored streltsy with their short bardiches into a ‘push of pike’ (the collision of two opposing pike squares) would have been a recipe for disaster: these were arquebusiers, first and foremost.

Via Wikipedia, a soldier of the streltsy. Note how he wears effectively no armor, being primarily a musketeer.

In a sense, streltsy weren’t all that remarkable. Most missile troops on the battlefield carried some kind of ‘backup’ weapon in case they found themselves in close combat. Look at artwork of 16th and 17th century musketeers and you will see that nearly all of them carry a sword, even though their general lack of armor is usually a clear indication that these troops do not expect themselves to be in close-combat. The streltsy weaponized their arquebus rests, which is an interesting innovation, but doesn’t actually give them radically different capabilities from a western musketeer who carries a sword (neither of them would be expected, for instance, to stand up to cavalry). So while the in-game streltsi are true hybrid units – at home in melee or at range – the actual streltsy were dedicated musketeers who carried a bardiche mostly as a backup weapon.

Via Wikipedia, a 1664 woodcut showing a musketeer. You can see both his sword as a backup weapon, but also his musket-rest. Like the streltsy, he is unarmored.

True ‘hybrid’ units of this sort are hard to find, in fact. The Roman heavy infantry carried javelins (the pilum) and later darts (the plumbata), but these were hurled in preparation for a charge and fairly clearly secondary weapons to the legionnaire’s gladius or later spatha. Missile troops of all types generally carried sidearms and we do occasionally see archers in some degree of armor during the Middle Ages, but these were still clearly dedicated missile troops; they carried melee sidearms much the way a tank crew carries pistols, rifles or shotguns: for emergencies when something has gone wrong. Elite Persian infantry, like the Immortals, seem to have been expected to engage in both melee and ranged fighting, but it is hard to conclude they could hold their own against dedicated shock infantry unsupported, given that Persian infantry tended to fare very poorly when in close combat against dedicated shock infantry (e.g. hoplites) and Persian successes (and there were Persian successes) often relied on avoiding head-on-head collisions with more heavily armored hoplites. So we can’t say forms of ‘hybrid’ infantry didn’t exist – they did – but they were less common; far more common were missile infantry that carried backup weapons.

But if true hybrid infantry was relatively rare historically, how might infantry fill this capability gap?

Composite Infantry

I should be clear I am making this term up (at least as far as I know) to make a contrast between what Total War has, which are single units made up of soldiers with identical equipment loadouts that have a dual function (hybrid infantry) and what it doesn’t have: units composed of two or more different kinds of infantry working in concert as part of a single unit, which I am going to call composite infantry.

This is actually a very old concept. The Neo-Assyrian Empire (911-609 BC) is one of the earliest states where we have pretty good evidence for how their infantry functioned – there was of course infantry earlier than this, but bronze age royal records from Egypt, Mesopotamia or Anatolia tend to focus on the role of elites who, by the late bronze age, are increasingly on chariots. But for the early iron age Neo-Assyrian empire, the fearsome effectiveness of its regular (probably professional) infantry, especially in sieges, was a key component of its foreign-policy-by-intimidation strategy, so we see a lot more of them.

That infantry was split between archers and spear-and-shield troops, called alternately spearmen (nas asmare) or shield-bearers (sab ariti). In Assyrian artwork, they are almost always shown in matched pairs, each spearman paired off with a single archer, physically shielding the archer from attack while the archer shoots. The spearmen are shown with one-handed thrusting spears (of a fairly typical design: iron blade, around 7 feet long) and a shield, either a smaller round shield or a larger ‘tower’ shield. Assyrian records, meanwhile, reinforce the sense that these troops were paired off, since the number of archers and spearmen typically match perfectly (although the spearmen might have subtypes, particularly the ‘Qurreans’ who may have been a specialist type of spearman recruited from a particular ethnic group; where the Qurreans show up, if you add Qurrean spearmen to Assyrian spearmen, you get the number of archers). From the artwork, these troops seem to have generally worked together, probably lined up in lines (in some cases perhaps several pairs deep).

Via Wikipedia, Assyrian Archers and Shield-Bearers deployed in pairs from the Lachish Relief (c.700-681 B.C.) originally in Ninevah, now in the British Museum.
Note how each shield-bearer also carries a weapon (they can be hard to see with the state of the relief and the background).

The tactical value of this kind of composite formation is obvious: the archers can develop fire, while the spearmen provide moving cover (in the form of their shields) and protection against sudden enemy attack by chariot or cavalry with their spears. The formation could also engage in shock combat when necessary; the archers were at least sometimes armored and carried swords for use in close combat and of course could benefit (at least initially) from the shields of the front rank of spearmen.

The result was self-shielding shock-capable foot archer formations. Total War: Warhammer also flirts with this idea with foot archers who have their own shields, but often simply adopts the nonsense solution of having those archers carry their shields on their backs and still gain the benefit of their protection when firing, which is not how shields work (somewhat better are the handful of units that use their shields as a firing rest for crossbows, akin to a medieval pavisse).

We see a more complex version of this kind of composite infantry organization in the armies of the Warring States (476-221 BC) and Han Dynasty (202 BC – 220 AD) periods in China. Chinese infantry in this period used a mix of weapons, chiefly swords (used with shields), crossbows and a polearm, the ji which had both a long spearpoint but also a hook and a striking blade. In Total War: Three Kingdoms, which represents the late Han military, these troop-types are represented in distinct units: you have a regiment of ji-polearm armed troops, or a regiment of sword-and-shield troops, or a regiment of crossbowmen, which maneuver separately. So you can have a block of polearms or a block of crossbowmen, but you cannot have a mixed formation of both.

Via Wikipedia, a bronze ji.

Except that there is a significant amount of evidence suggesting that this is exactly how the armies of the Han Dynasty used these troops! What seems to have been common is that infantry were organized into five-man squads with different weapon-types, which would shift their position based on the enemy’s proximity. So against a cavalry or chariot charge, the ji might take the front rank with heavier crossbows in support, while the sword-armed infantry moved to the back (getting them out of the way of the crossbows while still providing mass to the formation). Of course against infantry or arrow attack, the swordsmen might be moved forward, or the crossbowmen or so on (sometimes there were also spearmen or archers in these squads as well). These squads could then be lined up next to each other to make a larger infantry formation, presenting a solid line to the enemy.

(For more on both of these military systems – as well as more specialist bibliography on them – see Lee, Waging War (2016), 89-99, 137-141)

The Cathay faction of Total War: Warhammer III also models a fantasy version of this system, but like Three Kingdoms doesn’t include them as composite formations, although it does encourage the player to keep contact infantry and missile infantry in close formation (but in separate blocks).

Pike and Shot

This is a particularly acute problem for direct fire missile infantry, which in Total War games generally means firearm infantry. Whereas Total War allows bows and crossbows to fire at high arcs with high effectiveness – which they generally shouldn’t – firearms, in exchange for their higher damage have to fire on flat trajectories and thus need uninterrupted lines of sight, meaning they cannot simply be tucked into the back rank to fire over the front ranks.

Instead most players use some form of a checkerboard formation to combine contact and direct-fire missile infantry (and often archers too). Checkerboard formations were not entirely unknown – the Roman triplex acies is essentially a checkerboard – but they did not work the way that they are used in many Total War games. Whereas in a Total War checkerboard, the ‘lanes’ are used to allow missile units to fire without being on the front rank, the ‘lanes’ of the Roman triplex acies were used for maneuver, particularly for withdrawing the skirmishing screen (the velites) and, if things went badly, the front lines of infantry back behind fresh rear ranks. By contrast, the core use of checkerboards in Total War isn’t lanes to withdraw forward troops, but rather getting enemy units ‘hung up’ on front line formations so that ranged units can fire at them through the gaps, taking advantage of the fact that individuals in a Total War formation cannot push too far from the rest of their unit and so if just a few soldiers get ‘caught’ on the edge of an infantry formation, it will prevent the entire block from just pushing through the gap.

A fairly standard ‘checkerboard’ Total War formation with Reikland. I play almost exclusively single-player, but my understanding is that in a multiplayer setting, you’d want the intervals between units to be wider than this. Still, contrast this battle array with the one at Nieuwpoort (1600) below to see the difference.

But slow-firing arquebus infantry were exquisitely vulnerable on early-modern battlefields, particularly to cavalry which might cross the entire effective firing range of these weapons so fast the arquebuses might only have one or two shots. One solution we’ve already noted was to put them behind wagons and other moveable barricades – the obstructions would prevent a cavalry charge on the firearms.

For a more mobile option, we can look to the tactical system that emerged in Western and Central Europe during the early modern period: pike and shot. Pike and shot formations often employed a range of soldiers (including swordsmen and halberdiers) but the bulk of these units was made of soldiers with pikes (typically armored with at least a helmet and a cuirass) and soldiers armed with firearms (muskets or arquebuses; typically unarmored with a sword as a backup weapon), that is the ‘shot.’ The concept was deceptively simple: a large square of pikemen created a safe base for the ‘shot’ to fire from. Early on, the ‘shot’ mostly existed to support and screen the pike, with the expectation that the pikes would be the decisive offensive weapon seeing off enemy pikemen in what was termed the ‘push of pike.’ As firearms got better and more importantly firearm drill got better, the pike increasingly became the supporting arm to the shot. The ratios of pike to shot in these units changed accordingly, initially favoring more pikes, later more shot.

Via Wikipedia, the armies at the Battle of Nieuwpoort, 1600, depicting two phases of the battle. Here you can see somewhat later tercios, deployed with their ‘sleeve’ of shot. on the left (the Anglo-Dutch forces) while on the right the Spanish make use of a more traditional bastioned square with a thin sleeve.

And so far this doesn’t sound entirely different from Total War ‘checkerboarding,’ but as always the devil is in the details. Actual cavalry, after all, could have rushed through the gaps of any sort of checkerboard, unlike the cavalry formations of Total War which get ‘stuck’ as described above.1 Indeed, this is why the pike fought in squares: it was assumed the cavalry was mobile enough to strike a group of pikemen from any direction and to whirl around in the empty spaces between pike formations, so a given pike square had to be able to face its weapons out in any direction or, indeed, all directions at once.

Instead, pike and shot were combined into a single unit. The ‘standard’ form of this was the tercio, the Spanish organizational form of pike and shot and one which was imitated by many others. In the early 16th century, the standard organization of a tercio – at least notionally, as these units were almost never at full strength – was 2,400 pikemen and 600 arquebusiers. In battle, the tercio itself was the maneuver unit, moving as a single formation (albeit with changing shape); they were often deployed in threes (thus the name ‘tercio’ meaning ‘a third’) with two positioned forward and the third behind and between, allowing them to support each other. The normal arrangement for a tercio was a ‘bastioned square’ with a ‘sleeve of shot:’ the pikes formed a square at the center, which was surrounded by a thin ‘sleeve’ of muskets, then at each corner of the sleeve there was an additional, smaller square of shot. Placing those secondary squares (the ‘bastions’ – named after the fortification element) on the corner allowed each one a wide potential range of fire and would mean that any enemy approaching the square would be under fire at minimum from one side of the sleeve and two of the bastions.

Via Wikipedia, a bastioned square with a ‘sleeve of shot,’ training diagram c. 1600. Note that by this point the muskets were far more numerous and central than the pikes.

That said, if drilled properly, the formation could respond dynamically to changing conditions. Shot might be thrown forward to provide volley-fire if there was no imminent threat of an enemy advance, or it might be moved back to shelter behind the square if there was. If cavalry approached, the square might be hollowed and the shot brought inside to protect it from being overrun by cavalry. In the 1600s, against other pike-and-shot formations, it became more common to arrange the formation linearly, with the pike square in the center with a thin sleeve of shot while most of the shot was deployed in two large blocks to its right and left, firing in ‘countermarch’ (each man firing and moving to the rear to reload) in order to bring the full potential firepower of the formation to bear.

Indeed it is worth expanding on that point: volley fire. The great limitation for firearms (and to a lesser extent crossbows) was the combination of frontage and reloading time: the limited frontage of a unit restricted how many men could shoot at once (but too wide a unit was vulnerable and hard to control) and long reload times meant long gaps between shots. The solution was synchronized volley fire allowing part of a unit to be reloading while another part fired. In China, this seems to have been first used with crossbows, but in Europe it really only catches on with muskets – we see early experiments with volley fire in the late 1500s, with the version that ‘catches on’ being proposed by William Louis of Nassau-Dillenburg (1560-1620) to Maurice of Nassau (1567-1625) in 1594; the ‘countermarch’ as it came to be known ends up associated with Maurice. Initially, the formation was six ranks deep but as reloading speed and drill improved, it could be made thinner without a break in firing, eventually leading to 18th century fire-by-rank drills with three ranks (though by this time these were opposed by drills where the first three ranks – the front kneeling, the back slightly offset – would all fire at once but with different sections of the line firing at different times (‘fire-by-platoon’)).

Via Wikipedia, William Louis of Nassau-Dillenburg’s diagram of his countermarch technique.

Coming back to Total War, the irony is that while the basic components of pike-and-shot warfare exist in both Empire: Total War and for the Empire faction in Total War: Warhammer, in both games it isn’t really possible to actually do pike-and-shot warfare. Even if an army combines pikes and muskets, the unit sizes make the kind of fine maneuvers required of a pike-and-shot formation impossible and while it is possible to have missile units automatically retreat from contact, it is not possible to have them pointedly retreat into a pike unit (even though in Empire, it was possible to form hollow squares, a formation developed for this very purpose).

Indeed if anything the Total War series has been moving away from the gameplay elements which would be necessary to make representing this kind of synchronized discipline and careful formation fighting possible. While earlier Total War games experimented with synchronized discipline in the form of volley-fire drills (e.g. fire by rank), that feature was essentially abandoned after Total War: Shogun 2‘s Fall of the Samurai DLC in 2012. Instead of firing by rank, musket units in Total War: Warhammer are just permitted to fire through other members of their unit to allow all of the soldiers in a formation – regardless of depth or width – to fire (they cannot fire through other friendly units, however). That’s actually a striking and frustrating simplification: volley fire drills and indeed everything about subsequent linear firearm warfare was focused on efficient ways to allow more men to be actively firing at once; that complexity is simply abandoned in the current generation of Total War games.

Reducing Complexity

Now for fans of the Total War series, it is in some ways not difficult to see why the complexities of composite infantry and synchronized musketry have been abandoned or were never attempted at all: the complex internal movements and maneuvers are finicky and ‘break’ easily. That problem was compounded by the shift from sprites to 3D models: if a pixelated, low-rez sprite needs to skip part of its reloading animation or slide awkwardly into position, most players won’t notice, but if a fully modeled 3D soldier does so, it would look absurd. But then without that ability to ‘fudge’ and without the flexibility of a unit composed of actual humans, getting the complex drill to go off properly ends up being difficult.

Players who played Empire: Total War will know this difficulty well. Fire-by-rank drill ends up being an extremely powerful unit ability, but getting it to ‘go off’ right requires a lot of ‘baby-sitting’ – units have to face just right, be positioned just right and so on. There is nothing so frustrating as watching a 100+ unit of muskets fail to execute its fire-by-rank attack because a handful of fellows on the corner were reloading at the wrong time and for some reason this meant that the whole unit had to stop and wait for them. Meanwhile, fire-by-platoon exists in Empire: Total War but is entirely broken, actually being less effective than fire-by-rank because the game can’t handle reloading by platoons, with the result that each platoon fires one after the other, but everyone waits to start reloading until after the last platoon has fired. Fall of the Samurai brings in ‘kneel-firing’ but it too is finicky and liable not to work except under utterly perfect conditions.

And it seems clear to me that at some point, Creative Assembly simply looked at these complexities and decided they weren’t worth it and abandoned them, much like they largely abandoned naval battles and the more complex trade and economic models of their earlier games. And I think, frankly, that is a shame. It leaves a noticable gap in terms of historical ways of fighting, because ‘composite’ infantry executing synchronized fighting drills were both common, but also tended to be the best and most lethal sorts of infantry in their day: this was the fighting style of the trained, elite professionals.

I also think that it is a challenge Creative Assembly ought to return to. It is by now a commonplace that the Total War franchise has been stuck in something of a rut. Total War: Rome II (2012) and its standalone expansion Attila (2014) both frustrated many fans; Thrones of Britannia (2018), Three Kingdoms (2019) and Troy (2020) all received criticism for having limited unit variety, armies that felt ‘samey’ and for not innovating enough off of the basic formula that, in the purely historical games, Creative Assembly hasn’t really ‘nailed’ since Shogun 2 back in 2010. Total War: Warhammer has avoided this fate in part because the fantasy setting lends itself to a lot of unit variety, but it isn’t hard to see with factions like Kislev and Cathay that the developers are struggling to make factions that play sufficiently differently from one another to keep things interesting (though I’d argue they mostly succeed; TW: Warhammer III would be an excellent addition to the franchise if it weren’t for how much of a slog its primary campaign is; all eyes now turn to the final combined campaign, ‘Immortal Empires’ to arrive some time in the future).

One way then to break up the formula would be to begin experimenting again with trying to get those more complex systems to work. It is really striking looking at the series that Empire: Total War really was the peak of Creative Assembly’s ambition: a global map, naval battles, a complex trade simulation, multiple forms of government, elections and yes – synchronized infantry drill (though not with composite units). Their games sense then have steadily reduced that scope of features, pruning down to a core that works well (you can see that core in particular working well in TW:Warhammer and Three Kingdoms), but lacks variety.

Creative Assembly needs to take risks again, so this is my suggestion: go back to Empire. I’d love to see the series put the resources into tackling the more complex combat of the 1500s and 1600s (Empire originally had a start-date of 1700, so this would be an expansion of scope), incorporating both composite infantry units (that is, the pike-and-shot system) that actually work as composite units with the ability to alter formations as necessary to respond to changing conditions and use synchronized firing drills.

Because to innovate in a franchise like this, Creative Assembly doesn’t need gamey ahistorical gimmicks: there’s plenty enough complexity in actual historical warfare to allow the series to continue moving forward and staying fresh and at the same time, Creative Assembly has the resources to actually tackle these more complex fighting systems at a higher degree of visual fidelity and detail than other, valiant efforts to model the period.

And that’s it for this week, though I should note we’ll be back to Total War not too far in the future to talk about how it (mis)represents the command of pre- and early-modern armies, as a window into looking at how trying to wrangle an army like that on the battlefield actually worked.

  1. That ‘sticking’ behavior in turn is pretty clearly a necessary response to the inability to make composite infantry formations work in Total War. Without it, it would be casually easy to ‘push’ cavalry through or around blocking infantry to get to enemy ranged units (and indeed, this was obnoxiously easy in some Total War games and is still a problem with ‘monstrous infantry’ in Total War: Warhammer)

208 thoughts on “Collection: Total War’s Missing Infantry-Type

  1. Not directly relevant, but my experience with Total War games, (Which, I should mention, is limited to Rome and Medieval 2, as well as various mods and DLCs for those games) has a historical inaccuracy with them that drives me absolutely nuts. Someone seems to have heard that the majority of deaths in a battle occurred in the chase after a rout and conflated that with the notion that most people who routed and fled were cut down as they ran. Battles are enormously too bloody for realism, and while I don’t have records from my playing to go through it item by item, my gut sense of major battles between relatively even armies is one that ends with the winning army taking 20-25% losses and the losing army taking somewhere between 85%-100% losses.

    Battles like Cannae were famous because total wipeouts like that were very rare. And the overly bloody and decisive battles have downstream effects. You don’t get anywhere near as much frictional losses when campaigning. Some of that is because the game doesn’t really model it except for sieges, but also because campaigns are so much shorter than their historical counterparts, which is tied to the fact that one or two wins will completely eliminate your enemy’s field armies and then you can just mop up their cities most of the time.

    I don’t actually have much of a point, just ranting. Oh, and as another aside, there was this old Lord of the Rings RTS, I think it was called Battle for Middle Earth, which allowed for the creation of composite formations like what’s being described here. BFME came out in 2004, so you’d think it would be at least theoretically possible for Creative Assembly to copy the idea. Although IIRC, the composite formations in BFME were kind of weak, so you’d have to make the mechanics of battle work too.

    1. Also your military is composed entirely of professionals; disbanding a bunch of units or an entire army stack doesn’t suddenly boost the taxable/productive population of your empire or even that region. Those three thousand-some humans who are no longer marching under your banner just sort of cease to exist.

        1. Certainly would be interesting to have a game where the “disband unit” option, rather than simply despawning, popped up a prompt with a sliding scale of options between “pay out retirement pensions, on time and in full” at one extreme and “100% unit strength becomes uncontrolled and hostile immediately” at the other, with various increments of downgraded and/or delayed banditry in between.

          1. The recent wargame/4X Shadow Empire has a “Recruits” item, which is required to make Troops that compose a Unit.

            (Which can be indeed composite, with up to 6 different Troop types, including transports !)

            To draft lots of Recruits, fast, you need to have a loyal Populace to recruit from, and/or tell your Zone Governor to pay Recruits extra for sign-up, and/or have a high “Psychology : Fist” Regime Profile, and/or a high Cultural Adaptation (to your “main” culture) Score for that Zone, and/or use a Stratagem costing Bureaucratic and Political Points and requiring your Interior Council Director to make a high enough Leadership roll.

            This will leave less Populace to do Public or Private jobs – especially noticeable if you use lots and lots of cheap infantry. (Whose deaths, if high enough in a short time, will impact Populace’s Happiness and Loyalty.)

            Before being able to Raise a Unit with Troops, your Logistical Network will need to use Strategic Logistic Points to bring these Recruits to a/the Strategic HeadQuarters. And while staying there, they need to be fed with that HQ’s Food (and IIRC also paid the usual Soldier Salary ?)

            At low levels of militarization (and/or low Cultural Adaptation Score), and high Zone Militancy, there’s also Zone Militia, where all of this, including the raising of Units with mishmashed Designs, happens automatically, without your control.

            When Scrapping Troops, the 100-people items will be returned as that Zone’s Population. If the unit is not in one of your Zones, they will join one of the Free Folk Towns of that Zone.

            However, AFAIK, in all of this attention to detail, there’s *still* one issue : typically one of the most effective ways in ShEmp to defeat a Unit is to lower its Entrenchment, Readiness (= cohesion ?) and Morale and not leave it any hexes to retreat to. At which point, most of the Troops will quickly surrender (during a combat). But also AFAIK, they will then just disappear from the map, not even joining the Free Folk somewhere ! (or some kind of prisoner camp)


          2. That would be the historical norm. Veterans tended to be hard to re-integrate, accustomed as they were to violence. One solution was to send them off on crusade or to some other war far away. You could be sure most would not come back.

        2. I’m not sure what game it was, but one of the games in the Medival/Rome/Medival2 Era, hat disbanded units show up, in the mercenary tab, of that region..

          1. Pretty sure that was a thing in Rome: Barbarian Invasion, which had a lot of demobilized legionnaire merc types. Also I think in Rome recruiting troops reduced population and disbanding them increased it.

          2. > Also I think in Rome recruiting troops reduced population and disbanding them increased it.

            I remember about this mechanism being used as a way to move population between provinces

          3. It was Medieval (the original) which had disbanded units becoming mercenaries. Barbarian Invasion has veteran legionary mercenaries, but they’re generated randomly like all mercenary units, rather than being created by disbanding your regular units.

      1. In Rome I they don’t disappear – one of the quickest way of getting the Marian reforms is to hire large numbers of peasants, send them to your capital region, and disband them. On the old large unit scale, each would account for 240 extra population (while also depleting your population from territories you recruited them from). This was especially useful if you continuously enslaved settlements, as those populations would be transferred equally to all your provinces with governors, where you could then focus them into a single region (or just have a single governor in the region you wanted).

      2. This is actually not true for the original Rome. Settlements have populations that go down when you recruit soldiers and up when you disband, in addition to (unrealistically fast) natural growth. You could exploit this to move population from overpopulated or fast-growing provinces to smaller ones you wanted to grow.

        1. Well, we can consider that growth not as the actual population growth of the region, but as the population that is ‘legible’ to the state, so it’s not necessarily unrealistic.

        2. I think the population growth is reasonable if you interpret it as urbanization increasing the population of the major legible settlements that are easily taxed and conscripted, rather than the population of the province as a whole.

    2. I have had the exact same concern as you Adam–I only ever played Rome, Empire (and TW1) but I was always troubled by the incredibly ahistorical casualty numbers (it also seems like casualties for the victorious sides tended high–often like 95% casualties for the losers, 50% for the winners). When you compared them to casualty statistics on wikipedia they were eyepopping. From a gameplay perspective too it was a bit annoying because losing a battle, outside of rare circumstances, usually ended the life of that field army (although the game did help you a bit by letting you skip the rout of your own forces if you choose to end the battle).

      Is there anything in the historical record that explains how normal soldiers escaped from the rout? I assume the reason in real life that you could not use your cavalry to run down every single fleeing infantryman is because in real life it would become night or they would enter the brush and it would just be too hard? And presumably infantrymen would run too slowly to pursue fleeing infantry (who had likely abandoned some of their heavier weapons and armour)? I have heard of battles (like Towton) where it sounds like the pursuit continued for an extensive period of time but even in that case, the casualties for the loser are still below 50%.

      1. There might also be some cases where the losing side has enough cavalry left around that the victorious side’s cavalry needs to remain reasonably concentrated, lest dispersed elements be picked off in detail. Restricting their ability to fan out would seem to vastly slow their ability to run down fleeing soldiers.

      2. I’ve never read an academic account of a rout and trying to survive routing, but based off of a number of battle descriptions, I would hazard a guess as to the following factors that led to incomplete pursuits of routing foes.

        First off though, we do have accounts of infantry pursuing other infantry and bringing down at least some of them. It was a fairly common feature in hoplite warfare, and classical period fights between different Greek poleis rarely had cavalry fielded in any significant fashion.

        But even assuming there is a complete rout and the victorious force does have enough cavalry to initiate a pursuit, there are several countervailing factors that might prevent them from hunting down everyone. In no particular order:

        1) Unlike TW games, routing soldiers are not completely helpless. Granted, they’re far more vulnerable than soldiers still in some sort of functioning formation, but you can’t have the TW situation where you have 2 cavalrymen chasing 100 infantry and will eventually kill all of them without any harm. There is a definite risk of those infantrymen rallying at least long enough to kill or drive off too-weak pursuit, which in turn means that pursuing forces tended to stick together.

        2) Adding to the above, most soldiers are far more interested in making sure they stay alive than they are in killing the enemy. Killing the enemy who is actively threatening them and their friends will of course help the primary goal of staying alive, but actively seeking enemies who are in the middle of running away in a panic is an action that will increase your personal risk as a soldier in the victorious army, not decrease it. (Unless and until the routing soldier from the defeated army is rallied and comes back at you another day, but that might or might not even happen, and in any event, thinking that long term ahead is not exactly assured). So pursuit might not even happen, or be half-hearted when it is, except from the most motivated of victorious soldiers.

        3) Communication limitations. Most of the time, commanders would set themselves up a good distance behind their main line of battle. And if your side won, chances are the bulk of your forces have advanced from that position, with the “command post” might or might not advancing with them. Actually sending messengers to formations that were doing something else pursuant to the battle plan, possibly collecting them up if they’ve gotten dispersed and sending them to go chase after a fleeing mass of enemy soldiers takes time, and that gives the routing soldiers time to get away. (And of course, TW games completely eliminate these problems)

        4) Looting. Lots of armies, once victory was clear, stopped to loot. Once that happened, getting them to go chase enemy soldiers would be difficult at best. Doubly so since throughout a lot of history, equipment was self-provided, which usually meant that the more lightly armored troops (who would presumed to be the best pursuers) were also the poorest soldiers in the army, and the most to relatively gain by nabbing something off of the corpse of a fallen soldier.

        5) Pretty much every battle produces wounded, and in a related vein to #2, once it became clear that the action was over for the day, you’d see wounded soldiers wanting to tend their own injuries; their comrades helping them, especially if they were too badly hurt to easily move; and command elements organizing triage. All of that is manpower and attention taken away from pursuing the enemy. Where mounted troops are involved, also expect care to go to the horses and any injuries they might have. Good horses are expensive.

        Now again, I haven’t done any major research on this. I’m just rolling stuff together from dozens of half-remembered sources and applying my own skein of logic to it. But I suspect these are at least some of the limiting factors for how much a pursuit could actually do to destroy an enemy.

        1. Another point is simply that battles were often long and tiring, and even if you had cavalry units still around they’d still have a limited amount of stamina left for pursuit tasks.

        2. The first thing victorious soldiers do is head for the enemy camp – that’s where the loot and drink and women are. Getting them to leave that and pursue is a major job. A second point is that even when a battle-line breaks, often some elite unit will stand and cover the retreat (eg the Sacred Band at Chaeronea). They get wiped out, but more of their fellows survive. A third is that fleeing soldiers scatter in all directions. As the victors disperse in chase the odds can change (eg a group of the defeated can notice that they now outnumber the pursuers). Finally, good cavalry were trained to stick together, in case of a rally by the defeated. This limited pursuit – but limited the chance of a sudden reversal (the Parliamentary and then the Royalist forces both learned this the hard way in the English Civil Wars).

          1. The first two were why Sheridan was able to turn a Union rout into a Confederate rout in the Shenandoah valley — and without even the elite unit being wiped out.

        3. Another thing that applies — occasionally — is that fleeing soldiers may be faking it to draw you into a trap. See the Battle of Cowpens.

        4. The godlike command&control TW allows players is first and foremost the biggest impediment to what commanding an actual battle is like, even in the modern age no commander has that much information.

          Leaving that aside, there are a handful of composite units in 3K which I hope CA continues to iterate on in future historical titles. Warhammer is a bit difficult to rely on for how CA is adjusting its formula because 90% of the units are based on GW’s tabletop gaming system which leaves CA much less room for introducing its own concepts and locks in hybrid units rather than allowing composite units.

          If CA wants to lower the casualties in battles AND make battles slightly less decisive without hugely increasing the # of battles it could make 2 important changes;

          1. How armies are recruited and maintained is incredibly unrealistic in TW and really not necessary. Simply it is the traditional formula but is ripe for being shaken up.

          Each faction should have a basic army that costs 0 upkeep but is really just the ruler and his family/friends + whatever men they can press into service, remaining in controlled territory this army has morale bonuses and costs nothing. Using this basic army offensively leads to decreased morale and increased upkeep costs depending soley on the skills of the ruler leading the army in person. Not many kings would be satisfied with such an army since it requires their presence to be effective and limits the size of territory they can control.

          The campaign side of the game would be about constucting infrastructure to support a better army that can fight offensively by enlisting competent leaders, that can lead armies without the king, unlocking more types of units and supporting those units so they have good morale and won’t run away if they think they are likely to lose.

          Defeating an army wouldn’t automatially win a war as it might only wipe out a couple of more professional units who fight longer taking more losses before routing but the bulk of an army would be able to retreat and fight again especially when it was in friendly territory. Successive defeats would hugely weaken the quality and morale of a factions armies, with deaths in the royal family or of famous commanders having higher consequenes (death vs capture & ransom being strategically important choices, lower morale or more money to support infrastructure/armies).

          This would result in the opening battles of a war setting the tone but not winning the war outright most of the time unless a faction is really small and only has a single army and 1 territory to retreat into. Defeating an enemy in successive battles or capturing/killing enemy leaders would weaken their morale and lead to gains while the bulk of a faction;s fighting forces might suffer light manpower losses the ruler would be more willing to negotiate up until the final battle where the existence of the regime is at stake (final battle in the enemies capital or last remaining province the morale debuffs from suffering losses up until then would be removed).

          2. Looting- CA has hugely increased the bonuses from winning battles in more recent TW games as loot can completely surpass regular income if a player maintains momentum and wins victories every turn but this comes at essentially zero cost to the player- it is simply a bonus given for winning a battle which increase even more the idea that winning a single battle is 100% a win and losing a battle is 100% losses- this increases the temptation to save scum and increases the grind of a campaign since to keep the ‘challenge’ going CA simply ensures there will be literally multiple 100s of battles in a campaign.

          Looting should be a strategic choice, units stop within 50 meters of defeated enemy units and begin to loot which means they can’t pursue fleeing enemy. If pursuit is given, units cannot loot. The choice to loot or not should also reflect in army morale for 3-5 turns following a battle. When the strategic imperitive is to not allow enemy units to escape and offer further resistance in a future battle player can order the pursuit but this means missing out on loot for that battle and lowering the morale of the army for a few turns. Highly disciplined, more professional units will suffer less from lower morale since they have relatively high morale but mercenaries, militias or low discplined units fight mostly for loot and being forbidden to loot might even make them mutiny.

          1. If you want historical accuracy, looting shouldn’t be an operational choice, and whether units loot in the first place would depend on their discipline.

            In general, commanders don’t want their troops to turn to looting en masse. It damages discipline, both in the immediate moment when you might want them to be pursuing the enemy or regrouping to support other units still engaged, and in the aftermath when the troops are scrapping over their ill-gotten gains, or otherwise trying to spend or fence them and not paying attention to their duties; the majority of the loot gathered by troops will be dispersed, rather than commandeered centrally, which is bad both operationally and for your own greed; and indeed some commanders simply viewed the whole exercise with distaste (see for instance Wellington after Vitoria).

            Indeed the lure of loot is something that should really be factored in not only once a battle is won but during it too. It has been a frequent complaint throughout history with less well-disciplined troops who win their immediate combat chasing the enemy off the field into the baggage train and looting rather than manoeuvring to finish off the enemy who were still fighting. During the (first) English Civil War, this was arguably *the* decisive factor in the Royalists’ inability to make gains in the first two years of the war, because their consistently superior cavalry would defeat their opposite numbers and then run off to loot, leaving the infantry to fend for themselves.(It’s notable that when Fairfax and Cromwell formed the New Model to improve Parliamentary battlefield capabilities, looting was expressly forbidden for all ranks).

            But it could still be a *strategic* choice, albeit an indirect one: by giving players a choice between whether to recruit more professional, better-disciplined units in the first place, with factors established at the recruitment stage (and perhaps refined through existing veterancy mechanics) determining how likely a unit is to turn to looting and – indeed – how responsive they are to commands generally (which I imagine may be one of the things Bret talks about next week).

            At the moment, and with the exception of some rare cases like frenzied or berserk units, discipline is kind of lumped in with morale (in the way that tabletop Warhammer has done since its fourth edition): units are pretty much all equally responsive on the field while they’re controllable at all. While there is some overlap between the two factors, really they ought to have visibly separate effects.

          2. Something on point (1) that I’ve noticed in recent TW games is that if a unit loses around 90% of its combat strength, but remains on the roster, the replenishment mechanic will replace the lost troops with troops of equivalent experience. Especially in a game with limited unit variety like Troy where veteran units (especially missile troops) are worth their weight in gold, this is a massive factor, because there’s no particular incentive to shield your veteran units from casualties: so long as the unit survives the battle at all, you’ll make back the losses (often with significant replenishment from recruiting enemy prisoners).

            In older games this wasn’t an option, partly because unit replenishment was either less significant or absent altogether. If your veteran unit suffered significant casualties, you had to merge the unit with existing units, which would dilute the overall experience of the unit. Level-9 troops couldn’t just be replaced: once lost, you had to rebuild your troops’ experience through combat.

            This is, of course, rather more similar to how it “should” be from a simulationist perspective. Oft-overlooked in common discourse about the OG Pyrrhic victories is that Pyrrhus actually suffered fewer casualties than the Romans in both battles: what made them crippling was not the number of casualties but that the fallen were disproportionately experienced and high-quality troops who he couldn’t replace. In TW, that just doesn’t happen any more.

            There has, as Bret notes in the article, been a steady tendency of “dumbing down” TW games, in my experience, since Empire: the mechanics have been simplified, the number of options reduced, etc. and it’s hard to see what the payoff for that is. It’s not like it has led to a noticeable improvement with regard to things like AI, pathfinding, or indeed unit collision… which mostly *also* seem to have got worse over the same period.

          3. As CA is at least borderline competent, I trust they would never implement the terrible decision of requiring the player to defeat the same army over and over repeatedly in order to finally get rid of it.

            If we want factions to be more persistent and difficult to destroy, killing and capturing *commanders* could be more troublesome (the way it sort of is in Three Kingdoms). I’ve long wanted the Shadow of Mordor Nemesis system to apply to Total War.

        5. 2) Adding to the above, most soldiers are far more interested in making sure they stay alive than they are in killing the enemy. Killing the enemy who is actively threatening them and their friends will of course help the primary goal of staying alive, but actively seeking enemies who are in the middle of running away in a panic is an action that will increase your personal risk as a soldier in the victorious army, not decrease it. (Unless and until the routing soldier from the defeated army is rallied and comes back at you another day, but that might or might not even happen, and in any event, thinking that long term ahead is not exactly assured). So pursuit might not even happen, or be half-hearted when it is, except from the most motivated of victorious soldiers.

          Plato has an interesting anecdote about this in his Symposium (Alcibiades speaking):

          “There was another occasion on which his behaviour was very remarkable, in the flight of the
          army after the battle of Delium, where he served among the heavy-armed, I had a better
          opportunity of seeing him than at Potidaea, for I was myself on horseback, and therefore
          comparatively out of danger. He and Laches were retreating, for the troops were in flight, and I
          met them and told them not to be discouraged, and promised to remain with them; and there
          you might see him, Aristophanes, as you describe, just as he is in the streets of Athens, stalking
          like a pelican, and rolling his eyes, calmly contemplating enemies as well as friends, and making
          very intelligible to anybody, even from a distance, that whoever attacked him would be likely to
          meet with a stout resistance; and in this way he and his companion escaped; for this is the sort
          of man who is never touched in war; those only are pursued who are running away headlong.”

      3. Hoplites normally escaped by dropping their shields, which enabled them to outrun their pursuers, still encumbered by heavy shields. Hence the expression, “Return with your shield or on it it,” i.e., return victorious (with your shield) or dead or severely wounded (borne on the shield as a stretcher).

      4. Lots of good responses here already, but I wonder how much of this discrepancy is due to the fact that in total war units must still rout *as a unit,* which can be chased down as a group, whereas in real life people can scatter in various directions and become much harder to pursue. Especially since in order to pursue, the pursuing army must itself break up to some extent and become vulnerable to counterattack thereby.

        1. Why *must* they though ? Couldn’t there be a “2nd level of rout”, at which point the unit ceases to exist (but survivors might reform after the battle). Or is it because this would require a problematic “scatter to pursue” order for units ?

          1. Even just sub dividing the unit so they can rout in a few stages, would help.
            So maybe some men hide, play dead or hang back and the unit does not collapse all at once but slows an breaks apart over time.

          2. Presumably its because of how the pathfinding is calculated… it would be much more cpu intensive if suddenly one unit became 64 one-man units, and it would become more of a micro chore to pursue them. I’d be interested to know more about that too.

          3. Routing units don’t really maintain any sort of formation and can get scattered quite a bit. It’s actually really annoying, because the unit chasing them stays in formation and seems to aim towards the unit standard rather than the closest individual model.

        2. In Warhammer 2, routing units will sometimes disperse. I don’t think they’re programmed to stick together, each individual makes a beeline for the edge of the map and that means they usually all end up heading there together. If you knock them about with cavalry or chariots they can get scattered.

      5. I think a large part of the answer is that armies historically didn’t tend to rout in the way they do in TW. The way TW battles work, the two armies square up, fight, and continue to fight until all units in one army are routed (or the timer runs out).

        In real life, battles would likely be much more circumspect. Once it became apparent that a battle was unwinnable, commanders might order their troops to disengage, but this wouldn’t necessarily result in a full-scale fleeing from the field, abandoning all defences. Rather, those units that remained in good order would withdraw in better order, dissuading the kind of pell-mell pursuit you see in TW. The effects of morale are also a bit more gradual IRL than in TW: in TW a unit is generally either in good order or it’s routing. IRL you’d see units that were refusing to advance (or just advancing so slowly that they might as well not be) because they didn’t like the look of what they were advancing into. That could apply both to armies that are losing, with its troops refusing to charge suicidally just because their commanders haven’t given up the fight, and also armies that are winning, with troops who have seen off the immediate threat not really fancying charging after enemies who might still be capable of putting up a fight.

        Add to that the general effects of looting, as otherwise discussed in this thread, and battlefield “fog of war”: commanders didn’t know what was going on at all times everywhere on the field, and their troops certainly didn’t. Troops might be aware of what’s going on in their immediate environment but they won’t necessarily know the position along the rest of the line (especially once gunpowder enters the fray and you have gunsmoke affecting visibility as well as general dust, weather, etc) This makes it a lot easier for retreating troops to get away, because the kind of coordination needed to “shepherd” herding troops in the way you see it in TW and the inclination of the troops and officers required to carry it out – who might not know that they have won the battle until hours after the decisive moment! – would rarely have been present in the way TW depicts.

        Plus, of course, men fleeing for their lives often tend to run faster (especially once they’ve discarded their heavier equipment) than the guys chasing them, unlike in TW where often pursuers seem to move faster).

        1. Depends on period too. Once a line of spearmen breaks, it’s run or get surrounded (and cut down). Battles can be quite decisive, in that the loser suffers major losses. C18 and C19 musket warfare was much less decisive, as small units could retreat, covering others, and each took time and order to overcome. So losses were much less and battles often indecisive. It was fairly usual for an army to rally some distance back, and be ready to fight again in a few days.

        2. One thing that I think plays into this a lot is the “losing after a retreat is a wipe” mechanic. Unless it results in a significantly better strategic position, usually because the attacker can’t follow, retreating from a projected defeat often puts you in a worse position than before, even with no losses having been taken. As a result, you’re incentivized to roll the dice on winning the projected defeat rather than attempting any sort of retreat, especially if it’s some sort of retreat while in contact. To be fair to the other side of the argument, however, a retreat under fire is probably not nearly as meaningful in the pre-Napoleonic settings, but it is functionally impossible regardless.

    3. It been a while but seem to recall the original shogun total war captured the dynamic you suggest. Both in a larger battle field (I really hate how RTW II massively scaled down the battle space) so you could actually also hide units and just send out some light cav to skirmish and often run out the battle clock without accepting a battle. But also that unit broke and ran and if you did not have fast/fresh troops to pursue you did not rack a huge kill total.

    4. I think only Battle for middle earth 1 had it. But I do remember certain units being able to merge with other units and serve as support. Like tower guards with archers, and then the tower guards would use their shield to protect the archers. Although I liked doing that, I dont remember it being too useful. It often seemed more useful to have the shock infantry just go out there and fight. That said I was fairly young when I played it so I didnt dive into the details.

  2. This all bodes poorly for the pike & shot era TW game I’ve been hoping for…

    It’s a real shame that CA has been steadily abandoning the more complicated aspects of the games, and is a large part of why I have been falling off their newer games. Three Kingdoms was a big step in the right direction, at least on the campaign level, though I do find battles a little lacking.

    “There is nothing so frustrating as watching a 100+ unit of muskets fail to execute its fire-by-rank attack because a handful of fellows on the corner were reloading at the wrong time and for some reason this meant that the whole unit had to stop and wait for them.”

    Yes, but it’s all worth it because there’s no feeling in the game better than watching a well executed one-two-three punch of volley fire scythe down an entire enemy unit at close range. Except for blast of timely canister shot. Or exploding enemy 1st rates with rocket ships. It’s a game with its moments is all I’m saying.

    Empire was a really ambitious game and, though it failed to completely deliver on its promises, I think it’s worth them trying again. But I fear we’ll never see another game like it from CA. The Warhammer series has instead been pushing towards much more arcadey action and a weaker focus on tactics.

    1. Empire was fantastic when it worked. Unfortunately it only worked about 30% of the time.

    2. Also worth considering: TW games have showed that their preference is for unit sizes to be no more than about 150 men on the maximum unit size. The half-sized later tercios were about 1,500 men (on paper). So a TW tercio would probably be about 1/10th the size of its real-life counterpart. If all its constituent parts were scaled accordingly, the front garrison, which would normally be 80 men, would instead by only 8 men at maximum unit size. Or just 2 men on minimum unit size.

      Some formations just don’t scale down well. My recollection is that there was a minimum number of men needed to form square in E:TW and, on smaller unit sizes, units very quickly hit that threshold after just a few casualties. Grenadier companies were so small that they weren’t even given that ability by default. I think that was because, on smaller unit sizes, their maximum strength was less than the threshold needed to form square.

      1. Amusingly, Geoffrey Parker quotes a Spanish commander in the 80 Years War. His 80 man company had been reduced over the years as lack of pay, casualties and desertions took their toll. Still, he recounted, “we formed up in the regulation manner: two musketeers in front, five pikes next and a washerwoman behind.”

      2. This suggests to me that it could be reasonable to make one of these composite units as a collection of discrete single-weapon units: 20 units multiplied by 150 men each gets you … well, actually, about one full tercio.

        The sticking point here is that, at least among the recent ones, which are the ones I’ve actually played, the engine feels like it has absolutely terrible support for macro-formations. There’s a nod in that direction with the six different arrangements in Three Kingdoms, though, so maybe one could manage to do a bunch of different pike-and-shot bastion formations in a game that was dedicated to the pike-and-shot era. Assigning one of those formations tends to result in fairly arbitrary unit-placement choices, though, so it would be a real change for the macro-formations to not end up having your units all get re-sorted by whatever order they are on the tooltip at the bottom.

        1. I’ve thought about that. They could give pikemen a “pike square formation” and arquebusiers a “mangas” formation, they have a command to sort of staple them together. They’d have to give arquebusiers a “run into the square” command, which would only be enabled when they’d be “tied” to a pike square unit, I guess. But having seen TW units attempt to interact with each other, I’m not sure it would make for a good experience.

          There’s also the problem that armies in TW games are traditionally limited to 20 units (or 21 in 3K). How many squares do you need in order to feel like you have a real army? Tercios never deployed into a single square, they usually deployed in 2+ lines, each of several squares. If you’re using 10 units to make 5 squares, is that enough? That’s also using up half your army on infantry, leaving only 10 slots for cavalry, artillery, and any other units you want (like arquebusiers deployed in skirmish mode).

  3. A fascinating suggestion. Having played Rome I and Empire during a brief period of obsession with the Total War series, though, CA really has struggled to portray historical realities in a gamified way. To a certain extent, Paradox games manage to dodge this simply by not dealing very much with tactics at all. Still, you’re quite right in saying that CA is one of the few studios that can actually afford to deal with video-game tactics in a truly innovative way.

    P.S. I was quite surprised that Total War was today’s focus though! Were you actually planning to write up something on Expeditions: Rome this week, or was that just something you were doing on the side?

  4. Interesting to read about composite infantry, especially Neo-Assyrian. It reminded me of how Teucer and Ajax combine at Iliad 8.266-272:

    “Teucer came ninth, flexing his curved bow, taking his place behind the shield, of Ajax, Telamon’s son. Ajax would slide his shield aside, and Teucer would spy his chance. Sending an arrow flying through the ranks, his target would fall down dead where he stood. Then Teucer would scurry back, like a child to its mother, taking shelter again behind Ajax’s shining shield.” (translation from

    It’s obvious from the simile that there’s no honour in Teucer’s fighting style (“like a child to its mother”), but clearly Teucer is a much more effective fighter for the cover Ajax offers him. Is this the only example of such battlefield tactics in Homer?

    1. I’m pretty sure Pandarus and Aeneas do something similar, although I don’t have my copy of the Iliad handy to go looking for a passage.

  5. Per the military manuals, the medieval Romans used composite formations for their infantry and their cavalry, with the former in a 4:1 heavy:light proportion (heavy being spear-and-shield, light being bow), with the archers in the center rank of five. They’d typically deploy at double depth (4:2:4), with the explicit advantage of being able to double again (8:4:8) for stronger formations or split to the minimum (2:1:2) if the line needs to extend.

    The heavy cavalry wedge incorporated front ranks with maces, flanks with spears, and the center and rear with bows (again, scalable depending on numbers and circumstances).

    Later innovations for the infantry include the incorporation of super-heavy spearmen to counters enemy heavy cavalry, originally deployed ahead of the rest of the army but pretty quickly pulled back and integrated into the infantry line.

    The closest I’ve come to replicating such formations in Total War games IIRC was in the original Barbarian Invasion, with exceedingly thin lines of Easter Archers sandwiched between thicker lines of Legio Lanciarii. Most TW games have issues with archer units firing while too close to friendly units, resulting in the archers either refusing the shoot outright or causing friendly casualties to the rear ranks of the adjacent unit.

    Dwarf ranged infantry in the TW:Warhammer series are also hybrid troops, always heavily armored and with shields and axes for when the range closes. They’re one of the factions that can pull off missile-only armies with greatest expectation of success.

  6. It’s odd, because it seems like units with a “stance system” where they switch between two different weapons are insanely common in RTSes. Is it just too hard to animate when it’s a block of infantry rather than a single unit?

    1. For one, composite infantry is not “all one type of weapon all at once.” It’s different parts of the same unit having different weapons and all fighting at the same time. Second, the way Total War handles military units is that the entire body of troops is doing the same thing at the same moment. So if you press the “bring out guns” button _the entire unit_ stops what it was doing to bring out their guns — not just the portion currently unengaged with the enemy.

      1. I distinctly remember hoplite phalanxes in Rome1 where a few soldiers were fighting with their swords against enemy units on the flank. So I don’t think they necessarily have to act in unison. They end up doing so, most times, agreed, but it’s not a hard coded thing.

        1. They’ve always had the technical capacity to do that kind of thing, problem is that it tends to fuck up with the logic and animations and such.

        2. They did a major engine change between Medival 2 and Empire. This might have been necerssary for the line fire infantry in Empires, but they do not seem to be able to do things the way, they did back in Rome 1, and Medival 2.

          1. Though unlike what Bret seems to hint at, Rome 1 and Medieval 2 were already 3D…

          2. @Peak Singularity
            Yes, the graphic engine was already different, but the underlying engine that computed units, indivudal soldiers, damage, hits etc, was pretty much still the same.

            At least, that what I remeber some CA employee told, in an interview, back in the day when Empire was anouced.

          3. We’re talking about graphics and collision issues here, that would have changed radically going from 2D in Medieval : TW to 3D in Rome : TW.

      2. I was thinking that rather than switching between “your units are holding guns” and “your units are holding pikes,” you’d have “your pikes are in the front” and “your guns are in the front” or other formation changes as the “stances.” Which would be functionally the same – the guns aren’t going to be shooting when they’re tucked inside a square – but would be more realistic than having all your units by “hybrids.”

        1. That’s how the hybrid units in 3K works: Toggle one button and they’re all mixed in, toggle the stance on and the front ranks kneel behind shields with polearms braced and the back ranks take up firing positions.

          1. Do they have distinct polearm and archer entities or if your front ranks die do the troops behind them pull out polearms?

    2. I think the bigger problem is that the AI has to play the unit reasonably well. Or at least, not totally moronically. The more clever the unit is, the more glaring the AI’s faults.

      No wonder they decided to just add dragons, everybody loves those and they don’t even need pathfinding.

  7. I would love to see a Total War: Empires 2. You could even have the Wars of Latin American Independence as a late-game event to shake-up too-stagnant empires!

  8. I may be reading this wrong, but did the Romans in the Imperial phase have combined units like the Han did? It was mentioned near the end that combined units are the mark of elite armies, but as far as I can tell from this post it seems the Romans just let their skirmishers leave after a point and focused on the heavy infantry clash.

    1. In Field Of Glory 2, granted a phase based game, many of the later Eastern armies have stuff like 50% archers or w/e. Many Roman units have 50% swordsman modifiers or other such modifiers. Lots of units have light spear/swordsman and so forth. And these are separate from heavy/medium infantry modifiers. Also Romans have Impact Infantry because of Pilums vs Deep Pikes for Greeks.

      There’s some evidence that many times soldiers in ancient armies could all use slings properly although they didn’t always carry them. Macedon I think had cross training that way. And a few important battles where mass slings were used.

  9. Minor clarification on the invention of countermarch fire. You’re right to say that William Louis of Nassau-Dillenburg was key to popularising such tactics in early modern Europe, but it’s worth noting that he actually claimed to be following precedents from antiquity; specifically, the theories of Aelian and Byzantine Emperor Leo VI.

    Now, I’m of the school that thinks classical influences on the Orange-Nassau military reforms have been exaggerated by previous scholarship (particularly that which overestimates the practical impact of neo-Stoicism and the works of Lipsius, who was so keen on Roman examples that at one point he suggested abandoning firearms altogether and replacing them with javelins), but in this case it does seem pretty clear-cut; William Louis explicitly cited his ancient inspirations in the letter of which you’ve included a picture. So, in order to fully substantiate the claim that “in Europe [countermarch fire] really only catches on with muskets”, it could be useful to expand on Roman/Byzantine practice. I’m an early modernist, so I’d be interested to hear from someone with more early-period expertise about the extent to which the ideas of Aelian and Leo actually left the pages of their manuals and made it onto the battlefield.

    1. Modern Spanish historians have also argued that by the time of Maurice’s reform the tercios had already being using the countermarch for a time.

  10. Not sure why my first comment didn’t post, but yes I’m working on a game that I hope will put ideas such as fire by rank, fire by file, caracoles, and pike+shot formations into actual gameplay mechanics. Your blog posts were a major inspiration for these mechanics

  11. Interestingly enough, Medieval 2 game very very close to having composite infantry in practice, even if not an advertised game mechanic.

    On release, pikemen in the game were universally decried as too weak and ineffectual because units tended to be a little too loose and as such they would too easily lower their pikes in favour of their swords… until players discovered that double-layering your pike units would actually allow one unit of pikes to support the other!

    This quickly led to exploration of other combinations such as layering a unit of pikes (on top of, but slightly behind) a unit of crossbows/archers/gunpowder infantry, which would allow the gunpowder infantry to fire while still being within the hitbox of the pikes, protecting them from cavalry. Swordsmen could also be mixed in instead to make a horribly effective melee formation.

    This was especially effective with the first gunpowder unit unlocked, which had half-plate cuirass and decent melee attack with their swords, allowing them to both provide some high morale impact volleys before activating melee blender mode.

    While it’s not the same as true composite infantry, it was interesting that a unit could be so ineffectual until used in close combination with others. Doubly interesting because it doesn’t seem to have been deliberate, pikes never functioned that way in previous or future games.

    1. Strictly speaking, the issue was with the triggers for switching between primary (pike) and secondary (sword) weapons. While unit spacing didn’t help, it was not the main problem.

      You can actually easily experiment with this yourself, as unit spacing is a single number in the unit_export_descr-file (it was called something to that effect, it’s been a long time since I modded MII). You need to unpack the gamefiles to access the file, but it can be easily edited with notepad.

      You’ll notice that while changing unit spacing has a slight effect, it is negligible compared to simply baby-sitting the unit and repeatedly issuing attack-commands whenever they drop their pikes.

      This means the pike is very awkward to balance through modding. The only way to make them reliably continue using their pikes without baby-sitting is to delete the secondary weapon, which makes them overwhelmingly overpowered due to how fast the pike attack animation is (and without a secondary weapon to switch to, they’ll continue using the same animation at point blank range, leading to absurd animations where the entire pike phases through the enemy, while the pikeman fists his opponent to death).

      Some mods try to deal with this by giving pike units very, very poor stats, others by simply accepting that pike units sit in this awkward space where they are incredibly powerful when given a lot of attention and used in very particular ways, but terribly inefficient otherwise.

    2. I do note that it’s not entirely absent from newer games either: In warhammer it’s often a good idea to put a chaff unit “on top” of a unit of monstrous infantry or elite squishy infantry (like Depth Guard) to absorb hits, while the units will be able to hit back

  12. “The result was self-shielding shock-capable foot archer formations. Total War: Warhammer also flirts with this idea with foot archers who have their own shields, but often simply adopts the nonsense solution of having those archers carry their shields on their backs and still gain the benefit of their protection when firing, which is not how shields work (somewhat better are the handful of units that use their shields as a firing rest for crossbows, akin to a medieval pavisse).”

    They don’t have this in Warhammer, but they *do* have this “properly” in 3 Kingdoms (which uses a later iteration of the engine) where some units can adopt a “mixed infantry” formation, usually with polearms and shields in the front and archers behind them.

    1. To clarify, in the base game this is restricted to a single unit (the Azure Dragons) but a few more are added in the DLC (I seem to remember bo Cao Cao and Yuan Shao getting variants in their DLC)

    2. Yeah, Three Kingdoms has composite troops available in the endgame, Azure Dragons and… Jade Dragons? The game usually ends before you get them, but they exist.

  13. I never played Empire; I had heard it was buggy. I jumped from Mediaeval II to Napoleon (for which I had to get a new rig).

    On the matter of running down routers, the AI doesn’t do that. As player, if you defeat the enemy you have a choice of running down the routers, or accepting victory; if you don’t want them to form a new, more experienced unit, it seems wise to run them down. In real battles, your chance of dying increases dramatically once you turn your back on the enemy – whether you are an individual or a unit, so the “killability” of routers seems realistic.

    In the Napoleon game, the best units for running down routers are lancers and similar light cavalry units. I believe that was basically the purpose of lancers; and if they trained units specialised in running down routers, then running them down must have been at least not rare.

  14. I look forward to the article on command and how it doesn’t work like in TW, but in fairness to CA there’s also a series possible where you show how awesome the TW series is as a learning tool for some of the basics of battlefield warfare by the things it has consistently did better than its predecessors, in terms of the centrality of formation cohesion, morale, and need for combined arms. It may be a very incomplete simulation with dodgy AI, but It works well enough that as a player of TW Medieval II it is obvious that a heavy spear/pike formation that could blast its enemies at medium to close range would be well nigh unbeatable, which suggests that the basic battle model does have much to commend it.

    1. The “problem of command” is really the central point: There has been indie/gimmick attempts to “properly” model the fog of war (Radio Commander, I believe one attempt was called?) but it makes for a very different game experience, some weird hybrid games like Sacrifice (where you have a “commander” unit that you basically control in 3rd person and can only really command units around him) does something similar almost accidentally but doesen’t allow for the entire “sending and recieving messages” thing.

      There’s a whole bunch of other things, too, but the omniscient (or nearly so) commander capable of instantly reacting and giving orders is probably *the* most unrealistic aspect of Total War games.

      1. I’m not sure whether composite units would work from a gameplay perspective; you’d tend to disproportionately lose your front rank and the numbers wouldn’t match to maintain the formation, with likely limited opportunity to corral it back into a workable formation manually. It could attempt to autosort into a new arrangement but I’d expect that to fall apart into a messy blob of entities getting stuck on each other unless the new engine massively improves individual model behavior. From a gameplay perspective it’d probably be better to represent composite infantry as two separate regiments (maybe getting fancy and having them count as one regiment on the campaign map) and be able to link them so the archers fall back through the melee regiment rather than hither and yon.

        I’m curious about exactly why hybrid infantry weren’t generally a thing, and how various fantasy factors might impact that. Obviously you can’t draw a bow while holding a big shield, but was it impractical to carry both or use a bow in armor? Was it training, especially in non-pike eras? Was it just not affordable to get everyone a full set and better to have a hundred each rather than 150 hybrids? I’m particularly thinking about how most of the non-Kislv non-terrible* WH hybrids are elves and dwarves, who are long-lived, wealthy artisans with declining populations.

        *Yes, Free Company Militia, I’m referring to you

        1. One thing I do note in Warhammer is that *in practice* hybrid units tends to be ranged units that can defend themselves in melee. There are a few circumstances where you want to charge in (generally if the enemy is all missile units and it’s more coste effective to get into melee where you beat them rather than trade fire evenly) but it’s rare. And as pointed out *that* wasn’t actually that uncommon.

          Also, the way the Warhammer tabletop game tended to work was that (non-elite) units generally had similar stats across their race (generally with 3 being “average ” for strength, thoughness/weapon skill, etc) so eg. a unit of crossbowmen and a unit of swordsmen had the same “base stats” (all units wer eassumed to have a “hand weapon”, sword, mace, etc. as part of their basic statline) and the reason a unit of swordsmen would generally fare better in melee was all down to formation (generally deeper, each “rank” adding a bonus to combat resolution) and other equipment (armour, or in some cases special weapons that gave extra bonuses under certain conditions) note that unlike in TWWH units could only fire in a limited number of ranks (exactly how many depending on the type of weapon and unit, generally it was 1 rank, 2 if you deployed on a slope, bows getting an extra rank compared to guns and crossbows but being significantly less damaging)

          Lothern Sea Guard were somewhat special in that they were a missile unit that could also bring spears (most missile units being restricted to only using the basic hand weapon/unmodified profile) Dwarf quarrelers could bring two-handed weapons in a similar way. Note that because of the way their respective weapons work you could deploy Lothern Sea Guard in a way more like a standard infantry unit ( a deeper block of 3+ ranks) since because of the spear and bow combo they didn’t suffer as badly from the back ranks being unable to fire.

          1. Honestly what I mostly use the hybrid infantry for is my standard two-line formation, which normally has archers behind a line of melee troops with a few units of cav to catch flankers. When I’m playing elves I’ll often replace my front spears with Sea Guard; they provide shooting on the way in and then engage the enemy while the archers or Sisters of Avalorn arc shots overhead. Likewise with Kislev I won a whole campaign with the Ice Guard doomstack in the same array. Or as dwarves I use the thunderer’s pseudo-hybrid status to have a gunline and then pass warriors up through it while they’re engaged.

            In Medieval 2, and particularly a mod where one of the factions was actually pike-and-shot shading into bayonet era, I remember actually being able to pull off having gunners in front and fall back through the pikes as the enemy closed to melee, but in Warhammer when I try that with the handgunners it just takes too long to get the whole unit behind the front rank of the melee troops, and the same stickiness that makes the checkerboard work means it’s damn near impossible to get a unit to disengage from melee.

        2. I’m not sure how much individual units getting stuck on each other is an issue these days, see for instance this Zero-K/Spring pathfinding video… (from 6 years ago already !) :

          (Also, with Total War’s formation, maybe they could use the trick of the “allied unit” hitsphere being smaller than the “enemy unit” hitsphere ?)

          1. Standard RTSes have much better pathfinding, but Total War’s pathfinding for individual soldiers falls apart under a variety of circumstances and passing one unit through another is a particular mess. I’m guessing having 4000 soldiers limits the sophistication of the algorithm for each one.

            An engine update might fix that, but I’d still be worried about a mixed unit going funny when it takes disproportionate casualties and 50 archers and 50 shieldbearers becomes 40 archers and 20 shieldbearers.

          2. guy, I’m still somewhat hopeful – these days, between faster processors and potential multithreading –
            (which I honestly wouldn’t have thought possible for pathfinding, yet in the following video he seems to be able to reach up to 500% core usage – even though ATI was kind of misleading in their advertising – so a FX-9590 is closer to being a 4 core than an 8 core CPU)
            – that these days we can have thousands of units doing decent pathfinding, with many of them being in close range with each other !


            And the TW series seems to be due for a full engine upgrade – the first game using the current TW Engine 3 (Warscape Engine), Empire: Total War, came out 13 years ago ?

  15. From what I have read Roel Konijnendijk does make the argument that Persian infantry could stand against Greek hoplites, using as evidence for example thiw Herodotus passage: “They fought a long time at Marathon. In the center of the line the foreigners prevailed, where the Persians and Sakai were arrayed. The foreigners prevailed there and broke through in pursuit inland, but on each wing the Athenians and Plataians prevailed.”

  16. One thing that is hard to talk about is that (especially from the middle-ages) we have plenty of lists of equipment, but not neccessarilya good idea of how they actually fought: Eg. lots of scandinavian laws have lists of equipment every free man should keep in case of warfare (obvious caveats about how much that is wishful thinking, and how much uit actually happened, but Visby seems to indicate that it was at least not *entirely* fantasy) and a lot of them includes both a missile weapon (bow or crossbow) and some combination of spear, sword or halberd. (note: basically now weapons at Visby, the tehory is they were stripped by the victors but the heat meant they didn’t have time to strip the corpses of armour before throwing them in the mass grave)

  17. Question about the musket-axe. CA claimed it was a real weapon in Kislev’s roster reveal, but the best I could find on it was a picture on Wikipedia’s page for “Combination weapons” and a Youtube video from “Shadiversity.” Should I be thinking of these as experimental or novelty weapons? Or some sort of weird variant bayonet?

    1. It’s one of those novelty weapons (and even then the CA version is grossly oversized) but there are the occasional weird attempt to combine things like that.

      Those that I have seen have been attached to (admittedly fairly hefty) pistols rather than arqebus though.

  18. In fairness to CA, I think they just ran into the limitations of computer programming. Other commenters mentioned how buggy Empire’s fire by rank/by platoon feature was, when you had different people within the same unit performing the same motions at different times.

    Now imagine having to program a unit that has multiple kinds of infantry in it, each of which will look different, have different attack animations, different stats, and have all of that going on at the same time. I’m not even sure if that’s possible to do.

    1. Eh, as someone who programs for a living (not games, but I’ve dabbled in them before), that’s perfectly doable. It wouldn’t be *easy* (which probably goes a long way towards explaining why it hasn’t been done yet—companies have finite resources, after all), but it’s definitely not impossible. If you can program a single unit in the first place, it’s not a huge leap to running through an array of units, which is what the computer’s already doing under the hood every frame anyway. It’s just a matter of coding up the logic to coordinate groups of units, and while it’d be a lot of tedious work it’s not conceptually difficult.

      1. I suspect it’s more to do with CA’s longstanding inability to get path-finding working adequately than any theoretical difficulty in working out the logic- their very bad path-finding, which is likely a much harder task than actually cycling the animations/unit logic, results in all sorts of issues. I mean, sometimes you need to micromanage units walking over a grassy hill and engaging the enemy just so they don’t sprint head first into a nearby wood due to some finicky intricacy of the system.
        That’s likely what limits their attack abilities- they can’t have any situation where some are finishing an attack or reload animation as others are walking, because it would likely cause the path-finding system to collapse entirely. You’d give people an order, they’d start out, but half of them wouldn’t move, and then the entire unit would likely end up suspended out of formation forever, because getting them to re-form beyond an obstacle is something the system both already struggles with, and is a gameplay issue because even when actually working correctly, it takes far too long.

        1. That’s certainly plausible; colloquially, I hear path-finding is hard in terms computational power needed, and it gets much harder the more units you have trying to recalculate their paths every tick. With the number of units in those screenshots, that actually might start running into the limitations of what can be accomplished on a consumer CPU without slowing the game down to sub-15 FPS speeds.

          1. There’s also a foresight problem to the pathfinding question: Units are following each other, so you might be able to have follow-on models step into the space they vacated, but only if the people in front are somehow actually out of the way.

            To pick on a widely-played example from long ago, Starcraft units rather dramatically fell out of formation within seconds of moving, because they were constantly recalculating to walk around their compatriots, who were also moving. This was sort of bad for Marines, who had a small collision box and would usually have one or two stragglers wandering off, but it was spectacularly bad for Dragoons, which were so big that they needed to go up ramps single-file, resulting in the last ones getting frustrated and wandering off in all kinds of unexpected directions.

            I’m sure it only gets worse if you’re worrying about animations, too.

  19. One total war: shogun/military history question: why no shields in Japanese kit?shields seem pretty good vs archery. Total war would have you believe there were mass infantry archers, and you would then expect there to be shields used—or were there actually quite a bit less archery in reality, so shields were of less utility?

    1. There were quite a lot of archers in reality, and in fact Samurai were hybrid cavalry most of the time, with bow as the main weapon, naginata as a secondary weapon, and katana for emergencies. Japan did actually have shields, but the most commonly depicted warriors are samurai, who can’t use shields while shooting and have high-quality armor in any case.

    2. You do get shields in Japan, but they are typically large shields that are carried into battle and then put on the ground to make up a sort of makeshift wall. Smaller man-portable shields such as for example the round or kite-shaped shields that were typical in Europe were rarer.

      The most popular weapons in Japan – the yumi (longbow), yari (long spear) and naginata (polearm) all require two hands to wield. Notably, these are also the main weapons of the warrior elite, the samurai, as guy notes. The mounted elite of Medieval Europe, on the other hand, tended to rely on the lance and developed techniques that allowed the lance to be used one-handed together with a shield (contrast with e.g. Macedonian companion cavalry that typically also charged with lances, but did not use shields).

      Infantry armed with pikes and other polearms typically did not use shields in Europe, just like their Japanese counterparts. In later periods in Medieval Europe you also see the deployment of increasingly large pavise shields that crossbowmen used for cover, especially in sieges, somewhat similar to how shields were typically deployed in Japan.

  20. I also do note a competing influence here: Warhammer is well… A *warhammer game* and Warhammer has tended to go with single-unit types. (though occasionally with characters, special units, musicians, or other stuff sprinkled in, depending on edition) CA is at least on some level tied to the IP here (though they clearly have some leeway to do their own thing, GW has final say)

    1. But the Warhammer miniature game always had composite units. Just to name some examples I have in the back of my head: Ogre Maneaters, Skaven Giant Rats (with drivers), Goblin Fanatics, Empiral Detachment (sleeves of shooters or heavy hitting units with beidhanders, or polearms)

  21. Whilst the history lesson is interesting, bashing a game repeatedly for being a game and not a classroom, simulator rather frustrates me when I see other people do it.
    Of course, the moment I see a depiction of a succubus with horns in a roleplaying game product such as PATHFINDER, it’s entirely different, and I want to know why the game hasn’t statted up the succubus as having access to a ‘headbutt’ attack of some kind…

    1. Given that we’re talking about a seduction-focused shapeshifter, I’m guessing the “horns” are cosmetic. Magical cosplay, effectively, optimized for looks rather than structural solidity. That said, Pathfinder’s succubus stats do include barehanded attacks comparable to short swords. Only major potential complication I can think of from reskinning “claws x2” into “horned headbutt x1 and tail lash x1” is whether they remain usable while the succubus’s hands are otherwise occupied – for example, by wielding manufactured weapons.

    1. If they did, heavy infantry would carry pistols as back up weapons only.

      Heavy cavalry in Europe started using pistols instead of lances in the Renaissance. Cavalry can fire pistols reasonably accurately on the move because it’s the horse doing the running, not the rider. Muskets weren’t used because it would be almost impossible to reload on horseback and you can’t carry a bunch of loaded muskets in holsters.

      Heavy infantry have the opposite problem. Trying to shoot a pistol while you’re running is a lot harder. Infantry charging with early pistols would either have to stop to shoot, losing the morale and impact value of being on the move, or mostly miss and get skewered by the enemy. So why use a pistol when you can carry and reload a musket which will shoot much further and more accurately? Stand still and shoot your enemy at a distance.

    2. As seifihughf says, pistols would be strictly sidearms, and realistically used principally at very close range anyway. The effective accurate range of a musket was generally around 100m, which when you consider how fast a running person moves, isn’t all that far. Even the elite troops of the 19th century, prior to breech loading, would get at most two shots off against a person running at them full tilt. (Of course, they would start shooting before the enemy entered that range).

      The effective range of an early pistol would be even shorter, such that by the time you’ve fired a shot, discarded the pistol and grabbed your melee weapon, you’re probably already under direct attack and in trouble. Shooting at very close range, like the guys in rear ranks shooting over the shoulders of their comrades, might have some value, but only questionably more than just using a pike (or a musket).

      So pistols were probably only questionably useful for infantry in formation combat, and while some experiments can’t be ruled out, infantry being equipped with pistols as standard (as opposed to muskets) wasn’t something that I’m aware of in the regular troops of any European army (including the Ottomans). In a disorganised brawl, however, they become more useful, because they give greater effective range than a sword without being as unwieldy as a pike or spear. This might be why they became and remained more popular in naval combat, where formation fighting was much more difficult.

      Also, of course, heavy armour tended to be expensive to produce. While the best armour could stop gunshot, outfitting whole regiments in bulletproof armour would have been ruinously expensive even if it were tactically desirable. The gunpowder era sees a decrease in the amount of armour worn by the average soldier: during the Thirty Years War your pikemen might wear a helmet and breastplate, but your musketeers wouldn’t even have that. By the Nine Years War, the majority of infantrymen were unarmoured musketeers. Aside from the period in the late Middle Ages/early modern period when dismounted men-at-arms were in vogue (before gunpowder was a major feature on European battlefields), the people who could afford the best armour tended to fight from horseback.

      Cavalry, of course, are a different matter: cavalry using pistols as their *primary* weapon were a common feature of battlefields during the 16thm 17th and 18th century (and possibly the 19th too, though I know less about post-Napoleonic armies). Earlier in the period they tended to stand off and attempt their equivalent of the infantry fire-and-countermarch (the caracole, a manoeuvre of dubious effectiveness during this period), but the tactic supposedly introduced by Sweden in the TYW was to charge in and fire pistols at effectively point-blank range.

      Even among cavalrymen, though, armour was in decline during this period. Famously, Arthur Haselrig (one of the “five birds” MPs) raised a unit of cuirassiers at his own expense during the first English Civil War, armoured head to toe. They were nearly invulnerable, but also largely ineffective, being regarded by almost everyone as anachronistic and obsolete, and were notably routed by the Royalists’ more modern cavalry at Roundway Down. Nobody else repeated the experiment during that war, and while cuirassiers in some form remained in service until the 19th century they were never *so* heavily armoured again (that I’m aware of).

      1. I’d argue that the Haselrig’s Lobsters where quite effective at with standing charges and there head to knee armour (greaves and sabatons had been largey done away by then.) gave them considerable staying power as a unit, there rear guard action at Ripple Field saved a lot of the army.

        1. The Lobsters did have a couple of successes. Ripple Field is debatable: their invulnerability was an asset in covering the Parliamentary retreat but at that point the battle was already lost and the Lobsters suffered terrible casualties. Even at Lansdown, where they had a robust defensive position, they apparently distinguished themselves but were unable to resist a reckless assault from much more lightly-armoured troops largely unsupported by cavalry of their own.

          Overall I think the assessment that they were obsolete and anachronistic remains fair, simply because nobody witnessing them in action felt the need to get some cuirassiers of their own. Even their major success, at Cheriton, was an attack on unsupported infantry that could probably have been pulled off by regular harquebusiers. Indeed the “Ironside” cavalry was much more effective than the lobsters had been, with rather lesser armour.

  22. Shame you’ve missed mentioning Grand Tactician Civil War seeing as it does model firing quite well while also including elements like order delay — with runners modeled to boot — formations with individual command initiative — you give a division a move order and an aggression level and it trots off to carry out the task as best it can — and a great deal of effort spent on mechanics that make the player a good deal better at understanding why commanders did what they did and how hard it can be to enforce your will on a battlefield with multiple corps engaged.

  23. Some cursory research suggests that shields were commonly used in Japanese warfare for all the obvious reasons and their absence in western media is a pop history anachronism. Most of the relevant video games are set in the Sengoku period and portray the infantry kits as something roughly analogous to pike and shot, and then that gets generalized to the entirety of pre-gunpowder Japan.

    It’d be kinda like looking at a picture of a Tercio Square and maybe some cavalry in full plate and using that to form the conclusion that shields were absent from European history.

  24. Mr. Devereaux,

    I have read your blog for some time and quite enjoyed it, and I think this is a good piece as well.

    However, I do have a criticism of this particular line:

    “The Roman heavy infantry carried javelins (the pilum) and later darts (the plumbata), but these were hurled in preparation for a charge and fairly clearly secondary weapons to the legionnaire’s gladius or later spatha. ”

    This characterization of the nature of Roman legionaries in combat has been repeated for decades, maybe even centuries, so it’s really no surprise that you would phrase it that way. Unfortunately, it’s wrong.

    More recent scholarship on the “face of battle” among the Romans has instead placed a central emphasis on the use of missiles (Pilum and other tela) as the primary weapon, with the gladius as the back-up, rather than the other way around as was often thought.

    I would strongly recommend the following papers:

    “The Face of Roman Battle”, by Philip Sabin (2000)

    “Roman Republican Heavy Infantrymen in Battle (IV-II Centuries B.C.)” by Alexander Zhmodikov (2000)

    “Not so different: individual fighting techniques and small unit tactics of Roman and Iberian armies within the framework of warfare in the Hellenistic Age”, by Fernando Quesada Sanz (2006)

    “Pilum and Telum: The Roman Infantryman’s Style of Combat in the Middle Republic”, by Jordan Slavik (2018).

    The combination of pilum and gladius for the Roman legions functioned something like the musket and bayonet for later soldiers of the 18th and 19th centuries. A gladius charge, like a bayonet charge, is decisive in that it puts opponents to rout, but most of a legionary’s time combat was most likely spent on throwing missiles instead.

    A key primary source example is Julius Caesar’s account of the Battle of Ilerda, from the Civil War. At Ilerda, Pompeian and Caesarian legionaries fought for five hours before the Caesarian legionaries drew their swords, as per Caesar’s own account.

    1. So…as someone with a PhD in Roman history who specializes in the Roman army of the Middle Republic and wrote my dissertation on the military equipment of the third and second centuries BCE…I have my own views on these points and do not entirely agree with the current shift towards emphasizing skirmish, particular for the Middle Republic (as opposed to the Late Republic) when the Roman army had dedicated skirmishing troops.

      And, you must pardon me…but it’s Dr. Devereaux.

      1. Hello Dr Devereaux,

        I’ve heard it suggested that the pilum had a variety of forms throughout the Roman period and that this suggests pilums had a variety of uses. Is it possible that there were two types of pilum and one was a spear and the other a javelin?

        I only suggest this because of how historical peoples often didn’t classify weapons to the same degree we do i.e. the invention of the term broadsword/arming sword to describe a cruciform sword typically used by knights. Back then it would just be a sword and a shield a shield since there wouldn’t be a need to specify. Similar to how lance was sometimes used to describe an infantry spear rather than a cavalry well lance.

        Sorry if I sound presumptive, i’m not a historian by any count and certainly have less knowledge on this matter than yourself. Hope you have a nice week and are able to keep on top of everything.

      2. Dr. Devereaux,

        Well fair enough, every scholar has their own views and interpretations.

        I do think that the skirmishing-focused models has so much utility, matches the observations of crowd and combat psychology available to us in modern times, and answers so many questions about the functions of the Roman triplex acies and the many peculiar aspects of the Roman way of warfare that it seems to me to be the most plausible theory at this time.

        That being said, one can go overboard with emphasizing the use of the javelin. “Musket and bayonet” is a good analogy, because the bayonet charge was often the decisive moment of a 18th or 19th century battle. The bayonet charge is the moment when one side or the other will rout. Similarly, I think the gladius charge was the moment when the legionaries either broke their opponents, or themselves fell back. The gladius is still the decisive weapon in the legionary’s toolkit, even if it not the one they use most often in combat.

        I don’t think that they could have spent very much time physically exchanging blows with sword and shield. I think a lot more time was spent on throwing javelins, and at a stand-off range getting themselves amped up for a charge. I may not be a professional academic myself, but I am a fencer, and I know well how physically and mentally exhausting it is to exchange blows in close range, even in play with blunts. How much more so with sharp steel, and actual death on the line?

        Of course, Livy and other primary sources tells us about battles which were won “at the first shout” or “at the first shock”, the enemy breaking and running before any blows could even be exchanged. In that sort of situation, the older classical model of “Everyone throws their pila and then immediately charges” makes a lot of sense. If the enemy is already routing and running away by the time your sword is drawn, then you don’t need to be spending much time in actual close combat and most of your swordwork will be cutting down people who aren’t fighting.

        The sticky bit about that classical model occurs when the opponent actually stands and fights, and indeed the Romans had to win many pitched battles against formidable peer opposition. They often had to actually fight through opposing infantry who stood up to them, including in frontal combat with the Macedonian phalanxes. How exactly they are supposed to fight against determined spearmen or pikemen has always been a mystery to me, and why they would choose to prefer swords to spears is one of the many very peculiar things about Roman military practice. The skirmishing model I think is a necessary update to the academic understanding of the Republican Roman armies to make sense of these mysteries.

        1. Livy isn’t really a primary source (unless the subject of research is historiographical practice): he’s just a historian like any other, who has the virtue of closer chronological proximity. I’m not aware that he had any military experience at all and even if he did it wouldn’t be of the “triplex acies” system, nor can he realistically even have talked to many people who fought under the triplex acies system, given the time periods. Most of his history is based on earlier histories that are now lost, and he is also known for “printing the legend” on occasion.

          He is a useful source, but there is a limit to how closely he should be relied on, especially when it comes to military affairs.

          1. Even if authors like Livy aren’t technically primary sources, they’re universally referred to as such in the modern historiography.

          2. There is a big difference, though, between heavy infantry manoeuvring and countermarching to rotate and refresh their front lines, and skirmishing.

        2. I think that Bret’s comment that the javelins were “clearly secondary weapons” is a bit too strong, but I also doubt that the javelins were used for skirmishing by Roman infantry. Following based on writings by wargame designers and historians I’ve read – I will try to read the sources for the “new view” provided but haven’t the time and access yet.

          As to how the Romans fought through pikes and spears, that’s why the javelins are important. A volley of heavy armour or shield piercing javelins just before the two formations come into contact will kill some, break up the dense line of points as others are slowed or have to step around fallen comrades, even just make some troops stop and raise shields. The Roman infantry draw swords (still on the run, a short gladius on the right side of the body is especially handy for this) and can break into gaps in the pike/spear formation, or have momentum for bashing into enemies with their shields. It would be very similar to the use of pistols by heavy cavalry being discussed in another thread, where pistol and sword cavalry would beat lancer cavalry by shooting just before contact.

          Now while I’m open to the idea that Roman infantry spent more time throwing javelins at the enemy than we conventionally think, the idea that they were *skirmishers* seems ridiculous when considering the armour and especially shields they were carrying.

          Skirmishers with bows or slings can shoot enemy infantry from a distance. Skirmishers on horses with javelins or bows can approach much closer to shoot because they can retreat much faster than infantry can pursue.

          Javelin throwers on foot are in much more danger. (Unless you are central / southern American or Australian with an atlatl or woomera for extra range, but that’s not really applicable to Eurasia.) You have to get much closer to throw effectively, and you are running towards them as you throw.

          So if the enemy infantry don’t like this and decide to charge you, javelin skirmishers have to turn and run, fast. And I cannot think of a worse piece of military equipment for this scenario than a big heavy shield with a single central hand grip. The armour isn’t as bad as some people think for running, but all that weight on the end of one arm? It would be like trying to jog with a suitcase.

          We know what ancient Greek javelin skirmishers and medieval Breton and Catalan javelin skirmishers looked like: little or no armour and small shields. Roman legionaries? Nope IMHO.

          1. Another reason to doubt that the hastati and princepes acted as skirmishers is that the pre-Marian legions actually had dedicated skirmishers, the velites, who indeed typically* had a small shield and no armor.

            *They bought their own gear so I expect there was some variance.

          2. Scifihughf,

            In Book 2 of Polybius’s Histories, he discusses a battle in which the Roman consul made a tactical error by deploying his legions with their backs to a river, and thus leaving them no space for gradually falling back, which Polybius identifies as particular to the Roman way of war:

            “The Consul Flaminius being thought to have mismanaged the battle by deploying his force at the very edge of the river-bank and thus rendering impossible a tactical movement peculiar to the Romans, as he left the lines no room to fall back gradually. ”
            (Histories, Book 2, Chap 33)

            Later, in Caesar’s Commentaries on the Civil War, he discusses fighting against Pompeian legionaries in Spain who had adopted a fast moving, skirmishing-heavy style probably inspired by fighting Iberian peoples. He contrasted this with his own troops, who stuck by their standards and kept their ranks more strictly:

            “The manner of fighting of those soldiers was to run forward with great impetuosity and boldly take a post, and not to keep their ranks strictly, but to fight in small scattered parties: if hard pressed they thought it no disgrace to retire and give up the post, being accustomed to this manner of fighting among the Lusitanians and other barbarous nations; for it commonly happens that soldiers are strongly influenced by the customs of those countries in which they have spent much time. This method, however, alarmed our men, who were not used to such a description of warfare. For they imagined that they were about to be surrounded on their exposed flank by the single men who ran forward from their ranks; and they thought it their duty to keep their ranks, and not to quit their colors, nor, without good reason to give up the post which they had taken. ”
            (Civil War, Book 1, Chap 44)

            However in Caesar’s description, even his Caesarian troops who adhere to their formations more strictly are implicitly capable of giving up a position which they had taken, if there is good reason to do so, although they feel it is their duty to hold their positions.

            One of the characteristic functions of the Roman legion, and it’s triple line battle formation, is the ability to move fresh troops up to the forward edge of combat and pull worn out troops back. This requires the Roman battle-line to be capable of both forwards and backwards movement in combat. Adam Anders in his 2015 paper “The Face of Roman Skirmishing” argues that back-and-forth movement was a key component of a Roman soldier’s combat techniques.

            The point of having a heavy shield and armour, I would argue, is to make the Roman legionary more confident in getting close to the enemy battle line, through the hail of missiles the enemy would doubtlessly be throwing at them, and making close, powerful throws with his pilum, or of pressing home a gladius charge which could actually drive the opponents off and rout them. If you don’t like calling them skirmishers, then call them line infantry. As said above, I think musket and bayonet is a pretty good analogy for understanding how the tactical system of the pilum and gladius worked together.

          3. Replying to Eric, since there’s a limit on nesting.

            I am now wondering if the differences between the old and new views are perhaps being exaggerated? I’m an older tabletop wargamer and that’s where I get most of my knowledge, not from video games or Hollywood. There’s not much in your comments that wasn’t generally accepted in the 1980s.

            Romans fought in shallower but wider formations than hoplites or pikes? Yes. The Romans were able to refresh their front line troops and this was a key to their success? Yes, although last C nobody was really clear on how. Romans can give up their position and are more flexible than a phalanx? Yes. Romans couldn’t spend much time in hand to combat due to fatigue? Yes, because nobody can.

            “Skirmishers” is a confusing word because to me this is not a just a particular role or tactic but also dictates formation and equipment. Unlike modern or even 19th C infantry, a Greek peltast or Roman velite is carrying very different gear when they skirmish. So yes, I’m much more comfortable with “line infantry” to describe Romans.

            The major difference seems to be that Ye Olde School like me think of Romans as always charging and throwing javelins as part of the charge. The newer view, oversimplified by me, seems to be that Romans are more likely to throw javelins, then decide whether to charge or not.

            I do intend to read the sources you provided, and thanks.

          4. I’m dubious they would build their strategy around reusing javelins the enemy threw at them, because the enemy could just not throw javelins at them if they think they’ve got a better relative edge in non-javelin components of warfare. I’m sure they did when they had the opportunity, but as noted javelins often broke on impact, sometimes intentionally.

            I am also really doubtful hand-throwing rocks was a major component of their strategy. From what I understand it’s not terribly effective against riot police presenting a shieldwall so I doubt it’s much good against heavily armored classical infantry. Slings are quite another matter but it doesn’t seem like legionaires typically used them in field battles.

    2. “A key primary source example is Julius Caesar’s account of the Battle of Ilerda, from the Civil War. At Ilerda, Pompeian and Caesarian legionaries fought for five hours before the Caesarian legionaries drew their swords, as per Caesar’s own account.”

      If those five hours were spent skirmishing, that suggests that each legionary took about five minutes to throw two javelins and then cycle the next guy in the formation to the frontlines to throw his. Or else that entire legions took turns skirmishing with each other while other entire legions just hung out and drank lemonade. It’s easy to imagine units being engaged for five hours in this kind of low-key combat, trying to get the upper hand before committing to a decisive melee, but it’s impossible for an individual legionary or even an individual century to spend even a fraction of that time actually stuck in. They didn’t have enough ammo.

      1. Chamomile,

        I do think that the majority of time “in combat” was in fact spent out of reach of the opponent’s weapons, a sort of mutual stand off as both sides got themselves amped up to actually get ‘stuck in’.

        Secondly, presuming that the legionaries couldn’t spend five hours skirmishing with only two pilum per man assumes that all the legionaries would throw their pilum in quick succession, one after another, and then be left without any ammo at all.

        In fact, a battlefield was littered with potential missiles to be thrown, not only from throwing javelins back at the opponent but also throwing stones or other objects plucked from the ground.

        I also think that for a battle-line of Roman legionaries drawn up with some ranks of depth, only the front-fighters are really able to make accurate and powerful aimed throws of their pilum. I don’t think the “mass volley” model of pilum throwing holds up with the accounted times spent in battle using missiles in primary sources like Caesar, so it seems likely to me that a Roman maniple or cohort instead gradually discharged its pila over time, with people moving to the front, throwing pila, and then filtering back to the standards.

        If you combine gradual use of pila, most of combat being a stand-off out of range (Imagine groups of rioters and riot police having a stand off, but not actually exchanging blows yet), and re-use of missiles or use of improvised missiles like rocks, then the legions of Ilerda spending five hours in skirmshing starts to make a lot more sense in spite of the two pila per man attributed to the legionaries.

        1. So to back up your claim that the primary weapon of the legionary was the javelin, you posit that they spent all their time either 1) not engaging the enemy at all and thus using no weapons or 2) throwing rocks and other debris, which are not javelins, and also aren’t actually super common. Seriously, go outside, walk to a random spot in a field or a forest or anywhere, and look around to see how many usable missiles are within a few feet of your position. The answer is usually “none” unless you are standing in or near a stream bed where the water has worn the soil away from the rocks. Video games depict loose rocks as an infinitely available resource, but in reality finding rocks big enough to be used as a throwing weapon is going to require breaking formation to go scrounging for them.


          “I also think that for a battle-line of Roman legionaries drawn up with some ranks of depth, only the front-fighters are really able to make accurate and powerful aimed throws of their pilum.”

          What did you think I meant when I said “cycle the next guy in?”

          1. Chamomile,

            Livy actually tells us that Roman soldiers swore an “that they would not quit the ranks save to fetch or pick up a weapon, to strike an enemy, or to save a comrade” (Livy Book 22, Chap 38).

            To fetch or pick up a weapon is listed alongside striking enemies and saving comrades as a valid reason to depart the ranks. This testifies I think to the importance of retrieving spent missiles from the battlefield to be able to use again. Some of those might be rocks, when suitable, but a lot would be throwing the enemy’s javelins back at them. Caesar in fact tells us that the Gauls at the Battle of the Sabis threw Roman javelins back at them, and Livy describes missiles being used all throughout even very long battles, so I think the re-use of javelins was pretty common.

            And yeah, I do posit that only a minority of time in battle was actually spent actively engaged with the enemy. A minority of that minority might be hand to hand combat. There are physical and psychological limits to the human soldier in combat, and the old Hollywood image of people stabbing each other for hours on end just isn’t sustainable.

            If you watch footage of rioters and riot police facing off today, there are brief periods of fighting interspersed among lulls and stand offs. I think the battlefield also had its lulls and stand offs. This goes a long way to explaining how battles can last for hours when we know human beings don’t have the physical endurance for hours of active combat.

            Of course, a stand off isn’t necessarily neutral either. That time can be spent maneuvering and jostling for position or advantageous ground, and a lot of the grand tactical maneuvers of Antiquity make a lot more sense when you posit that the battle lines were in fact separated at a stand off for much of their actual battle time.

          2. “Livy actually tells us that Roman soldiers swore an ‘that they would not quit the ranks save to fetch or pick up a weapon, to strike an enemy, or to save a comrade'”

            Do you think that Roman legionaries were constantly breaking ranks in order to drag wounded comrades to safety or otherwise save them from the enemy? That the formation immediately disintegrated in melee as every legionary broke ranks to strike their enemies? If not, why would you assume that retrieving a weapon must have happened continuously, when both of the other two things mentioned are valid but *rare* reasons to leave a formation?

            “There are physical and psychological limits to the human soldier in combat, and the old Hollywood image of people stabbing each other for hours on end just isn’t sustainable.”

            Please, elaborate upon which Hollywood movie’s five hour long battle scene established this myth.

            Nobody ever claimed legionaries remained in melee for hours at a time (such a claim runs into the exact same math problem as the claim that legionaries spent most of their time hurling javelins – are we to believe that two men can exchange blows for 5-10 minutes on average before either of them gets a decisive blow on the other?).

            Your claim is that missile weapons were the primary weapon, which means any amount of time spent at a stand-off is not evidence in favor of your claim. Modern footage of rioters and riot police facing off – to whatever extent those provide valid insight into ancient battles at all – are evidence against your claim, because they show opposed forces staring at each other for a long time and not exchanging large amounts of missile fire.

            And the idea that Romans spent a lot of time retrieving javelins hurled at them falls down because it requires the enemy to have javelins to hurl at you. This only happens when 1) Romans fight other Romans, which didn’t become commonplace until long after the javelin was a staple of the arsenal, or 2) Rome’s enemies are throwing their own legionary’s javelins back at them in sufficient numbers to re-equip the expended legionaries. Since legionaries only have two javelins in the first place, this means the enemies must be throwing Roman javelins back at Romans in nearly as great a number as the Romans were throwing them in the first place.

            And since the Romans are clearly not inflicting severe casualties with these skirmishing attacks (if even one in every ten javelins thrown actually wounded or killed an enemy, the Romans would inflict 20% casualties on a same-numbers enemy formation before the melee began even if they didn’t retrieve and reuse a single javelin, which you’re claiming they did continuously), that means that the typical result of the Romans throwing their “primary weapon” at an enemy is to equip the enemy with that weapon.

          3. It should be noted that at least some versions of pila actually bent on impact, making them impossible to throw back.

          4. Chamomile,

            Javelins were extremely common weapons in Antiquity! Almost all armies used them! Certainly most of Rome’s enemies did: The Samnites and the Iberians notably so, but the Gauls did, javelineers were present among the Hellenistic armies, Polybius tells us that Hannibal equipped his troops with Roman arms after the Battle of Trasimene, and so forth.

            Most battles were opened by skirmishes between the light troops of both sides before a clash of the battle lines, meaning that any javelins left on the field from wounded or slain skirmishers, or thrown and fallen short between the lines and still intact, could indeed be picked up and re-used by legionaries.

            We also have plenty of primary source mentions of troops in battle re-using thrown javelins and returning them at their enemies. Caesar tells us that the Gauls did this to the Romans at the Battle of the Sabis:

            “But the enemy, even in the last hope of safety, displayed such great courage, that when the foremost of them had fallen, the next stood upon them prostrate, and fought from their bodies; when these were overthrown, and their corpses heaped up together, those who survived cast their weapons against our men [thence], as from a mound, and returned our darts which had fallen short”
            (Commentaries on the Gallic War, Book 2, Chap 27)

            Polybius tells us that the Roman velites took special care to sharpen their javelins to an extremely fine edge, so that the edge would be blunted after use and wouldn’t be returned at them, indicating that the enemy throwing your javelins back at you was a fairly normal tactical concern at the time:

            “The wooden shaft of the javelin measures about two cubits in length and is about a finger’s breadth in thickness; its head is a span long hammered out to such a fine edge that it is necessarily bent by the first impact, and the enemy is unable to return it. If this were not so, the missile would be available for both sides. ”
            (Histories, Book 6, Chap 22)

            As for Roman movement in battle: Caesar and Livy and Polybius all tell us that the Romans did indeed fight in formations and kept their ranks and stood by their standards, but the primary texts also indicate to us that the Roman method of fighting allowed for a lot more individual leeway and movement than other, more strictly close order formations would do so. Caesar’s men at the Battle of the Sabis were impeded when they got too clustered together, and he tells us that to make effective use of their swords he had to order them to spread out again. Polybius tells us that each Roman soldier moved as an individual, and required more space as a result than contemporary Macedonian soldiers did in their phalanxes.

            The other factor in Caesar’s account of the duration of combat at Ilerda is that both the Caesarians and the Pompeians were able to reinforce the front line and relieve tired troops with fresh cohorts. How could that be done if they weren’t at a stand-off state, separated by a no man’s land in which each side would enough breathing room to move tired troops back and fresh troops to the front? This is one of the notorious mysteries of Roman warfare, how the rotation of front lines worked, and I think it only really can be answered if you assume a mutual stand-off state at the front line, and a model of combat in which ranged skirmishing is more prevalent than hand to hand exchanges.

            Of course, ranged skirmishing may be more prevalent than hand to hand combat, and yet also not necessarily a continuous state of affairs. A century might move up into range to throw some javelins, backpedal out of range again if met by fierce resistance and it doesn’t appear a charge will be successful, and so forth.

          5. “Javelins were extremely common weapons in Antiquity! Almost all armies used them!”

            Almost all armies had dedicated skirmishers who used javelins. Almost none of them had javelins as part of their troops’ standard kit. If the javelin was the legionary’s primary weapon, it would obviously have to be usable against at least most of the enemy army, not just exclusively their skirmishers.

            “Polybius tells us that the Roman velites took special care to sharpen their javelins to an extremely fine edge, so that the edge would be blunted after use and wouldn’t be returned at them, indicating that the enemy throwing your javelins back at you was a fairly normal tactical concern at the time:”

            I…yes? Which is evidence against your assertion that the Roman strategy counted on the enemy returning javelins in order to keep troops armed for extended skirmishes? Why did you think this was evidence in favor of your position? Your position is that Romans counted on enemies throwing their own javelins back at them in order to keep troops skirmishing for hours despite carrying only two missiles. They can’t do that if their javelins are specifically designed not to be returned!

            “The other factor in Caesar’s account of the duration of combat at Ilerda is that both the Caesarians and the Pompeians were able to reinforce the front line and relieve tired troops with fresh cohorts. How could that be done if they weren’t at a stand-off state, separated by a no man’s land in which each side would enough breathing room to move tired troops back and fresh troops to the front?”

            Okay, I’m gonna copy/paste a quote from my previous response here:

            “Nobody ever claimed legionaries remained in melee for hours at a time (such a claim runs into the exact same math problem as the claim that legionaries spent most of their time hurling javelins – are we to believe that two men can exchange blows for 5-10 minutes on average before either of them gets a decisive blow on the other?).

            Your claim is that missile weapons were the primary weapon, which means any amount of time spent at a stand-off is not evidence in favor of your claim. Modern footage of rioters and riot police facing off – to whatever extent those provide valid insight into ancient battles at all – are evidence against your claim, because they show opposed forces staring at each other for a long time and not exchanging large amounts of missile fire.”

            Go ahead and re-read those two paragraphs as many times as it takes for the meaning to sink in.

          6. Chamomile,

            I don’t understand your argument. How does the existence of lulls or stand-off periods in a battle argue against the predominance of ranged combat in terms of a legionary’s time spent in combat? If they’re at a stand off, outside of range, then a small movement forward will bring them into range

            I don’t quite understand why stand-off periods mean that a legionary isn’t spending most of his combat time at distance from their opponents, which is what I’ve been arguing all along.

            Most of a given day of battle isn’t actually spent in direct combat with the enemy, and at any one time only a minority of any army is actually in contact.

            The majority of a legionary combat time being mostly exchanges of javelin throwing doesn’t mean that they’re spending actual literal hours on end throwing javelins. It means that there’s comparatively more javelin throwing than sword fighting, but both of these activities are a minority of actual time in a battle.

            Also, almost all of Rome’s enemies made extensive use of javelins themselves. The Gauls and Iberians and Germans most certainly did, the Carthaginians most likely did. The only people who definitely didn’t were the Macedonian phalanxes, against whom the legions experienced considerable difficulties in frontal combat.

            Additionally: Almost every army of Antiquity opened battle with skirmishing from their light troops, those dedicated skirmishers you point out. Where do you think their javelins go after those opening engagements? Do they disappear into thin air? No, every javelin that misses or falls short will be left on the ground, and every javelineer who falls wounded or killed will leave his ammunition there also.

            A Polybian period legion had 1,200 velites. How many javelins did each carry? There’s various suggests of 5 to 7 per man, let’s assume it’s 5. That would be 6,000 javelins carried by that group. If even a hundred of the velites are wounded/killed in the opening skirmish, that’s up to 500 extra javelins left on the battlefield. That’s not counting the enemy’s javelins from wounded or dead skirmishers, or any javelins which had fallen short or missed between the lines and were still in condition to be used.

            And even pila could apparently often be re-used and thrown back, because Caesar says that the Gauls did that to his own troops!

            So, no, I’m not too concerned that each legionary has only two javelins. Javelins would be very common, and they would litter a battlefield after the opening engagements of light troops.

          7. “I don’t quite understand why stand-off periods mean that a legionary isn’t spending most of his combat time at distance from their opponents, which is what I’ve been arguing all along.”

            No it isn’t. You started out by arguing this:

            “More recent scholarship on the “face of battle” among the Romans has instead placed a central emphasis on the use of missiles (Pilum and other tela) as the primary weapon, with the gladius as the back-up, rather than the other way around as was often thought.”

            Then you switched to arguing that they spent most of their time in stand-offs, pretended you’d been arguing that all along, and hoped no one would notice the difference. Time spent using no weapons is not time spent using missile weapons.

            Also, if you’re going to try and make a “do the math” post, you should probably, uh, do the actual math first. You talk about how a Polybian legion would leave hundreds of javelins behind by the end of skirmishing! Hundreds!

            But a Polybian legion has (by the lower estimates) three thousand heavy infantry in it. You estimate that the number of javelins left behind on the battlefield is less than the number required to arm even one-third of the heavy infantry, and then declare that this proves that the heavy infantry spent all their time throwing javelins. If literally every single javelin carried by every single veles wound up in the hands of the heavy infantry, it would not even be enough to give all of them two extras. In addition to the two they started with, these three-and-a-half javelins would take less than a minute to throw, and are specifically designed not to be returned to them by the enemy. And no, the fact that enemies were *ever* capable of returning Roman javelins does not change the fact that the Romans designed their javelins to be as expendable as possible.

      2. Is it possible that the Romans could have simply got support troops or camp followers to bring up fresh javelins throughout the battle? Granted I can’t think of any primary sources mentioning such a practice, but then again, the primary sources fail to mention a lot of things that were apparently just taken for granted (which is why we’re having this discussion in the first place!).

  25. All the pike and shot formations that you state are historical but not possible in Total War series were, as far as I recall, quite properly implemented (and indeed were the core of) in the 2000’s era Cossacks game series from an Ukrainian developer. E.g. you could have shot outside of the pike square and then retreat them inside of it. Apparently there’s a modern remake Cossacks 3, but I have not tried it.

    1. Regarding Cossacks 3, don’t bother. It is a sequel in name only and a pretty lackluster game.

  26. I actually own a matchlock musket and have reenacted a little, so it’s been really disappointing to see Creative Assembly leave this massive historical/tactical gap in their game library and instead go off making games about orcs fighting mummies. I actually tried the first Warhammer game but without the historical background I bounced off it pretty hard. I have tried several mods for Medieval Total War Two such as a 30 Years War mod and an English Civil War mod. While fun, by the time I was playing them Medieval Two was quite old, as were the mods, and this resulted in allot of what is best termed “jank”. For whatever reason no one ever made a pike and shot mod for Empire even though that would be the best game for it.

    The author here brings up a good point when he says that the reason CA never made a pike and shot game was due to the complexity involved in doing it justice. I had not really considered that but it seems likely. Otherwise why not make Empires time frame from 1550 or 1600 to 1800 instead of just 1700 to 1800. Here’s to hoping CA breaks out of their recent funk and does so by going back to making ambitious historical war games and fills that pike and shot hole in their library.

  27. In context of Total War: Warhammer – I can’t say which edition of the game it was (and there are like eight or nine to pick from), but I remember that the tabletop game actually had a nod to mixed-formation infantry. Namely, the Empire had a specific rule that allowed to field an infantry unit with “detachments” – for example, you could field five dudes with muskets attached to a twenty-dude block of melee infantry. Pretty much, you could not field an actual mixed infantry square, but as a consolation prize you could at least keep the bastion.

    It’s worthy to note that the Empire is the closest of all Warhammer factions to a real Early Modern military.

    Which brings me to another point. This is a fantasy tabletop wargame, with at least one faction closely modelled on a real Early Modern state. As a fantasy game, it can be excused for not being modelled too closely, if you catch my drift. But, surely, there is at least one _historical_ tabletop wargame that does mixed infantry correctly?

    Also, I had a real good thought on asymmetrical warfare, but by now it’s like three weeks past the relevant blog post. Alas.

    1. Yes, though I like to add, that as far I remeber, detachment mostly were used for heavy hitting two handed sword-, or polearm troops, attached to spear, or pikemen.
      The pikemen would take an enemy charge, and the detachment, would try to engage the enemy units flank.

      1. In Swiss and landsknecht practice, the halberdiers and two-hand swords men were a small unit grouped around the banners. When the pikes crossed, the press often became so great the long weapons were useless. They then came forward to break up the mass, and also acted a banner-guard.

    2. Sure, pick up any tabletop ancient / medieval / renaissance wargaming rule set from the past few decades and you should find rules for composite infantry formations. If you can find multiple sets of rules you will find that there are multiple interpretations of what “correctly” means.

      For those who are not particularly interested in the fine details of wargaming, I can recommend the Wargames Research Group “De Bellis Antiquitatis” ancient/medieval ruleset, which everyone calls DBA. It is short, readable, and the first few pages explain the design philosophy, what the game is trying to model, and what the game designer regards as the important differences between troop types and roles. Even if you end up disagreeing, it’s a good starting point.

  28. In the original Rome disbanding units *does* increase the population of the region they are disbanded in, whilst raising units decreases the population in that region. The population that is added or removed is based on the unit scale you’re playing at; the amount added or subtract is the unit size listed on the unit card.

    I think it’s the only Total War game that does this — maybe earlier games than Rome do? — even the remaster makes it an option that you have to explicitly turn on.

    Horrible population crashes destroying the economy from over-recruiting can become a thing if you’re playing on one of the larger unit scales. I think the reason they took it out is the AI isn’t set up to deal with the mechanic at all, meaning that if you know about it and can avoid the downsides (population crash) and use the upsides (ability to move population about) you get a big advantage over the AI they didn’t intend.

    Not that I played Rome I extensively, I’ve just watched Many A True Nerd’s youtube playthrough too many times.

  29. God, I’d love an Empire II that gives the Qing, Mughals and Persians their due. In fact, my dream TW would be an Empire where they take a leaf out of the Warhammer trilogy- focus on one region at a time, gradually combining the map.

    One innovation from Three Kingdoms that I think has potential for the historical games is the dividing of armies into thirds, each under their own commander. Now, in that game it’s mainly used to field lots of characters to set up wuxia rivalry, and that’s very fun.

    But I’d like even the straight historical games to do something with it- instead of one general effortlessly coordinating the whole army, make sub-commanders matter. So you can be Alexander charging off with the companions while Parmenion holds the centre, for example.

    You’d have to get away from the Three Kingdoms model where the units assigned to one commander can be deployed wherever; if your colonel is on the left wing, you’d want to make it so his units have to be positioned within a certain radius of him. That way you’d naturally get interesting divisions on the field.

    1. Speaking of commanding mattering, the space 4X Sword of the Stars 1 (also sometimes called “Total War in space” !) has the fleet reinforcement mechanic, where you need to have at least one non-wrecked command ship to be able to field a decent number of ships in the same (timed) battle (and also to do formations and have reinforcements warp near the CnC and in correct order).

      So once your fleets get big enough (and until you get larger CnC ship sections), you’re incentivized to split them in order to not have ships sitting “wasted” in the reinforcements list –
      (which is a tricky balance in itself – you do NOT want to run out of combat ships, or god forbid, of replacement CnC ships, and then have your poorly armed and fragile tankers and repair ships to be forced to join the battle !)
      – but use instead those extra ships to carry multiple strategic attacks on different star systems in the same strategic in-between-turns.

      An example :

  30. I like when you go over historical use of troops. Can you write something about use of reserve forces/conscripts in the early modern to modern era vs the use of professional forces? This subject has been interesting to me lately

  31. Speaking of composite heavy infantry you didn’t mention one of the most historically obvious examples; the ottoman janissaries. My understanding is that janissaries were all expected to know how to use the bow, later musket but were also expected to act as shock infantry when the moment arrived. That said I don’t believe that janissaries using a bow in battle would switch over and pick up an axe when the enemy came in close. My understanding was that the they were given a specific role for a battle. Would be an interesting element to a medieval total war game to have units where you can select the type of weapon they will be armed with before a battle.

    1. AFAIK janissaries primarily carried a bow or later musket, but had decent armour and helmet and carried swords. The “shock” role was that they were just about the only Ottoman infantry who were disciplined and had good morale, so could be relied on when say the Sultan needed troops to assault fortifications.
      They weren’t composite infantry because (nearly) all carried the same equipment. In field battles the Janissaries needed to be behind chains or wagons or other obstacles to protect them from cavalry charges, unlike the composite pike and shot formations.

      1. So they didn’t actually all use the same equipment during a battle or siege. As you noted they did all generally use a bow or musket and for a period of time they would divide their unit to have both firearms and bows going at the same time. The firearms had the power but lacked, the rate of arrows an archer could put down range, so they would use both together at the same time intermixed with each other. If you so choose to spend the time look up Ottoman illustrations from the Süleymanname where it shows depictions of Janissaries from the Siege of Rhodes.
        You see them sporting pole-arms, firearms and bows. So I would definitely consider them to be composite infantry.

        1. Since we’re referring to Total War as not having composite infantry, just having multiple weapons present on the battlefield wouldn’t count. Wikipedia thinks the palace guard specifically carried the halberd but that has no citation and I wouldn’t know where to check that.

          The siege of Rhodes image I think you’re referring to shows a sheath on one of the musketmen, which makes me think they would in fact switch over to swords when the enemy got close. Which wouldn’t be at all unusual; most archers had at least a knife just in case.

        2. Spent some time reading through Armies of the Middle Ages Vol 2 by Ian Heath; Firearms a Global History by Kenneth Chase; Iran at War by Kaveh Farrokh.

          Agree that Janissaries could have bow or musket, although looking at the (Ottoman) pictures of the siege of Belgrade and battle of Mohacs in the 1520s the units of Janissaries are either all musket or all bow, not mixed. For disciplined infantry trying to do volley fire it would be much more practical to have everyone with the same rate of fire and range.

          Janissaries are not the kind of composite infantry Bret is discussing because AFAIK they never tried to be “cavalry proof” infantry. The Assyrian, Achaemenid Persian, and European pike & shot formations all have enough spears/pikes to hold off cavalry *and* provide a significant hand to hand advantage over other infantry.

          The majority of Janissaries are described / painted by Europeans, Ottomans, and Iranians as carrying bows or later muskets and a sword or axe. Some may have been wearing armour under those robes, and some carry medium sized round shields, big enough for sword fighting but not so big that they’d interfere with shooting if slung on your back.

          In the 15th C the bow and arrows were paid for by the Ottoman state, later on the muskets were supplied by the state. No mention of official pikes or polearms. On the battlefield their primary role is ranged combat rather than melee, and wherever possible the Janissaries are protected from cavalry by trenches, camels, chains. (As mentioned earlier the Janissaries *are* shock infantry in sieges, because they’re by far the most reliable available to the Ottomans.)

          When they don’t have fixed defences Janissaries are vulnerable to European cavalry charges, or even to Iranian infantry who in 1745 were explicitly ordered not to shoot it out with the Janissaries but to charge after one volley. That’s not what we’d expect from composite infantry formations of the kind Bret is describing, who regularly move around the battlefield and attack other infantry.

          Still, there are definitely Janissaries with halberds. Maybe they are like the NCOs in the English 18th C army who also carried halberds? The NCOs were supervisors rather than shooters, and a halberd made them easy to identify while still being handy in a melee.

  32. The thing with the Streltsi is that in older Warhammer lore they were exactly like the historical ones. Look up the Storm of Chaos stuff and the Kislev mini-dex that came with it. Streltsi there had guns and bardiches. The new lore on current Streltsi is that their new guns is a thing that was enforced by adaptation. Not to mention the axe-guns they use are based off actual historical axe-guns that were very popular among slavic nations. With the Polish Hussars being the most fond users of this. Of course the blade, instead of being put where the muzzle is, is put where the butt of the gun is so to use it as a 2-handed axe.

    Now I shall wait for Nippon and their Samurai made into powerful hybrid units.

  33. …so where does the glaive go when you need both hands to fire the bow?

    Presumably, they hold it behind their back for a second and then let go, hoping you won’t notice that there’s no conceivable harness or scabbard that you could stick a weapon that size into with that sort of motion. You know, like half of fantasy heroes do with their swords.

    1. TBH, I assumed that the glaives are conjured weapons, since they’re made of enchanted ice. In that case, I’d expect it to literally turn into the bow. (But also, I assume the animation is just “Look over there! A three-headed monkey!”)

    2. Makes you wonder what Persian Immortals did with their spears while they used their bows. According to Herodotus they even carried shields as well!

      1. Maybe they laid their spears down on the ground, and picked them up again if they needed to move/the enemy got too close? Or perhaps they only actually carried one or the other into battle, like the Assyrian infantry.

    3. Some cavalry had harnesses which could be used to carry spears, leaving both hands free for archery (Procopius mentions such equipment at the beginning of De Bellis). Though I guess trying a similar thing with infantry would be too unwieldy.

      1. But spike to put it in the ground, or slave child to hold it for you. Persians had child labour right?

  34. I haven’t played any of the Warhammer Total Wars, but I do think it’s worth mentioning that there was an attempt to model the kind of composite unit Bret mentions in *tabletop* Warhammer – specifically with the Empire. There you had a “parent unit” of melee infantry and could attach “auxiliary units” to it which were dependent on the parent unit’s command, and had a number of specialised manoeuvre options: for instance, they could execute the “stand and shoot” command against units charging their parent unit (an option otherwise unavailable to units who weren’t being directly attacked).

    It was a bit clunky and the actual composition of the units was all to cock, not least because the Empire never actually had pike units on the tabletop (only spear units, though pikes featured prominently in some of the fiction), and because the rules encouraged the auxiliary unit to be another melee unit rather than missile troops. From what I saw, it was common for the main unit to be made up of “swordsmen” and the auxiliaries to be “militia” who had a bonus hand weapon and therefore a second attack (another nonsense, for several reasons, but if I get stuck into Empire army list composition I’ll be here all day).

    In 5th edition, the Dogs of War – the only tabletop army that *did* ever get pikes – had a unit which combined crossbowmen and pikes, in what seems to have been another attempt to model tercio-style formations. I think the Dark Elves also had a combined arms crossbow-and-spear unit for a while. I can’t remember how either of those worked, though.

  35. Didn’t TW: Three Kingdoms have a high-tier unit that was an actual composite of polearms and missile troops? I don’t know how they worked though, because the game was always basically over before they came online.

    1. There were a couple of units like that in 3K. The green food line of the reform tree allowed one of those, I think Azure Dragons, to be recruited in something like 25 turns (5 techs) and it could last nearly the entire campaign with only high tier dedicated ranged or melee able to defeat it. It was popular in coop matches, I remember people making armies of just 2 retinues stacked with 4 of those + 2 trebs in one retinue and 2nd retinue of 2 polearm infantry + 4 cavalry which could easily defeat 99% of AI armies of 3 retinues with light losses.

        1. Yes. Azure Dragons are especially useful because they can be recruited by any type of commander. Normally, only Strategists can recruit any kind of ranged unit that isn’t the crappy basic Archer Militia. 25 turns also isn’t that long for these games, and you can do it in less than that if you capture the temple at Pengcheng or are Cao Cao. Being able to fill all your armies with crack archers is definitely worth it.

          1. Technically commanders can recruit the regular/non militia archers and crossbows as well. And that’s before we get into Nanman and Yellow Turban general-types.

        2. It depends on where you are, since it’s part of the green tree it boosts peasantry income, which is mainly located in the Central Plains (so Cao cao and Yuan Shao basically)

  36. I feel like a levy system could work in a TW style game, a pool of units that semo-randomly generates in a settlement getting larger and better over-time (to reflect the buildup of arms and armour due to wealth). But that when mustered into an army started depleting local food and pulled population from the settlement.

    And overtime these units would degrade at a speed depending on factors such as victories/defeats, leadership and if you decide to actually pay them. with units that were disbanded returning to their original settlements and returning the pop, but the units in the local pool would have to slowly build up again. refleacting the local economy recovering and allowing their levy to rebuild.

    Although a system like this would need a lot of fleshing out and control through buildings and policies, like an armoury giving the armoured billhooks. Or enforced archery shifting the proportion of longbowmen produced.

    1. That actually wouldn’t be too much of an extension from warhammer 3’s demon factions; Khorne and Slaanesh can generate temporary armies and Nurgle has a recruitment pool that fills over time.

    2. That’s how Thrones of Britannia worked, sort of. You had a pool of units that generated over time (based on techs, various policies, etc.) that you could then recruit (and would take a few turns to replenish to full strength)

      Generally it was divided into “Levy units” that replenished quickly, “retinue” units that were more expensive but also didn’t replenish as fast and “Elite” units that could take 10-20 turns to replenish but were much more powerful.

      1. I feel a combination of a court system with retinue units being tied to that instead could be interesting. i.e. investing in your personal guard gives you a larger number of elite units, but replenishing them is expensive if they take casualties. It could be tied into a legitimacy system similar to Dr Deveraux talks about courts and how kings needed to rule.

  37. I’d love to see an analysis of Pike and Shot: Campaigns in here. Maybe too small of a game, or too abstracted for the purposes of your blog, I dunno, but it’d be cool. There aren’t that many games that cover the period and, as you said, it’s pretty much the big remaining gap in TW history (not that they did Bronze Age warfare justice, but at least technically they covered it).

    1. They could have done the fertile cresent or looked at the late bronze age with egypt vs mesopotamia and massive chariot battles. instead they did a psuedo 3K/warhammer mix that didn’t really add anything new. They could’ve looked at chariot mounted infantry or integrating different infantry types.
      Instead they couldn’t make up their minds with troy and produced a weaker game for it.

      1. I do note that there are some fairly interesting things they did with Troy beneath all the fantasy stuff. It’s one of the few games where lighter infantry has an actual tactical role, for instance, and the terrain modifiers being much more severa had some interesting implications.

  38. Another great step back in TW games is the lack of enforced unit variety. As long as your coffers can afford it, it is entirely possible, and actually preferable, to have armies full of one type of high-end unit (the so-called doomstack).

    I think it was only Medieval 2 (a high water mark if there ever was one) that had a system where your elite units could be recruited only in limited amounts. So while you could recruit as many peasant mobs as the city/castle could carry, you could only recruit one high-end knight unit, and you would need to wait for a few turns before another one became available. Sadly, this mechanic, along with splitting settlements into economy-focused towns and military-focused castles, was abandoned.

    It’s last vestiges of this were, I think in Shogun 2 FOTS (high point number 2), where elite sharpshooters and foreign troops were limited. The highest-end japanese line infantry, however, could be (and would be) fielded in endless armies.

    1. It wasn’t just the elite units that were limited; all units were limited, although lower-tier ones had a high enough unit cap and fast enough regeneration rate that you could usually build as many as you wanted.

      Empire/Napoleon had a similar mechanic to Shogun 2 FOTS, wherein you could only build a limited number of certain kinds of units, and not always the best ones, either. Spain, for example, had a unit that was Irish exiles, was basically slightly different line infantry, and you could only build four of them.

      1. It doesn’t, actually.

        Tomb Kings in Warhammer 2 have an enforced unit cap, but Med 2 used a different system where unlocking access to a unit only meant that your barracks would ‘grow’ up to a certain amount of that unit over time to be available for recruitment. This meant that you couldn’t, for example, do the Tomb Kings thing of, “oh balls, I lost an elite stack of Ushabti. Welp, I’ll have it retrained in like three turns from this province where I stacked extra recruitment slots.”

    2. It’s actually still true to a certain extent in both 3K and Warhammer. Though not as comprehensivel as med 2.

  39. I believe Total War originally broke off from the classic RTS systems of the likes of Age of Empires. In those games composite units were easy to set up with the grouping feature, but classic Total War started simple with monolithic groups of units. A shame, there could have been a timeline where the series continued to scale up the classic RTS combat system but the appeal of cinematic graphics and scale doomed that experiment.

    1. Pure RTS is niche these days, but had quite a lot of improvements over the years :

      “Zero-K[/Spring] Circle Guard”

    2. I don’t think TW is RTS scaled up in that sense, rather it’s an adaptation of the classic wargame. And the element they introduced (and that RTS has had trouble with) is loss of control/morale. (there are some games, like Company of Heroes or Dawn of War that has morale in a classic RTS format, but you notice they also tend to use squads as your basic unit type, not individual units)

      1. Also, in those games you don’t actually lose control of those units unless you hit the “retreat” button to send them sprinting home. Instead they take significant damage and/or movement penalties to represent huddling in cover. IIRC the Dawn Of War devs said they’d experimented with having squads rout in early builds, but players hated losing control of their units.

        1. In DoW 1 cover (and “anti-cover” like water crossings) automatically apply bonuses (maluses) to units in the right location.

          Morale is a separate mechanic, and I’m pretty sure that you (mostly ?) lose control of low morale units : IIRC they stop fighting (unless forced into melee ?) and start running away ?

          1. They don’t; they get a debuff that significantly reduces their damage output such that in most cases you want them out of the fight to regroup but they still take orders normally. There may be a “retreat” button that makes them leg it for HQ (definitely present in 2) where they do stop shooting entirely and can’t be controlled until they arrive, but that’s manually triggered.

          2. Hmm, I might be just misremembering AI opponent’s squad behaviour then ?

          3. Likely; I remember DoW2 a lot better than 1, and in that one the campaign AI, especially in some of the earlier missions and for orks or Tyranids without synapse creatures, is liable to hit the run button when suppressed. It pops up a yellow exclaimation point and they run for a base structure at maximum speed. The player can do the same thing, and the running cannot be canceled once activated.

            There’s some variance in the exact effects between various DoW and CoH games and I’m not entirely sure which is which, but in some of them being suppressed grants a considerable buff to ranged damage resistance because they go prone.

      2. DoW/CoH units have significant (for the genre) independent action as well; taking cover on their own and some other things

  40. To add onto what you say in the Reducing Complexity section (as well as a comment I remember from an earlier article regarding how units coordinated), this problem is aggravated by the implicit assumption of direct control that so many games make. In strategy games, for example, soldiers aren’t independent actors whose needs you, as their leader, must negotiate; they’re more or less perfect extensions of your own will, there to carry out your actions with as much efficiency as the game will allow. However, making that work – especially in a real time and/or multiplayer setting – requires abstraction and reducing the degree of complexity you’re asking the player to manage, so even where the point of the game is navigating a highly complex series of systems (as in a lot of strategy games), some aspects of these systems simply can’t be represented without contradicting that basic assumption about how the player approaches the game.

    In that regard, I’m kinda interested what you’d think of Frozen Synapse. The premise there is less about direct control than it is about devising a plan and seeing it carried out, which I imagine is, in some ways, closer to how military strategy works in practice.

  41. The Atreides composite unit of spears and swords was interesting and seemed plausible with Duniverse type body shields. The spears keep the enemy off balance while the swords slip through.

  42. “how the Ottomans and Mughals protected there musketeers” – their
    “Their games sense then” – since

  43. An interesting variant on RTS genre that allowed the player to build arbitrary composite units from multiple unit taype was Kohan: Immortal Sovereigns. Your “units” were built from mix-and-matching sub-unit types, and they would manage their own “positioning.”

    It’s been a while since I played, so I don’t recall the exact details

  44. The most frustrating thing about Total War’s lack of composite units to me is that Warhammer Fantasy Battles had composite units. Some are for factions that never made it to the video games, such as Pirazzo’s Lost Legion which has pikemen and crossbowmen and is clearly inspired by the tercio. The Lizardmen had many composite units on tabletop such as mixed Kroxigor and Skink units, as well as salamanders requiring skinks to herd them into battle.

    Not only are they not reflecting history well, they are also not reflecting my favourite army well.

  45. I wonder that you don’t mention one of the best examples of composite infantry (to my mind), Alexander’s planned reformed phalanx, which was sadly unrealized because of his death. As far as I can tell, the plan was for a phalanx with the front ranks being phalangites, with Persian archers/javelins in the ranks behind, and phalangites bringing up the rear. That sounds like pike and shot well before that idea ever came into being. See Alexander Nefedkin, A Note on the Later Phalanx of Alexander the Great

    1. His successors had plenty of time to experiment, and did not pursue the idea. The point of a phalanx is its weight and density, while archers need room. For that matter, the Persians had both heavy infantry (eg the Immortals) and archers, and did not try to integrate the two. Nor did the English try to mix the men-at-arms and the archers. All the experience is that they operate best as cooperating units, not a single mixed unit.

      1. You are probably correct in that Alexander’s successors would have used his plans for a “reformed phalanx” if it had actually been a good idea. However as Bret points out, Persian ‘Immortals’ were expected to fight both as archers and melee troops. Herodotus says that all Persian infantry “carried short spears, long bows, and arrows of reed”, which seems to concur with Achaemenid artwork of royal guards (

        1. My guess would be that they carried short spears because whatever they did with their spears while they shot wouldn’t work with long spears, which put them at a significant disadvantage in a hand-to-pike fight with phalanxes, but the fact that they’d shot at the phalanx beforehand made up for that enough that they didn’t decide to switch to phalanxes at that point in time; by my understanding after Alexander pretty much everyone switched to phalanxes.

        2. The Persian army drew on all 20-odd satrapies, with all their distinctive equipment. For that matter, they quickly realised the value of Greek heavy infantry, and used Greek mercenaries extensively. Still, no “spear and bow” units.

          1. As i wrote both Herodotus and Persian art portray ‘Immortals’ as “spear and bow” troops. In fact in his Catalogue of Nations Herodotus describes most Iranian peoples as equipped in this fashion, as well as the Aethiopians

  46. Ever since reading about the Battle of Neva at the beginning of the Great Northern War, I’m curious about the apparent, charge, volley, charge practice of Carolinian Infantry.

Leave a Reply