Building a Fantasy Army Part 8: Weapons and Tactics
‘Dis be the final part of the series. ‘Ere we go!
Weapons and Tactics
Weapons and armour
In antiquity and early Middle Ages, main weapon was spear. Both infantry and cavalry used spears to stab enemies. One exception was Macedonian Phalanx, which is one of (if not the) earliest pike formations in history. Byzantine Kataphraktoi meanwhile used maces, which could be swung or thrown, as their primary weapons. In later Middle Ages, main weapon of infantry could be a pike or a polearm (pollaxe, polehammer), and heavy infantry was supported by crossbowmen who kept pollaxe or polehammer as a backup weapon. In either case, density of troops – massing at the point of the attack – was an advantage, at least up to the point. Greek phalanx was 8 ranks deep, and Macedonian one 16. Ranged weapons were not used to produce casualties, but rather to inconvenience the enemy and break up their ranks prior to the contact.
Overall, weapons will adopt to the terrain, tactics, and enemies faced, and enemies will adapt to weapons, tactics, terrain and so on. Tactics also depend on weapons, terrain, enemies and culture. Horse nomads, such as Mongols and Huns, used bow as their primary weapon: they used it for hunting, and this naturally extended to combat. Bow had cultural significance similar to sword in feudal cultures; Armenians referred to Mongols as “the nation of archers”. Mongols were required to possess 2 – 3 bows and several quivers of arrows. Full kit was bows, two spears (long and short), an axe and ropes; swords were used only by the wealthy. Mongol regulations also required soldiers to possess nine pieces of leather armour, and armour for horses. Armour was acquired by trading with settled societies, and thus often resembled the patterns of armour used by said societies. Mongol lamellar armour was, in fact, Chinese armour, and they also brought Chinese siege techniques, tactics and gunpowder weaponry with them, as well as Chinese engineers. As a rule, nomadic societies generally avoid shock tactics as much as possible – nomadic way of life means low population density and inability to afford casualties. Instead, tactics are based on maneuver: hence why horse archers are so important, as they could utilize caracole, peltering the enemy with missiles until they either broke or gave chase. When they do utilize shock and siege tactics, it is always done through confederating settled peoples who can actually afford casualties inherent in such style of warfare.
For comparison, settled societies typically opt for shock combat and sieges. Fortified cities require investment to take, and this in turn requires long-term raiding – but density of settlements also means that raiders have to be able to stand up in pitched battle. As a result, heavy infantry and heavy cavalry are extremely important.
Armour itself heavily influences what weapons will be used. No armour, cloth armour and mail armour all mean that additional protection from missiles is required. This protection typically comes in the form of a shield, which in turn means that main weapons used will be one-handed: primarily spears, but also maces, bludgeons, and swords as a backup weapon. As mail became more prevalent, weapons adapted too – swords and spears became pointier, as did arrows. And while armour may be defeated by a solid spear or sword thrust, getting such a thrust against a moving target is very difficult; as a result, maces became more prominent, and cavalry adapted the technique of couched lance which could defeat the mail, unlike single-handed spear thrust from horseback which was a rule previously. In fact, there were accounts of knights in mail being thrown from horses by spear thrusts, much to the horror of Saracens they faced, or else hit by so many arrows they resembled pincushions. Aside from couched lances, armour could be overcome by targeting bony joints and by blunt-force weapons. This in turn meant that mail was reinforced with additional plates, starting with bony parts of the body (first head, then knees and elbows). A sword fight usually devolved into hitting the other guy in the head so hard he got a concussion. Drawings do exist of swords piercing or even cutting through armoured plate, but they are rare and may not be based on the eyewitness accounts. Overall, development of armour was driven (or rather, constrained) primarily by development of metallurgy; but weapons were developed and adjusted in order to defeat armour. This also impacted the military organization, as explained earlier.
While spear or sword trust can defeat mail if done properly, as could longbow at short enough range, no amount of man-generated force will completely defeat plate armour. Only a heavy, couched lance was capable of that, and even that depended on where and how it hit (hence the v-shaped deflector on the breastplate). Thus, with development of plate armour, main weapons become warhammers and especially polearms – pikes, pollaxes, polehammers. Neither swords nor older types of maces can be expected to reliably defeat plate armour, while at the same time increased protection of full plate makes shields much less useful. Thus weapons were needed that could concentrate high amount of force on a point by combining leverage with attack by a point – this is the purpose of a perpendicular spike on a bec-de-corbin, Lucerne hammer, war hammer and halberd. Cavalry lances also became much heavier, and lance arrest was added to the breastplate in order to increase the energy transferred during the clash. When such weapons were unavailable however, stabbing joints and bashing the head to cause a concussion remained a viable tactic.
Plate armour developed primarily among melee-centric cultures. Reason for this is simple. Lamellar armour consists of many small plates, which offer better flexibility and also better protection against piercing damage. An arrow will not pierce either a solid plate cuirass nor a lamellar cuirass (it might pierce thinner plates such as sides of helmet). But lamellar cuirass is more flexible, which means that it has more give before it is pierced – and a damaged lamella can be simply taken out and replaced. Plate armour is far better when it comes to protecting against trauma damage and specialized armour-piercing weapons such as polehammers and bec-de-corbin. Lamellar armour also has many more gaps in coverage than either mail or plate armour do, and its construction means that it gradually weakens under repeated blows, much like mail does. While these flaws are not much of a problem in missile-centric warfare, they can be a major issue in a hand-to-hand combat. In fact, lamellar armour was highly popular among steppe cultures, such as Avars, Mongols and Turks, partly for logistical reasons noted earlier, and partly due to prevalence of missile warfare among these cultures which made gaps in armour much less important while placing a premium on defense against piercing attacks.
Likewise, shields also depend on what armour is used. Ancient hoplites used massive shields because armor only protected torso and portion of the head, but upper arms and legs especially were comparatively unprotected, and even armour itself – being often of bronze – was not reliable. Later, mail armour also could not stop arrows or spears which would split rings open, which is why Norman cavalry for example carried massive kite shields, while Byzantine infantry of the period carried even more massive round shields. As plate armour spread, so did shields become smaller. 15th century infantry in plate armour either did not use shields at all, or used pavises for protection against missile fire – latter was because infantrymen often (if not typically) utilized 3/4-plate armour, with full protection of upper body and upper thighs, but no protection whatsoever for the legs. Against arrow-shooting enemies such as Ottomans, protecting legs was a good idea – but areas actually covered by armour required no additional protection.
First image is a reenactor in Norman 11th century armour. Second image is a group of 11th century Byzantine soldiers, while third image shows 15th century Hungarian infantry in combat.
Pikes were predominant infantry weapon in 16th century, but their usage requires discipline. They were thus used predominantly by various mercenaries, as well as by later (16th century) professional infantry. For a feudal army, various polearms were preferable – such as pollaxes, polehammers, halberds, partisans, guisarmes, and so on. Hussites replaced function of a pike block with a wagenburg, which allowed polearm and crossbow infantry to oppose feudal cavalry. This was later picked up by Hungarians, who also had no access to professional infantry. Militias also preferred to use halberds and other easier-to-use weapons. Earlier in Middle Ages, Byzantine thematic infantry relied primarily on spears and swords; pike formations only (re)appeared as army professionalized.
Equipment depends heavily on the style of warfare waged. Pitched battles favour heavy infantry – specifically pikemen – and heavy cavalry, with exact focus being determined by terrain. However, since most such battles happen in mountainous or otherwise limited terrain, there is a tendency towards pike infantry for societies which favour pitched battles. Open country both allows easier evasion of battles and usage of cavalry, thus favouring raiders. If main style of warfare is raiding, then primary focus will be on cavalry – heavy cavalry if pitched battles are also frequent, light cavalry if they are not. If warfare is primarily based on ambushes, light infantry and light cavalry will dominate (depending on the terrain).
Feudal cavalry-based system was made possible, and then reinforced, by the increase in size of horses. Horses which made up destrier bloodline came from the Low Countries, and was then mixed with horses from central Asia and Arabia. By 15th century, these horses could be well over 700 kg in weight, able to carry a knight and also heavy horse armour. At the same time, return of pikes to the battlefield and appearance of guns meant that cavalry had to utilize sophisticated combined-arms tactics, thus leading to professionalization.
As far as fantasy races go, their situation would heavily influence tactics. Elven need to adopt guerilla-style warfare in forested areas would lead to widespread usage of spearmen and archers, or else crossbowmen. However, if they are truly slender and weaker than humans, their ability to use bows against armoured opponents would be compromised.
Dwarves fight mostly underground, which means that they would be a heavy infantry force utilizing weapons such as warhammers and axes. When fighting on surface, however, they would need to compensate for their lack of reach. This means that they would opt for pikes and bows, as former would provide very long reach while latter would make use of their physical strength. Their strength and relatively high mass for height would give them an advantage in direct physical contests. Lack of height however would limit practical length of pikes, and slow movement speed will in any case make dwarven phalanxes not very effective in offense.
Werewolves would be highly effective at assassinations, as they could infiltrate enemy army in human form before transforming, as well as using their wolf form to track the target. They would also help with night operations, pathfinding, scouting and reconnaissance.
Trolls would be used as heavy shock troops, punching holes in the heavy infantry formations, though this might not be as effective against massed pikemen.
Despite a very common depiction of an undisciplined rabble charging at another undisciplined rabble with no thought to order, premodern battles were typically very structured and ordered affairs fought by trained soldiers. Cohesion was the key: a side which kept cohesion better typically won. Archers and heavy cavalry were both used to disrupt cohesion, although by different approaches. Archers generally did not cause mass casualties, but arrow impacts did cause injuries and exhaustion, as well as psychological pressure. This meant that the side with fewer or worse archers would often arrive to the point of contact exhausted, unnerved, and out of order. And since close combat required on morale and maintenance of order, side with superior archers had a distinct advantage. Heavy cavalry on the other hand disrupted cohesion by the power of their charge. Infantry, if arrayed deep enough and with adequate weapons, could physically stop a charge. But this depended on high levels of morale, and on terrain. Successive charges by heavy cavalry, especially on flat ground, would eventually break through. And once a break in the line was created, infantry would collapse as it depended on the mutual protection. This psychological and physical dependence on mutual protection in battle was also what made flanking attacks so devastating, and it was true for cavalry and infantry both. But cavalry, by virtue of its mobility, was better able to avoid being outflanked. Not all cavalry forces possessed the discipline to perform successive charges, however: it was not a rare occurence for a cavalry to charge at the opposing cavalry or infantry, through the enemy formation, and keep charging straight off the battlefield. Overall, a contest between heavy cavalry and infantry depended on three things: training, morale, and terrain. If training and morale were equal, infantry would win on uneven terrain and in forrests, while cavalry would have advantage on the solid ground. Otherwise, a side with superior training and morale would win. During Middle Ages this equation typically favoured heavy cavalry, as knights and men-at-arms were the closest thing to full-time professionals that feudal Europe had. In early Modernity however it favoured infantry, as heavy cavalry usually came from the ranks of nobility, and unlike knights they could not observe discipline to the level of now professional infantrymen. But only with reliable gunpowder weapons did infantry achieve definitive advantage over cavalry – yet even then a lot depended on the terrain.
Greek hoplite armies were probably the worst in terms of tactics used. They were incapable of taking fortresses (for reasons shown later), and their battlefield tactics were extremely simple, to use a polite term. In essence, hoplite warfare consisted of smashing a square block of armoured men into another square block of armoured men, and seeing which one will break first.
Crusading armies of 11th century and later displayed significant tactical sophistication. They relied on scouting, espionage, and other forms of intelligence gathering, to inform their decisions. When on march, typical medieval army was organized into three divisions: vanguard, main body, and the rear guard. The vanguard held the forward position in the field, while the rear guard protected the supply. Main body would be further divided into left wing, center and right wing.
“Frankish” knights were extremely disciplined, able to remain standing under constant missile bombardment before mounting a decisive charge. Usaban ibn Munqidh complained of his enemies, writing that “Of all men, the Franks are the most cautious in warfare” – a high praise indeed, considering that he may well have faced Byzantines at some point. In fact, success of a heavy cavalry charge depends on the strict discipline. Heavy cavalry had to charge at the right time, against the right target, and maintain tight ranks throughout. It is thus not surprising that “Frankish” cavalry charge was respected and feared, with Anna Comnena commenting that it could break through the walls of Babylon. At Muret in 1213., greatly outnumbered French force commanded by Simon de Montfort burst through the enemy army, killing King Peter of Aragon and annihilating his army. Knights were also fully capable of being recalled from the charge and reorganized for further assaults. This included tactics such as feigned flight, which are often thought to have been the exclusive province of the steppe nomads. With time, heavy cavalry tactics became even more complex, evolving into a combined-arms operations complete with missile and light cavalry support; latter was especially true in Poland and Hungary, which faced Ottoman threat.
While it is true that Western feudal armies often did have discipline issues, these consisted almost exclusively of lower-level commanders disobeying a high-level commander, with good example of Battle of Nicopolis. Common soldiers were extremely disciplined, and there are next to no accounts of Western infantry collapsing in disarray on contact with disciplined Eastern troops such as Janissaries. When such collapse did happen, it took a long time to reach – and infantry often kept fighting even after being abandoned by cavalry and facing a certain defeat, as happened at Battle of Kosovo in 1448. Infantry played a key role in medieval battles long before the 15th century, and practice of dismounting heavy cavalry to stiffen infantry line was used from 6th century campaigns of Belisar and Narses, through 11th and 12th century, and all the way to 15th century, in both Western and Eastern Europe.
Discipline combined with foot crossbowmen was a key in countering mounted archers which were common in Eastern armies. Byzantines knew this, yet Romanos sent away his archers and most experienced tagmata to attack city of Chiliat. Without archers and with only half of his army, he was easy prey for Alp Arslan’s mounted archers.
Tactics differed as they adapted to the enemies being faced and on the troops present. When high quality infantry was available – which was actually most of the time, even in medieval Europe – then typical deployment was with cavalry on the wings and infantry in the center. This was used generally against the like opponents, as it protected infantry from being outflanked. But tactics could vary significantly between disparate opponents. When settled peoples faced horse nomads repeatedly, they usually were aware that they are likely to get outmaneuvered. As a response to this disadvantage, two main approaches were used: infantry square and wagon fort. Both protected and enclosed cavalry within a defensive perimeter of infantry. But for these tactics to be effective, infantry had to be provided with a large number of missile troops to prevent the enemy from simply picking them off at the distance, as happened to Romans at Carrhae and Manzikert. If infantry could win missile exchange with cavalry – almost a certainty, given enough archers – then enemy cavalry would be forced to commit to a charge. As noted, successive charges actually stood a good chance of breaking the infantry. But if square was set up properly, enemy would not get a chance to mount successive charges: a counter-charge of heavy cavalry, sheltered behind the infantry, would outflank and destroy cavalry which was now heavily engaged and likely carrying out a tactical withdrawal.
Above shows (in order): tactical deployment of Burgundian army under Charles the Bold, deployment at Battle of Agincourt, Byzantine infantry square, and Black Army of Matthias Corvinus
Europeans in Outremer adopted tactics much more cautious than was typical in the Western Europe. This is an adaptation to nomadic horse archer tactics, and can be observed in other societies which faced such enemies, such as Byzantines, Hungarians and Poles. Additionally, limited manpower meant that a defeat in the field would inevitably cause a fall of a large number of forts and cities – as such, armies only marched to relieve a siege, which was in fact the case at Hattin as well. When the army was assembled, it would take up a strong defensive position and wait for a Muslim attack. If Muslim army dispersed in search of plunder – which was often the case – then it could be destroyed piecemeal. On offensive, strategy revolved around besieging cities while avoiding field battles. Infantry was used defensively, on a Byzantine model. Oftentimes deployed in a square formation, infantry would wait for the enemy to attack and engage him with missiles and with support of their own missile cavalry. When enemy came close and became entangled with infantry, reserve formed of heavy cavalry would engage and attempt to outflank him. Even in raids, it was unusual for Christian force to immediately attack Muslim one; typical response was to withdraw to a hilltop, observe the events and try to exploit opportunities that did not present risk. Both sides very frequently used concealed forces, feigned retreats and other military strategems. At Arsuf, Crusaders spent days holding under arrow fire before launching a counterattack and winning one of the most decisive victories in the entirety of the Crusades. But due to cavalry-heavy nature of Muslim armies, decisive victories were extremely rare.
Fortress assaults are a special branch of tactics. The simplest way to take a fortress is by a siege – but it is by no means simple, as it involves building siege works, securing supplies (lest besieging army starves before besiegers do) and securing against possible relief force. This means building fortifications facing both the besieged fortress (circumvallation) and the possible relief army (countervallation), with besieging army being encamped between the two lines of fortifications. Assaults are usually mounted only if defenders are taken by surprise or after preparation had been done. Former, being quick from-the-march assaults, are both simple – usually no more than scaling the wall with ladders and ropes – and almost certain to fail. A serious assault can be mounted once the fortress had been invested. But before assault is attempted, oftentimes other ways of taking a fortress will be investigated: many fortressess fell because a gate was left unlocked by accident, or someone within was bribed to open the gate for the attacker. If that failed, next phase was the preparation for the assault: this meant extended bombardment with siege machines and/or sapping efforts, in order to degrade enemy defences. Assault itself will be a hugely complex undertaking, involving suppression fire (from both missile infantry and artillery – latter being used to decrenelate the wall, assuming accurate enough weapons are available), siege towers (as access points and/or firing platforms to suppress defenders) and possible attempts to undermine the foundations of the wall by digging a series of tunnels underneath. All these efforts have to be well coordinated in order to produce the desired result. Escalade in particular can only be attempted if enemy had been caught asleep, or if defenses had been suppressed by artillery. What is common is that all these attacks require good coordination between elements of the army, as well as the presence of the corps of engineers. As such, typically only professional armies will opt for such tactics. An army that is both highly professional and large may also build an earthen rampart to gain access to the wall.
Less professional armies may opt for investment, or simply not besiege anything in the first place and try to wear down the enemy by raids. Latter was far more likely: even simply sitting near or around the city still required well-developed logistics, as well as the ability to defeat enemy’s attempts at interference (by either forces being besieged within the fortress / city, or by any possible relief force). If such an army did mount an assault, it was typically a hasty assault – aiming to surprise the enemy with ladder escalade (which typically failed). A larger and more organized medieval army could (and did) build weapons such as trebuchets and siege towers. While the massive (and extremely organized) Roman army used ramparts as access points and siege towers as artillery platforms to protect the construction of ramparts, medieval armies had no organization necessary to construct the ramparts. Because of this, siege towers themselves became access points, which is a role that made it to the popular culture (e.g. Siege of Minas Tirith).
Type of artillery used also depends on the sophistication of the army. Roman-style torsion artillery (such as ballistae, onagers, scorpios etc.) is the most complex in terms of both technical construction and tactical use. It depends on the sinew springs to move the arm(s) of the machine, and many types of torsion artillery (all but the onagers, essentially) were accurate enough to hit individuals. Such artillery was used to supress the defenders and decrenellate the wall. Simpler but highly effective device was trebuchet. Traction trebuchet was pulled by a team of people, and was used as a suppression and anti-personnel device. Counterweight trebuchet was far more powerful (if more complex) type of trebuchet, and was possibly the first type of artillery capable of bringing down curtain walls (though larger torsion devices may have been capable of the same). Gunpowder artillery, if available, is likewise capable of bringing down curtain walls, and far better at it. But at the same time, it poses major logistical problems – lack of projectiles and gunpowder often limited employment of cannons far more than did availability of cannons themselves.
Everything changes if magical attacks and similar destructive weapons (e.g. gunpowder) are available. In what way it changes will depend on how powerful the attacks are, and how – and whether – they can be defended against. Adaptations can be mechanical, by developing a defensive capability (such as shields against arrows, or pikes against cavalry). Tactics will also be changed to minimize impact of new weapons and tactics (note: even if weapon is not new “in universe”, it may be new relative to historical model – such as giving ancient Romans ability to throw fireballs). Lastly, offensive weapons may be developed to offset the new capability, leading to mutually assured destruction.
If some kind of defence is available (e.g. forcefields), then army may concentrate itself under the defensive umbrella, which will limit offensive and maneuver options. Any offensive moves are likely to be carried out by cavalry supported by mages, assuming cavalry is swift enough to avoid worst of the magic attacks. Depending on how magic is performed, defences may instead focus on countering or disabling the spells before they are cast. They could take a form of magic dampening fields or directional suppression effect. If spells require a spoken incantation, preventing the opponent from speaking (such as by causing him to inhale smoke) can be highly effective. Written magic could be used to enable permanently present spells for purpose of magical defence. Offensively, it could be used to reinforce weapons with magical effects, or to produce magical weapons which rely on writing – such as magic tomes.
If no such defence exists, then the army will spread out – and the more lethal and destructive magical attacks are, more spread out the troops will be. This happened in real life as well: introduction of rifles and explosive artillery shells made linear tactics unviable, and by the mid-19th century armies were dispersing. But if it does happen, then dispersed order will make infantry essentially helpless against the cavalry – which means that infantry becomes nearly useless outside of trench warfare (which requires squad tactics and is basically what RPG parties excel at). But due to low population density trench warfare can only be used in sieges, while most of war will consist of cavalry raiding and counter-raiding. In a field battle, infantry may well prefer to stay in tight formations, as long as magical attacks are not too dangerous; but even so, attacks will help cavalry by creating holes in enemy ranks. If attacks are too dangerous to remain in lines, then infantry will have to rely on cavalry for support. And if attacks are too dangerous for even a dispersed formation to be reasionably survivable, warfare will turn into medieval version of World War I trench warfare. Teleportation, if available, would completely change the tactics and even strategy, depending on the scale and limitations.
If mages can heal friendly troops, then they would be positioned in the back of the army. Depending on how quickly healing happens, mages may or may not be embedded within combat units (but again positioned at the back, as they would be a priority target). If mages can be used to send and receive orders quickly, they would also be a priority target. All of this could make an army far more effective – and far more aggressive, as soldiers would not have to worry about wounds as much, and commanders would be able to send orders more quickly.
On larger scale, magic could be used to affect battlefield itself, such as casting protective spells over an entire army or a city. If powerful teleportation spells are present, they could allow basically instantaneous tactical or even operational movement by whole units. An invisibility spell would allow large-scale ambushes even in the open terrain. Some spells could be used to alter the battlefield itself: fire magic and various constructs (of earth, water, plants) could be used to funnel or cut off enemies, or else create instantaneous fortifications for one’s own army. Enemy soldiers could be washed off by a sudden flood, or else blinded by a fog bank set up specifically against them.
Illusion magic could be supremely effective on both tactical and strategic level. In battle, it could be used to hide units while making the enemy chase after the apparitions, thus opening enemy’s flank. How successful it would be at that depends on how lifelike illusion is: soldiers that “clip” through the terrain or else march twelve feet off the ground are not going to convince anyone. Strategically, hiding several elite units and then making it appear as if they are present elsewhere can completely misdirect the enemy as to one’s intentions, plans and even goals. This is an old trick: from wars of Julius Caesar to World War II, deception was a key element in warfare. Ottomans conquered Bosnia as easily as they did because they were unexpected: Sultan had agreed to a truce just before he started the conquest, decieving king Stephen as to his intentions. During World War One, Entente deceived Germans as to a direction and location of a major attack merely by misreporting the position of their elite Canadian and Australian units. Likewise, major part in the success of the Operation Overlord was the fact that Allies had succeeded in completely deceiving Hitler about their planned point of invasion.
Both illusion and teleportation magic could be used to facilitate guerilla-style, hit-and-run attacks. If mages can render enemy soldiers blind or deaf – by burning out their retinas or rupturing the eardrums – entire units could be rendered combat-ineffective in short order by but a few mages. If healers are present however, and can heal said injuries, attacks would have to be repeated and sustained to have an effect – creating a high degree of risk for attack mages.
Problem with magic is that, since it is supernatural, it has neither set limitations nor a model to use from the real world. The only limitations are what the author says they are, and thus effect of magic can be felt either nowhere, or everywhere, or anything in between. It may be a mere addition to classical way of waging war, or a profound transformation. And if it is latter, both modelling the flow of the battle and writing its effects become extremely difficult: a mere handful of powerful mages may fundamentally change the nature of warfare.
Psychic powers are also technically magic. A telepath may range from reading person’s thoughts to completely controlling their mind – controlling actions, rewriting memories, implanting illusions, disabling brain functions and so on. Telepathy may also be used to help coordinate units on local or even global scale, depending on how powerful it is – it would, in essence, give 20th century radio to a premodern army. Of course, it is possible that such abilities could be countered in various ways.
Aerial warfare in fantasy typically involves magical creatures such as dragons, gryphons, pegasi and so on, but it may also involve mundane-but-essentially-magical devices such as airships and helicopters.
When it comes to flying beings, dragons especially, they are typically too large and too heavy to actually fly under normal laws of physics. Thus, there is always some magic involved – but because they usually have wings (but not always – Chinese and Japanese dragons being a good example of that), laws of physics and aerodynamics must still apply, at least to an extent. This means that dragons will not be able to hover as if they were helicopters, or if they did, only for a short time. Dragon will not be able to hover for minutes at the end, observing the battlefield: in fact, the necessity of exertion required to get a very heavy body into the air would mean that dragon would glide more than anything else. Hovering, plummeting, twisting a turning would be out of the question, though hovering on an updraft may be possible.
Even within these constraints there may be differences. Long and pointed wings are good for speed, while long and broad wings are good for endurance. But neither of these are good for agility: a flying animal (bird, dragon, whatever) that is optimized for agility will be low endurance, and though potentially fast, will not be able to maintain that speed for long.
Size of a dragon means that it will be heard possibly even before it is seen, as each wingstroke will produce a loud thud. Thus it might be possible to differentiate different sizes and even breeds of dragons merely by their wingbeat (good examples of this in practice appear in the Temeraire novels by Naomi Novik). Still, above points can be applied: a dragon optimized for reconnaissance will have long and broad wings with short tail, due to need for endurance. Different breeds of dragons will utilize different tactics, and may – if they serve in the same military – be given different roles.
Dragons will have developed appearance which would reflect their hunting strategy. A dragon that uses acid or fire as primary form of attack would likely be lightly built, and either swift or agile. A dragon that does not have such forms of attack, or does not rely on them as the primary weapon, would likely have powerful jaws and leck, and potentially legs, for grabbing the enemy. This would potentially be done in a manner of a bird of prey (specifically Falcon), with dragon using legs to imobilize the victim before delivering a killing bite. But even there, an emphasis on one or the other would be noticeable, and would play in dragon’s aerial tactics.
Dragons however would be vulnerable to diseases – anything affecting heart and/or respiratory organs may render a dragon combat-ineffective. This means that they would have to be accompanied by surgeons (meaning doctors), as well as large quantities of medicine. Doubly so if they are brought to unfamiliar climates, as they would easily catch diseases.
Dragons would also have to learn actually flying, much like birds do – newly hatched birds have to be taught how to fly by their parents, much like human babies need to be taught how to walk. This also means that they would develop different flying and aerial combat styles and tactics, and this has to be taken into account. Unless dragons were members of a unit from a very young age and thus taught together, unlearning their habits might be impossible, forcing the commander to improvise and adapt.