Size of Armies Part 2

Advice for Fantasy Writers

This article is by Toni Šušnjar.
Note: This is Part Two of a two-part essay. Here is the link to Part One.

Feudal Army

Feudal army is a complex beast: it may include landed troops, mercenaries, town militias, conscripted peasants, or any combination of these. Thus estimating size of a feudal army will require estimating size of each element of the listed, and the size of field army will never be an exact proportion of the total military force available – or often even able to be estimated.

Landed troops are perhaps easiest to estimate – but even that is not simple. In 15th century Germany, it was generally held that 20 hufen (plots) were required to support a man-at-arms (armored cavalryman) while 10 hufen were required for a lightly armored cavalryman (typically a mounted crossbowman). Hufe was a single peasant’s plot or 121 406 m2. In theory, a state with million men and 85% rural population should have been able to raise 12 750 cavalry, as per the calculation below:

  • 1 000 000 men * 0,85 = 850 000 men = 170 000 plots;
  • lance (gleve) = 1 man at arms + 2 light cavalrymen (note that historically size of lance varied – this is just an example) = 40 plots
  • 170 000 plots = 4 250 lances = 4 250 men-at-arms + 8 500 light cavalrymen

This gives a total of 1,275% of population. Accounting for inefficiencies in the system can thus lead to an easy “rule of thumb” of 1% of population for a feudal army.

But cavalry is merely one portion of a feudal army, and even then, above numbers were not a hard-and-fast rule. In 15th century Hungary during the Ottoman wars, mounted archers could be raised at rate of one per every 20 or 10 peasant plots, depending on the threat level. Foot gunners could be mustered at one per 20 plots. In 1397, Holy Roman Emperor (and Hungarian king) Sigismund had required raising one mounted archer per 20 tenant plots, but in 1435 this was shifted to three mounted archers per 100 tenant plots. In 1454 decision was made to raise four mounted archers and two footmen per 100 tenant plots. In 1464, king required nobility to equip men-at-arms at ratio of one per 12 tenant plots – but Slavonia was expected to contribute at ratio of one per 20 plots (likely due to devastation of the Ottoman raids). In 1513, decision was made to raise one armored horseman per 36 tenant plots. Diet in Tolna in 1518 required mustering one foot gunner per every 20 tenant plots. In 1522, one light horseman or arquebusier was required per every 10 plots.

Overall, assuming that each plot had 5 people, proportion of troops could vary from 0,6% to 2% of the population. This however was often only the affected populace, as in feudal system nobility would often refuse to participate in the war unless they themselves were threatened.

In addition to these troops, feudal forces also consisted of retainers, mercenaries and others whose numbers are basically impossible to estimate for certain. Cities were significant source of infantry: in 1224, Hungarian King Andrew II required city of Klokoč to supply 15 armored cavalrymen (loricatos) and 200 infantrymen (pedites). By contrast, Venice, having conquered Croatian coastal cities by force, did not allow locals to enter city guard and instead relied on foreign mercenaries; locals were employed solely on an ad-hoc basis. Venetian crew of 87 men in Šibenik consisted nearly entirely of foreign mercenary crossbowmen.

Tribal army

In a tribal army, basically all men of the tribe are warriors. This gives a general rule of thumb of 20% of the population as warriors in barbarian societies. On the flip side however the warriors will be disorganized and still not very numerous due to small size of such societies. Thus a barbarian tribe of 100 000 men may have 20 000 warriors, but tribes larger than 100 000 would be an exceedingly rare sight.

Size of Field Army

Size of the field army should not be confused to that of total military. While military size indicates available manpower under arms, size of field army is limited by ability of the army to supply itself in the field. Due to limitations of most premodern armies, this limit was dictated by foraging: in most conditions an army larger than 20 000 men simply cannot be maintained for any length of time from foraging alone. Deployment of armies much larger than this required building up supply depots in advance, which was an enormous logistical undertaking.

However, the “in most conditions” above is the key. Number of troops could be significantly increased in good conditions and densely populated area. During the American Civil War, some 30 000 men could have foraged successfully while making forward progress. When accounting for supply wagons and other support elements, this will have meant 35 000 men. Even there however armies on march tended to be split in columns of 10 000 – 20 000 men.

As usual, Clausewitz is a good fallback:

If we bear in mind that in a community consisting even as it does in great towns, of consumers only, there must always be provisions enough to last for several days, we may easily see that the most densely populated place can furnish food and quarters for a day for about as many troops as there are inhabitants, and for a less number of troops for several days without the necessity of any particular previous preparation. In towns of considerable size this gives a very satisfactory result, because it enables us to subsist a large force at one point. But in smaller towns, or even in villages, the supply would be far from sufficient; for a population of 3,000 or 4,000 in a square mile which would be large in such a space, would only suffice to feed 3,000 or 4,000 soldiers, and if the whole mass of troops is great they would have to be spread over such an extent of country at this rate as would hardly be consistent with other essential points. But in level countries, and even in small towns, the quantity of those kinds of provisions which are essential in war is generally much greater; the supply of bread which a peasant has is generally adequate to the consumption of his family for several, perhaps from eight to fourteen days; meat can be obtained daily, vegetable productions are generally forthcoming in sufficient quantity to last till the following crop. Therefore in quarters which have never been occupied there is no difficulty in subsisting troops three or four times the number of the inhabitants for several days, which again is a very satisfactory result. According to this, where the population is about 2,000 or 3,000 per square mile, and if no large town is included, a column of 30,000 would require about four square miles, which would be a length of side of two miles. Therefore for an army of 90,000, which we may reckon at about 75,000 combatants, if marching in three columns contiguous to each other, we should require to take up a front six miles in breadth in case three roads could be found within that breadth.

Professional Army

Roman and Byzantine armies can, as usual, be considered to be professional armies of antiquity and medieval world. And here we see some general rules. Specifically, while there are occasional massive armies in the 80 000 – 120 000 men range, majority of field armies tend to be in the 10 000 – 25 000 range. Incidentally, two legions with support units would have 20 000 combat troops (4 x 5 000).

Roman and Byzantine writers are generally in agreement here. Vegetius believed that 24 000 men was the maximum for a large field army on campaign during the Principate. Maurice believed that 36 000 was the upper limit, and 5 000 to 15 000 was average.

Of course, reality was more varied than such idealistic prescriptions. An ad-hoc response force could number as few as several cohorts (few thousand men at most), while largest armies could go as high as 200 000.

Territorial Defense Army

Author of the “On Skirmishing” considers 3 000 cavalry to be a large force – which for a raiding party it indeed is. Opposing Arab raiding force is assumed to be 6 000 cavalry plus a body of infantry. Largest Byzantine army mentioned is also 6 000 cavalry. Considering the typical balance of infantry to cavalry, army of 6 000 cavalry may imply an overall army size of 12 000 to 24 000 men. Considering the nature of warfare in the period, smaller army size with greater proportion of cavalry is more likely.

Later, as Byzantine army professionalized and went onto the offensive, overall number of troops as well as proportion of infantry in the army increased – creating essentially a throwback to an earlier time. Both Praecepta Militaria and De Re Militari describe an army of either 12 000 or 16 000 infantry with 6 000 to 9 000 cavalry. Overall, the largest field army in either of these works is some 25 000 men. This may seem to be a small force compared to the largest of Principate or Dominate armies, but numbers are misleading. For one, Byzantine armies had more cavalry, and higher quality cavalry, than their ancient counterparts did. For another, size of the army overall had decreased. Praecepta Militaria was published in 965 AD, and at that time Byzantine army may have numbered some 150 000 men overall. Thus 25 000 men in 965 AD was proportionally the same as 100 000 men in 312 AD (~17%).

Feudal Army

When it comes to feudal armies, field army sizes could vary widely just as everything else about feudal states. How large a field army was depended entirely on how well-organized a state was. When it comes to a decently sized, well-organized medieval state, we have an example of 15th century Hungary. Taking the list from here (compiled from From Nicopolis to Mohacs and Povijest Hrvata), we see that most Hungarian armies were below 10 000 men:

  • 2 000 – 5 000: 17
  • 6 000 – 10 000: 20
  • 10 000 – 15 000: 7
  • 16 000 – 20 000: 4
  • 21 000 – 25 000: 3
  • 26 000 – 30 000: 4
  • 30 000 – 40 000: 2
  • 41 000 – 50 000: 0
  • 51 000 – 100 000: 3

As can be seen, majority of significant armies deployed by Hungary during the Ottoman wars were in the 6 000 – 10 000 range, and armies below 10 000 men account for over 60% of major field armies deployed. Armies in the 2 000 – 5 000 range may in fact be more numerous when considering that chronicles may have simply missed greater proportion of smaller engagements, not recording them because they were of lesser interest. If one includes minor raids and skirmishes with forces below 2 000 men – which were likely the vast majority of engagements during the Ottoman wars, but left comparatively little evidence – this proportion increases even further. Armies on the highest end of the scale (50 000+) are also numerically questionable, and may have been exaggerations of the contemporary chronicles. This is especially likely since the entire military force of Kingdom of Hungary was not much greater than 60 000 men. In fact, most likely number of troops available to the kingdom was 55 000 – 85 000 actual soldiers, though a total of 150 000 – 170 000 troops could have been raised through general mobilization (exercitus generalis or insurrectio generalis which originated as a general call to arms of all nobility, but by the 15th century had degenerated to including peasants and other troops of questionable quality). With Hungary having population of some 3 000 000 – 5 000 000 in the 15th century, we again see that actual troops numbered anywhere between 1,2% and 2% of the population. But such a force could never be deployed all in one place. Even at Mohacs in 1526, with fate of the kingdom hanging in the balance, Hungary was not able to deploy more than 25 000 men in a single army, and number may have been as few as 20 000. By contrast, Ottoman army facing it numbered some 65 000 soldiers.

Tribal army

While tribal armies may be able to raise significant proportion of populace to arms, they typically lack the large demographic basis as well as the organizational requirements to raise large coherent forces. Thus, the previously noted 20 000 men in overall army is about the maximum practical field army size, and even this is only true in the best organized tribes.

Mongol army, while technically a “tribal” army, does not really fit the definition and was really a professional force provided by a tribal society. It was in fact organized much like the Roman army was, with basic units being tuman (10 000), mingan (1 000), and further into 100 and 10 men. This Mongol nomadic core was further supplemented by auxilliary units from other states as the Mongol Empire grew – much like the Huns, Mongols too relied heavily on settled peoples for their infantry and siege expertise. Mongol conquest of China was in fact carried out predominantly by Chinese infantry and siege corps in Mongol service. Same can also be said for the much earlier Huns. In both cases however the tribal core always remained (relatively) small: Mongol army itself never went much beyond 100 000 men, even with the absorption of other nomadic tribes, and typical field army was 3 – 5 tumans.

Other factors to consider

Beyond the raw numbers, one needs to consider composition of the army as well. An all-cavalry force will have far greater logistical requirements per man than an all-infantry force. An infantry army may have a mule per eight soldiers (as was standard in the Roman army). By contrast, a heavy cavalryman will typically have a combat horse (charger), a riding horse, and a pack mule. A light cavalryman may have only one horse and a mule. Moreover, none of these numbers really account for the food that soldiers and horses eat – and horses eat a lot. All of this means that larger proportion of cavalry will typically result in a numerically smaller army.


Mark Whittow – The Making of Orthodox Byzantium

Warren Treadgold – Byzantium and its Army, 284 – 1081

Vjekoslav Klaić – Povijest Hrvata

Tamas Pasafalvi – From Nicopolis To Mohacs

Further Reading

About the Author

Toni Šušnjar is an amateur historian and fantasy enthusiast with interest in ancient and medieval history as well as Medieval and High Fantasy, especially warfare. He also writes War Fantasy blog.

Featured Author
Latest posts by Featured Author (see all)

Leave a Comment

Please log in to your forum account to comment