New Acquisitions: How Fast Do Armies Move?

Hey guys, this is just a really quick post to address a question I’ve seen lurking around thinking about my previous post on the logistics of the loot-train battle. No pictures today, just some quick text on the topic. A lot of readers were throwing up question-marks on how I was figuring marching speeds. I realized that I was taking for granted what would be understood in my field – the standard rule-of-thumb marching speeds.

A lot of readers questioned two things: first that Cersei reports riding from somewhere (King’s Landing?) to Winterfell in “a month” being discordant with the long march time from Highgarden to King’s Landing and second that 12-miles-a-day was much too slow. Both of these get into “how do we calculate how fast an army can move?” Let’s start with infantry on the march.

Infantry on the March

The intuitive and a touch clever method is to take normal human walking speed – around 3mph – multiply it by walking hours per day (maybe 8) and go with that. This makes intuitive sense, but if large army logistics made intuitive sense, they wouldn’t be hard, and as Clausewitz says (drink!) “War is very simple, but the simplest thing is very hard.” Logistics is very hard. So why don’t armies move at c. 25 miles per day?

So let’s think about – in very general terms – what needs to happen and in what order for a large body of infantry to march. Everyone wakes up and starts to get moving (probably around 5am). Breakfast need to happen, which may require making fires. Tents need to be struck and stowed along with all of the gear in the baggage train and individual soldiers need to stow their own equipment. All sorts of small tasks add up to eat away parts of the morning. Then everyone needs to get gathered and ready to march.

And now – because you are a large body of infantry, you wait. Let me explain – let’s take a nominally full strength (c. 3,000 men) American civil war brigade, marching on a road 13ft or so wide. You can get five men (a little cramped) into a single row on that road, meaning that the infantry itself stretches 600 ranks deep. Unlike in the movies (which love ultra-compact marching formations because it looks cool) you need a few feet of seperation between rows for best effect (WW2 US Army guidelines specified 2-5 yards), let’s assume each soldier occupies about 5 feet in the marching order. So the infantry is 3,000ft long (914m; nine football fields). We also need unit separation (between the regiments, it’s important to avoid ‘accordioning’ on the march and facilitate control; WW2 army regs suggested 100 yards between companies, 50 yards between platoons, so these could be quite large), so let’s round up to 4,000 ft (1219m; 13 football fields).

But we also have tents, food supplies, spare ammunition and all sorts of other of what the Romans would have called impedimenta (sidenote: if you are thinking, “well, but a pre-gunpowder army doesn’t need this; 1) arrows take up space and 2) camp entrenching supplies – the Romans marched heavy). How many wagons, pack animals or porters you need varies – the Romans seem to have often moved with a mule for every six-to-eight men, plus the army’s siege train. A good rule-of-thumb I’ve seen for American Civil War estimates is around 20 wagons per thousand, so 60 wagons. Rule of thumb in the ACW is 80 wagons to a mile of road, so our wagon train ought to take up around another 4,000ft.

(Sidenote: you can see why logistics gets complicated fast. Even in explaining a rule of thumb, I have to resort to rules of thumb, or else we have to inventory all of the stuff an infantry brigade needs, and all of the food they need and then parcel it out by wagons (and then figure for the mule-or-horse teams for the wagons) and on and on. Fortunately for the historians, this sort of work was done by the armies at the time and written down, so we tend to use their staff-work.)

So the entire force is probably a bit more than 8,000 ft long – 1.5 miles. In practice, there’s actually a lot more space eaten up in separation (between wagons, between men) so it would be longer, but I don’t want to get lost in the details. And this is for just 3,000 infantry – we have no cavalry with their many spare horses (three per man as a typical minimum) or god forbid a siege train which might involve hundreds of wagons.

Here’s the thing: the last man in the line cannot start marching until that entire formation has marched past him. The more men you are sending down the same road, the worse the problem gets. For our single brigade, the wait is probably close to an hour. For an army of – say – 10,000 men, you’re now talking about the last man waiting in the camp while some five miles of army marches past him (that’s about 2-3 hours of waiting). Moreover, the front of the army has to stop where the last man will stop for the night (since the army camps together, because to do otherwise is dangerous). And remember, that last man may have started marching hours after the first man, putting him miles back down the road. So another two or three hours where now the front of the army isn’t moving even though the back is. And worse yet – the entire army is limited to the speed of its slowest element – either waiting for it to move forward in the morning or waiting for it to catch up in the evening.

Well, getting started ate quite a few hours, but at least we’re going to move at a constant speed all day right? Of course not. These are humans – they need to eat (lunch), drink and relieve themselves. Men will fall out of line because they are sick or because they sprained an ankle or because they’re tired of marching and faking it (many army guidelines put the medics at the back of the marching column for this purpose). To add to this, wagons get stuck in the mud, mules and horses get stubborn or lame (that chance may seem low, but remember we’re dealing with thousands of animals – small percentages add up fast when you have a few thousand of something).

For reference on how much time this can eat up, 1950s US Army marching regulations (this is again FM21-18 “Foot Marches”) suggest that “battle groups or smaller” (800 men or less, generally – so small, fast-moving infantry) can “under favorable conditions” (read: good, modern paved roads in good weather) make 15-20 miles in a continuous eight hour march. A forced march – marching longer than 8 hours and at a higher than normal pace – can cover more ground (c. 35 miles in a day in some cases) but such a pace will wear out an infantry force fast.

At the end of the day, the army needs to arrive at its planned camp site long enough to make camp. Cooking needs to be done. Food that was foraged by flanking units needs to get to the camp, be recorded and stored (or processed and eaten) – speaking of which, note that we haven’t even discussed flankers, scouts and foraging parties. Wages may need to be paid, paperwork needs to be done. In many armies, the camp will need to be fortified – the Romans built a wood-palisade fortified camp every night on the march. And then everyone goes to sleep around 9pm. And that, to be clear, is when everything works like clockwork – which it never does.

For a large army, the breaking camp, waiting to begin marching, waiting for the last man to arrive, dealing with pack animals and wagons slices a few hours off of that eight hour march routine. All of which is why a normal large body of infantry moves something closer 8-12 miles per day than the 24 miles (8 hours x 3.1mph) per day implied by Wikipedia’s Average Human Walking Speed.

Historians doing studies of campaigns thus tend to use these sorts of rule-of-thumb speeds without much feeling the need to explain why armies move so slow because I think they expect that most of their readers are either fellow historians or former soldiers and in either case, already know. These rules of thumb, in turn, derive from staff planning in the age when armies still mostly walked to war (especially the 1800s and early 1900s): those staff office planners would have (and presumably still do have) elaborate tables of how many men can move how fast over what sort of roads in what kind of weather – because bad staff work multiplied over massive armies can mean catastrophic logistics and timing failures (see: Frontiers, Battle of the (1914) for examples).

If anything, for a medieval army of conscripts, fresh from a successful battle, with a long supply-train moving off of the main roads, 12 miles per day is actually quite fast. Large armies with lots of wagons often strayed into single-digit marching speeds. And, to be clear, marching speeds are highly variable based on terrain and the rest. For Jamie straying off the Roseroad, terrain would be deeply unfavorable for his wagons, slowing him down further.

Getting to Winterfell on Horseback

Ok, and what about Robert’s royal progression making it to Winterfell in “about a month?”

According to the standard reckonings, Winterfell is 1,600 miles from King’s Landing (this is based on using the official maps and creating a scale from the known length (100 leagues = 300 miles) of the Wall), but we actually don’t have any dialogue suggesting that the Royal Party went directly from King’s Landing to Winterfell – it’s possible the ride was shorter (sidenote: these kinds of travel distances are part of why the sort of kingdom Robert has – where rule requires personal relationships like this – doesn’t get this big historically).

Robert’s party may be large by our standards – a few hundred individuals, all mounted – but by army standards, it’s a tiny force. And small groups of cavalry, if they ride hard and bring spare horses (and they do) can move very fast. As I’ve noted before, the Mongols manage strategic movements over long periods up to 60 miles per day, but then, they are the Mongols. And the books note specifically that the pace was unlikely to be too rushed, since Robert had brought his entire family.

There are a lot of things we could do to make this work. If we had Robert ride quite hard, he can make it in maybe 35 or 40 days, which we might say is “about a month.” We might also have Robert start from somewhere further North (perhaps Darry or the Eyrie or even the Twins), since the sort of personal kingship that Martin seems to have envisaged (but not always planned for) historically required very frequent travel.

But I think all of that is over-complicating the obvious, which is that – taking much of the worldbuilding of Westeros together – it is clear that while Martin is great at many things, managing scale in his world is not one of them. I realize that one of the claims that ASoIaF fans like to make for their series is that it is ‘more grounded’ and ‘more real’ than other medieval fantasy (Martin occasionally makes this claim too), but speaking as a specialist in pre-modern armies and economies (particularly Roman ones), it isn’t. In practice, then, I think the mostly likely thing is: the story-teller messed up and misjudged the time it would take to cover the distance.

(A word: if anything, Tolkien’s first-person experience of war and his deep grounding in medieval and Anglo-Saxon literature means that – while Tolkien does not stop to discuss Aragorn’s tax policy – Middle Earth tends to fit within the ‘zone of the possible’ much more neatly than Westeros. Armies in Middle Earth move at normal speeds, the economic and population systems (and the human terrain with them) make sense, characters behave within the confines of religious belief and social custom, and so on. Martin is a fantastic story-teller, but I think his claims to a ‘realer’ medieval fantasy are hollow; but then, he’s telling such a good story – should he ever finish it – that I don’t think he needs to make that claim.)

As for the show, Benioff and Weiss are even worse at managing distance and time and understand how the various parts of a pre-modern society work even less, which means they tend to take Martin’s minor miscalculations (a two-month trip becoming a one-month trip) and magnify them (armies instantly teleport over continental distances).

Maybe somewhere down the line, when I have a bit more time, we can get a bit more into the weeds on how fast armies move and the complications of logistics that entails. For those who don’t want to wait and instead want some book recommendations, some good places to start for thinking about these problems are:

  • J. Landers, The Field and the Forge: Population, Production and Power in the Pre-industrial West (2005)
  • J. P. Roth, The Logistics of the Roman Army at War (1998)
  • P. Erdkamp, Hunger and the Sword: Warfare and Food Supply in Roman Republican Wars (1998)
  • D. W. Engels, Alexander the Great and the Logistics of the Macedonian Army (1980)
  • G. Parker, The Army of Flanders and the Spanish Road (2004, 2nd ed.)

I especially recommend Landers, which offers a truly excellent overview of nearly all of the sticky, messy problems of creating, mobilizing, moving and paying armies in the pre-modern world (he stretches from Rome to the early modern in an effort to show the (very real) continuities in warfare). Unfortunately, the book is pretty expensive (as is normally the case with these sorts of studies) so maybe check your local library for a copy.

23 thoughts on “New Acquisitions: How Fast Do Armies Move?

  1. The really funny thing about the GoT *show* getting logistics so badly wrong is that a significant portion of location shooting *is* logistics, to the point that there are always at least 3 departments dedicated solely to sorting out logistics (Catering/Craft Services, Locations, and Transport). If they’d ever paid attention on the rare occasions when they were on set, they might have clued in to what a behemoth anything on the move really is.

    1. Even casual viewers noticed that Daenerys could not rescue the wright hunting party in time unless her dragon was able of flying at an absurd speed. The loot train battle is far from the worst that the show has to offer in that regard.

      The show’s producers are probably quite aware of these problems, but prioritized getting done with the plot. For the later seasons, characters were separated and reunited a lot.

  2. I really appreciate this blog, you’re doing an excellent job. Was introduced with your Minas Tirith series. With regards to Tolkien and Logistics, I’m curious on your thoughts with regard to the Rohirrim. Would a post on their on-page conflicts relate somewhat to the awaited post on the Mongol Hordes, or are they in their own category?

    1. Glad you’re enjoying the content!
      I plan to do a full study of Helm’s Deep the same way I need the Siege of Minas Tirith, when I have time (late-fall/early-winter is a busy season for me).

      That said – and you can actually see back in the comments for the Siege of Gondor Posts – I don’t think that the Rohirrim are much like steppe nomads. The thing about steppe nomads is that the whole package – subsistence patterns, the mobility, the style of warfare, the very ecology of warfare (that is, the horses) – are all tied together. You cannot have one without the other. The Rohirrim have stables with stable fed horses which *by itself* largely walls off a logistic system that looks like the Mongols (whose horses are smaller, weaker and able to completely feed themselves by grazing). Effects like that ripple through the entire system of fighting at all levels (i.e. tactics, operations, strategy). You can’t ‘half-Mongol it’ something that the Yuan dynasty and the Ilkhanate both seem to have discovered somewhat quickly.

      So the Rohirrim are, I’d argue, best understood as an agrarian cavalry force, which has a logistics situation much more like our loot train army and much less like our Dothraki, although with some oddities that we’ll discuss when we get there.

  3. I started with your posts on war elephants but am now perusing everything! I plan on looking the Landers book, so thank you for the recommendation.

    Is there a book or two you can recommend on military logistics in pre-modern China? This is pretty specific, but I am especially interested in the differences between the logistics of supplying an army in China and Europe. Do rice cultivation or the generally difficult terrain of western China have substantial, general effects?

    1. I have to admit, this is pretty well outside my specialty, but it is something I want to do. But I’ll need to do some research to be able to talk about it, and currently time is short, so while I want to do this, it may be a while.

  4. One interesting consequnce is that the the optimum size of unit to move and fight seems to be close to 10,000, a size fairly constant over a very long period (Roman legion plus auxiliaries, Mongol urdu, Napoleonic or modern division).

    On marching, the fastest sustained march pace I know of is Nicholson’s move from Peshawar to Delhi in the Mutiny. His Pathan infantry covered over 1200 kms in just under 60 days in the Indian summer, fighting two battles along the way.

    1. And adding – one sub-force, the Frontier Guides, covered 580 miles in 22 days (Hoti Mardan to Delhi) in the month of Ramadan (they were Muslim) and the height of summer. With arms carried and rations and ammunition in packs. Mostly hill-tribes men from the North-West Frontier.

  5. Found you via Brad Delong.
    the old says goes: Amatures talk tactics and Experts talk logistics.
    ACW saying: Get there first with the most.
    I have read a lot of women’s diary’s of the American westward migrations. Family history and all that.
    On good roads 8 to 12 miles was done. They also did not go single file if possible – dust..

  6. Hi Bret, I came to ask a question. I’m curious what the roads might have looked like at different ages when they weren’t full of soldiers. Did people have to wait all day for the armies to get out of the way so they could continue their transport of goods? Just like movies make the mistake of not showing baggage trains, do they also make the mistake of not showing commuters, herders, convoys, caravans, etc? I realize the answer would be different in different times and places!

  7. Very interesting read. One minor quibble: American Civil War infantry marched in columns of four regardless of how wide the road would be due to their fighting formation of two ranks deep. This probably also applies to any other army that fought in two ranks. Road distance would be 250 yards for 1000 men at regulation spacing of one yard per man, but most likely would be two to three times that due to lagging, etc. Paddy Griffeth’s “Battle in the Civil War Generalship and Tactics in America 1861-1865” has an excellent chapter on moving an army. There is a website – marching pace of the British Army has a page covering movement rates for WW1 formations. The original Volley and Bayonet ruleset had a section on the length of road columns for various troop types. Again, a very entertaining rad.

    1. Interestingly, when an early Simulations Publications Inc board wargame tried to make the point about road space – Lost Battles: Operational Combat in Russia (1971) – it was a flop partly because of being too much work to handle.

      Something like divisional artillery required four clear hexes ahead and behind them if they wanted to move by road = traffic jams were inevitable unless you planned very carefully.

  8. I just ran into your blog from Lawyers, Guns, and Money’s Friday Natsec Roundup (they had a link to the first post of the Fremen Mirage).

    Since yesterday afternoon, I have read several tens of thousands of words – thank you for having such an interesting blog!

    With this post, I wonder – how fast can units travel if there are some boats in play (either by river or sea)? This must have varied significantly with changes to naval technology over time.

    Also – to break into science fiction for a bit – even if you do not have to worry about travel time, it should not be negligible. Further, I would expect that one still cannot escape the operation trap of having to coordinate ones forces (one of the major issues for Rogue One for me was having support appear before the operation was done – it felt like it should have been too late)!

    Again – thanks for the great writing!

  9. D. W. Engels s good, but he really needs to put out an updated version on his widely cited work. He has some really bad mistakes in his citations and uses outdated data on classical horse tack and wagon capacity. Most amusing he seems to think horses don’t/will not graze at night.

  10. Bret,
    Just one, lone proofreading correction for this short post:
    Breakfast need to happen -> Breakfast needs to happen

  11. Interesting and well done, except for one glaring mistake. Your Civil War marching example makes the road wide enough for marching “5 in line,” which is nice for math but not something any unit of the day would EVER do. Infantry was structured to march by 2s or– most of the time– by 4s. A ‘column of 4s’ was the standard marching order and the formation from which everything else they do to change formation comes from. So your march column is even longer than your example states.

  12. “Unlike in the movies (which love ultra-compact marching formations because it looks cool) you need a few feet of separation between rows for best effect”

    Amusing that Laurel & Hardy’s Bonny Scotland is more realistic about this than the works concentrated on in this blog!

Leave a Reply