#army size

Robb Stark and the Northern host

The numbers given for Robb’s army in the ASOIAF wiki are actually what started me on the process of providing my own estimates for the various armies, as I think it makes a few significant methodological errors in the process, which has a knock on effect for how Robb’s decisions need to be interpreted. Primarily, I disagree with the numbers given for Robb at Moat Cailin and the breakdown of infantry vs cavalry.

Robb's initial Northern force consists of 19,500 men (AGOT, Catelyn VII), with slightly more than 1/4 being heavy cavalry (AGOT, Bran VI gives the figures of 3,300-3,400 heavy cavalry out of an initial muster of 12,000). This would suggest ~14,500 infantry and ~5,000 heavy cavalry in his Northern host.

Starting with with Bran VI, I have a slightly different interpretation of Maester Luwin’s breakdown of the heavy cavalry. His explanation of the numbers to Bran is that there are “Three hundred, perhaps four [hundred]” knights “among three thousand armored lances who are not knights.” The original calculator has taken this to mean that there are 300-400 knights in addition to the other 3000 heavy cavalry, and this is a perfectly valid interpretation of the sentence. However, I believe that Luwin’s use of “among” signals that the knights should be taken as part of the 3000, not as an addition to them for a couple of reasons.

Firstly, GRRM uses “among” in an inclusive sense elsewhere in AGOT. When Tyrion thinks on the Field of Fire, for instance, King Mern is “among” the 4000 men burned, not the 4001st man to be burnt, and Ned Stark thinks on the fact that he’d never spoken to Gregor during Greyjoy’s Rebellion because he was just “one knight among thousands”. So, while Luwin’s phrasing can be interpreted as meaning that the knights are there in addition to the other heavy horse, I think GRRM’s general use of “among” as an inclusive term means that they are intended to be included among the 3000 heavy horse, not separated from them.

Secondly, while the proportion of heavy horse gathered at Winterfell would be quite large even with the knights as part of the 3000, the general trend of what individual contingents we see is of very low proportions of heavy horse; just 13% of the Karstark forces is heavy horse (300 out of 2300), Roose has 500 mounted men (14%) out of 3500 men just before the Red Wedding (ASOS, Catelyn VI) and, in spite of having the highest number of landed knights, the Manderlys only send 240 horse (16% of 1500 men) to join Robb at Moat Cailin (AGOT, Catelyn VII). This general trend of low numbers of cavalry suggests to me that, when the numbers of cavalry can be taken in two ways, the smallest option should be taken.

This leads to the second major disagreement I have with the wiki’s army size estimates: the assumption that the 1500 Manderly men in Catelyn VII are in addition to the 18 000 men that Robb already has, not the contingent that brings Robb’s forces up to 18 000. This is a number that I’m more confident of; Robb’s statement that his "whole army” is 18 000 men strong comes well after the Manderlys have arrived, and the use of ravens for communications likely means that Robb knew how large the Manderly force was going to be shortly after they left White Harbour.

The reduction of Robb’s army to 18 000 total men and the reassessment of typical contingents from slightly over 25% cavalry to 13-16% cavalry also requires a lowering of the total cavalry from ~5000 at Moat Cailin to a maximum of 4000 cavalry. The number of foot, however, doesn’t need to be significantly reduced.

There are some references to Robb having 20 000 men (AGOT, Catelyn IX; ACOK, Catelyn II), which would better fit the wiki’s figure of 19 500 men. The problem is that we also hear of there being six thousand men with Robb at the Whispering Woods (AGOT, Catelyn X), which is far too few if we go with the wiki’s number of 5000 Northern cavalry for Robb. Between the Freys, the Mallisters (see below) and the other scattered Riverland remnants who join Robb, he’d have near enough to 7000 cavalry by the time of the Whispering Woods, and he wouldn’t be outnumbered by Jaime by 3:1, but by 2:1 (AGOT, Catelyn X).

My way of resolving this conflict is to suggest that both instances of people saying Robb having 20 000 men were exaggerations, not accurate accountings of size. In the first instance, Catelyn is angry with Lord Frey and is threatening him with the size of Robb’s army. In this scenario, claiming that there are “twenty thousand men” outside of Lord Walder’s walls is easier to say than “eighteen thousand men”, being a rounder number, and is also more threatening because it increases the size of Robb’s army. In the second instance, Renly is summarizing reports he has received of Robb’s strength, and there is plenty of room for errors to creep into these reports and for Renly to be generous in his estimates of Robb’s strength, as it only emphasizes the disparity in strength between Robb and Renly if Robb has fewer troops than Renly allows for (which he does).

On the other hand, Robb’s statement that he has 18 000 men occurs long enough after the Manderlys have arrived for him to have factored their force into his calculations and Catelyn is present with Robb at the Whispering Woods and should have a good idea of his strength. This is supported by Galbart Glover’s estimate of Jaime’s 14-15 000 men outnumbering Robb’s force by three to one. If Lord Glover is assuming that Jaime has 15 000 men, and Robb 6000 men, then Jaime outnumbers Robb by 2.5:1 which, rounded up, would be three to one. Additionally, Catelyn may have been overstating Robb’s numbers in her mind for exaggeration and contrast between his bodyguard of 30 and his army of 6000, neither of which she is certain can protect him from harm. Robb may well have had 5500-5750 men rather than 6000, which pushes the degree to which his is outnumbered by closer to three to one.

I think this proposal reconciles the contradictions in the text and offers support for my “low count” of 18 000 men. After all, if Galbart Glover is rounding up to 3 for effect, then why can’t Catelyn or Renly do the same?

Tywin’s army at the Green Fork

The size of Tywin’s army at the battle on the Green Fork is one of the few armies that we have a mostly reliable estimate for. Chella estimates it at twenty thousand men, “by their fires” and, based on the area the camp occupies, Tyrion thinks that she can’t be far wrong (AGOT, Tyrion VII). This is further reinforced in the next of Tyrion’s chapters where, after having settled in, he calls it a ”vast host twenty thousand strong”.

The one minor caveat here is that we don’t know how many men were with Addam Marbrand scouting Robb’s approach. The Neck is something in the order of 370 miles away from the Ruby Ford and the crossroads, which is 15 days of hard riding and 21-22 days of fast, horse preserving, riding. Marbrand is definitely operating quite near the Neck, as a number of his outriders are killed immediately after Robb debouches into the Riverlands, so there was an expectation that he would be facing down Robb’s army at some point and that means a sizeable force for protection during the retreat downriver to Tywin. There’s also evidence of skirmishing between Marbrand and Walder Frey’s forces, which relates to another point: the scouting force needed to be large enough to give Walder reason to pause and not immediately attack if he saw an opportunity.

Given the speed of communications between Marbrand and Tywin and the distance between the two forces, it’s likely that at least one castle with a resident Maester had been taken by Marbrand near Frey lands in order to reduce the delay in communication. While we haven’t heard of a such a castle being taken, it’s the only way that the message for Marbrand to return to Tywin, burning and harassing the Northern army, so soon after it entered the Riverlands.

Medieval scouting parties could range between dozens of men all the way up to several hundred men, largely depending on how far they were operating from their own army and how much opposition they expected to face. One of the largest forces was William Felton’s scouting/deeds of derring-do party prior to the Battle of Nájera, which had somewhere between four and five hundred men, approximately 200 of whom were men-at-arms and the remainder were mounted archers. This is probably the minimum size of Marbrand’s scouting force, although it would be mostly cavalry as mounted archers barely exist in ASOIAF, and I suspect that six or seven hundred men would be more likely based on the need to capture a castle for communications and then skirmish during the retreat.

Combined with Tyrion’s force, Marbrand’s scouting party might well raise Tywin’s army to 21 000 men rather than 20 000, especially if Tyrion and Chella were being conservative in their estimates of the army size and it was more than 20 000 strong.

This caveat aside, the important thing about Tywin’s army is that, unlike most armies in ASOIAF, we can create a fairly accurate breakdown of the army’s composition thanks to the detail GRRM goes into in describing the Battle on the Green Fork. He lists a total of 6800 cavalry (4000 on the right wing, 300 in the center and 2500 in reserve) and 2500 infantry (in reserve). From these figures, the mention of what the other troop types are and the overall size of the army, we can reconstruct the rest of the army.

Starting with the conventional 20 000 number to be conservative, we can remove 9300 men straight away, leaving us with 10 700 men. Of these, perhaps a thousand form the “swarming mass” that makes up the vanguard, as it’s likely that the “thousand other voices” screaming back at Gregor excludes Tyrion’s clansmen - any fewer and it’s hard to describe it as a “swarming mass” of low quality cavalry. This leaves us with 9 700 men to account for.

The center of Twyin’s army is made up of “squares” of pikemen flanked on either side by “three long lines” of archers, with “rank on rank” of men-at-arms behind the archers and a reserve of 300 knights. This is where we fully enter the realms of speculation.

To begin with, we don’t know what exactly is meant by “squares” of pikemen. Medieval pike columns could contain as many as 6000 men, while early modern squares might have as few as 100, so any size estimate is going to involve a lot of guesswork. I’ve gone back and forward about how to interpret this phrase and have finally decided to treat the “squares” as distinct units for the purposes of command, but tactically in a similar manner to 15th century dismounted men-at-arms. As a result, I’m estimating their strength as 250 men - to be deployed 10 deep - and their total number to be 3000 men (12 pike squares). This allows for a reasonable width of line in the center (300 yards), but also adequate depth to resist cavalry charges.

Estimates for the archers and men-at-arms follow directly on from this. The archers need to be in thousands because, as Philippe de Commynes points out, they’re only useful when employed in such numbers. Given that fully half of Jaime’s infantry are archers, I’m inclined to assign 4000 archers and 2700 men-at-arms to Tywin so that his overall archer to footsoldier ratio is about 1:2. The archers, each occupying a space of three feet, would occupy 667 yards on either side of the pikes and the men-at-arms, each occupying four and a half foot, would have the same frontage, but double the depth (6 ranks).

In all, I believe a conservative estimate of Tywin’s numbers, not counting Tyrion’s Clansmen, to be 7800 cavalry, 3000 pikemen, 2700 men-at-arms, 4000 archers and 2500 other infantry of an indeterminate type.

The Mallister forces

The number of men Jason Mallister can field, in spite of his importance during Robb’s initial campaign and later at the Battle of the Fords, has yet to be mentioned. While I’ve seen a number of estimates, with numbers as low as 1500-2000 men, the best estimate I’ve seen to date has been @racefortheironthrone‘s estimate in his Chapter-by-Chapter analysis of A Game of Thrones. 

However, some interesting information in Fire and Blood has caused me to rethink the relative military power of House Mallister, and I believe I’ve also found a semi-reliable method of estimating their forces. The particular passage can be found in the chapter titled “Aftermath - The Hour of the Wolf” and runs thus:

House Tully was unique amongst the great houses of Westeros. Aegon the Conqueror had made them the Lords Paramount of the Trident, yet in many ways they continued to be overshadowed by many of their own bannermen. The Brackens, the Blackwoods, and the Vances all ruled wider domains and could field much larger armies, as could the upstart Freys of the Twins. The Mallisters of Seagard had a prouder lineage, the Mootons of Maidenpool were far wealthier, and Harrenhal, even cursed and blasted and in ruins, remained a more formidable castle than Riverrun, and ten times the size besides.

The main point of interest is that the four military powerhouses of the Riverlands are the Brackens, the Blackwoods, the Vances and the Freys, while the Mallisters are excluded from this list. Instead, they are singled out as having a better pedigree than the Tullys which, although not excluding them from having marginally larger armies than their overlords, does quite firmly give them fewer men than House Frey.

In re-examining the question of Mallister manpower, I was inspired by @racefortheironthrone‘s comment that Jason Mallister must have had substantial forces, as his men covered four fords by themselves. Since we know roughly how many infantry Edmure was working with (8000, ACOK Catelyn V) and the minimum number of fords (12, ACOK Catelyn VI), we can get a reasonable maximum number of Mallister strength by assuming that all fords had an equal number of defenders and that there were only twelve being defended. This provides a maximum number of 2664 infantry and, based on typical Riverland horse to foot ratios, about 890 horse, for a total of 3554 men.

These are, however, maximum numbers. While the ford nearest to Riverrun is very broad and at least 50 yards wide (ACOK, Catelyn VI), it’s likely that the other fords were much narrower and more easily defended. It’s also probable that there were more than a dozen fords, and that Tywin was simply trying to break through a broad stretch of the river rather than overwhelm every ford. 

Based on the number of men stationed at the Blanchetaque by the French during the Crecy campaign (3500) and probable English numbers by that point (12-13 000), I would say that medieval commanders expected that someone attacking a ford needed an advantage of at least four to one to succeed. As Tywin can’t have had above 16 000 men at this point in the campaign and, being a cautious man who preferred to significantly outnumber his enemies, likely aimed to have an advantage of 5:1 over the defenders, the minimum number of men for House Mallister can be calculated as 1070 infantry and 357 cavalry, or 1427 men total.

This means there’s a difference of 2100 men between the maximum and minimum estimates, but I think we can discard anything close to the minimum number. 

Firstly, the Mallisters are clearly a powerful house, even if they aren’t one of the four military powers of the Riverlands. There’s no reason to suspect that they are significantly less powerful than House Tully, which must be capable of raising at least a couple of thousand men, and they may still even exceed the Tullys in military might given House Tully’s typical weakness compared to their lords.

Secondly, the Mallisters saw off what, if not an attack by the main Ironborn fleet, was at least a major assault by the Ironborn during Greyjoy’s Rebellion and were placed in charge of what was probably one of the largest fords across the Red Fork. You don’t fend off even part of the Iron Fleet or get selected to defend a ford at least fifty yards across unless you can bring a lot of men to bear in the defense of these fords.