What happened when Cavalry charged Cavalry?

[eques]

As a companion to the discussion on cavalry charging infantry (and to a conversation I was having at my club) what actually happened when the cavalry wings of the 2 sides met head on?  Did they literally charge full pelt into each other?  Did the horses physically collide?  Did they play chicken with each other?  Or was it more a case of slow manouvering, with pairs of horsemen (and the wings as a whole) trying to outfox each other?

[Patrick Waterson]

Much of this would depend upon the cavalry, and their training and tactical repertoire.  Numidians would respond to a charge by about-turning and departing at speed, waiting until the chargers called a halt and then returning to pepper them with javelins.  Parthians and Iberians (as described by Tacitus) would loose arrows at each other, then charge in for a corps-a-corps struggle when they ran out of ammunition.  Crassus' Gallic cavalry at Carrhae charged straight into Parthian cataphracts and attempted to seize their spears and grapple them from their horses.

So it helps if we specify which cavalry types are engaging whom.  Classical javelin-armed cavalry would not behave in the same way as cataphract lance-armed cavalry.  Perhaps it would be worth looking up a few accounts of battles and asking what exactly goes on between the cavalry of two particular armies or nations.

Good question, though.

[eques]

Was thinking of Cannae, Trebbia, Issus, Gaugamela : regular melee cavalry on the flanks charging frontally at each other then fighting until one side won.

[Patrick Waterson]

This gives two Carthaginian vs Roman and two Macedonian (and Thessalian) vs miscellaneous Achameneid.  Extrapolating from what I know (or think I know) of the practices of the powers involved, plus (more importantly) the actual source descriptions, we get the following picture:

Cannae
On the Roman left, the Allied Italian cavalry, trained to melee face-to-face, are bewitched, bothered and bewildered by the Numidian cavalry swarming away every time the Italians look like charging and then swarming back as soon as the charge pulls up.  Appian suggests some concealed Numidians appeared on the Roman (i.e. Allied Italian) flank, which would have been a deterrent to folowing up the main body too determinedly.  Net result: stalemate, with very few people hurt.

The Roman right was very different; here Hannibal massed his Iberian and Gallic cavalry and sent them in against the Romans (and perhaps also the Extraordinarii, although the impression one gets is that these were on the left with Varro).  The Romans stood and fought, the result being practically an infantry melee on horseback, and as the Romans began losing they actually dismounted to try and give themselves extra steadiness.  Polybius III.115:

"But as soon as the Iberian and Celtic cavalry got at the Romans, the battle began in earnest, and in the true barbaric fashion: for there was none of the usual formal advance and retreat; but when they once got to close quarters, they grappled man to man, and, dismounting from their horses, fought on foot. But when the Carthaginians had got the upper hand in this encounter and killed most of their opponents on the ground,— because the Romans all maintained the fight with spirit and determination,—and began chasing the remainder along the river, slaying as they went and giving no quarter; then the legionaries took the place of the light-armed and closed with the enemy."

An interesting observation is that "there was none of the usual formal advance and retreat" (anastrophēs kai metabolēs), suggesting that anastrophe (wheeling around, of a horse) and metabole (exchanging, as by successive sub-units) were standard and expected tactics and procedures.  Also interesting is that both sides eschewed these standard tactics in favour of an all-out knock-down-drag-out fight.  The Carthaginians were trying to force a rapid conclusion with their superior numbers and the Romans had no ground to give without exposing the flanks of their army.

Having forced this conclusion, Hasdrubal took his cavalry en masse behind the rear of the Roman infantry and appeared on the flank of Varro's cavalry wing, which promptly took off at speed, harassed by the Numidians.  Hasdrubal then delivered his blow against the Roman infantry rear (which more properly pertains to another thread).

Trebia
The Trebia was more of a 'standard' cavalry action, although the Carthaginian preponderance in cavalry plus effective missile support (from slingers) rather skewed matters.  (On an elephant note, Hannibal committed his pachyderms against the Allied Italian infantry and not against the Roman and Italian cavalry.)  In this battle, the Roman (and Italian) cavalry did yield ground when overmatched, with unfortunate consequences for the poor Itlaian allies on the infantry wings.

"As soon, therefore, as their advanced guard had retired again within their lines, and the heavy-armed soldiers were engaged, the cavalry on the two wings of the Carthaginian army at once charged the enemy with all the effect of superiority in numbers, and in the condition both of men and horses secured by their freshness when they started. The Roman cavalry on the contrary retreated: and the flanks of the line being thus left unprotected, the Carthaginian spearmen (sc. peltasts) and the main body of the Numidians, passing their own advanced guard [i.e. working round the flanks], charged the Roman flanks: and, by the damage which they did them, prevented them from keeping up the fight with the troops on their front." - Polybius III.73

The Roman (including Italian) cavalry seem to have backpedalled in what we might characterise as a skirmishing retreat; most of them got away from the battlefield, indicating they had not been very closely engaged. 

These battles show that while there was an underlying doctrine of manoeuvre, it could be abandoned at the drop of a hat, and cavalry unaccustomed to Numidians could be upset by too much manoeuvre.

[Patrick Waterson]

Now a look at Issus and Gaugamela.

Issus
At this battle, a few thousand Thessalian and Greek cavalry engaged a stated 30,000 Achaemenids attacking across a shallow river, and held them while the battle was won elsewhere.  Exactly how they did so is not specified, but the Persian cavalry would have been constrained by the narrow battlefield to deploy in very deep formation, so numbers in contact would have been about equal.  Tactics and techniques are not described; my own assumption is that the Thessalians, using the xyston, were well able to hold their own against the Achaemenid cavalry, who although better armoured were armed with the four-foot palta, a short javelin-spear.

Gaugamela
This has one of the most intricate and involved cavalry actions in our period.  The action on the Macedonian left is barely covered, and that only by Curtius, who tells us that Mazaeus 'pressed hard' against Parmenio and this pressure slackened when Darius fled, allowing Parmenio to mount a cautious counterattack which helped Mazaeus decide upon departure.

The Macedonian right fought a remarkable action against the Persian cavalry left, with odds similar to those at Issus.  Darius, seeing Alexander moving his army right to avoid Darius' trap, ordered in his chariots and cavalry, the cavalry on the Achaemenid left moving out en masse to sweep round the Macedonian right.  To perform this, they seem to have faced left and begun moving in what would have resembled a huge column.  Alexander sent out Menidas and his mercenary cavalry to intercept the Persians (actually mostly Scythians and Bactrians) by attacking the head of the column.  This worked; the Achaemenids redeployed to fight Menidas and were driving him back by weight of numbers and armour when Alexander added the Paeonians, here, as in all of his major battles, committed in an unambiguously shock role, together with the Greeks (xenous).  The Greeks, mercenaries and Paeonians gained the upper hand by anestrepsan, which suggests they fell back, drawing the Achaemenids on in some disorder, and then launched a coordinated charge which sent their opponents reeling.  (Curtius suggests Parmenio did something similar on his wing towards the close of the battle when Mazaeus' pressure slackened.)

Alexander improved upon this by promptly sending in the prodromoi/sarissophoroi (again, like the Paeonians, used in an unambiguous shock role against heavy opponents), catching the disordered Achaemenid cavalry left in flank and routing it off the field.  This left him clear to perform his own stroke through Darius' infantry left right into the centre where Darius himself was.  The results can be seen on the Alexander Mosaic.

Achaemenid cavalry seems to have conducted itself in a single large mass in which various contingents ebbed and flowed.  (There was one exception: 1,000 Indian and Persian cavalry were detached, or sent, to rescue the royal heir and incidentally help themselves to the Macedonian baggage; these performed their mission to the letter, forgoing opportunities to attack the phalanx in flank and rear.)  Alexander's cavalry fought by squadrons, able to act independently and also to launch massed coordinated charges.

General
On the minor detail front (speed of charge, what happens when men and mounts reach bump-together distance, etc.) we get little if any detail.  Curtius notes that Alexander's Companions were striking at their opponents' faces, which seems to have been Alex's preferred modus operandi, especially against armoured opponents.  Of interpenetration there seems to have been little or none; formations appear to have remained coherent and cohesive and, at least for the Macedonians, under control even after engaging.

[Justin Swanton]

In the examples Patrick gives I don't see any case of two cavalry bodies charging full-speed into each other, and IMHO I can't imagine horses being persuaded to charge into contact like this as horses (and riders) know they would come out of it the crippled losers. Cavalry seemed  to have halted when in sparring distance and then letting psychology and perhaps superior weaponry decide the issue.

It's possible though that in some cases the horses remain in constant movement by ranks charging and then countermarching, allowing the ranks behind them to charge in turn. Asklepiodotus describes squares in which the men in a file were three times further from each other than from men in adjacent files, whilst the formation preserved the shape of a square:

...the Greeks modified the squadron formation by making it an oblong in mass, while giving it to the eye the appearance of a square. For they drew up the riders with a front of sixteen and a depth of eight, but they doubled the interval between the riders because of the length of the horses. And some made the number of men in length three times that of the depth and then tripled the interval in depth, so that it again appeared to be a square... - Asklepiodotus 7: 4

Which gives this configuration:



Notice that if the front rank countermarches to the rear the next rank has the distance to get up a good speed before contacting the enemy. It doesn't mean they charge into contact but it does create psychological intimidation and maybe the perceived threat of a disastrous contact. It also allows the rider with the longer lance to fatally wound his opponent as his weapon penetrates his opponent with the speed of his charge. An idea.

[Erpingham]

It is worth re-reading the Contemplating cavalry thread from earlier this year, where many of the same questions were explored.

[Nick Harbud]

I just read Justin's article in Slingshot 320 on the physics behind infantry stopiing (or at least slowing) a charging horse.  In calculating the number of ranks required to stop such a horse, he relies upon conservation of kinetic energy.  This is not the correct basis for such a calculation.

In accordance with Newton's First Law of Motion, momentum is conserved between colliding bodies, but not necessarily kinetic energy.  This is easily demonstrated in the classic experiment where one sticky body on a frictionless carriage is collided with a similar stationary sticky body.  Following collision, both bodies move at half the velocity of the single body prior to collision.  Momentum is conserved, but energy is not.

Applying this to the putative horse charging several infantrymen successively, what tends to happen is that the infantryman 'sticks' to the horse for a short period and is accelerated to around the same velocity as the horse, whose speed is consequently reduced in accordance with the Great Newton's observation above.

In the table below, I have taken Justin's figures for horse speed, mass of horse, etc.

Infantryman mass	70 kg
Horse initial speed	11.11 m/s
Horse mass	250 kg

I have also assumed that the horse and infantryman will be in contact for approximately 2 metres, during which time the latter will be accelerated to the same speed as the former.  From this can be calculated the acceleration and, hence, the force on the infantryman.

	Horse Speed	Horse Momentum	Combined Velocity	Infantry Momentum	Contact Time	Infantryman Acceleration	Infantryman Force
Rank	m/s	kg.m/s	m/s	kg.m/s	s	m/s2	N
1	11.1	2777.5	8.7	607.6	0.2	37.7	2636.8
2	8.7	2169.9	6.8	474.7	0.3	23.0	1609.4
3	6.8	1695.3	5.3	370.8	0.4	14.0	982.3
4	5.3	1324.4	4.1	289.7	0.5	8.6	599.5
5	4.1	1034.7	3.2	226.3	0.6	5.2	365.9
6	3.2	808.4	2.5	176.8	0.8	3.2	223.3
7	2.5	631.5	2.0	138.1	1.0	1.9	136.3
8	2.0	493.4	1.5	107.9	1.3	1.2	83.2

So what is the outcome?  Well, Justin calculated that the infantryman needed to experience a force of at least 307.8 N to be bowled over by the horse.  From the table it can be seen that the horse will knock over the first 5 infantry, but can probably be stopped by the sixth.  Any volunteers for the first 5 ranks?

[Andreas Johansson]

That looks like a Guardroom piece waiting to happen.

250 kg sounds like a pretty small horse even in our period, though?

[Duncan Head]

250 kg sounds like a pretty small horse even in our period, though?

At https://digitalcollections.sit.edu/cgi/viewcontent.cgi?referer=https://www.google.co.uk/&httpsredir=1&article=2074&context=isp_collection 250 kg is cited as a typical weight for a Mongolian Darkhad horse; so lighter than you'd expect a cataphract to ride, perhaps, but not unreasonable.

But shouldn't we add the mass of the armed rider on top?

[Andreas Johansson]

At https://digitalcollections.sit.edu/cgi/viewcontent.cgi?referer=https://www.google.co.uk/&httpsredir=1&article=2074&context=isp_collection 250 kg is cited as a typical weight for a Mongolian Darkhad horse; so lighter than you'd expect a cataphract to ride, perhaps, but not unreasonable.

I did say "pretty small", not "unreasonably small" I don't think steppe types on ponies are the most relevant for hitting solid infantry at speed. From a brief look online there seems to be considerable dispute about what medieval chargers weighed, but modern horses of similar height weigh 400+ kg, and warhorses presumably weren't more gracile than racehorses.

But re-reading Nick's post I see he says he took the number from Justin's piece (which I don't seem to've got around to reading), so if anyone it's him I should nitpick.

[Nick Harbud]

I just tried plufgging some numbers into my spreadsheet. 

Increasing the weight of horse by 10% increases the force on the infantryman by 22%.

Conversely, increasing the weight of the infantryman by 10% decreases the force by 12%.

All of which emphasises the importance of having a sturdy mount or big chaps in the front of your shieldwall, depending upon which side you find yourself on.   

[Justin Swanton]

I just read Justin's article in Slingshot 320 on the physics behind infantry stopiing (or at least slowing) a charging horse.  In calculating the number of ranks required to stop such a horse, he relies upon conservation of kinetic energy.  This is not the correct basis for such a calculation.

In accordance with Newton's First Law of Motion, momentum is conserved between colliding bodies, but not necessarily kinetic energy.  This is easily demonstrated in the classic experiment where one sticky body on a frictionless carriage is collided with a similar stationary sticky body.  Following collision, both bodies move at half the velocity of the single body prior to collision.  Momentum is conserved, but energy is not.

Applying this to the putative horse charging several infantrymen successively, what tends to happen is that the infantryman 'sticks' to the horse for a short period and is accelerated to around the same velocity as the horse, whose speed is consequently reduced in accordance with the Great Newton's observation above.

In the table below, I have taken Justin's figures for horse speed, mass of horse, etc.

Infantryman mass	70 kg
Horse initial speed	11.11 m/s
Horse mass	250 kg

I have also assumed that the horse and infantryman will be in contact for approximately 2 metres, during which time the latter will be accelerated to the same speed as the former.  From this can be calculated the acceleration and, hence, the force on the infantryman.

	Horse Speed	Horse Momentum	Combined Velocity	Infantry Momentum	Contact Time	Infantryman Acceleration	Infantryman Force
Rank	m/s	kg.m/s	m/s	kg.m/s	s	m/s2	N
1	11.1	2777.5	8.7	607.6	0.2	37.7	2636.8
2	8.7	2169.9	6.8	474.7	0.3	23.0	1609.4
3	6.8	1695.3	5.3	370.8	0.4	14.0	982.3
4	5.3	1324.4	4.1	289.7	0.5	8.6	599.5
5	4.1	1034.7	3.2	226.3	0.6	5.2	365.9
6	3.2	808.4	2.5	176.8	0.8	3.2	223.3
7	2.5	631.5	2.0	138.1	1.0	1.9	136.3
8	2.0	493.4	1.5	107.9	1.3	1.2	83.2

So what is the outcome?  Well, Justin calculated that the infantryman needed to experience a force of at least 307.8 N to be bowled over by the horse.  From the table it can be seen that the horse will knock over the first 5 infantry, but can probably be stopped by the sixth.  Any volunteers for the first 5 ranks?

Mmmmh... not quite. Have a look at the horse collision videos. The people hit by the horses are not borne along with them at their speed, certainly not for two metres, but are knocked flat immediately. They don't actually move from the spot where they were standing. So what you need to determine is the amount of force necessary to topple a standing man when hitting him in his middle, rather than the force necessary to get the entire human body moving at a speed just less than the horse's (which is slowed down by the impact).

But notice that even using this system, increase the weight of the horse slightly as suggested by Duncan and you can easily reach 7 ranks. I doubt that the rider contributes anything - if the horse slows down too much he is just pitched forward off the saddle.

@Andreas: I took 250KG as the lower end of the weight of horses in Antiquity, in order to produce results that were conservative and not exaggerations.

[Erpingham]

But notice that even using this system, increase the weight of the horse slightly as suggested by Duncan and you can easily reach 7 ranks. I doubt that the rider contributes anything - if the horse slows down too much he is just pitched forward off the saddle.

It's an interesting mechanics question.  I think you have to consider the weight of the combined object but the object is in two parts, not strongly attached, so forces involved in deceleration could cause the object to separate into its constituent parts (i.e. the rider fall off).

Re Horse sizes, IIRC Duncan turned up some figures on Nisaen horses when we were discussing Persian logistics which put them in the 350-400kg range.  There is plenty of information about the height of ancient horses on line but not a lot on weight, probably because horses of the same height can vary quite a bit in weight depending on breed.   

[Justin Swanton]

What appears to be established is that a decent-sized horse can burst through an infantry line 8 ranks deep (a typical depth) unless the infantry do something about it. My guess is that they originally deployed in great depth before later switching to ranks bunched up together, othismos-like, which turned them into a stable mass a horse couldn't knock over.