I've heard that in most battles prior to the introduction of gunpowder weapons, the casualties were usually very low (around 5% even in long battles) prior to the moment when someone's formation was broken and the side stopped really fighting and just ran away (or was fighting without the benefits of proper formation if the side couldn't retreat), and that most people killed in battle were killed by the pursuing victorious force.

Is it true? Is it a matter of debate between historians? Is it basically true but exagerrated? Or is it considered a myth?

  • 3
    From Wikipedia, but marked as "Citation Needed": "In most medieval battles, more soldiers were killed during the retreat than in battle, since mounted knights could quickly and easily dispatch the archers and infantry who were no longer protected by a line of pikes as they had been during the previous fighting."
    – jbabey
    Commented Dec 9, 2012 at 17:10
  • 1
    Were there any well-documented battles that ended with a controlled withdrawal?
    – kubanczyk
    Commented Dec 12, 2012 at 22:18
  • 1
    @jbabey: thanks. Essence of my question is whether the statement in the wikipedia article is disputed or agreed with by most historians/ archaeologist/ history reenactors. Find a source or two and it's a perfect answer for me.
    – Pavel
    Commented Dec 15, 2012 at 19:32
  • 1
    Where did you hear this? What is the source? In particular where did the 5% figure originate? Citation please. And I think that the scope of "prior to gunpowder" is far too broad in time and space.
    – MCW
    Commented Apr 4, 2013 at 18:00
  • @kubanczyk: Caesar's withdrawal at Gergovia after the assault failed, for one. Commented Dec 15, 2013 at 19:03

9 Answers 9


It all depends on which era you are talking about.

There are mostly 3 different medieval eras; early, high, and late middle ages.

In Medieval Warfare: A History there is a chapter that talks about it. In the early and partly high middle ages, people were few and only the nobles and the clergy fought in wars.

Those two groups were pretty important and, as time went on, heavily armored and well equiped. In impoverished enviroments it was also better to take ransom, rather than lives. So battles were fought mostly on small scale, and casualties from the battle itself were few. The percentage according to the book (if I am not mistaken it was 5% exactly) was that low for the battle itself - excluding marching deaths, illness, and other factors.

Later on though, during the high and late middle ages, the European war-machine became more sophisticated, population increased, and the middle class arose and took part in non-equestrian battles using pikes and other less glorious but just as effective weapons. This led to large scale battles with more casualties on the ill equiped. So yes, as TomAu and fred2 said, there was a time when war took a great toll.

And this is especially true during the late middle ages and the introduction of gunpowder.


In Rome at War, Nathan Rosenstein provides a very careful study of mortality rates in the Republican Army from 200-168 B.C. The overall mortality rate strictly attributable to combat is estimated to be 2.6 percent of soldiers per year (125). Overall mortality is estimated at 4.75 to 5.45 percent of soldiers per year, with non-combat mortality amounting to 1.9 to 2.6 percent of soldiers per year. However, because 1.5 percent of conscripts would have died from disease even if they had remained civilians, the "excess mortality attributable to warfare" was 3.25 to 3.95 percent of all soldiers annually (136).

However, these figures include legions that were in the field but did not engage in major battles. The average mortality rate for legions in combat was around 5.6 percent (124). And defeats were around 4 times as costly as victories: victories saw mortality rates of around 4.2 percent of participants, while defeats saw mortality rates around 16 percent (118). In general, Rosenstein finds that mortality rates due to both combat and disease were lower in the Roman legions than in 19th century mass warfare (125-126).

Incidentally, Rosenstein warns against relying on estimates taken from accounts of one or two battles. There is heavy selection bias going into chronicles. He notes that "figures fall broadly into two clusters--those that are very high and many (like the eighty who died at Pydna) that are strikingly low" (23).


As a specific source for the casualties suffered in one particular ancient battle, in 57BC in Caesar's Gallic campaign, here is a quote from The Conquest of Gaul:

In Chapter 5 (of my translation) Failure in the Alps (57 BC), 4th paragraph, Caesar writes:

... But what told against them [Caesar's forces] was that the enemy, when exhausted by prolonged fighting, could retire from the battle and be relieved by fresh troops, which our men could not do on account of their small numbers; not only had tired men to stay in the fighting line, but even the wounded had to remain at their posts without any chance of respite.

... and Baculus - the chief centurion who was disabled from several wounds in the battle against the Nervii - came running to Galba ... and told him that their only hope of escape was to try their last resort, a sortie through the enemy lines. ...

Suddenly they charged out from all the gates, without giving the enemy a chance of realizing what was happening or of preparing to meet their onslaught. It was a complete reversal of fortune: the Gauls who had counted on capturing the camp were surrounded and cut off. Of the forces that had taken part in the attack - known to number over thirty thousand - more than a third were killed; the rest fled in terror and were not allowed to rest even in the mountain tops.

It is clear from this description that most of the casualties on the losing side (the Gauls) occurred after their morale broke and they turned to flee. Also that total Gallic casualties were in the vicinity of 30%, including those suffered both before and after the morale break.

  • 2
    Caesar may have inflated some of the numbers to emphasize his accomplishments (others in a similar position did so). Do you know of a secondary source which confirms them?
    – Drux
    Commented Mar 6, 2014 at 14:13
  • 3
    @Drux: For two thousand years Caesar was regarded as completely unreliable solely based on the complete impossibility of his description of Alesia. Then the ruins were discovered, and found to be in complete conformance with Caesar's description of them. To the best of my knowledge, no-one has caught Caesar out in any significant exaggeration that can be independently verified. Commented Mar 6, 2014 at 23:16
  • 1
    Hmm, I have in front of me a book that says Caesar's claim was he went to Britannia with 600 transport ships and 28 war vessels, and that these numbers are now known to be inflated.
    – Drux
    Commented Mar 7, 2014 at 17:44
  • 1
    @Drux: By what, or whose, evidence are the claims known to be inflated? A reference would e nice so I could follow up. Commented Mar 7, 2014 at 23:03
  • @Drux: Remember that Caesar's claims of building 25 miles of fortifications at Alesia (in barely 30 days) were regarded as hopeless exaggeration for nearly 2,000 years; until 25 miles of fortifications were discovered by archaelogists, just as described by Caesar. Commented Mar 7, 2014 at 23:53

I think the answer would be 'it depends'. Three late medieval battles might serve as an example (artillery was used, but probably not in a way that had much bearing on the result). The Battle of Sauchieburn in 1488 pitted an unpopular king James III of Scotland against rebels led by his son. The king himself was killed, but there is a suspicion that those who could sneak away and avoid risking their lives for him, did so. The Battle of Bosworth where Richard III was killed seems somewhat similar. As much a chaotic punch-up as a classic battle, where you get the feeling those on the losing side did not really want to be there. There were many deaths, but percentage-wise it does not seem to have been great.

Contrast that with the Battle of Flodden in 1513, where a very popular James IV invaded England with 30,000 men. The army was confident, believed in their leader, did what they were told, and got massacred by the infantry wielding English bills (basically a machete on a stick).

  • 6
    Interesting stuff, but how does it answer the question? Commented Dec 15, 2012 at 18:03
  • I know there are exceptions from the rule I'm asking about. In a long battle of two armies with good morale determined to win or die, which seems to be the case of Battle of Flodden and many later battles, there could be great casualties without fleeing. Or if one army (or both) manages controlled retreat (most inconclusive battles, like Battle of Meggido between Egyptians and Hittites, 1285 BC), most casualties are in open combat, not while fleeing. What I want is anything that proves or denies the "small casualties before the battle is lost in most battles" pattern. I must agree with Felix.
    – Pavel
    Commented Dec 15, 2012 at 19:47
  • 2
    I'd say any generalisation this broad is almost bound to be inaccurate, especially when you're talking about a time scale of thousands of years, in completely different places,with completely different weapons. If 3 British battles fought within 30 years can have very different levels of fatality, imagine it over 3000! My answer then is 'there is no answer'. It's ipso facto easier to kill someone who has their back to you and is running away in disarray, particularly before the invention of guns, but there are so many other factors in any battle that affected the result and mortality rates.
    – fred2
    Commented Dec 15, 2012 at 21:35
  • I get @fred2's point, but I'm not really convinced there isn't a better way to answer this question.
    – o0'.
    Commented Dec 16, 2012 at 1:19

Medieval battles could get very bloody. For instance, in the Battle of Falkirk, Scotland, the Scots lost perhaps 2,000 (out of 6,000) in the battle, and another 1,000 or so in a well-organized retreat. English losses were proprotionately lighter, perhaps 1,000 out of 15,000. Wiliam Wallace vs. Robert Bruce: Why Did One Win and One Lose?


In ancient times, it usually happened that either one army enveloped the other, or it completely blocked its path of retreat by attacking its rear or flank after its own cavalry defeated the enemy cavalry. Thus, if the surrounded side lost, it was completely cut down, as happened at Cannae.

Given that in medieval times the difference between combat branches was deeply blurred, that there was no actual cavalry or infantry, the battle consisted of a direct, frontal clash between the whole of the opposing forces, the path of retreat being clear. In case of defeat, the losers fled, whilst the winning commander lacked the effective authority to order his troops to pursue the enemy.

Moreover, the men fought not in order to kill the enemy, but in order to capture him alive, in the hope of obtaining a handsome ransom. This inclination, Hans Delbruck notes, weakens substantially the warrior spirit, which must stay put in annihilating the foe. The same renowned military historian notes that he was yet to find an instance of medieval battle where one side purposefully detached a squadron so that it may land a surprise attack after the battle started. This is the only imaginable case where an army's retreat path is blocked, thus leading to it's total annihilation. In the absence of such scenario, medieval battles were essentially limited in bloodshed.

  • Whether retreat path is blocked or not affects casualties after the end of real fight, when one side flees. My question focuses more on casualties before this moment.
    – Pavel
    Commented Feb 11, 2014 at 8:57

There is actually a word for battles where (at least) one side takes heavy losses before leaving the battlefield: "Pyrrhic victory". Note that the wp list is by no means exhaustive, e.g. the Battle of Beneventum from the Pyrrhic Wars is not mentioned, although the losses on the victorious side there look also quite significant.


One must come to the realization that being armored and at the same time swinging a heavy metal or bronze sword all the while slamming into other men hacking, stabbing etc., would be tiring even for a badass of his day and one more thing to truly grasp the realness ever try to neatly slash or stab your way into a thick leather coat much less chainmail or hard leather let me a sure U its not like in the movie's at all taking all this into account one would think it be mere minutes before duress and fatigue overcame the lines meaning that to overkill or to kill for that matter (remember there's no Hollywood blades slicing through armor like a knife through hot butter) would take lots of energy for either side of the lines of war to hack and stab repeatedly away just to kill one or many armored, and armed men fighting back! It doesn't add up so I definitely can side with there being low death rates in war at certain period's in times of old.


About the early ancient era... especially about those famous battle of the ancient Greeks, but the period of Ancient Near East too...

Our only sources are written sources by somewhat biased historians from ancient times who, to be fair, seem to exaggerate a bit. The first recorded "battles" in history seem to be nothing more than a fight between brutes, rock throwing, broken arms and legs, and the victor is the one who had more men left standing. All of this lacked a tactical approach which we often put an accent to when we think about those battles. This is certainly the case in the Eastern empires, but of course, military developed over time so it's hard to put a clear line over when there were battles as we all imagine them, and when they were just fights of brute force. The battles of the Ancient Greeks probably weren't as epic as we think they were, but they were not for underestimation of course. But yes, the casualties were, more often than not, very minimal.

  • 4
    "This is certainly the case in the Eastern empires" I think you underestimate the Persians.
    – Jeroen K
    Commented Mar 8, 2014 at 23:07

Not the answer you're looking for? Browse other questions tagged or ask your own question.