I have recently read about two ancient battles that were won by a much smaller army and/or lead to very different casualties:

  • Defeat of Boudica - according to Roman accounts only 400 Romans died as opposed to 80,000 of Boudica's forces, thus resulting in a 1 : 200 casualties ratio. However, the figure is considered unrealistic.
  • Battle of Alesia - casualties were about 12,800 Romans and 250,000 from Gallic Confederation, resulting in 1 : 20 casualties ratio.

Are these the most "efficient" battles in the history?

I am interested in significant battles there were fought between at least 10,000 soldiers.

  • Comments are not for extended discussion; this conversation has been moved to chat.
    – MCW
    Commented Nov 11, 2019 at 15:15
  • Hmm, do you count massacres if the victims were armed?
    – C Monsour
    Commented Oct 18, 2020 at 22:02

16 Answers 16


It's actually quite difficult to point to a specific answer because, as your Boudica example shows, in many cases casualty estimates are not very reliable. The 80,000 figure given for Boudica's forces is considered to be a rhetorical device and not an actual factual number - a problem that plagues many pre-modern accounts.

Another difficulty is measuring what counts for casualty. Historically, a great part of war losses comes from disease and illness, rather than direct combat. Moreover, there is no unified standard for counting what passes for "wounded" in historical records.

Nonetheless, here are some contenders:

Operation Desert Storm

Among larger set piece battles, this is probably the clear winner. Coalition forces suffered about 1,155 casualties (379 deaths and 776 wounded). Iraqi casualties are less clear, but early claims of 100,000 fatalities are wildly implausible. A more reasonable estimate would be 26,000 killed or wounded with about 83,000 captured. Even with this conservative estimate, the ratio remains an astonishing 1:94. If prisoners are excluded, this ratio falls to a still high 1:22.5.

Desert Storm also contained a number of heavily lopsided engagements, most notably the Highway of Death during which Coalition forces suffered no casualties, giving an effectively infinite ratio.

Battle of Omdurman

The Anglo-Egyptian-Sudanese force took about 430 casualties (48 dead and 382 wounded). The army of Abdullah al-Taashi lost around 30,000, including 12,000 dead, 13,000 wounded and 5,000 captured. This gives a ratio of 1:69, or 1:250 if we only count fatalities.

Battle of Majuba Hill

Probably the most lopsided result in a relatively evenly matched engagement. The Boers lost only 1 dead and 5 wounded, but managed to inflict 92 fatalities, 134 wounded and 59 captured on British forces. This is a ratio of 1:47.5, or 1:92 if we only count fatalities.

It is very rare for pre-industrial battles to achieve ratios like Desert Storm or colonial skirmishes at any scale, although some total disintegrations of the defeated side did occur. In such cases casualties are not necessarily combat deaths or injuries or captured, but often also from exposure during a long rout or simply desertion. Some contenders are:

Battle of Tumu

A 230,000 or 500,000 (estimated actual size vs paper strength) strong Chinese army led by the Ming Emperor was destroyed by a Oirat cavalry force of around 20,000. Precise casualties are not known, but Chinese records indicate that the expedition was entirely annihilated. Even assuming they costed the Oirats extremely dearly, the ratio couldn't have been any lower than 1:23 or so, and might be (but probably isn't) as high as 1:166 going by Wikipedia figures.

Battle of Fei River

A Jin force about 80,000 was able to crush an 870,000 invading Qin force through inducing a panicked rout. Again detailed casualty numbers are unavailable, however the Qin Emperor was recorded to have returned with only 20-30% of his original force. Traditional sources indicate minimal Jin casualties, though some online sources claim 5,000 with no clear evidence. Conservatively estimating a Qin 600,000 casualty, most of which incurred during the rout, and assuming a more reasonable 10,000 lost for the Jin, that still gives a ratio of 1:60.

Mongol Invasions

I feel this deserve to be mentioned. During the Mongol Empire's expansion into Europe and elsewhere, they fought several militarily wildly successful battles that annihilated entire armies. For example, the Battle of the kalka River, or the Siege of Bagdad, and the Battle of Liegnitz. The ratios are particularly lopsided because the Mongolians could chase down routed enemies with great lethality. However, there are usually no reliable estimates for Mongolian casualties, so no real ratio can be calculated.

Other honourable mentions:

Battle of Kircholm

One of the most successful heavy cavalry charges of all time. The Polish-Lithuanians are said to have suffered about 300 casualties (100 dead and 200 wounded); the Swedish army however lost between 7,600 to 8,900 men, most killed or executed as they were routed. That's a ratio of 1:25-30.

Battle of Rorke's Drift

Example of lopsided victories for a heavily outnumbered force. About 150 British and colonial soldiers held off a 3,500 strong Zulu force, inflicting 351 confirmed kills for 17 fatalities in the process. Including 15 British and an estimated 500 Zulu wounded, the ratio is 1:26.6.

  • 1
    It doesn't hit the size minimum, but the Battle of Hill 448 had a similar ratio at about 1:20 (and most of the US deaths came at the hands of a sniper after the VC forces had withdrawn).
    – T.E.D.
    Commented Nov 8, 2019 at 17:17
  • 6
    IIRC, Desert Storm US deaths were about only 120-150 combat-caused, with the bulk of the larger overall number being caused by friendly fire rather than enemy action. Commented Nov 8, 2019 at 19:51
  • 7
    @ItalianPhilosopher The US had 148 combat deaths in Desert Storm, yes. It's another illustration of why casualty comparisons is fraught with problems.
    – Semaphore
    Commented Nov 8, 2019 at 20:00
  • 2
    Incredibly involved answer. You clarified (very well!) that casualties technically includes wounded. It's only recently been added (thanks, grumble, to common-usage) to include "deaths only". There was quite the kerfuffle about this regarding Iraq war stats, which led me into digging up meanings. Commented Nov 11, 2019 at 15:10
  • 1
    @mart The former was against military targets, the latter civilians.
    – Semaphore
    Commented Oct 20, 2020 at 6:48

Pearl Harbor attack:
Discarding the materiel lost, the USA had 2,335 killed and 1,143 wounded, while Japan had only 64 killed and 1 prisoner. That makes a ratio of 1:56.
If one adds in the materiel probably the difference is quite bigger.

Battle of Sidi Barrani:
During operation Compass, the British had 624 losses, while Italy had 2,194 killed, 2,286 wounded and 38,300 prisoners. That makes a ratio of 1:69.

Many naval battles might fulfill the request, because it is easier to understand that one side has many losses, while the other none. For example, the sinking of the battleship Bismark (5 losses for UK and 2,310 losses for Germany), or the [battle of Cape Matapan].5 (3 killed for UK and 3,315 losses for Italy, ratio 1:1100!)

  • 1
    Most naval battles did not involve 10,000 people. Commented Nov 8, 2019 at 17:48
  • 6
    does pearl harbor count as a battle? should we include other terrorist style attacks on unprepared civilians/soldiers? if so i can think of cases where one side suffered only a few casualties but thousands of innocents were killed Commented Nov 8, 2019 at 19:50
  • 5
    Mindwin : Naval battles that I cited had 10,000 sailors, as requested by the OP. USER_8675309 : If both sides figth, by definition that is a battle (the wiki page of Pearl Harbor attack has the template of a battle). Japanese losses were not accidents, but result of american defense.
    – Santiago
    Commented Nov 8, 2019 at 20:26
  • The sinking of the Bismark is borderline: the infobox in the Wikipedia article says the British had seven destroyers, but there are only five destroyers named in the narrative, giving a total of around 9700 sailors between the two sides.
    – Mark
    Commented Nov 8, 2019 at 20:40
  • @USER_8675309 For instance, according to Wikipedia the September 11 attacks have a 1:156 ratio. Commented Nov 9, 2019 at 18:44

Battle of Gaugamela in 331 BC between Alexander and Darius III's Persians.

One of Alexande's great victories. Even if we take the highest estimate for Alexander's casualties (1,500) and the lowest for Darius III's army (40,000), it still gives a ratio of 26.67 to 1.

Battle of Carrhae in 53 BC between the Romans and the Parthians.

This is the famous battle where Crassus came to a sticky end. The Romans suffered about 20,000 killed (according to Wikipedia) or 24,000 killed (according to Ground Warfare: An International Encyclopedia, Volume 1). The Parthian casualties were minimal (according to Wiki). Actually, Wiki gives 38 but even if we consider 1,000, the ratio is 20 or 24 to 1 killed.


1988 - Battle of Afabet

Fought between an Eritrean People's Liberation Front (EPLF) force of around 15,000 and an Ethiopian army numbering 20 to 25,000, this was a huge defeat for the latter.

Ethiopian casualties were over 18,000 killed or captured while EPLF casualties amounted to only 394 (125 killed, 269 wounded, giving a ratio of at least 45 to 1.

1798 - Battle of the Pyramids

On the 21st of July 1798, Napoleon's army scored a decisive victory over a Mamluk army. Although the total size of the armies is disputed, the French numbered at least 20,000 while the Mamluk estimates range from 12,500 to 40,000.

French killed amounted to just 29. Estimates for Mamluk killed on French Wikipedia range from 1,200 to 8,000. The ratio of Mamluk killed to French killed was therefore at least 41 to 1 Total French casualties (killed and wounded) were 289 while, accordion to Napoleon himself, Mamluk casualties amounted to 20,000 (mostly likely a very generous guess), giving a ratio of 69 to 1.

1504 - Battle of Cochin

Fought between Portugal and their Cochin allies against Calicut and lasting from March to July, this was actually "a series of confrontations" but is worth a mention as there were no reported fatalities on among the Portuguese and 'neglible' among the Cochin while their Calicut opponents lost 5,000 killed in action and a further 13,000 to disease.

260 AD - Battle of Edessa

A Roman disaster to rank alongside Carrhae mentioned in JLK's answer. Roman losses were at least 60,000 (with Valerian having the dubious honour of being the first Roman emperor to be captured). Sassanian casualties were "minimal".

422 BC - Battle of Amphipolis

Fought between the Athenians and the Spartans in 422 BC during the Peloponnesian War, this was a comprehensive Spartan victory and the killed ratio reflected this: about 600 Athenians to just 8 Spartans, a ratio of about 75 to 1.

This battle was also notable for those involved and for the fact that the commanders of both sides died: the Athenian general Cleon and the Spartan Brasidas (who lived long enough to see the victory). Brasidas was the outstanding military officer of the first part of the Peloponnesian War; an ossuary believed to contain Brasidas' remains is on display at the Archaeological Museum of Amphipolis.

Amphipolis also saw the end of the historian Thucydides' military career, while the philosopher Socrates was among the Athenian survivors (according to Plato's dialogue, Apology)).


Tsushima 1905. 117 vs 4380 dead. Fleet engagement, so the Japanese + Russian numbers of men, combined, probably were over your 10k threshold, but that 4380 looks like it was most of Russian fleet.

The Invicible Armada lost 15000 men in 1588, but most of that was also due to weather on the retreat trip around Scotland. English losses were about 400. The wikipedia entry is unclear about the ratio as it tallies up English losses due to disease the following year. But the overall picture of the initial battle + retreat is one of the most lopsided victories in history.

Wikipedia:SpanishArmada and ThisDayInHistory:SpanishArmadaDefeated

Some of the early battles of Barbarossa, say July-Oct 1941, consistently had some really lopsided numbers as well, when you tally up the POWs in the encirclement phases, but nowhere as extreme as some other battles others will cite. Yet, this happened at such a scale that it merits mention as well, even if the ratios aren't as high.

Kiev - Aug 1941

German: 100K casualties (from which 95K are wounded). USSR: 615 KIA+POW (plus 85K wounded)

The reason I am breaking out wounded here is that you can have a reasonable expectation that some/most of the wounded will be back on the front in some capacity later on. Not so with a KIA or POW.

(the same logic applies to tank battles, whenever you read loss ratios - whichever side occupies the territory after a battle can salvage some of their tanks, but the side that withdraws loses any immobilized tank).

If you assume 50% can heal and fight again, German losses are about 50K vs 650K+ Soviet side. So, from 1-7 to 1-13

I admit, this is a bit of a cheat in the context of this question, but its effect was certainly true in practice and it worked the other way too - the 300K POWs at Stalingrad were a full loss to the Nazis. Tactically - irrelevant, strategically - quite relevant.

Suomussalmi, Russia-Finland Winter War 1939

  • Finnish 1750 casualties from 11500 troops.
  • Russia 15000-25000 from 45-55K troops.

All the more impressive since the artillery, aviation and tank advantage was massively Russia's.

Last, but not least, how could we forget everyone's favorite Hannibal?

Cannae 216 BC

5700 Carthagians KIA, 67-85K Roman KIA

Lake Trasimene 217 BC

1500-2500 KIA Carthage vs 15K KIA + 15K POW Rome

  • 3
    Tsushima is easily over the 10k mark. If you add in the number captured, the Russian losses are about 10,400, and a quick count of Japanese ships engaged gives a total complement of around 15,000-20,000.
    – Mark
    Commented Nov 8, 2019 at 20:49
  • And if you include the losses the Russians suffered while getting their "fleet" to Japan it's even worse for them (that journey is probably the greatest display of ineptitude in command of a military operation ever put on).
    – jwenting
    Commented Nov 12, 2019 at 5:08
  • @jwenting oh, come on, that's unfair. at least they didn't lose their battle with those fearsome British trawlers. en.wikipedia.org/wiki/Dogger_Bank_incident Commented Nov 12, 2019 at 6:06
  • @ItalianPhilosopher I agree, and they also had great success against the hundreds of invisible Japanese torpedo boats in the Bay of Bisquay.
    – jwenting
    Commented Nov 12, 2019 at 7:47

The battle of Aquae Sextiae (102BC)

As with all ancient battles there is a wide variety of figures thrown out for the casualties and participants for this battle.

Roman Total Forces = 40-50k (6 Legions plus auxiliaries) & Roman Casualties = 500-1000

Teutones/Ambrones Total Force = 50k - 150k & Tuetones/Ambrones Casualties = 30k - 120k

upper end ratio would be around 300:1 (150k/.5k) and lower end would be 30:1 (30k / 1k)

I would imagine the truth is somewhere in the middle of those ranges.

After doing some more research I found out about 2 naval battles in WW2, the Hi-81 convoy battle and the battle of the Bismark Sea. I do think these would count using the OPs qualifications as both had over 10k participants.

Battle of Bismark Sea - 2–4 March 1943

The Battle of the Bismarck Sea (2–4 March 1943) took place in the South West Pacific Area (SWPA) during World War II when aircraft of the U.S. Fifth Air Force and the Royal Australian Air Force (RAAF) attacked a Japanese convoy carrying troops to Lae, New Guinea. Most of the Japanese task force was destroyed, and Japanese troop losses were heavy...

...The battle was a disaster for the Japanese. Out of 6,900 troops who were badly needed in New Guinea, only about 1,200 made it to Lae. Another 2,700 were saved by destroyers and submarines and returned to Rabaul.[63] About 2,890 Japanese soldiers and sailors were killed. The Allies lost 13 aircrew, 10 of whom were lost in combat while three others died in an accident. There were also eight wounded. Aircraft losses were one B-17 and three P-38s in combat, and one B-25 and one Beaufighter in accidents. MacArthur issued a communiqué on 7 March stating that 22 ships, including twelve transports, three cruisers and seven destroyers, had been sunk along with 12,792 troops

Allied deaths = 13

Japanese Empire deaths = 2390

Ratio = 222:1

Convoy Hi-81 - November 15–18, 1944

Convoy Hi-81 (ヒ-81) was the designation for a formation of Japanese transports that carried soldiers bound for Singapore and the Philippines during World War II. The transports were escorted by a large force of surface combatants including the escort carriers Shinyo and Akitsu Maru which were sunk in the Yellow Sea by American submarines. Over the course of a four-day convoy battle in November 1944 nearly 7,000 Japanese were killed in action while the Americans sustained no casualties.

Ratio = infinite


Additionally there was the battle of hi-71 which was another convoy battle which occured from August 18–19, 1944. This battle resulted in 60 killed for the Allies and 8k killed for the Japanese Empire. Resulting ratio is:


#2 (07242020) I was reading about the Italian-British naval actions during WW2 and I came across the Battle of Cape Matapan It had over 10,000 participants and the result was 3 dead on the British side and 2,300 on the Italian side giving a ratio of around



The Battle of Agincourt (1415)

French strength:       12,000 - 36,000 (modern estimates)
English strength:       6,000 -  9,000

French casualties:     approx. 6000 killed
English casualties:    112 - 600 killed

Strength ratio (F:E):  min 4:3,  max 6:1
Casualty ratio (F:E):  min 10:1, max 53:1

Numbers above from Wikipedia entry

History.com entry which validates English strength number (actually says the English strength would be half of 11,000) and the French casualty estimates.

New York Times article reporting on recent research which numbers the English at between roughly 6000 and 8600 by the time of the battle, and the French at somewhere between 12,000 and 24,000.


Wikipedia:Battle_of_Cajamarca, assuming that contemporary accounts were accurate.

Conquistador Francisco Pizarro ambushed the Emperor Atahualpa. The Spanish suffered no deaths and one wounded. The Incans suffered 2000 dead with 5000 more taken prisoner.


There is a historical record of the battle with 50 thousand to 0 score.

According to Procopius, during the war with Vandals, there was a battle near the Bourgaon (Libya) in which the Roman general Solomon defeated the army of Moors.

And there perished in this struggle, among the Moors fifty thousand, as was declared by those of them who survived, but among the Romans, no one at all, nor indeed did anyone receive even a wound, either at the hand of the enemy or by any accident happening to him, but they all enjoyed this victory unscathed.

Cited from Procopius, Loeb Classics edition, vol 2, Book IV (the Vandalic war-2), xii, 25-26. (Greek with parallel English transl., available online).

Procopius was a secretary of the Roman Commander-in-chief Belisarius, and he was participating in the events he described. He is generally trusted by modern historians. (You can even see from the cite above how carefully he states the facts:-)


The Battle of Omdurman, 2 September 1898, between British and Egyptian forces under Lord Kitchener avenging the death of Chinese Gordon against Sudanese forces led by bdullah al-Taashi.

Despite being outnumbered by about 2:1 (26,800 vs ~52,000) the British inflicted an estimated 12,000 killed on the Sudanese forces for just 48 (or perhaps 47) killed, a ratio of ~250:1. Counting wounded and prisoners the totals are 30,000 vs 382, still a respectable ratio of over 76:1. As Kipling famously noted:

Whatever happens, we have got
The Gatling gun,; and they have not


Salsu River in 612 is a strong contender, made all the more notable by the victor's forces having been massively outnumbered.

In the course of the Second Goguryeo-Sui War, a 300,000-or-so-strong Sui army attacked approximately 10,000 Goguryeo cavalry at the Salsu River and were almost completely destroyed, with only 2,700 (less than 1% of the original force) escaping with their lives; the massive losses suffered in the battle would prove to be one of the factors leading to the collapse of the Sui dynasty. Casualties on the Goguryeo side were unknown but presumably light, given that they decisively won the battle.

Due to the lack of information concerning Goguryeo casualties suffered in the battle, an exact kill:loss ratio cannot be calculated; however, given the number of soldiers they started out with and the losses they inflicted on the Sui army, it would be mathematically impossible for their kill:loss ratio to be any less than ~29.7:1 even if they were completely wiped out, and, since they decisively won, and their casualties can, thus, be presumed to have been fairly light, the actual kill:loss ratio must have been, correspondingly, even greater (potentially much more so - for instance, if there were actually 2,500 Goguryeo casualties, the true kill:loss ratio would be 118.9:1).


Another colonial battle: Battle of Namutoni (1904) in German South West Africa (today's Namibia):

Four German colonial soldiers (Schutztruppe) and three German reservists defended the Fort Namutoni with 1500 rounds of ammunition against about 500 Ovambo warriors. The Germans held out all day and retreated during the night without casualties, the attacking Ovambo had 68 killed, 48 missing and 20 wounded.

The abandoned Fort Namutoni was destroyed during the following days and rebuilt two years later. Today it is one of the campsites in the Etosha national park.

  • 3
    The question is looking for battles with at least 10,000 participants.
    – Mark
    Commented Nov 8, 2019 at 20:43
  • 1
    Their strength was as 10,000 because their hearts were pure. :-) Commented Nov 9, 2019 at 2:20

The Battle of Blood River easily wins this award.

Casualties were: 3 wounded, 0 deaths vs. 3000+ dead, i.e. a casualties ratio of infinity.

  • 4
    "Casualties" typically does include the number of wounded.
    – Yozarian22
    Commented Nov 9, 2019 at 16:31

Caernarfon Castle was defended by a force of 30 (I found a number of references that state 28) against an attacking force of both the Welsh and French with siege engines and battering rams. (1403) It's even mentioned on the netflix series "Secrets of Great British Castles".

I don't think it's known exactly the size of the attacking force; However the book "Castles, Battles, and Bombs: How Economics Explains Military History" states that they were able to inflict 300 casualties. This gives the very worst ratio of 1:10 ... but a potential best of 0:300.

There is also the battle of Agincourt https://en.wikipedia.org/wiki/Battle_of_Agincourt

  • 112-600 killed vs 6000 killed + 700-2000 captured giving it the low end of 1:11 to the high end of 1:71.

I would also suggest your definition of "significant" being 10k men is a little wonky - in that battlefields have changed significantly with time - a Roman legion was 6k men - most battles would be unlikely to involve more. This doesn't make the battles any less significant...


Another pretty one-sided battle was the battle of Casilinum in southwestern Italy in 554 AD, between roughly 20000 East Romans under Narses and similarly many Franks under Butilinus. According to Agathias (who is probably exaggerating a lot) the Romans lost 80 men and the Franks lost all except 5 men. This although one part of the East Roman had initially declined to fight and only arrived on the battlefield after the Franks had started their initial charge.


To add to the other answers, I want to mention the Battle of Tsushima.

The Russian losses were 4,380 dead and 5,917 captured, i.e. 10,297 in total.

Japan's losses were 117 dead.

The ratio is 1:88.

  • 5
    This is already given above in Italian Philosopher's answer.
    – Allure
    Commented Nov 11, 2019 at 1:52

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