(Note by the Saker: Tatzhit Mihailovich submitted this article with *many* embedded photos and videos and I asked him to replace them by links to the media in order not to overload the page and to make the formatting easier.  Make sure to check his YouTube and LiveLeak channels which are chock-full of very good stuff!)

Seeing through the doublethink: Primary evidence on losses of the combatants at Donbass

by Tatzhit Mihailovich


Both sides in the Ukrainian civil war hugely overestimate enemy losses and conceal most of their own. Therefore, there is pressing need for a fact-based estimate of the loss of life.

Lostarmour.info database lists over 1000 destroyed and captured UAF armored vehicles. In previous similar conflicts (Chechnya, Syria, Iraq) the number of soldiers killed per armored vehicle destroyed was between 10 and 30. Therefore, we can conservatively estimate nationalist losses to be about 14 thousand KIA. There are also about 4 times this number of wounded, about a third of which will be severely permanently disabled. This is far greater than the (widely discredited) official Kiev claims of around 1500 KIA and 4500 WIA.

Rebel losses are lower, but probably closer to Kiev estimate of ~8000 KIA than to the rebel claims of 1000-1500 KIA; can be roughly estimated to be 5500 KIA and about 4 times this number of wounded.

Civilian losses due to direct enemy action are not concealed by either side and the estimates of about 5-6 thousand are probably correct.


Recently, the Washington Post published a piece estimating the real inflation of Ukrainian currency to be 272%, 10 times higher than the Kiev government admits to (28%).


Unfortunately, unlike  hard facts about prices and currency exchange rates, there is no easy way of seeing when casualties are underreported, especially when all mass media is controlled by the government and warzone reporting is tightly controlled by military officials.

Video: Ukrainian Journalists Protest Against ATO PR Department


It’s not that true statistics are known and concealed by the government; it’s that at best, they are not interested in them at all, and at worst, working to falsify them.

Here is some evidence that losses are in fact reported incorrectly:

There is of course a plethora of primary evidence from the troops and sources close to the fighting, like this video –

Video: Pro-Kiev volunteer Yuri Kas’yanov on real losses at Debalcevo https://www.youtube.com/watch?v=vHTIKvBMAe8

Transcript: “A lot more people died that the General Staff, Security Council, and the President say… I have people that I trust 100%, that have been there, that tell me things. One of my comrades-at-arms, he just got out of the hospital, he was breaking out with a column of a hundred men. I read the [government] reports – 19 KIA [out of entire 3,000-strong battlegroup]. He tells me out of a hundred of them, only 14 got out. All others are dead…”

Politicians say the same: Congressman Lyashko, who came 3rd in presidential election and heads 5th largest party in Rada, states the government concealed over 8,000 losses by September.

Video: Lyashko’s – Poroshenko burned our boys to hide the losses



Ukrainian majority party MP Olinik states he was told not accepting official casualty figures makes him the “opposition”, so now he has to choose – if he is in opposition to the Poroshenko government, or in opposition to reality

Translated statement:


MP Filatov writing to Poroshenko – “When you are talking about six killed in Debaltsevo, you are ruining the confidence in you and in the state. … Read what the people who are collecting the corpses in the fields are writing. We can handle the truth. Don’t make us live in a lie.”

Translated statement:


All of the above politicians are hardline nationalists and Russophobes, by the way. Opposition politician Shufrich claimed similar figures – one of the few things both sides agree on.

Untranslated, Shufrich claimed real losses were 14,000 by November


There are also hacked government documents saying losses are underreported tenfold, unconfirmed but from a previously reliable source:


Of course under-reporting is widely exposed in various media, even in some pro-Kiev outlets that have a measure of independence:


However, we can not use rebel claims either: most of them tend to be wildly exaggerated when it comes to enemy forces, as is always the case in war, and being split into several loosely connected groups means rebels can not even give a reasonably complete account of their own losses.

Video: LPR brigade commander Mozgovoi on casualties of war

From the video above, famous LPR Brigade commander Alexei Mozgovoi: “Neither side will ever report true casualty figures. Losses are colossal on both sides.”

So, there is obviously a pressing need for a reliable, data-based estimate of the real loss of life in this conflict.


Let’s see see what numbers have been reported so far:

Kiev government estimates their losses around 1550 KIA


(varies depending on source, General Staff tends to quote noticeably higher figures than President and Security Council – although should be the other way around, because General Staff does not normally report paramilitary losses).
UN estimates the death toll at around 5000 people total.



It is hard to get primary data on casualties as bodies are normally quickly removed from the battlefield and buried, usually by multiple sides (UAF, DPR, LPR, “recovery” NGOs, local civilians), and often identity or affiliation can not be determined with any degree of certainty.

Therefore, it is much more reliable to count destroyed military vehicles: they often stay on the battlefield for weeks, their affiliation is easily identified via model, external markings, and position, and they are much more conspicuous and commonly photographed than human remains.

Lostarmour.info is a website that catalogs destroyed and captured vehicles from all sides of the Ukrainian conflict (also some artillery, planes, MLRS, etc., but the collection is not so exhaustive).

It only lists vehicles that have been captured on video/photos and conclusively identified; there are viewable photos of every vehicle, unique IDs, most are mapped, etc.

Links to “destroyed” and “captured” databases



Unfortunately there is no English version (I tried to make a translation for them, they took it, but never had the manpower to put it up and maintain). However, the website is easy to understand, here is the key showing what the various sections are:


*as a side note, under the “Off-Topic” section there is also the “Шушпанцеры” (WeirdPanzers) section, which collects pictures of various civilian vehicles armored for military use and light armor with weird weapons bolted on. 381 vehicles so far – basically looks like “Mad Max:Road Warrior” times 50. Highly amusing:


But for the purposes of this article, we are just interested in the numbers of standard armored vehicles lost by UAF and the Donbass militias, easily counted by anyone using just the flags in lostarmour.info database (be sure to click “show all vehicles” if you just see a few out of ~680 destroyed and ~340 captured).

As of March 3rd, 2015, a total of 627 destroyed UAF armored vehicles were caught on camera, and 76 destroyed militia vehicles. Also, 376 UAF vehicles were captured, while 27 rebel vehicles were captured back.

————  UAF / Donbass militias

Destroyed   627 /  76

Captured       376 /  27

Based on these numbers and data from previous conflicts with more reliable reporting, we can extrapolate a rough estimate of the real number of casualties in the conflict.

To get relevant data from previous conflicts, we need to find similar armies – armed with outdated Soviet equipment and Cold War-era Soviet tactics, poorly led and trained.

Article on Ukrainian army shortcomings:


Obviously, the prime example would be Russian army itself, circa 1993 – first Chechen war; then countries like Iraq and Syria.

First, let’s determine the absolute minimum ratio of personnel losses to armored vehicle losses in a military campaign under these conditions.

Russian losses in Chechnya  in 1993 and Iraqis in 1990 are ideal for that: both armies had a lot more armor than Ukrainian forces (where there is a constant lack of running armored vehicles – e.g. complaint by Poroshenko’s advisor http://www.liveleak.com/view?i=baa_1422989162 ), and a lot better conditions for losing armor – one, by running unprepared into an insurgent army that spent 3 years training and arming to destroy tanks, the other, by facing overwhelming enemy air power in a desert.

Iraqi KIA losses in Desert Storm are variously estimated between 25.000 and 100.000 KIA, with 5-6 thousand armored vehicles lost. This gives us (using middle estimates for casualties) roughly 10 Killed In Action per destroyed Armored Vehicle (termed “KIA/AV”).


Russian losses in Chechnya are not well reported, since both sides tended to conceal their losses and Russians would usually evacuate their destroyed armor from the battlefield; a variety of sources lead me to believe a 15KIA/AV estimate is reasonable, although it could be much higher (though the siege of Grozny involved mass armor losses, the subsequent counterinsurgency campaign did not). The absolute lowest possible KIA/AV in any war is probably the Russian armored/mechanized troops that entered Grozny and were ambushed from all sides; for example, the hardest-hit Maikop brigade reportedly lost 800 men and 100+ armored vehicles, giving us a ratio of 8 KIA/AV.

On the other hand, Syria is a disorganized civil war, much like Ukraine, where both sides often rely on lightly-armed paramilitaries but the government has many armored vehicles, also like Ukraine. Although obtaining good estimates from the disorganized fighting is impossible, it is clear that the KIA/AV ratios are higher: for example, government losses are estimated at 40-80.000, whereas the estimates of destroyed government armor are at around 1800, giving us KIA/AV of about 30.


Based on this, we can now make an estimate of Kiev regime losses.

It is obvious that KIA/AV in Ukraine is higher than the absolute minimum of 8-10, as the government forces are not using as much armor as Iraq-1990 and Russia-1993, nor are the rebels as adept and well-equipped for destroying tanks as US Air Force or prepared tank ambushers in Grozny.

On the other hand, Ukrainian civil war is less violent than Syria – it does not have as much of the sectarian violence aspect (with the exception of a few hundred Jihadists working for Kiev http://www.liveleak.com/view?i=aab_1425325987),  so the crews of disabled or surrendered vehicles are often spared, at least when it comes to the rebel’s treatment of forcibly conscripted government soldiers. Accompanying infantry are likely spared as well during mass surrenders (in video below, government soldiers surrendered 6 tanks and were reportedly allowed to leave riding 2 BMPs).

Video: Surrendered Nationalist Tanks at Starobeshevo


This makes for fewer dead soldiers per destroyed vehicle, and far fewer – per each captured one. Unfortunately, Kiev government usually orders its soldiers not to surrender armored vehicles even when they are fully surrounded (Ilovaisk encirclement, Debaltsevo encirclement, etc.), so more often than not many men die before the rest surrender the heavy weapons.

Video: Breakout Attempt at Ilovaisk (see video description on Youtube for details)

Article: Report from Debaltsevo, plus strategic overview of the winter campaign

Overall, a conservative estimate would be about 20 KIA/AV per destroyed armored vehicle, and 5 KIA/AV per captured one.

Based on just the vehicles in lostarmour.info database by March 3rd, 2015, that gives us a figure of 14,420 KIA for the nationalist forces.

Interestingly enough, this figure is in the same ballpark as the above-mentioned claims of Congressmen Lyashko (at least 8000 “hidden” plus 1000 official casualties by September) and Congressman Shufrich (14000 dead by November).

As not all destroyed vehicles have been filmed and some heavy fighting has happened since the Congressmen’s estimates were given, the real figure is likely somewhat higher by now. The highest estimate that could be considered remotely feasible is probably the “24,000 nationalist KIA” figure put forth by the DPR spokesman Eduard Basurin.
Another interesting estimate was put forth based on the number of heavily wounded evacuated from the frontlines to central hospitals in Dnepropetrovsk and Kharkov, as well as Odessa and Kiev.


By compiling records on the number of heavily wounded soldiers paid for by the army, the author arrives at the figure of about 12,000 UAF medevacs by late January, and at least the same number of lightly wounded who went to local hospitals.

Due to poor medical care, the proportion of heavily wounded to killed is about 2:3, so 12,000 medevacs corresponds to about 8,000 KIA. However, this figure does not include  several thousand injured paramilitary fighters, who are not on state’s payroll, nor does it account for UAF soldiers that were left in encirclements and were impossible to evacuate. We can get an idea about the number of abandoned soldiers from the fact Kiev sources routinely say that 6-8 thousand UAF soldiers are missing in action – KIA likely make up a sizeable part of that (even if the majority are deserters). Overall estimate ends up in the same 12-15 thousand KIA ballpark.

The reason this more direct estimate was not used as the focus of this article is that WIA numbers were apparently obtained confidentially and could not be independently confirmed, unlike the easily verifiable public database of lostarmour.info. The reason that this estimate was still included is that it’s rather conservative, much lower than the rebel claims, and competently written, implying that the author really knows Ukrainian hospitals.

This analysis and official data both imply that on top of ~14 thousand KIA, the human cost also includes 20+ thousand severely disabled nationalist soldiers (and if we include those with less severe disabilities and mental health issues, the number is likely several times greater).

As far as the rebel losses, extrapolating them from the low number of armored vehicles they lost would be a mistake: the rebels had little armor all up until Minsk-1 truce, and even in the winter campaign, tended to use it more cautiously and competently than UAF. Their salvage/repair crews also seem to do a lot better job (as salvage is their major source of vehicles and spare parts), also they do timely retreats instead of being encircled, etc. Basically rebels rarely lose armor, so their KIA/AV can be 50 or higher.

A better estimate can probably be based on UAF losses: although the outnumbered rebels tend to perform better than the UAF due to a combination of higher motivation and better leadership, they still take comparable losses, likely around one-third those of UAF during the summer (when they were mostly defending urban terrain against a poorly trained army) and one-half those of the UAF during the winter (when they were attacking dug-in UAF troops; the losses during the advance were likely closer to 1:1, but eventual victory enabled the rebels to wipe out a lot of the opposing units). Therefore, we can estimate the total rebel losses to be around ~ 14,000 * 40% = 5600 KIA.

Remember that the rebels are outnumbered and have a hard time replacing their losses, so a much higher estimate than ~6000 KIA (and ~12000 disabled) is not logical – simply because the rebel armies are ~35,000 men total, so if they took say 12,000 KIA it would mean more or less every rebel was killed or wounded. Much lower estimates, like official figures of ~1000, do not match the number of lost armored vehicles or widely reported “heavy casualties” in some large battles like Donetsk airport and Debalcevo.

As for the civilian casualties, deaths due to direct enemy action are easy to estimate, as the sides have no incentive to conceal them. Poroshenko’s estimate of “5638 civilians killed’ meshes well with DPR’s Human Rights Deputy data of “2251 dead in DPR alone”” and Pushilin’s figure of “7,000 dead, 80% of them civilians”.

Untranslated links: Poroshenko and Pushilin



One final very important category, excess deaths outside the frontlines, is not estimated here.

However, deaths/shortened lives due to over a million people being made homeless, economic ruin (average pension in Ukraine is has been reduced to about $50/month, and IMF-imposed utility price hikes mean over half of it may go towards utilities), lack of imported medications, huge increases in crime, etc. etc, although less visible, may represent the greatest loss of life in this conflict.


Conflicts not referenced:

USA /Western losses in recent wars are not relevant because the vast majority of the time US was not fighting an armored war, but suppressing local guerrilla movements, not to mention US army’s far better training compared to UAF.

For reference, US losses in Iraq are about 20 KIA/AV (as armored vehicles were rarely used), and 7 wounded per KIA (great medical care).


Moldovan, Georgian, Azerbaijani losses in post-Soviet civil wars, though obviously extremely similar to Ukrainian conflict in terms of training and composition of the sides, are not referenced  because both sides generally had little armor, so armor/KIA ratios would be very high. In addition, much like Ukraine, reporting was extremely unreliable and biased on both sides.


Worth noting that following the success of LostArmor, nationalist activists launched the website LostIvan.com, purportedly to track Russian soldiers in Donbass. Although there are certainly plenty Russian citizens fighting for the self-defense militias, and possibly even Russian soldiers, the website is a poor resource as it is plagued with fakes, insufficient information, and trying to pass off private citizens / retired veterans as soldiers.

Typical “source” for LostIvan:


For example, in the first few listings I just randomly inspected, one soldier’s presence in Ukraine was “confirmed” by the fact he posted one (1) photo of a… pack of Ukrainian cigarettes. Considering the number of Russian troops on Ukrainian border and in Crimea, similar “evidence” can probably be found for tens of thousands of soldiers. In short, they really need a decent moderator team.

On Russian troops in Donbass:

I did not talk about “Russian Army” KIA, as there is no way to distinguish Russian veterans who left the army years/months/days ago from active service members, especially since there are plenty of sources making fake “Russian soldier in Ukraine” social network accounts, pictures, etc.

There is a shadowy force called variously “Northwind”/”organized volunteer”/”vacationers” that could be either organized veterans trained in Russia, soldiers who left the army to volunteer in Donbass, or straight-up Russian army units. The difference may be more of  a legal technicality since Russia (like most countries) doesn’t have a law against citizens participating in wars abroad (only against getting rich doing so, or committing war crimes), so technically even soldiers on leave would be within their rights to go fight for the Donbass militias.

Western volunteers in Syria:


Zaharchenko actually said that a few of the 3000-4000 Russian volunteers in the Donbass are soldiers on leave (interview was widely misquoted by the likes of The Telegraph to say there are 3-4 thousand Russian soldiers in the Donbass http://www.telegraph.co.uk/news/worldnews/europe/germany/angela-merkel/11060559/Serving-Russian-soldiers-on-leave-fighting-Ukrainian-troops-alongside-rebels-pro-Russian-separatist-leader-says.html ).

Another reason is that Northwind, whatever they are, take little losses: They are only committed in surgical strikes against weak points, are much better trained than nationalist fighters, and prefer to fight at a distance. If we go by the widely publicized Dordzhi Batomunkuev’s interview (which was published in an extremely pro-Western source, looks severely altered, and may be fake), his battalion only suffered 4 WIA / no KIA in the battle for Debaltsevo, and considering the scale of the engagement, it would have been the biggest/only Russian unit participating. Overall Northwind losses are probably in the range of several dozen WIA/KIA.


Before anyone gets triggered by this word and makes the inane statement that Russian actions in Chechnya are no different than those of Kiev government in the Donbass, let me point out the basic difference: Russia’s mythical “right” to Chechen soil was never the reason for the first or second Chechen campaigns.

In fact, Russia did NOT invade Chechnya when it declared independence, or seriously try prevent its nation-building.

The war started 3 years later, and the reasons for it were genocide of ethnic Russians in the republic (~20,000 killed, the rest had to flee), crime-based economy of the region preying on Russia, strategic concerns over oil transit and production, and supporting the constitutional government in its civil war against dictator Dudaev.


Also, while Russia’s methods in Chechnya were often deplorable, compare Chechnya 12 years after Russian invasion to Iraq 12 years after American invasion… or even Ukraine one year after nationalist takeover.

The Essential Saker: from the trenches of the emerging multipolar world