Who’s Keeping Count? How Militaries Estimate Death Tolls


    1.0 Introduction

    On 25 February 2024, Ukrainian President Voldymr Zelensky announced that Ukrainian soldiers had inflicted a death toll of 31,000 since Russia’s full-scale invasion two years prior. Since February 2022, official announcements on Ukrainian losses have been exceedingly rare. 

    The Ukrainian Ministry of Defence has, however, kept an official toll of Russian losses in Ukraine, releasing casualty numbers daily on X (formerly Twitter). As of March 2024, the MoD estimates that Ukraine has “eliminated” 427,840 Russian personnel. [source] The MoD casualty figure does not distinguish between killed and wounded, nor have the Ukrainians revealed the methodology they used to reach this number.

    With the war still raging in Ukraine, these announcements raise some broad questions about how deaths are being counted on the battlefield:

    • How has Ukrainian intelligence reached such a precise number?
    • How accurate are their death toll figures?
    • What is the benefit of sharing death tolls with the world’s press?

    In 1346, people famously depicted Edward III pointing and individually counting the dead at the Battle of Crécy. [source] In the Industrial Age, however, counting the cost is no longer as straightforward as simply tallying bodies after a pitched battle. The modern battlefield, especially with Ukraine, is as complex as it is vast. Industrial heavy weaponry also has a tendency to destroy human bodies beyond all recognition, further complicating counting and identification.

    Edward III counting the dead at the Battle of Crecy, 1346. Painted in 1410 [source]

    1.2 The Task At Hand

    Calculating the exact number of fallen soldiers is, however, an essential moral and strategic task for militaries. Knowing the exact number of fallen soldiers is an essential moral and strategic task for militaries as it allows commanders to measure the success or failure of their actions.

    In an attritional conflict, it is imperative that commanders understand how many losses their actions inflict and at what cost. Here, accurate intelligence is imperative and drives decision-making. However, trying to collect this data on the modern battlefield is complex, perilous and, at times, an impossible activity.

    Obtaining accurate information is further complicated during wartime by political and military actors who wish to manipulate or conceal the true number, either for propaganda or to protect morale. Furthermore, official ‘casualty’ announcements do not always distinguish between killed, missing and wounded. To clarify, in military terms, personnel considered “casualties” are: [source]

    • Killed in action 
    • Wounded in combat who later die of their injuries
    • Wounded
    • Listed as missing
    • Prisoner of war

    Analysts must, therefore, rely increasingly on novel techniques and methodologies to estimate the human cost of war.

    This article will explore:

    1. Why is it important for militaries to keep death tolls?
    2. What methodologies do analysts use to calculate death tolls?
    3. Why do discrepancies arise between different death tolls?

    1.0 Why is it important for militaries to count their own losses?

    While there are obvious moral reasons for a military to count their losses, there are colder and mathematical incentives as well.  

    Leaders and commanders need an understanding of their force strength and composition before decision-making. As conflict leads to deaths and a loss of equipment, commanders need constant updates to make good decisions about where and how to apply their forces.

    “Determining the number of casualties and fatalities suffered in militarised conflicts is important for conflict measurement, forecasting and accountability” Benjamin J. Radford et al, 2023.


    In an attritional fight between two opponents, the party who can inflict the most damage at the lowest cost marks out the clearer path to victory. It is therefore imperative for militaries to accurately measure rates of attrition in order to enable commanders to formulate coherent plans and strategies. [source]

    Combat loss is also an integral part of the concept of combat effectiveness. When an analyst is assessing the capability of a unit, staffing levels and troop quality are powerful indicators of how likely said unit is to achieve a given task or mission. Without an accurate intelligence picture of a unit’s combat effectiveness, a commander cannot reliably concentrate their forces to defend or attack their opponent. [source

    Ukrainian President Zelensky looks over Russian casualties near Bucha. The site was host to a Russian massacre of Ukrainian civilians.
    Ukrainian President Zelensky inspects destroyed Russian equipment near the town of Bucha. [source]

    And finally, claiming to have out-killed your adversary has a strong propaganda value. Since humans struggle to accurately comprehend large numbers, they can only compare death tolls either directly to their adversaries or to other conflicts [source]. Therefore, the adversary can aim numbers as potent psychological weapons at the morale of the adversary [source].    

    2.0 How do analysts calculate death tolls?

    Calculating the death toll among friendly units is, in theory, straightforward. Most units regularly report back their numbers of killed, missing and wounded. This complication arises only when a unit degrades to where it can no longer provide accurate figures, resulting in rough estimates for the numbers. [source]

    However, recording how many casualties the adversary has inflicted is far more challenging. Casualties in long-distance or in-direct engagements are challenging for units to assess, even with cheap surveillance drones: 

    “They may think they’ve killed 20 soldiers when in reality there have only been five deaths. Generally, the units on the ground always believe – in all good faith – that they’ve inflicted more casualties than they actually have.” Cédric Mas, a military historian, explained to Le Monde in 2022.


    Estimating the death toll of a military force, either as an adversary or an outsider, is at its core an all-source analysis exercise. Building an accurate picture requires the collation of intelligence from various collection methods.

    2.1 Imagery Intelligence (IMINT)

    Just behind the frontline lies the first clue for analysts: military cemeteries. 

    In the Russo-Ukraine War, utilised high-resolution satellite images to log the expansion of military grave sites. For example, following the Siege of Mariupol in May 2022, imagery analysts highlighted a sudden increase in the number of freshly dug graves at the nearby Starokrymske Cemetery. [source

    The quality of the satellite imagery enabled the analysts to identify and measure individual grave plots within the cemetery. Using this data and comparing it with images months prior to the conflict, analysts could calculate a low-end estimate for the death toll in Mariupol.

    Analysts have also used similar methods to calculate Russian war crimes.

    2.2 Technical Intelligence (TECHINT)

    Whilst heavy weaponry tends to destroy any evidence of fallen soldiers, the number of destroyed or captured material left on the battlefield is often a good indicator of the ferocity of violence. 

    Military forces during the Vietnam War relied on captured equipment as a more reliable metric for measuring the number of enemy combatants killed. This was especially the case following the 1972 “Body Count Controversy” where a Newsweek article revealed that US units were often killing or including the deaths of non-combatants to inflate “body counts”, with the number being seen as an indicator of a successful mission. For example, during an operation in 1968, they reported killing “11,000 enemy combatants” in action, but later investigations revealed that they had only captured 748 weapons. [source]

    Using captured and destroyed equipment as a metric does, of course, have its limitations, as loss of material does not always equate to loss of life. Nevertheless, material losses and the ferocity of violence are useful metrics when estimating the ratio of killed to wounded.

    Bodies are exhumed near Izyum. Russian troops executed Ukrainian civilians in the region. The exact number of causalities is not yet known in Russian occupied Ukraine.
    Bodies are exhumed near Izyum to confirm the number of Ukrainian civilians killed. [source]

    During the Russo-Ukraine War, data on captured and destroyed equipment is openly available and updated almost daily. In the digital age, it is also now possible to rapidly geo-locate and visually confirm these losses, providing analysts with more reliable and contextual data. 

    For example, since the Russian invasion in 2022, Oryx, a Dutch intelligence firm, has logged 19,000 destroyed, damaged, abandoned and captured pieces of Russian and Ukrainian equipment. Each data point undergoes dating and visual confirmation by Oryx, with either photographic or video evidence.

    2.3 Open Source Intelligence (OSINT)

    While far less glamorous than satellite reconnaissance and capturing weapon caches, openly available information regularly provides the most accurate and contextual information on military losses.

    Despite OSINT appearing to be a recent development, using open information to estimate death tolls is nothing new. Historians of 20th Century conflicts for example have typically relied on archival records such as staffing records and surviving first hand accounts to piece together losses in certain engagements. [source]

    However, without accurate unit-by-unit information, observers of the Russo-Ukraine war have had to turn to alternative databases to extrapolate death tolls. For example, Mediazona, in collaboration with BBC Russian Service, has been estimating Russian deaths by observing excess deaths in Russian men under the age of 50. A method popularised during the COVID-19 pandemic, the measurement of excess death allows analysts to see how many extra deaths have occurred higher than the average that would be expected. [source]

    Mediazona’s methodology relies on tracking the Russian probate registry, a public record of inheritance cases. Based on this data, Mediazona estimated that based on the excess mortality rate in Russian men under the age of 50, 75,000 Russian men have been killed in Ukraine. Of this number, Mediazona were able to identify and name 44,654 individuals. [source]

    A list of notable databases that track global military death tolls past and present [source]:

    • UCDP (1989 – present) [source]
    • Project Mars (1816 – present) [source]
    • Militarized Interstate Events (1816 -present) [source]
    • Correlates of War  (1816 – present) [source]
    • Peace Research Institute Oslo (1946 – 2008) [source]
    • Conflict Catalog (1400 – 2000) [source]

    3.0 Why do discrepancies arise?

    When Zelensky announced that 31,000 Ukrainian soldiers had been killed, he was quick to highlight that an estimated 180,000 Russians had also been killed up until February 2024. Moscow has only on one occasion publicly revealed the death toll of their troops in Ukraine. In September 2022, Kremlin Defence Minister Sergei Shoigu floated the figure of 5,937. During the same period, from February to September 2022, the CIA estimated the Russian figure to be higher at 15,000. The Ukrainians estimated the number even higher, at 55,000. [source]

    From the highest to lowest Russian casualty estimates, the discrepancy is almost a factor of ten. How do we account for this difference?

    3.1 Motivations for Inflation

    As aforementioned, political and military actors have incentives to manipulate death toll figures both up and down. In terms of public perception, the death toll alongside territory gained are two of the foremost markers of military success or failure. With territorial gains scarce, both Kyiv and Moscow have been quick to publicly report their adversaries death toll whilst remaining shy about their own. 

    A 2023 report by PNAS showed that both sides in Ukraine were overestimating the death toll of their adversaries. Russian sources overestimated the Ukrainian death toll at rate of 4.3:1 and Ukrainian sources overestimated the Russian death toll at rate of 2:1. [source]

    A destroyed Russian T-90 Tank
    A destroyed Russian T-90 Main Battle Tank. Destroyed equipment is often a good indicator of the ferocity of fighting which correlates with death toll. [source]

    Russian overestimations, have also translated into underestimation of their own losses according to PNAS:

    “Bias does not necessarily imply intentional misrepresentation but rather any systematic over- or underestimation relative to our estimated loss values. When looking at Russian military deaths, we find that, for every loss suffered, Russian sources report only 0.3 losses. … We find no evidence of systematic bias in Ukrainian reports of Ukrainian military deaths.” PNAS, May 2023.


    The UK Ministry of Defence has suggested that it is unlikely figures are being manipulated by high level Russian officials. Instead a “long established culture of dishonesty” is preventing bad news filtering up the chain of command. This has according to the MoD, left Kremlin mouthpieces with a “low level understanding about total casualty figures”. [source

    These claims by the UK MoD line up with judgements from a leaked 2023 US intelligence report that revealed the FSB had accused the Russian Ministry of Defence of undercounting losses in Ukraine. The FSB claimed that officials were intentionally not counting Russian National Guard units and mercenary groups. This was to bring down the total reported to the Kremlin. [source]

    4.0 Conclusion

    Accurately estimating let alone calculating death tolls in armed conflicts is a complex but necessary task. From a purely military standpoint, death toll has an intelligence value and is key to developing and adapting strategies. Without an accurate understanding of death tolls, a disparity is created between a commander’s intentions and a force’s real capabilities, increasinging the likelihood of mission failure.

    In the public imagination, death tolls are arguably the primary metric for judging wartime success. Unlike territory lost or gained, death tolls are directly comparable to the adversaries regardless of whether a force is attacking or defending. Weaponising the adversaries’ higher death toll is a potent means of destabilising morale and holding their officials accountable.

    Jake Cremin
    Jake Cremin
    Jake Cremin is an Intelligence Analyst specialising in the Russo-Ukraine War and Western Defence. Jake holds a Masters in Intelligence and Security Studies from Brunel University London as well as BA in Military and International History. His research interests are Western Defence, West African Security and Terrorism.

    Table of contents


    Get the weekly email from Grey Dynamics that makes reading intel articles and reports actually enjoyable. Join our mailing list to stay in the loop for free!

    Related contents