At least 47,000 Russian soldiers killed in Ukraine. A data investigation by Mediazona and Meduza

The Russian state instigated an aggressive war and now conceals its mounting cost from its own citizens. Military casualty data is classified, with those daring to report on the number of fallen soldiers are prosecuted for spreading “fake news” about the army.

The last time the Ministry of Defence disclosed any casualty figures was in September 2022, when Sergei Shoygu announced that a total of 5,937 soldiers were killed in action. This figure is in stark dissonance with reality, as by that juncture, our own list of names, compiled through publicly available data, already surpassed the official count.

In any military conflict, each side is predisposed towards exaggerating the losses of the enemy whilst downplaying its own. Therefore, accepting the Russian army casualty estimates, as offered by Ukrainian officials and Western media citing intelligence sources, proves problematic due to the inherent challenge of verification.

Mediazona, BBC News Russian, and a team of volunteers have have been meticulously maintaining our own list of fallen soldiers since May 2022. We gather data from social networks, local press reports, photographs of cemeteries and memorials. As of June 30, 2023, the count tallies at 26,800 names—and we know each individual name.

Clearly, these figures do not account for all losses—many deaths remain unpublicised. This was repeatedly confirmed by our volunteer teams visiting cemeteries, discovering new casualties not mentioned in obituaries, news reports, or even on social media. Furthermore, in early summer 2023, following our persistent updates of the casualty count, the Ministry of Defence requested regional Russian authorities to limit the publication of casualty data.

As we continued with this project, we utilised various sources, including the Federal Notary Chamber’s National Probate Registry. This is a public database, maintained and populated by notaries, recording information on inheritances that require state registration—such as cars, apartments, and parcels of land.

To locate a record in the registry, one needs to know the person’s full name. Usually, we didn’t use the register to search for new casualties, but rather to supplement and refine the existing data. A registry entry serves as unequivocal confirmation of a person’s death, offering precise details about their age and exact date of passing.

Now, in collaboration with our colleagues at Meduza, we have found a way to utilise this register to estimate the actual number of casualties: approximately 47,000 men under the age of 50 by the end of May 2023.

How the National Probate Registry helps uncover war casualties

The National Probate Registry (РНД, or RND in Russian) was launched on July 1, 2014. Public access to probate case data was made available in 2018.

We gained access to 11 million entries of probate cases opened between July 2014 and May 27, 2023. While this does not cover the entire registry, this is not necessary for our analysis: such a vast sample size makes it representative enough.

We compared the number of opened probate cases in our sample with the complete mortality data for pre-war years in Russia. On average, probate cases are opened for 30–70% of deceased individuals. This proportion increases with age, as over a lifetime, people gradually accumulate assets that may be claimed by heirs.

The substantial number of probate cases opened for minors can be attributed to the common practice in Russia of registering property in the names of children.

The main premise of our research is that records in the Probate Registry can be used to ascertain the overall mortality level. That is, we need to calculate the ratio between the number of entries we found in the registry and the number of deaths during the same period within the same age group.

By determining this ratio, we can learn about the excess mortality of Russian men in 2022 and 2023. When calculating it, we needed to consider not only the age of the deceased (as we have already established: the older a person is, the more inheritable assets they possess), but also other factors—for instance, social status.

Our calculations reveal that the deaths of prisoners, who constitute the largest group of war casualties, lead to three times fewer probate cases being opened compared to military personnel. At the same time, military personnel tend to leave assets to pass on slightly more often than their peers not associated with the army: we assess that it can be primarily attributed to military mortgages. We take all this into account.

The following spoilers outline our methodology, first in a simplified manner and then in more detail:

Sample: young men aged 20 to 24

Here is h0w we applied our method on real data about probate cases for men aged 20 to 24. We chose this group because the impact of the war on them is particularly significant—young people rarely die in peacetime.

Before the full-scale invasion, between 2014 and 2021, this group saw an average of 11±4 probate cases opened for women and 32.8±9.4 cases for men across Russia in a single week. Male mortality in this group has always been higher—this is known from demographic data.

In the early weeks of the war, the number of probate cases sharply increased. From February 28 to March 6 (the first calendar week after the start of the war), 117 men in this group died, and probate cases were subsequently opened for them. This is almost four times higher than the trend of previous years.

Meanwhile, in the same week, we found only 16 cases for women in our sample—indicating a significant shift in the ratio of cases between men and women.

Taking into account long-term trends, seasonal fluctuations, and the number of cases for women who died that week, in peacetime, we would only expect about 33.8 probate cases for men. This means that in one week, we get over 80 excess probate cases for young men of the appropriate age.

We know that after the death of military personnel of this age, probate cases are ultimately opened in approximately 60% of cases. Therefore, the number of excess cases should be multiplied by a coefficient of 1.73. As a result, we get an excess mortality for men aged 20–24 in that week of 155 individuals.

Since we’re talking about the very first days of the war, no adjustment is needed to account for the proportion of prisoners—there were none in the Russian forces at that time.

In the same way (but with an adjustment for prisoners from late summer 2022), we calculated the total excess mortality for all age groups of men from 15 to 49 years old.

By the end of 2022, we see approximately 25,000 excess male deaths, and by May 27, 2023 (the latest date for which we have data), there are 47,000 excess deaths.

Here's how the excess mortality calculated from the probate registry correlates with the obituaries we found during the same period. Please note the delay between the actual death and our identification and processing of the related reports.

Rosstat data confirms our findings

The method of calculating mortality based on the probate registry may seem overly complex or unreliable—but there’s a way to verify the findings using official data.

The main source of mortality data in Russia is Federal State Statistics Service (Rosstat) which receives and aggregates information from civil registry offices (ЗАГС, ZAGS). Over the wartime period, Rosstat has been publishing updates on overall mortality every month. The data can be used, for example, to calculate excess mortality during the pandemic. However, we can’t use it in our count, as Rosstat does not break the data down by sex and age, and military losses become “invisible”.

We sent an information request to Rosstat and in June 2023 received information on the total number of deaths for each year, including 2022, broken down by sex and five-year age groups.

Due to the impact of the COVID-19, it is impossible to calculate excess mortality by comparing 2022 with 2021: 2021 was the peak of the pandemic in Russia. By the beginning of 2022, it had not yet ended, and some of the increased mortality that year may also be related to the coronavirus.

To address this, just like when analysing the probate registry, we calculated the ratio of male deaths to female deaths.

The chart below clearly shows an abnormal surge in the ratio of male to female deaths in all age groups from 15–19 to 45–49. There is no such anomaly in older groups, with the ratio for 2022 simply returning to the pre-COVID level.

Calculations based on Rosstat data for 2022 (no later records available yet) give us 24,000 excess deaths among all men in the 15–49 age group. This almost exactly equal to our probate registry estimate (25,000). Rosstat yields a slightly higher estimate in younger age groups, and the National Probate Registry does so in older ones.

This means that the method based on interpreting the data from the registry allows us to correctly estimate excess male mortality. And if it gives fairly accurate results for 2022, it can be used for 2023, which is not yet covered by Rosstat.

Comparing both methods of calculating excess mortality (Rosstat and the National Probate Registry) with our casualty count shows how much of the real losses we can see in open sources.

By the end of 2022, Mediazona, BBC Russian, and volunteers found the names of 14,119 casualties—that is, about 58% of the 24,000–25,000 excess deaths.

This proportion depends heavily on age: for example, in the 20–24 age group, the count from open sources covers almost all the casualties, while in the 35–39 age group, it falls below 50%.

Therefore, if we try to multiply our count number by some coefficient to figure out the true number of losses, we have take the age into account.

Casualties are not just the deceased. Here’s what is known about the missing

Far from all the deceased make it into the general mortality statistics. If a soldier’s body is left on the battlefield, they will be considered missing in action—until the body is retrieved or exchanged, or until they are legally declared dead. Until that moment, their death will not be registered by either the civil registry office or the probate registry.

Ukrainian officials have repeatedly stated that they have “tens of thousands of bodies” of Russian soldiers that no one wants to retrieve—however, there is no independent confirmation of this.

In Russia, no open sources exist that would allow even a rough estimate of the number of people missing in the war. In Ukraine, for example, there is a registry of the missing, and in June 2023, it listed 23,000 people; most of them are military personnel.

The military themselves know the full data on the missing but this is not published. Nevertheless, some commanders’ reports with data on the missing in action have still made it into the public domain.

The most extensive set of such data is the “trophy” paperwork of the 1st Guards Tank Army (1TA), published by the Main Intelligence Directorate (ГУР, HUR) of Ukraine in May 2022. The intelligence officers posted a scan of a document compiled by Russian military listing the losses of the 1TA in the first weeks of the war (until March 15), including the wounded, missing in action, and those who surrendered.

The list counts 61 people as deceased and 44 as missing in action. One soldier from the list of the missing turned out to be alive—he was wounded and taken prisoner.

In the probate registry, we found half from the list of the deceased and a quarter from the list of the missing. Assuming that all the missing, except one, are actually dead, it means that Russia has not received about half of the bodies—and the civilian mortality statistics still know nothing about them.

Such an estimate cannot be extended to other units, let alone the entire front. The number of missing grows unevenly, and the rates change depending on the nature of the battles. In the spring of 2022, the 1TA was almost surrounded near Sumy and was forced to retreat—this is probably the reason why many bodies were left on the battlefield.

For the missing from the 1st Tank Army list who also appear in the probate registry, there is a huge delay between the actual time of death and the time of its registration. This suggests that their bodies were received during a later exchange with Ukraine.

Such delays occur for “peacetime” deceased—people go missing all the time. But in 2022, a surge of such cases was found in the registry.

We studied the frequency of such delays in the group of deceased men under 50 before the war and after its start (as usual, comparing data on them with the same age group of deceased women). It turned out that in 2022, with a delay of more than 128 days, 3,000 more deaths were registered than should have been according to the pre-war trends and compared to the cases for deceased women of the same age. For example, with a delay of more than 80 days, 10,000 more young men were registered as deceased than in peacetime.

Obviously, this is connected to the war—but it is far from certain that all these deceased were once missing in action. Delays in registration can also occur when identifying bodies.

The numbers above cannot be considered our estimate for the missing as it is impossible to calculate it this way. Our main conclusion is that some part of the missing, whose bodies are subsequently evacuated or exchanged, is still taken into account by our count.

In April 2023, the State Duma passed a law simplifying and speeding up the legal process of recognition of a missing serviceman as dead. So far, there are no signs of the application of these amendments: courts records do not show any increase in the number of such cases either in 2022 or in 2023.

What we know about the number of wounded?

Knowing the number of soldiers killed in war, one can estimate the number of wounded. It is most reliable to calculate wounded-to-killed ratio for conflicts that have already ended, provided that the loss data is accurate. For ongoing hostilities, it can only be estimated approximately, and there are several methods for this.

Method 1. Historical data

The most common estimates of the ratio of wounded to killed for industrial and post-industrial wars are “3 to 1” (or 3.5 to 1), that is, three wounded for every single killed. This average ratio was derived from an analysis of human losses during the Second World War and other wars of the 20th century.

The more recent the conflict, the higher the number of wounded and fewer the number of killed. Researchers explain: developments in the battlefield medicine, enhanced evacuation protocols, and improved individual protection made wars “less deadly”. Servicemen receive fewer fatal injuries, and medics are able to save more lives.

Christopher A. Lawrence, military analyst and executive director and president of the Dupuy Institute critisises the application of the “3 to 1” estimate to the Russo-Ukrainian war. Based on reports from the authorities of the self-proclaimed DPR and the statement by Ukrainian President Volodymyr Zelenskyy about the daily losses of the AFU, he assumes that the number of wounded on both sides should be higher—at least “4 to 1” or “5 to 1” in the case of Ukraine.

Method 2. Studying the losses of specific units

In the aforementioned list of losses of the 1st Tank Army, which was published by the Main Intelligence Directorate of Ukraine, 204 wounded and 60 killed are mentioned. This means that the ratio of wounded to killed is approximately 3.4 to 1.

It must be noted that there are dead among the 43 people who are listed as “missing in action”—and at the same time, some of those taken prisoner (there are 96 in total) could have been wounded in battle. In addition, the 1st Tank Army fought in conditions where evacuation and assistance to the wounded were difficult, i.e., the ratio should be lower than “usual”.

We also studied personnel documents of the 205th Separate Motorised Rifle Brigade. According to our calculations, from June 1 to September 20, 2022, the approximate losses of the brigade were 14 killed and 56 wounded; their ratio is 4 to 1.

Method 3. Counting the severely wounded by pay data

To understand the overall situation, it’s crucial to not only consider the ratio of wounded to killed, but also the consequences of injuries.

According to a research paper published in The Journal of Trauma, Injury, Infection, and Critical Care, over 50% of all wounded American servicemen in Afghanistan and Iraq returned to duty within 72 hours. Overall, based on injury statistics during various wars, about 75% of the wounded return to duty within 20 days.

There is also a group of severely wounded who leave service due to their injuries. In the context of war, this group is particularly important as it, along with the killed and missing, affects the army potential.

We can roughly estimate the number of severely wounded who left service due to injuries judging by the sum of compensations to the military from the state budget. This estimate is possible because we know the number of the dead.

There are several types of compensations to the families of the deceased, as well as to the wounded in Russia. Our focus is on payments established by No. 306-FZ law. This is not the same as the one-time payments of 5 and 3 million rubles introduced by Vladimir Putin’s decree after the start of the war: they do not specifically highlight severe injuries, plus the data is classified.

The 2011 law specifically mentions the deceased and those discharged from service due to injury. The amounts of payments under it are updated every year; in 2022, they amounted to 4,452,696 rubles for the family of the deceased and 2,968,464 rubles for the wounded.

Money is transferred to the region of the military unit where the deceased served. Until August 2022, regional branches of the Treasury published monthly reports on Budget Execution and Cash Management—then they stopped, presumably after “Novaya Gazeta Europe” attempted to estimate the scale of losses using these numbers. However, monthly data from March to the end of July is still available.

Sergei Krivenko, director of the Citizen.Army.Law human rights group, told Mediazona and Meduza that it takes at least a month to pay compensation to the families of the deceased, and at least two months for those discharged from the army due to injuries.

However, data from the treasury website indicates that some families of the deceased and wounded received payments as early as March, suggesting that certain compensations were issued without delay. Notably, in the first month of the war—a period of significant losses for the Russian army—a comparatively modest sum of 12.2 billion rubles was disbursed for compensation. This figure rose to 23.1 billion in April.

From the start of the war until July, a total of 85.6 billion rubles was disbursed. In the same timeframe, as per the National Probate Registry, approximately 9,000 men died in the war. Their families could have received around 40 billion rubles in compensation. This implies that the remaining 45.6 billion rubles would have been allocated to the wounded, with each receiving about 2,968,464 rubles.

This method allows us to estimate the number of severely wounded in that timeframe to be around 15.2 thousand, yielding a ratio of approximately 1.7 wounded to every 1 killed.

However, we lack precise data on the proportion of the wounded and bereaved families who had received their compensation by July’s end. Therefore, any ratio, including the 1.7 to 1 figure, should be approached with caution.

By adding the estimated number of severely wounded (calculated using this ratio) to the number of fatalities (47,000), we arrive at approximately 125,000 individuals in irrecoverable losses.

A joint team of reporters and editors from Mediazona and Meduza worked on this piece. Due to Russia’s military censorship laws, some authors are not willing to reveal their names.