Art: Maria Tolstova / Mediazona
For the second anniversary of the war in Ukraine, Mediazona and Meduza jointly release the second part of our investigation into Russian casualties. Using a method we developed six months ago, we reveal the number of Russian military killed in action as of the end of 2023.
Additionally, we offer a comparative analysis of the losses sustained by the Russian and Ukrainian Armed Forces, based on publicly available data.
Since spring 2022, Mediazona, in collaboration with the BBC Russian Service, has been tracking Russian losses using publicly accessible data. Our comprehensive and regularly updated casualty list now surpasses 44,600 names.
We’ve consistently emphasized that this number is merely the minimum known; the true death toll is undoubtedly higher, although previously undetermined. In June 2023, in a joint study with Meduza, we aimed to address this uncertainty. Through a combined analysis of our casualty list and the Probate Registry database, we succeeded in estimating the true mortality rate among Russian men.
On the war’s second anniversary, we applied this method again to assess the excess mortality rate up to the end of 2023. Our findings indicate that, by January 1, 2024, the war had resulted in the deaths of 75,000 Russian men under the age of 50. This figure does not include the military personnel from the units of Russian-controlled Donetsk and Luhansk People’s Republics.
This estimate has been added to our regular summary, which will continue to be updated.
The estimated 75,000 deaths among Russian men under 50 due to the conflict is a statistical figure that may vary, with the actual losses likely ranging between 66,000 and 88,000 men.
Additionally, our calculation method does not include the most recent months’ casualties. If the monthly casualty rate of approximately 3,900 continues, the current total is around 83,000.
The detailed description is included in our first publication, but here’s a quick refresher. Russia maintains a public Probate Registry, where notaries record inheritance cases for state-registered assets such as apartments, cars, land, etc. However, not every deceased person is included in the registry, as not everyone leaves inheritable assets.
In 2022 and 2023, there was a sharp increase in new inheritance cases for men, particularly younger ones. By comparing this surge against long-term trends and the ratio of male to female inheritance cases (female mortality was unaffected by the war), we can determine the number of “excess” inheritance cases.
The next, most challenging step is converting these excess inheritance cases into excess mortality. This is done using a detailed list of casualties. By checking this list against the registry, we can determine how often inheritance cases are opened for deceased individuals from specific social and age groups. For example, if we find that 60% of contract soldiers aged 20 to 24 from the casualty list are in the registry, we infer that the remaining 40% left no inheritance. To account for them, we multiply the number of excess inheritance cases in the period of interest by 1.6.
It’s crucial to emphasize: our method accounts for all excess male mortality up to age 50. The 75,000 figure represents not just those who left an inheritance but all casualties. Our estimation considers all possible factors: the social composition of the deceased, delays in notary consultations, death registration delays, etc. Even those who left nothing behind are counted. For more about our methodology, see Meduza.
By the conflict’s second year, the front lines had nearly stalled, yet the casualty rate remained high, even increasing significantly.
In 2022, characterized by several major Russian military offensives and retreats, including the early assault on Bakhmut, about 24,000 were killed. In 2023, this number more than doubled to 50,000.
The siege of Bakhmut stands as the most lethal single operation of the war. In the peak months from January to March 2023, Russian forces experienced up to 2,000 deaths weekly across the entire front. Nearly half of these were prisoners recruited by the Wagner PMC.
Following Bakhmut’s capture, the death toll decreased but remained above early war levels. A surge in fatalities was observed from October–November 2023, likely due to the conflict around Avdiivka.
Over two years, the composition of Russia’s military forces has shifted significantly. Initially, the invasion was led by professional soldiers. By the latter half of 2022, the ranks had swelled with volunteers, prisoners, and later, mobilised soldiers, who formed a substantial portion of the fatalities.
By the end of 2023, the estimated death toll by group was:
~ 21,000 contract soldiers
~ 19,000 prisoners
~ 16,000 mobilised soldiers.
This estimate of 75,000 only accounts for identified and buried individuals; we still don’t know almost anything about those still missing.
Current obituaries indicate ongoing delays in burial processes. Efforts to collect and identify bodies mean that some previously listed as missing are now accounted for in our tally.
In spring 2023, the State Duma enacted a law streamlining the process to officially declare missing soldiers deceased, leading to a noticeable increase in related lawsuits within military garrison courts, totaling 128 by year’s end.
Assessing the volume of these civil court cases is difficult, but an investigation by IStories suggests over 900 “excessive” lawsuits to declare people dead since the conflict’s start. The full impact of the new legislation, to be reflected in Supreme Court statistics for the latter half of 2023, is expected to be published in April 2024.
Military casualties include not only the dead but also the wounded, for which we still lack a reliable method to estimate numbers accurately. In our June 2023 study, we attempted to gauge the number of severely wounded by examining budget allocations and assessing the killed-to-wounded ratio in specific military units.
Assuming the ratio of “1 to 1.7–2” has remained stable, we estimate approximately 130,000 wounded, bringing the total Russian military casualties—killed and wounded—to just over 200,000. However, this estimate for wounded personnel is much less precise than our fatality figures.
For more details on our estimation methods and answers to potential questions about them (such as how we determine excess mortality from probate cases and account for people without heirs or social ties), see Meduza’s today publication.
In both our previous collaboration with Meduza and regular updates to casualty counts, we’ve seldom discussed Ukrainian casualties or compared them to Russian ones due to a lack of reliable data. Only recently has a source become available that permits a very approximate comparison of deaths between the Russian and Ukrainian Armed Forces.
Ukraine, like Russia, does not officially report its military losses. The most recent figure came from Ukrainian presidential office advisor Mykhailo Podolyak in December 2022, who estimated Ukrainian military fatalities at between 10 to 13 thousand. This number cannot be verified and thus is not suitable for a direct comparison with Russian losses, which we calculated from obituaries and Probate Registry data.
A valid comparison requires data gathered by similar methods, i.e., from open sources.
Several initiatives claim to track Ukrainian casualties. One such project, WarTears, was launched in late May 2022 by the pro-military Russian Telegram channel, Rybar. As of this writing, WarTears reports over 69,000 Ukrainian deaths. However, this database, available for download, lacks external verification links, rendering it unverifiable and unsuitable for analytical purposes.
Another Russian pro-military project, Lostarmour, also doesn’t publish source links; moreover, in their alphabetical list on the website, they have only reached the letter “O.”
A suitable, verifiable database has been compiled by UALosses, an anonymous project launched on December 31, 2023.
The UALosses website is registered in Saint Kitts and Nevis; the only contact information is a Twitter account. Mediazona attempted to contact the project through this account, but as of the time of publication, we received no response.
A volunteer for the Ukrainian “Book of Memory” project Herman Shapovalenko, believes that UALosses is “definitely not fake.” He notes that its creators managed to gather nearly 300,000 links from 3,500 different sources and emphasizes that their project’s data and UALosses’ largely overlap.
“Among the 40,000 entries [of their] first release, there were only about 1,700 duplicates, which speaks to a high quality processing of the primary data,” adds Shapovalenko.
As of the publication of this story, the UALosses list contains over 42,000 people who have died since February 2022 through February 2024. Each soldier has an entry with their name, general information, and links to sources.
To verify the UALosses database, we took a sample of 400 random entries from it. This is sufficient: with this amount, we can assess all 42,000 entries with a 95% probability (plus or minus 5% for error). We manually checked each of these 400 entries to see if the card included a source link and whether the source actually mentioned the death of that particular soldier. Based on this review, we deemed the database reliable.
All 400 entries featured active links; 385 entries (or 96%) referenced Ukrainian sources, including media reports, funeral coverage, local authority announcements, and memorial projects. 14 entries were based solely on data from the Lostarmour site; no confirmations could be found in open sources for these. Additionally, one individual in the database was identified as a deceased civilian, an acceptable margin of error for such a large database.
In assessing duplicates through various methods, we identified approximately 200 to 500 recurring entries, including false positives for individuals with the same surname or namesakes. Given the database’s overall size, this number is relatively minor and can be disregarded, especially since our comparison focuses on weekly rather than total casualties.
The completeness of such databases, crucial for understanding the full scope of military losses, is inherently challenging to assess, especially without demographic data or access to probate case databases, as is the case in Ukraine since the onset of the war. Despite these limitations, there are several factors suggesting the Ukrainian database might be more comprehensive than its Russian counterpart:
— In Ukraine, presidential decrees on posthumous military awards are publicly available and published in text form on official websites, facilitating collection and analysis. Conversely, similar decrees in Russia are classified and not made public.
— Ukraine hosts several significant projects dedicated to memorializing its fallen soldiers, such as the “Book of Memory” and the “Memorial” platform, contributing to a more comprehensive public record.
— Authorities in various Ukrainian regions maintain detailed projects about fallen locals, with local publications often posting comprehensive lists.
— Ukrainian sources tend to provide more complete information. For instance, without referencing the Probate Registry, birth and death dates were found for 55% of people in the Russian context, compared to 75% in the UALosses database.
— Russian local authorities face restrictions on disclosing detailed information about the fallen. Additionally, one of the main sources for such entries, personal posts on VK social network, often lack detail.
Given all these considerations, a rough comparison between the Mediazona/BBC database and the UALosses database is feasible, focusing on entries with known death dates.
This approach yielded 35,349 entries for UALosses and 35,049 for the Russian database, enabling a preliminary calculation of the loss ratio between the two armies. Additionally, in the graph below, we excluded dates after December 31, 2023, due to the inability to assess the Russian Probate Registry (or RND) and the shortfall from open sources.
From the database of Ukrainian casualties, we see that the Armed Forces of Ukraine experienced their highest losses at the onset of the conflict, with a significant peak in March 2022. Losses remained substantial through the summer of 2022, leading up to the Kharkiv counteroffensive and the Russian push towards Bakhmut.
The battle for Bakhmut highlighted a distinct difference in casualties between Russian and Ukrainian forces, accentuated by the scarcity of open data on Russian losses. Notably, Russia deployed a considerable number of prisoners, signed on six-month contracts, for the assault on Bakhmut.
In a February 23 interview with Fox News, Ukrainian President Volodymyr Zelensky suggested a casualty ratio of five Russians for every Ukrainian soldier killed. Even with the assumption that the UALosses database comprehensively lists all Ukrainian casualties—unlikely—and comparing it with the estimated 75,000 Russian fatalities against Ukraine’s 42,000, the derived ratio is approximately 1.7.
It is problematic to ascertain the precise loss ratio between Russian and Ukrainian forces based solely on a straightforward comparison of two open-source databases. Beyond the disparities between the lists, it’s crucial to acknowledge that in addition to the Russian military, units from the LPR and DPR (subsequently integrated into the Russian Armed Forces post-annexation, with little known about their casualties) also engage Ukrainian troops, contributing to AFU losses yet not accounted for in our estimations.
While speculative, it appears that the casualty figures for the Russian Armed Forces and the Armed Forces of Ukraine might be closer than widely assumed; it seems highly unlikely that they differ many times over.
Moreover, the Ukrainian casualty trend lacks the spikes that characterize Russian losses, attributed to “meat assault” tactics and the impact of Ukrainian long-range strikes on troop gatherings, such as the Makiivka strike on New Year’s Eve 2023. Conversely, Ukraine faces considerable challenges in repelling Russian advances, resulting in consistently high and sustained losses for the AFU throughout the conflict.
Editor: Dmitry Treschanin
Correction. The phrasing of the comparison between the RAF and AFU in the penultimate passage has been amended to more accurately reflect the “many times over” («многократно») nuance from the Russian original, as opposed to the broader interpretation used in the earlier version.
Support Mediazona now!
Your donations directly help us continue our work