This week, the independent news outlet Mediazona and the Conflict Intelligence Team (CIT), an open-source intelligence initiative, have raised significant doubts about the validity of alleged Russian military casualty records leaked by Ukrainian military intelligence. The data, published by the Ukrainian Defense Ministry’s project “I Want to Live”, claims that approximately 87,000 Russian soldiers lost their lives during the initial eight months of 2025. The publication of these figures has prompted scrutiny, with both organizations questioning the accuracy of the claims.
The CIT employed statistical analysis methods similar to those used in investigations of election fraud to examine the reported military loss records. Their methodology is based on psychological research indicating that when individuals generate random numbers, they often display unconscious biases toward certain digits. This tendency can become evident in the resulting data, raising concerns about its authenticity.
Statistical anomalies revealed in casualty data
In their analysis, the CIT uncovered several irregular patterns within the data. They observed that the distribution of the last digits of the reported figures deviated significantly from what is expected in a truly random set of numbers. Notably, there was an overrepresentation of the digits one and two, while sixes and sevens were underrepresented. This disparity was further supported by Pearson’s chi-squared test, which evaluates the degree of alignment between observed data and expected statistical distributions. The CIT’s findings indicated a mere 1.3 percent probability that the leaked data was produced by random chance.
When the CIT distinguished between primary (individual unit-level) data and aggregated totals, the inconsistencies became even more evident. In the raw data, the likelihood of encountering such a pattern by pure chance dropped to just 0.4 percent. The CIT surmised that the document’s author likely filled in figures for each unit separately before calculating overall totals. This aggregation process led to a more uniform distribution of final digits, except for zeros and fives.
Concerns from Mediazona regarding the data’s authenticity
The editor of Mediazona, Dmitry Treshchanin, expressed similar concerns about the reliability of the casualty figures presented by Ukrainian intelligence. He noted the possibility that the Russian military itself may have manipulated the records. Treshchanin stated, “It’s hard to conceive of a Russian officer who wouldn’t alter the records.” He further pointed out that glaring gaps within the leaked documents, such as a lack of data on desertions, cast doubt on the legitimacy of the figures.
From a bureaucratic perspective, other irregularities also stand out: the documents do not specify a clear recipient or defined purpose. There are no indications regarding who compiled the records, the time of submission, or where they might have been archived, raising further questions about their credibility.
The path to verifying casualty figures
Despite the skepticism surrounding the leaked data, Treshchanin emphasized that Mediazona and its collaborators cannot conclusively disprove or validate the claim of nearly 87,000 Russian soldier fatalities in Ukraine during the first eight months of 2025. He explained, “We won’t have enough certainty to assess these casualty figures until 2026, likely by next fall rather than at the onset of the year.” This caution stems from Mediazona’s methodology, which relies on a count of deceased soldiers based on published obituaries and estimates derived from public probate records.
The ongoing investigation into the authenticity of the leaked Russian military records underscores significant discrepancies that warrant further scrutiny. As the situation unfolds, the accuracy of these casualty figures remains uncertain, with implications that could resonate throughout the ongoing conflict.
