Mediazona and Meduza, Russian exile outlets operating outside state censorship, published a statistical estimate on 9 May 2026 placing 352,000 Russian soldiers killed in action by end-2025 1. The methodology is excess male mortality derived from Russia's national Probate Registry, which records deaths and inheritance filings. This is a new method running alongside, and not replacing, Mediazona's existing named-verified list 2.
The two figures measure different things. Mediazona's verified count had reached approximately 209,000 by mid-April at the named-attribution methodology : each death confirmed via open-source records with a name attached. The Probate Registry figure of 352,000 derives from comparing actual male mortality rates in military-age cohorts against the statistical baseline for that population, isolating the excess that cannot be explained by peacetime causes. The gap between 209,000 and 352,000 represents deaths that left no open-source trace: soldiers whose deaths were registered in civilian paperwork but never reported in named records, local news, or social media. Russia's military secrecy apparatus explains roughly 40% of the gap; the Probate Registry captures what the attribution method cannot reach.
The statistical methodology is not new to conflict analysis: Ukraine and independent researchers have used similar excess-mortality approaches for civilian death estimates across multiple conflicts. Applying it to Russian military casualties via the Probate Registry data is specific to this publication, the collaboration between Mediazona and Meduza, and the 9 May release date.
Russia classifies military casualty data as a state secret. The 352,000 figure covers the period through end-2025 only and does not include 2026 losses. The named-verification methodology continues to run in parallel as the attribution record.
