Estimating Casualties in the Ukraine War

[Vendetta]

Copying over some content from discussions in my own private group chat about what the real casualty figures are from the ongoing war in Ukraine. The Russian government isn't telling us anything useful (their last official update was in September and it acknowledged only 5,937 KIA), while the Ukrainian government is telling outrageous lies (as of this week they claim they've killed over 148,000 Russians). Ukraine's allies in the UK Ministry of Defense, a country with a sterling reputation for honesty and integrity in wartime propaganda, tell us the Russians at 40,000-60,000 killed and 175,000-200,000 casualties overall. Now we've all been raised on the story of the Three Bears and we know if one thing is too high and one thing is too low, then the middle thing is just right. NATO's claim is therefore the most believable. Or is it?

We've got a war going on between two groups of professional liars, but one band of amateurs seems to have missed the memo. Or rather, received it eleven months late. From February 24 up until December 16 of last year, the Donetsk People's Republic was regularly releasing casualty figures for its own forces. These were very strange figures, strange because they actually looked believable. After 12/16, someone was clearly given the order to stop doing this. On this final day of honest reporting, casualty figures for the DPR Armed Forces stood at 4,163 killed and 17,329 wounded.

So what can we do, knowing this figure for a small allied force and assuming it to be reasonably accurate? First, we can use it to infer something about Russia's overall killed to wounded ratio. For the Donbass, it is 4.16 men wounded for every 1 man killed. We know that the DPR troops are not as well provisioned or equipped as the Russian regular forces, and therefore will have more difficulty providing first aid to wounded men and evacuating them for treatment. Therefore it seems reasonable to assume the Russian killed to wounded ratio will be higher than 4:1 as well. How much better? No idea, but at least we have a floor.

Next, we've got an exact range of dates for these figures, 2/24/22-12/16/22, a span of 295 days. We can therefore calculate an average casualty figure per day. These figures come out to 14.11 killed per day and 58.74 wounded. We'll round KIA down and wounded up to make it even, 14 KIA and 59 WIA. We know the DPR forces have spent the entire war fighting in the Donetsk region, which has been the single most continually active front in the war. We know the DPR troops are comparatively underequipped, and we also know that they have been heavily employed as shock infantry while the Russian regulars provide them with supporting fire. All of this strongly suggests the DPR forces have suffered a higher casualty rate than the regular Russian forces.

So we also have a ceiling figure for overall Russian casualties…if we can figure out the ratio between the size of the DPR forces and the size of the Russian forces in Ukraine. Can we? We have some estimates to go on. They're vague and therefore we can't put too much trust in them. But conveniently enough, our cutoff date in December is before the new wave of mobilized troops began to arrive at the front at the end of the year. We can therefore use the pre-mobilization estimates, which were around 200,000 for the regular Russian forces and 40,000 for the Donbass militias. Conveniently, that's a 5:1 ratio.

Let's go with that and plug in those daily casualty numbers. Multiplying by 5, we get 70 KIA per day and 295 WIA. Then we'll multiply those by days of war (let's do the whole thing, not just until December, so 369 days): we get a total of 25,830 killed and 108,855 wounded. Now we add back in the DPR totals, expanded using their daily rates to cover the 74 days since they stopped reporting (5,207 killed and 21,676 wounded with that adjustment). And we add the LPR, who never reported their own figures, but whose losses BBC has estimated as 25% of the DPR's (thus 1,302 killed and 5,419 wounded). Grand total of Russia + DPR + LPR: 31,889 killed and 135,950 wounded.

That is if the Russian forces have been losing men at the exact same rate as the DPR. As I said before, I have reason to think their loss rate is lower. How much lower? No idea, but now we have a ceiling for a maximum reasonable estimate. But can we believe it? We are making several large assumptions with this math. It would be nice if we had a second source we could extrapolate from and see if it lands in the same ballpark.

Well, there is an ongoing BBC Russia and Mediazona project that counts soldiers' funerals through publicly available information in local newspapers and social media posts. As of this week, they confirm 15,136 deaths by name. So that gives us an an absolute minimum as well. Of course, they don't think they're actually finding all the funerals - but they have no way of knowing how many of them they're missing.

Can we guess? We do have one interesting bit of data: the men who disappeared in the sinking of the Moskva. Russia officially declared one man dead and 27 others missing. 17 of those missing men have since been confirmed dead through funerals. No word yet on the remaining 10. Can this tell us something about the overall ratio of funerals Mediazona is managing to find? Maybe, I don't know. What happens if we try it and plug in 18 out of 28 as the ratio? Multiply that by 15,136 and we get a total of 23,545 dead. If we add the Donbass dead that I calculated before, and extrapolate the figures for wounded using those same ratios from before, we get a total of 30,054 dead and 125,103 wounded.

I will then take the average of those figures I arrived at by the two different methods and arrive at a single overall estimate of 30,972 dead and 130,527 wounded (for a combined total of both Russia and the Donetsk republics). So how does this stack up against the figures NATO has been releasing for public information? The British Ministry of Defense gave their figures as ranges, so I will average those out as well. They claim 50,000 Russians killed and 137,500 wounded (subtracting their figure for KIA from the one for overall casualties).

So their averaged figure for Russian wounded is only 5% higher than mine, which falls within the range of their estimates. However, their figure for KIA, which is the really important one, since around 80% of wounded troops eventually recover enough to return to battle, is over 60% higher than mine. If we add 20% of the wounded to the KIA to arrive at a total for irrecoverable losses, my estimate is 57,077, and theirs is 77,500 - their figure is 36% higher. As for the numbers the Ukraine gives, just for a laugh - they claim they've killed 148,690, which is 480% higher than my estimate. Their figures are a joke.

So the single largest mistake that NATO officials seem to be making, if they actually believe these numbers they're telling us and they're not intentionally sowing disinformation, is that they are not getting the Russian killed-to-wounded ratio right. Going by the average of those British numbers - 50,000 killed to 137,500 wounded - they claim the Russian ratio is 2.75:1, which is worse than most World War II armies, and a death-to-wounded rate 33% worse than that of the DPR, who we know to be under-resourced in comparison to the Russian Army and should therefore be suffering a worse loss rate.

As for the Ukrainians…no idea, haven't gotten any useful leads to work from like there were on the Russian side. All I can say with confidence is just as high as Russia's, if not even higher. Artillery kills the most and the Russians have a lot of it. What's the highest believable number? If I had to guess, I'd take that 480% they were overestimating Russian KIA by and apply it to losses they admit for themselves, which are 10,000-13,000. Average is 11,500, times 4.8 equals 55,200. A total ass pull, but it sounds reasonable enough as an upper limit (1.78 Ukrainian dead for every Russian, anything 2:1 or higher I would certainly doubt). If we take an average of that as the upper limit and the Russian KIA estimate of 30,972 as the lower limit, we would get a Ukrainian KIA of 43,086. And if we applied the same killed to wounded ratio as we used on the Russian side, we end up with 177,170 Ukrainian wounded. Again, I emphasize that the methods I used to arrive at an estimate for Ukraine are an ass pull and should not be taken too seriously, but if you put a gun to my head and forced me to make a guess, that's what I'd give you.

[Belushi TD]

Ehhh....

It seems to me that this guy is trying too hard to underestimate the casualties. However, what the hell do i know?

Belushi TD