He doesn't actually.
This is what it encompasses. It is the total loss figure, military personnel affected by the war no matter in what way, there is no difference between the correlation of German war dead to total losses and US battle deaths to war deaths. Statistical constants exist for each conflict and century, while it is true that there are differences and it might be just a crude method, it can nontheless give us a rough estimate on the losses sustained. It can be defined as the equivalent, as an "upper boundary" so to speak.
You can even calculate it by using the operational losses. Let me demonstrate: US approx. 400,000 war related deaths (all fronts) , which results in 280,000 battle deaths (291,557 to be precisely). German example: 5,3 million as the upper boundary, this leaves us with 3,710,000 irrecoverables (a bit high but 3-3,5 is usually the accepted figure). Now for Ilenkov, the figure of 13,850,000, results in 9,695,000. I'll leave the rest to you.
No, what he refers is supposed to be total deaths. That's quite clear from the book.
Then he is lying, because those are irrecoverables as a consequence of battle (in fact, fallen + missing form the majority of this figure), the number of total deaths is usually above that for any belligerent. You see, the USSR is not a unique case, the laws of warfare also apply to the glorious motherland. A man of his position knows that well, so I have to assume that there are politics involved.
If what he is saying is true, then there are two options: a) the Soviet armed forces performed poorly and could not effectively protect their civilians, b) the RKKA did in fact suffer such incredible losses, which sheds light on their performance.
You see, if they lost 2-3 men for every Axis soldier on the EF and also sustained 4 casualties for every Axis soldier, it would be quite strange to observe a disparity in German (+ Axis minors) military to demographic and Soviet military to demographic losses (independent of the fact that Germany lost the war, factoring in further reprecussions, retaliation, camps etc etc).
It would simply make
no sense. If German irrecoverables resulted in 2,5-3 million for the EF, the Soviet figure
must be n times greater. If German military (and attaché) personnel total deaths resulted in about 5 million (Hillebrand and Overmans estimate) then the Soviet figure would be at least twice, perhaps even three times greater.
Should anyone still have doubts, then realize that about 40% of all wounded are no returns and further 10% can be decomissioned personnel. People might even die in hospitals in 46 or 47. Krivosheev could not even establish an exact figure, even if he attempted to.