Actually this isn't very far from the truth. By one order of magnitude at the most.
Back to the topic: if we argue strictly about modeling issues in CM, reading battle reports is hardly useful. Aside from the exagerrations, what you compare this way is the weapon systems _and_ the combat methods employed. The final results has very little to do with the actual tank vs tank capabilities.
If Stalin had given 30+ turns for the Soviets to take cities I guess the casualty statistics would have been fundamentally different.
IMO penetration tables, and single, verified cases should be examined. Nobody will know what was the reality back then, but probably it can be approximated (a little) better.