I've gone through these kinds of exercises before. The best way IMO to determine if there is a problem is to set up a small scenario with a mix of units and have them attack each other. Have the scenarios run only a few turns. Check the TOAW log file at the end of the scenario, and keep records of losses of whatever aspect you are looking at. Maybe do five to ten runs like this.
Then alter the small scenario. Increase or lessen the number of tanks and run five to ten more iterations. Again, keep records after each run terminates.
It is a tiring process that is not much fun but on occasion interesting in terms of what turns up.
At any rate, once you've produced enough data, you may be able to reach tentative conclusions on if there is a problem and have a better idea of what it is, should it truly exist. Then, present your findings in a thread for comments.
My own test scenarios have not looked at armor losses specifically, although I did one in which swarms of SAM vehicles were able to deal hard losses to World War II tanks on the basis of RBC (retreat before combat) actions. I pinned that result down to a too-high defense strength for those vehicles. I've tested armor losses against different kinds of infantry antitank rockets, and better infantry AT rockets killed more tanks. Note all these tests were in good weather and clear terrain. Anything approximating a regular scenario will introduce a fantastic amount of variation in the results because of terrain, weather, supply, etc.