Replenishment of the Imperial Army and losses.

[Sean Manning]

Aryaman's views seem correct, from what I know.

Thinking about factors affecting attrition ...

Troops who are regularly paid and content, have solid permanent shelter, are used to the local climate and diseases, and are in a basically peaceful area will suffer very few losses. Armies lacking these things can suffer losses almost as fast in peace as in a major war. I've seen some interesting figures from the wars of the Hapsburgs in Flanders in the sixteenth century somewhere. The Imperials suffered appalling losses in peace and war.

5% of losses, it's not so much.
50%/75% of losses for the losers would be possible if the army is entirely routed?
Cause most of the casualties were caused not in the battle, but during the rout.

Well, 10-15% losses are common for a looser who is not pursued, if I recall VDH's figures for fifth century Greek battles. 50-75% would only happen if one side could not retreat and was deliberately massacred. On the other hand, badly motivated troops often desert after a defeat, so a loosing army could suffer many more "losses" than "casualties" (dead and wounded).

A Dr. Walter Schiedel of Stanford wrote a paper on the demographics of the Roman imperial army. I think is is available publicly online, but I never read it properly after downloading it.