Some things here about ammo - I'm not keen on the idea of having to withdraw troops in order to resupply, sure in modern day fighting this is applicable but I think the idea of troops doing that in this era when people were fairly slow lumbering and fighting close to the resupply lines etc.
I recently read a book called Zulu Rising, a historical book on the Anglo-Zulu war. In there they describe that somebody would often run back to the QMs in order to collect more ammo (although apparently during this period -- UK Victorian -- it was extremely rare for troops to run out of ammo) and then run it back to the section and restock the men who were fighting.
So from this, I propose that resupply rate should act as a function distance... Meaning, from the company fighting and their distance there from the QMs trailer/wagon. Fairly close and you can expect a constant supply of ammo, but if they are further away, then there could be periods of time when the runner/runners fail to make it back quick enough to collect ammo and troops run dangerously low, maybe even so low not everyone is able to fire, and maybe one or two cycles of not everyone firing, followed by unable to fire.
If you're really ambitious, then maybe you could model the runner on screen running back to collect ammo and then making the journey back to the regiment/cohort he's part of, but I think getting the model for resupply right is more important (and mines only a suggestion btw).