infantry bonuses could work in interesting ways if you combine "composition" with terrain situations. Like i mean how tightly packed you are, how much formation you need and how centralized and drilled you are.

E.g. a primitive army is very loosely united, but very easy to recruit. a good shock value but terrible morale. because a loose army is harder to shoot at by a very structured army (in some terrains)& can easily overwhelm one with bad defense, but cant last that long.

i guess i need to define some categories here:

a) structure (or density): the battlefield deployment. think phalanx vs celtic tribesman. firing rows vs spread jungle rifle fire. a well structured battalion can "pierce" but be weak at the flanks if cav dont support it. it can also hit a tightly packed group easier, so a more haphazard infantry deployment means a wider target to aim at, &they can get that shock in. theres also lessened combat width for dense&better flanking for loose infantry. this is the main crux of my suggestion, cuz this also determines success based on terrain.
a less structured army can get around in a mountain or jungle easier, while a denser army will have trouble maintaining composition and footing. so the former gets discipline a penalty and morale penalty while the latter get a morale boost (jungle) or shock boost (mountain). but a denser army will get major bonuses in steppes and hills. this would curb blobs to a degree as well. of course this also depends on regional development but more on this later.

a structured army would have a higher training time than a looser one (formations), but this also ties in to level of westrrnization and access to guns. a newly westernizing nation would have penalties to training time if using denser armies than if using loose, and if they have guns, theres additional penalties on training time.

b) drilling: itd be a way of adjusting discipline at the cost of more cost, more training time. think of it as cohesiveness and unity. a veteran army knows how the other limbs tend to function, and this means better morale (trust) and slightly better discipline. so you can balance a dense, well trained but highly costly army with a loose, quickly raised horde, at least a little. i got a lot i can expand on here but ive written a lot already and dont know if youre interested.

of course, drilling means more manpower usage and higher maintenance. and to send out a well trained army, you need to supply it, leading into

c) logistics: now i know a logistics corp and strategic planning werent really a thing till modern times, but looking at the huge army the persians fielded (or how they tried to intercept alexanders supply route) shows this was a form of it. and besides, armies have to be fed, equipment needs upkeep, yada yada. supply limit is a fantastic way of doing this, and supply range works great. but this should also tie into prosperity and road work and infrastructure and development and yada yada. having scouts to meet with the locals to buy food for examples.

but more logistics means more army reform (tie in to westernization) and more cost, as well as more manpower. thered be attritition, with a denser army suffering more than a looser one, unless sufficient logistics exists. terrain matters once more. with more primitive nations, the further one is from the capital and lower the developement, the lower the supply limit. but combat ability should also be harmed. and since infantry is always the most numerous, theyd be the most affected.i mean, horses eat almost everywhere.

a loose cavalry doesnt really make sense tactically, so that applies to infantry most. drilling levels would have some aeffect on cav and artillery, but ballistics officers definitely werent a thing, so it was more aim and spray.

there could be triggered modifiers for the combinations between the 3 categories, and decisions for choosing the levels. and the terrain based combat modifiers would be generalized rules so its much less work.

im thinking about how to get the a.i. to know how to deal with this. my current leading theory is that the decisions would clearly link to combat traits so that its easier for it to parse it?

sorry for the huge writeup, or if this is in the mod already or in expanded infantry options. i just love jury rigging game mechanics so once i started i just went on for a while.