As I said, I really like your solutions... but my fear is that you're right about another thing: we'll see all this in EU15 or somewhere around it. I just hope that the devs are reading and are thinking of implementing at least something from this.Solution 1)
For each type, we should choose one implementation like in EU3 - just there are more slots. This would be beneficial because differing income/manpower balances could be reflected in the army composition. A big empire with no money to spare would field mostly light infantry (but lots of it), while a small trading state could afford heavy infantry - but not much of it.
- Early game - late game (with countinuity unless noted otherwise)
- Heavy infantry - formation (line) infantry
- Light infantry - skirmishers
- Pikemen - extinct (or grenadiers, countinuity dubious)
- Heavy cavalry
- Light cavalry
- Siege artillery
- Field artillery (only appears with houfnice, becomes serious with leather cannon)
We should get an number of slots to choose unit types. The number of available slots should slowly increase with time (land tech), from 3-4 at game start to ~8 by the endgame. For each slot, we could choose one unit type from all available, like in EU3.
For example, at game start, for each slot I could choose (from EU3 list): latin infantry, halberdiers, latin knights, chevauchée. Minus any already chosen for a different slot. This would obviously necessitate that different units have different costs, but would again allow different empires to field different armies - and probably even going so far as to having multiple types of the same group (formation light infantry and skirmisher light infantry, if really large but poor) which is the point of both systems.
Important for both
There should simultaneously be another unit categorization.
This would be another great factor in limiting conquests to more logical places and extents. Not only would it bring out the importance of centralization, but would also give the defenders an edge over the aggressors, especially in the early game (when it is needed the most).
- Feudal retinue (also called Banderial)
- Professional (standing army)
- Conscript or draftee
- Levies: mostly light infantry, hardly benefits from technology. Available in great numbers, but gives a serious WE strain if raised for longer periods. Its role is to prevent conquering too fast and far, and to keep borders more or less the same in the early game.
- Retinues: a buff for decentralized aristocratic states, mostly heavy troops. Basically free, but available only when the nobles are content with you. On the other hand, should you anger the nobles, this force is not only unavailable, but most likely ends up as the core of a particularist rebellion.
- Mercenaries: the troops of choice for a king with much money and no other choices. Varied in composition, salary, quality and loyalty. Generally on par with contemporary professionals, just more expensive and fickle.
- Professionals: high-quality and utterly loyal, the force all autocrats dream of. On the flip side, availability is highly dependent on administration (up from a thousand or so retinue in CK to the multiple tens of thousands by late EU timeframe) and keeping them in arms is costly (but still not as much as mercenaries are). Force-limit dependent on the bureaucracy (and monetary concerns).
- Conscripts: a late-game phenomenon. Similar in composition to the professional army, just even more determined by the administration. However, this is again a force which can be summoned up and dismissed as necessary - the last one with this property were the levies, obsolete since long.
I know this seems to be intimidating, but this is the simplest way to represent military matters in EU4 even remotely accurately.
To sum up, each 'unit' in an army should have two qualities: its equipment (whether it is heavy infantry or light cavalry) and what kind of troops make it up (mercs or levies). This is enough to simulate warfare well and also to guide the game down a believable path - one which is immediately acceptable as realistic in outcome, as well as devoid of 'arbitrary' modifiers.