The problem is unless you can design a system that properly simulates the constant structural changes to the way the Romans organise their armies, you have to stick with one chain of command system.
A regional chain of command, even if it is not entirely historical ( does Africa even have a Magister Militum?), would improve the gameplay and make it more challenging. At the same time, it allows more options for the players who wants to reorganise their armies. If the Roman Empire somehow manage to conquer the whole of Germania for instance, they should be able to create a new Magister Militum of Germania if they really want to.
The organisation of provinces and military command, namely how big is the size of the provinces and how many troops does the field armies have is something that a player can manage. This not only reduce the amount of micromanagement the player has to handle, but also force the players to make trade-off in regards to making decisions.
If I face more enemies attacking me by crossing the Rhine as opposed to the Danube, should I transfer more units to the Gallic Field Army from the Thracian Field Army? Would I be giving too much troops for the Magister Militum of Gaul? Or would I have hold back too much troops for him to sufficiently deal with the issue? Should I use a defence-in-depth strategy? Or should I be deploy all of my strongest units to the frontier?
These kind of decisions can be really fun to make as a player.