I don’t have any real answer that stays within the “without breaking it” part of your question, but one of my long-range future projects (if I ever find the enough spare time to get around to it) is to try to work out a completely different combat-resolution model for the game. The model would take into consideration what types of units (OOB and specialized extra) each player has in the formations which are fighting each other, but would do so differently than the current rules.
Apparently, there are strategic-level wargames (including professional military simulations) which resolve combat by simply comparing the gross force ratios of the formations which are fighting each other, without getting into too many fiddly details about the exact composition of each force. This sounds a bit dull to me, and I would imagine that these methods make the assumption that the two sides are similarly equiped by weapons at the same technological level – an assumption that’s simplistic even if you stick with WWII, given that how much difference there was between the weapons of 1939 and those of 1945. Still, the concept that the two formations fighting each other would be treated (for the purposes of combat resolution) as collective blocks rather than as a bunch of individual units is an intriguing one.
I’ve been thinking that it might be interesting to combine the two approaches, meaning both the individual-unit approach and the collective-block approach. Let’s say that the two land formations fighting each other in a given territory are X (the Allied force) and Y (the Axis force). Using reference tables of some sort, the players would start by calculating the respective total combat power of X and of Y, based on both the number and the types of units that each force is composed of. The tables would be based on the concept that each unit type brings a certain specialized capability to the overall force, and on the concept that the best option is usually to have a combined-arms force that blends these units together in particular ratios that are suited to particular types of operations (for example, with one ratio being optimal for offense and another ratio being optimal for defense). Once the total combat power of X and of Y have been computed, then combat would be resolved on a collective-block basis by comparing the values computed for X and Y, once again through the use of tables (perhaps with a degree of dice-based modification within a defined range, so that there will be an element of luck involved). Depending on the details of the system, the entire battle for that territory could be decided simply in that one step, or there could be a multi-step process in which casualties are applied, then the total combat power of X and of Y gets recomputed to reflect the casualties, then another round of combat takes place, and so forth.
As I said, this is just a rough concept that I haven’t had time to develop (and that I may not have time to develop in the foreseeable future), but I’m mentioning it because extra units such as Arreghas mentions could be worked into such a system; the units would bring particular capabilities that would affect to some degree the computed total combat power of a given force, but once the figures have been computed then the individual features of every unit type in a force stop being considered and it’s only the collective total combat power of the forces that gets used for combat resolution purposes.