Suggestion for the battle engine:
As the engine as it is now does not perform sometimes, and from what I understood is still a bit simplistic (i.e. firing ranges etc). I would suggest an overhaul and using one similar to that from Steel Panthers: World at War. Sure, you cannot model all the troops, but say, 5% usually would be enough (so instead of 1000, you get 50). Get a random map, in line with the terrain, simulate entrenchment, auto-deploy, run ai vs ai, and voila. With modern computer this should be quite all right to run in a second or two. And you have a *very* reasonable engine beneath the game. And finally all the micro-stuff about guns used etc would be of real use.
As Gary has been involved in SP:WaW, the engine itself should not be a problem, surely it lurks somewhere. Of course, the simulation in SP:Waw is not perfect, AI is not perfect, and so on but it is much better to what we have, and I believe much more realistic. Also, incorporating leader skills etc should not be a problem in that... One issue I can see is with unique guns (especially very big ones), which are very few and would not be caught into the 5% easily. Still, one can make those available for a given % of turns, so if the battle is 20 turns long in SP:WaW, say the siege mortars would be available for firing once... Other, non-unique guns would be more or less already well defined...