So it seems to me that you see the value of leadership, you just want it by default.
You're partially misunderstanding me. If tactical combat were not going to be in the game - if battles were going to play out like they do in GC2, where you have no control of the combat itself, then a leadership skill which improves your troops' combat AI would be a creative and innovative way of handling leadership. However, the moment you add actual tactical combat into the game, so the player has the choice of taking control of his troops directly, what was once a good idea turns into a terrible one. If the player has a hero with poor leadership skills in charge of an army, he would be down-right stupid to auto-resolve combat. The choice becomes the following: Do I auto-resolve, letting a deliberately stupid AI control my forces, or do I step in and take control myself, knowing that I am at least a good as even the best combat AI? The answer is obvious, and it would essentially force you not to use auto-resolve unless you have a high-leadership hero. Basically, it forces you to put in work and probably sacrifice to make auto-resolve usable. And auto-resolve should be usable from day 1, and with results as comparable to human control as possible.
Maybe instead of leadership points, a hero can earn leadership experience. For instance a new hero might only be able to effectively lead ten men. It doesn't matter if he is in an army of 100 or an army of ten, he is only actaully "leading" ten. After leading ten for a while and a few battles, along with XP he gains LP and is now able to lead 20, then 40, 80, 150, 300, 600, 1000.
What do you mean by "effectively lead ten men"? Do you mean that hero won't be able to head an army of greater than 10 units? Or that any units that go above that limit will run around like chickens with their heads cut off? Or that the hero will control 10 units, while the player controls the rest? I am absolutely opposed to any idea which would split control of your own forces between you and an AI, so I would not at all like the latter. The former situation would be ridiculous, I wouldn't like that one, either. The middle, however, where leadership would determine the maximum army size is better but Campaigner brought up a good argument against it. If you run into a string of bad luck, or are just in a really messy war, and most or all of your best generals are killed, then you are screwed, with no recourse. Even if you have a huge military remaining, you wouldn't be able to field it. Let's take the nasty war scenario - pretend it was a defensive war, and you manage to fight off the invader but lose your best generals in the process. Now, you have no way of protecting yourself against the next invader. He has 3000 units divided into 3 equal forces of 1000. You have 5000 units, but the largest army you can field is 500, the next largest is 300, and so on. You have the greater military force, but you are artificially limited to using a tiny portion of them at once.
Read below before responding to this, as what I have to say in response to the next part of your post is very much related to this.
If you recruit a hero late in the game, he should not be able to lead 1000 as effectively as your seasoned hero.
I agree completely. But that said, I should be able to field an army of 1000 without a hero at the front if I want. Kind of like in Medieval II: Total War - you could field armies without a general (a captain would automatically be promoted to lead them), but that force would be at a significant disadvantage against an opposing force led by a remotely competent general. A new recruit should be less effective at leading an army (whether that army is 10 or 10,000 strong) than a veteran - but this needs to be implemented in a way that impacts tactical combat and auto-resolve equally. Anything else would be broken.
And anything that splits control of my forces between me and an AI, which will have no clue what my plans are, would also be broken. Brad is great, but I don't think even he could make an AI good enough for that to work well. And having a trait like leadership which limits maximum army size is also somewhat broken, as it would bring back the Killer Stack problem along with some new ones. A general/hero with leadership should not be a requirement for any army of any size, but having one should confer combat bonuses of some sort to make whatever force under his command more effective. The conferred bonus could even be divided among all his subordinates, so having a huge army under a hero with a little bit of leadership will dilute the effect.