I don't think it was the point of your post but technological differentiation is a huge part of the medieval world. I think living in the 21st century has blinded you to exactly how different weapons were in times of yore. Wars were fought and won over the unique resources and weapon techniques of ancient kingdoms. From the Greek Hoplite to the English Longbow, slight differences and advantages were present not in each and every nation, but also in the minute factions of those nations.
You're right it was not my point to underestimate these differences, and I gave several examples. But I think tech trees are a very crude, wasteful and inflexible way to enforce these kind of differences. To take one of your own examples, the greeks had large armies of well drilled hoplites because they had a large cosmopolitan population and could afford to equip and train their citizens as hoplites. Archery was too specialised and horses were scarce. Whereas the persians had more of a feudal society in which only the aristocracy were at all skilled fighters. They also had access to horses and elephants because of their geographical situation. Neither had a built in predilection to a certain type of warfare, they were products of their societies and geography.
We have all this kind of stuff already: you could say the monsters and wargs are analagous to horses and elephants, and elite troops are equivalent to hoplites. The problem is that the game doesn't make use of them or make them competetive with each other to anywhere near the extent required. Even when I deliberately set out to pursue a monster strategy, I always end up getting normal troops because it's easier to do so and they're just more effective. Horses are barely worth taking. Creating balanced mechanics that players will actually make use of on a regular basis is much preferable to, for example, leaving the current monster recruitment system largely unchanged, and instead just piling on new mechanics (for example, giving a certain faction a bonus Dire Aardvark unit that only they can research and which exists outside of the current system).
The game is supposed to be a sandbox. If you're going to differentiate factions, surely it is better to do so in broad, nudging ways than with specific, proscriptive bonus buildings and tech trees. Inherent differences between factions are great, but they should be based on where a race has come from rather than where they're going: the latter should be up to you. Your Dire Aardvarks start to look pretty silly if your faction spawns in the polar wastes. Surely far better to give the faction a discount to monster recruitment, so they can get yetis instead. If done well, these kinds of mechanics will produce all the same results that i think you are asking for, but they will do so in a more flexible and organic way, that doesn't waste content by restricting it to a certain faction.
I have indeed seen the weapon data sheets, and it's a great step forward for adding depth. My concern is that the idea of turning elemental into GalCiv 2 style rock paper sissors (with piercing, slashing and blunt weapons instead of lasers, mass drivers and missile) isn't really a very good one. Firstly because when it comes to armour plate is generally better than everything, against everything, so to make it worse against bludgeoning than chain mail is just silly (and the alternative is that everyone just makes a bee-line to the plate mail, which defeats the whole point). Mostly however because I don't think damage types are really the most important thing for modelling warfare realistically. And realism is important, because mechanics that model a system with a little realism produce intuitive results. When you stab someone in the back you expect him to die because you stabbed him in THE BACK, not survive because chain mail is good against daggers. When people have to learn your system to find out how war works you're doing something wrong.
Mechanics like weapon reach (spears vs cavalry), devastating cavalry charges and the tactical flexibility of sword n' shield are much more important IMO to good tactical combat than damage types. And i haven't seen anything to give me hope in this regard. I expect damage types to pale in significance to the current ONE BIG ATTACK VALUE system. Indeed, we've seen screenshots of unit design that allow you to equip an upgradeable blunt weapon, implying that weapons research will be a matter of discovering endlessly bigger attack values, rahter than what it should be which is a system where attack values increase very slowly (men were no stronger with clubs than in the middle ages), and new weapons instead give you more tactical options like pikes with enormous reach, bows with longer range, couched lances for devastating cavalry charges, and the like.
