There's this obscure game called StarCraft - you probably haven't heard of it. It was one of those games that was so well designed that for years afterwards, most that came after its throne were either failed experiments or pale imitations, and even those that succeeded were just more of the same. Here are a few factions, they're unique but equal; here's a campaign where you fight each other faction then a civil war, with each level unlocking more stuff. Get unit x to position y, hold your ground for 30 minutes, insert tab A into slot B. You must construct additional... Mylons. Yeah, that'll do.
StarCraft numbed me to the RTS for years. Everything wanted to be it, but I'd already played it. Even to this day, I find very little to recommend from that era. In a shock twist, however, there's an exception in Warlords Battlecry 3.
It came out in 2004, as the last in a series of real-time strategy games spun off from the turn based Warlords series that utilised RPG elements as far back as the late 80s. It's one of very few RTS to sustain my interest for more than an hour or so, and has consumed me for weeks on end when I boot it up. The RPG RTS hybrid wasn't unheard of even when the the first Battlecry came out in 2000, but until the third, no other game had brought the genres together so well without also bringing the baggage along.
The joy here is in the execution rather than the premise, which is simple enough. Though there's a little backstory and a hint of flavour here and there, it's never prominent. We need very little reason to murder tiny electronic people, after all. Sixteen fantasy races, including all the usual suspects - undead, orcs, insects, elves subdivided into snobby/hippy/teenaged - are split into factions, given a selection of leaders, and pitted against one another, either in skirmishes or the campaign. The latter is a map covered in hubs, each home to a specific faction and a couple of battles or other encounters, in which you might buy items or hire mercenaries. You're free to travel about these as you see fit, and some can be replayed, with a few offering a permanent choice between factions to ally with, which allows you to play as their side and access different strategies and units for your retinue.
What's the retinue? It's a system that should have been made a genre standard years ago. At the start of the battle you get army setup points to spend on standard units for your faction or any units in your retinue, but every unit gains XP for killing, and once they've hit a few levels, you can bring them with you on any mission until they die. Depending on your hero's stats, this can even be the focus your strategy. At one point mine included a flying zealot with a flaming sword, a witch who laid spider eggs in her victims, two gigantic building-stomping ents, and a triceratops. Quite a few battles end early with a hell of a shock for some poor sod expecting squishy elves.
This, and the alliances, are as close to decision-making roleplaying as WB3 gets. It's not a character- or narrative-based game at all. The plot is hands-off, and the bulk of your time will be spent pootling around to uncover more of the map and level up your hero. Experience and units accrue even in skirmishes, which can be customised as you like, bringing to the foreground what WB3 is about.
It's not a heavily scripted, linear campaign like Spellforce, nor is it meticulously balanced like StarCraft. Battles are mostly freeform, and units are fairly autonomous too, despite individually tracked XP and many spells and special abilities, avoiding the micromanagement that turned Warcraft 3 into a bit of a chore.
Rather, it's more like someone squeezed some essence of Master of Magic into the free for all structure of Total Annihilation. You can't build the same army every time and be sure that it'll stand a fair chance, but you can summon an undead dragon that steals XP, build golems that spawn free kamikaze units, or make leprechauns fight a T-Rex.
As a result, factions in WB3 aren't meticulously balanced. Each has entirely unique buildings and units, and even similar ones will vary in cost, and in the significance of that cost to their faction. Say your spearmen hits harder than a rival's swordsman, but the latter are resistant to your piercing weapons. But while yours is inferior in raw stats, you only need gold to train yours, and you have no other use for gold, whereas your opponent needs stone and ore to train theirs. You'll have nothing to lose by burning useless gold, but they'll have to divert stone away from a building upgrade to keep up.
This applies to the faction level, too. The Empire, for example, start out with exclusively piercing weapons, while the insectoid Swarm have several low-level troops who take reduced damage from these. If the Swarm strike early it's a one-sided battle. Life is even tougher for the laughably weak Fey. Their troops are tiny, weak, and vulnerable to all physical damage - a common oaf with a sword, fully expected to die in droves on some demon's claws, will effortlessly tear apart Fey troops, who have little choice but to keep swarming and hope to hold out.
Then, fifteen minutes later, you'll notice a Fey scout with twenty times the xp of your soldiers, and the corresponding upgrade in stats. What's happening? Well, given time and resources (some of which they get for free every time their towers score a kill), the fey can upgrade all their units to a terrifying degree, and turn those campily prancing jokes into an unstoppable tide of magic-flinging death, led by dragons that generate free resources with every kill, and units that add gold to their coffers just by existing.
Economic warfare becomes more nuanced, too. There's no guarantee that you'll start near any of the resources you need. Dark Dwarves have almost no need of crystal, but are crippled without plenty of stone, whereas Orcs get free stone every time they destroy a building.
Making the most of these quirks and mismatches is where the RPG bit comes in. Your hero isn't a beefed up grunt that you daren't risk in any fight where they might be needed (there is a permadeath option wherein death ends the whole campaign and deletes your save, but any unit can get a lucky shot in that instantly mangles another, so good luck). Your hero's class defines which stats they can level up, which in turn give direct bonuses to them or their troops.
Take my Wood Elf Bard, who I'm going to call Cass and not the anatomically suspect name I actually typed in, presumably while drunk. She's loaded up with Dexterity, which translates to moving faster, hitting more often, and crucially, capturing mines faster. Even when the enemy retakes them, Cass converts them back again so fast that it's a net gain to simply keep doing it - for every minute their hero is tied up stealing my mines, she only needs 20 seconds to take them back. Of course, that comes at the cost of lower hit points and weak spellcasting, and giving the game a strategic emphasis, whereas different builds pull the game towards Nox territory, personally laying waste to small armies or casting spells that instantly wipe out thirty attackers. Exactly where it belongs on a scale that includes StarCraft and Dota is impossible to pin down.
All this speciation trickles down from one principle, really: screw balance. Each side is truly unique, and tuned well enough to play its own way, not to be equal. Heroes differentiate them further, and the retinue system and the need to roam and capture mines cut down on the early game tedium that wears out so much of the genre. The sheer variety of playthings on offer lends it a lot of long-term replayability. Its conventional layout and sprite-based graphics give it an unremarkable, old-fashioned appearance, but Warlords Battlecry 3 is comfortably one of the best base building games ever made.