Pet hypothesis: the “harvest-build-destroy” (HBD) subgenre of RTS games “died out” because they peaked so early.
Herzog Zwei, either the first RTS or a proto-RTS depending on who you ask, came out in 1989. Three years later, Dune II came out in 1992 and created what I’d call the first recognizable HBD RTS – harvesting resources, a tech tree, basebuilding, and different factions.
StarCraft came out in 1998. Age of Empires II came out the next year in 1999. Within seven years, the subgenre went from the first recognizable appearance of its core traits to its peak, releasing two games that still have large, active competitive communities and cash-prize tournaments decades later. After that, developers quickly learned that trying to release a new HBD-style RTS was tilting at windmills, and everyone would just see it as a crappy StarCraft ripoff.
also a smaller market, and hard to leverage graphical improvements in the way that shooters and sims could.
competitive multiplayer and online dissemination of strategies destroyed RTS.
playing with friends any RTS in the old days: lots of cool shit to figure out, tactics, strategy, FUN! the most fun you could imagine!
playing any RTS (except, sometimes, Blizzard games) in the modern era:
“the correct strategy is actually tank rush that hits at 5:53. look, everyone is doing a 5:53 tank rush.”
ironic, the genre killed by the meta.
I wonder if you could kill the meta by including so many random-per-game variables that you couldn’t decision-tree your way through all of them.
“You should rush tanks” not this game tovarich, the RNG pulled up the Problematic Logistics modifier for me so all my fuel costs are 20% higher. I’ll build cannon half-tracks and depend on scouting to keep them safe from enemy guns.
Additionally: Procedurally generated maps. Make them symmetrical around one axis to avoid gross imbalances. Probably still not great for a pro meta, but way better for the high-casual market.
God I would love that for many many games.