Roguelike or 2D action or tower defence game with evolving enemies making each playthrough\level harder and harder.
By evolving I mean a real evolution. Mating the most successful enemies with each other + random mutations added in.
This sounds like it could be a cool small game.
So the mating process would do stuff like exchanging body parts/abilities between enemies?
What I've found with evolution-centric games is that it's hard to determine the "most successful enemies", unless you make the player fight the same type of enemy many, many times over. This could be alleviated by making a fast-paced shoot-em-up game with a shitonne of enemies comin' at ya. The game would be arcade-y which may or may not be fun.
Or don't make it evolution-centric. Then what? A RL variation would mean less enemies -> less fights -> less data to determine the "most successful enemies". The game would more-or-less make randomly generated enemies throughout the game.
How do we gather more data? Within a complex combat system one could count how many times an enemy hits you and how efficient the type of attack was. The game would then spawn enemies which were stronger at those kinds of attacks (but weaker at others).
This wouldn't be an evolution with a mating process, as you've suggested, since it would use a more immediate and "greedy" algorithm to make harder enemies for the player. Is the game still considered evolution-based by your standards?
Let's say both of these games now exist. How does the evolution-centric game differentiate from a game like Alien Shooter? How does the RL-centric game differentiate from a game like Dungeon Crawl?
Okay. Your games throw enemies at the player using cool algorithms, while other games use predetermined units.
Will the player even notice the difference?
TL;DR: I wanted to develop a game that's using evolution/genetics more prominently for a while now, but I'm not sure how to make such a game FUN or otherwise unique for the player. Help me and let's make a game!