Is it worth $4? Maybe. Is it worth your time? Probably not.
Galactic Arms Race is built around a really cool core idea. Weapons, thousands (millions? trillions? bazillions?) of them are procedurally generated and evolve based on your preferences. As a tech demo for an advanced kind of procedural content generation, GAR is impressive. As a game, not so much.
The weapons, which are the sole focus of the game, are at the same time impressively varied and boringly repetitious. The variety of projectile patterns is astonishing: your shots may be concentrated straight ahead, in a spread pattern, out behind your ship, in a circle around it, in a snakelike wave pattern, in a seemingly random corkscrew, in a defensive wall in front of you, and other patterns that I don't have words for. The variety of colors, while maybe not astonishing, does seem to cover the entire range of human visual perception, so there's that. So what's the boring and repetitious part? For all the variety in their movements and colors, in almost every other way, every weapon in the universe
is identical. They all fire three shots. Those shots all do the same damage. They all have the same rate of fire. Their projectiles all have the same lifetime.
So although some weapons are objectively "better" than others, it's only because of their movement pattern. No weapon fires faster, or fires extra shots, or does more damage. I understand the academic
appeal of GAR (and indeed these guys are academics who have an upcoming article about it in an IEEE journal), but in my opinion it doesn't translate into a fun game
. I never really get to feel more powerful. Every sector looks the same and all the enemies are similar. I have to look at the numbers on the screen to see that I'm making any progress. I didn't get any new abilities until level 50, which took a couple of hours to reach, and when I did finally get that warhead launcher, I was totally underwhelmed. I quit playing shortly thereafter.
As an academic myself, I found some interest in reading about their cgNEAT algorithm and CPPN's (compositional pattern producing networks). But as a gamer, I have to say that its low price only serves to put GAR into that increasingly crowded category of Steam games that are not exactly a waste of my money, but are certainly a waste of my time. Pass on this one, and get a game that's an actual game, not a research project. Trust me.