in my few attempts to design games, i don’t think i fully appreciated the importance of strategic terrain. a system’s ability to represent anything is not a virtue

it’s long been observed that procedural generation can be a bit of a siren. the power to generate a million unique worlds may be no more compelling than a million technically‍-​unique bowls of oatmeal. it blurs together into indistinct noise.

when i design games, i so often want to be broad, to account for any possibility. i’ve made a few roguelikes, and my impulse is to make everything subject to variation. why pin down specifics when you can get more replayability if instead things vary inside a range?

this is a fallacy, i think. this came into sharp focus while playing the current builds of slay the spire 2, and being reminded of my years of playing the original.

one of the things you quickly get acquainted with, playing slay the spire, is the events. getting the right event can shape the course of your run. these events are memorable and specific. you can think about, oh what if i get the golden idol event, or what if i get the council of ghosts?

when i design a roguelike, it’s tempting to treat events as drawing costs and benefits from a bag of possibilities. more replayability that way, right?

but i think this can kill strategy. instead of there being a handful of events to consider your response to, there’s now a vague mist of what if i can slurp up some favorable event‍-​parts into my bowl of oatmeal.

this is what i mean by strategic terrain. if there are discrete, identifiable landmarks to reason about, you can make much more interesting strategic calls than if there’s just a blurry combinatorial space you have to vibe your way through.

but i do wonder what the rules of thumb for this should be. if 5‍-​7 is the magic number for working memory, what’s the magic number of elements of strategic possibility space?

i think there has to be a limit, because even if you have a bunch of hand‍-​crafted StS‍-​style events, at a certain point you have so many of them that it’s no longer proper strategic terrain, because you stop reasoning “i might get X event” and instead go “maybe i’ll get one of the events that do X”.

and slay the spire does have a bit of this. plenty of events that remove cards, or give you gold, or give you a relic. i don’t think there’s any avoiding it entirely.

but yeah. i’m thinking it’s a more interesting design direction to come up with a handful of interesting choices, rather than a framework to account a thousand variations.