I’m going to quote a post in whole that covered most of what I was meaning to write on this subject but far more succinctly than I imagined possible. Brenda Braithwaite’s post “Design Truth 1″:
Focus on second-to-second play first. Nail it. Move on to minute-to-minute, then session-to-session, then day-to-day, then month-to-month (and so on). If your second-to-second play doesn’t work, nothing else matters. Along these lines, if your day-to-day fails, no one will care about month-to-month, either.
This seems like an excellent imperative to good game design – especially a mechanics-based game. In counterpoint, (though I could quibble about “good” vs “successful” design) whole games are built on hooking players with long-term investment, be it emotional, social, or time (read: sunk cost fallacy), rather than refined short-term, low-level gameplay (see: grindy MMOs, Zynga), or some kind of story that players get invested in despite the gameplay (see: Final Fantasy games). I think an argument can be made for classifying games according to higher-level design philosophy. But yes, Dredmor’s core is certainly in the mechanics. Well; the mechanics and the insanity, which might count as “story” content though ours is decidedly nonlinear. But I digress. I’ll be doing a lot of that.
[It just struck me - I think this attention to the fundamentals of gameplay is what Jonathon Blow was getting at in a response to one of our earlier trailers. Strong iteration of mechanics is clearly something he is focused on, as evidenced by his work.]
But yes, to bring this back around, the Dredmor beta, at 0.90 coming on 0.91 very soon here, is just breaking through the stage of everything being horrible and horribly crashy and not-fun into a stage where we must consider game balance and pacing in larger, longer cycles of gameplay.
Let’s consider these cycles of gameplay as a subject in itself, from “second-to-second play” to “minute-to-minute play” on upwards. The smallest cycles repeat within larger cycles which repeat within yet larger cycles, all in differing combination. The dynamics which arise out of matching smaller cycles with different sets of larger cycles presumably create interesting gameplay (“emergent gameplay”, if you like). This does not require an ‘open world’; such an experience could be entirely linear and planned (the prime example perhaps being the various arcs at play in a film). In Dredmor’s case we’re somewhere between the two, probably leaning toward open world, what with the roguelike random dungeon generation. (As for digression, from the mention of films, there’s a hell of a lot that should be said about pacing in games. It’s certainly something we need to give some attention in Dredmor.)
Right, well: gameplay cycles can perhaps be usefully arranged by scale from lowest-level (built from the most basic game mechanics) to the highest-level (contingent upon the dynamics between all lower cycles working together) like so:
- An individual action (attack, manipulate object, get/drop item)
- Complete interaction with a single entity (slay the monster, pull a lever and gather loot, unlock/bash chest, defuse/step on trap)
- Clear a room (deal appropriately with all entities in a room, perhaps full of traps, perhaps a squad of monsters, or just steal all the gold)
- Clear a series of rooms (take on & complete a quest? negotiate the dynamic between themed rooms? This is a stage I’d like to do better with.)
- Clear a dungeon level (advance character stats, items, and wealth through the course of a dungeon level, then deal with a leveled-up set of monsters in the next level.)
edit: 5.5. “Improvise successfully against the unexpected” situations you meet across levels [for Brian] Not sure what this means just yet, but we should probably do something with it.
- Run the course of the game (Start as a weak, poor adventurer; Over the course of the game find the inner strength [read: phat loot] to defeat Lord Dredmor! Or die in a disgraceful dead-end, corpse surrounded by gibbering blobbies.)
Quantifying the game in this manner lets me break down the gameplay to see if it is actually succeeding on each level. Or, even, to consider what mechanics should be tweaked or added to enhance this or that level.
The real-world messiness of the Dredmor project did not of course allow us to take the ideal approach as suggested by Braithwaite, to start at the bottom and build up from a prototype of the lowest level gameplay. It happened more like we stuck a few globs on a rough armature that Nicholas dragged out of the back of his closet, expanded outward, poking, finding weak spots, and filling in; stepping back to consider the experience (of ourselves and others), then adding, cutting, or changing as seems appropriate; ripping out the heart of the game with a spoon then replacing it with some jury-rigged assemblage at the last minute, whatever. Such is theoretical game design hitting real life development.
Update: I just found this post by DanC (of Lost Garden), “Creating a System of Game Play Notation”, which fits closely with what I was getting at above – that is, a methodology for looking at gameplay. And he’s got a chart with a hierarchy of gameplay elements. Clearly this thinking has been around for a while.