World of Design: Flipping the Hourglass

Picture courtesy of Pixabay.​

Today we're going to discuss two ways of creating many things: not just games but stories and game worlds. That is, starting top-down, or bottom-up.

Many of the Worlds of Design columns start out “shotgun” (bottom-up) and may not be finished until years after they were started, but there are others that I have drafted from start to finish in one sitting (similar, but not quite, top-down). Can I tell now which ones were done one way, which another? No, because no matter how you get to a draft, there will be revisions and additions, just as there will be for any RPG setting or world.

The Gardner vs. the Architect​

Top-down generally means you start with your goal and then figure out how you're going to get there by fleshing out lots of details, often an outline in the case of writing, all at once. When you actually start crafting, you follow the outline closely.

Bottom-up means you start with individual pieces and gradually build something up; you may not even know exactly what the goal is until you get near it. George R. R. Martin (Song of Fire and Ice/Game of Thrones) writes that way. Martin calls how he writes the gardening method, and calls top down the architectural method.

The question is, do you start with your objective and work your way down or do you start with individual parts that you think are going to fit into something and work your way up to a whole? The bottom up version can be quite incremental and created over a long period as you add things to it, but you can also get in trouble with contradictions. The top-down method is more organized but requires far more effort initially.

Top-Down​

I recall one author, Maxwell Alexander Drake, describing at a Gen Con panel how he spends many months making a detailed outline for a novel, and then follows it exactly as he writes the book. That is an extreme of top-down.

There are less extremes of course; writers who start out with an outline may change course as they write, just as game designers may changes rules as they design a game. You will often hear writers speak about how their characters did something the writer didn't expect, and the plot changed accordingly: the character becomes a separate entity to them, plotting their own course and not adhering to the outline.

Bottom-Up​

You also have people like Martin who just start writing and go wherever the story takes them, more or less an extreme of bottom-up. You can design games the same way, perhaps more RPGs than board games or video games. This may vary according to your temperament, or may vary with the individual game.

I've done this. I've made what I call “shotgun notes” about games (or articles) and sometimes these would turn out to become something useful and sometimes wouldn't; or it might be months or even years later when I went back to those notes and came up with something useful.

A Spectrum of Design​

Like most situations with two extremes, people are somewhere in between, and may go to one extreme or the other (or both) in certain cases; as with how I write Worlds of Design.

Clearly, the same distinction of method applies to setting and world building (and to adventure creation). Some settings will “just grow”, others will be consciously created in major outline in a short time. Many GMs probably figure out details of a setting as they go along (bottom up, with the extreme being sandbox style play). But others may decide to start from scratch and write down a whole lot of information about the new setting (top down).

Bottom-up design relies heavily on improvisation while top-down design requires significant planning. Bottom-up design was common in early Dungeons & Dragons campaigns because there weren't published worlds to begin with, and the concept of a "milieu" where the adventures took place was developed through play. As the game matured and dungeon masters had time to develop their worlds, more and more-top down design became possible. A GM's comfort level with one or the other largely depends on their creative vision and comfort with the rules. It's a lot easier to develop a setting in real time (bottom-up) when the GM can easily whip up NPCs on the fly, while systems that provide rules for world-building lend themselves better to planning ahead (top-down).

I don't think either method is necessarily “better.” For some, the fun is in the surprises created by bottom-up design; for others, control and narrative can be satisfying as the details are filled in from top-down. The trick is to find what suits you and your players.

Your Turn: When you create a setting or an entire world for RPG, how do you go about it?

 

[dbm]

Like several posters already, I follow an iterative process - do enough top level prep to provide a ‘helicopter view’ of the campaign as a whole which could mean geography or themes or more likely some of both. Then drop into more immediately useful detail that will ground character creation and the first few sessions.

What ever happens in character creation and play feeds back into the big picture as we learn what is important to the players, PCs, or the antagonists. Do some informed work back on the high level and so on.

That is a practical implementation of YAGNI and EDUF principles from software development.

“You ain’t gonna need it” - don’t waste time on developing features you may never need or use. Focus on the foreseeable.

“Enough design up front” - do enough design to provide the bones of the solution and make major architectural decisions while keeping the details flexible so you can adapt.

 

[Barner Cobblewood]

Like Soloist I find the top down model is not that helpful at the table. I use a centre and periphery model. The original centre is shaped to suit PCs, and contains more detailed maps of both geography and the relations among community that inhabits it. This determines things like possible professions and skills, and mores and laws. PCs know all of this.

The periphery is then mapped out as seen by the centre, relations being more focused on resource and threat. This is not accurate, but is the way the periphery is seen. This sets up some basic economic relations, which provide reasons to go from the centre to the periphery, and explains the "treasure" centre-people expect to find, which by definition is something that is not from "home."

Peripheries are seen as monolithic from centres, which are seen as internally diverse, so that's what I map. I only need to deal with what the details of a periphery actually are when PCs have explored it. sometimes the maps are mostly right, other times not.

Then, if the PCs move (not travel), I set up their new "home" as a centre with a new periphery, which might include earlier centres.

 

[CF07]

It's fairly easy to get a game session to feel like a story by using scenes, which is why so many games call out running games that way. You don't need to plot things out, only throw opposition in the PCs' way.

So I don't die under prep burden I give PCs freedom of selection of a few choices and ask about what choice they want and prep that. I usually have a rough idea of what's in the other directions in case they decide to check out an option and then go with a different one. This is super important running on VTTs given how much work prepping can end up being. It's not like the old days of having everyone take a ten minute break while I draw on a square gridded whiteboard or even just a quick tactical sketch map on some notebook paper. This can work for RP sessions, wilderness encounters, and even some dungeons.

 

[overgeeked]

So I don't die under prep burden I give PCs freedom of selection of a few choices and ask about what choice they want and prep that. I usually have a rough idea of what's in the other directions in case they decide to check out an option and then go with a different one.

When running sandboxes I'll have some local stuff semi-prepped, say within a day's travel of the PCs' current location. I'll have generated points of interest for another day or two outside of that. And I'll have a couple of locations roughly prepped in the wider world.

When it comes to more focused campaigns, something with a theme and PCs with goals, it's so much easier to just prep NPCs and obstacles for their goals. Most everything else can be handwaved or improvised. Just follow the PCs around and throw complications in their way. It's so, so much easier.

The Game Master's Handbook of Proactive Roleplaying is great for that. Some of that I was doing already, but there's a lot of great advice in there.

This is super important running on VTTs given how much work prepping can end up being. It's not like the old days of having everyone take a ten minute break while I draw on a square gridded whiteboard or even just a quick tactical sketch map on some notebook paper. This can work for RP sessions, wilderness encounters, and even some dungeons.

I'm very much a VTT as digital white board referee so that 10-minute break would be plenty. If I already have the perfect map, I'll use it. If I have time to find a good map, I will. But it's not awesome map or nothing. A janky drawing is always good enough.

 

[Blue]

I love making settings, and each will have a strong theme that will let me explore in play something I couldn't in a generic setting. In other words, the uniqueness of the setting actively supports the plots I am doing.

But I'm still mostly bottom-up. First, I leave large blank spots on the amp to be filled in as the players or as plot demands. Second, nothing is true until it hits the table, so details of the setting (or plots, or NPCs, ...) can change as long as it doesn't take a retcon. But most importantly, is the one-two punch of (1) a setting isn't world design, it's place-to-adventure design which is a sigfnificantly different focus, and (2) I have no idea where my campaigns will end up, I follow player interest and character actions (and inaction). So I might need to invent an invading country for an oncoming war or whatever.

Firmly puts me in the bottom-up design category as described even if not "pure". I like having planning overplots and arcs that I can at least foreshadow -- but they can often morph in ways that doesn't affect what has happened at the table.

 

[Von Ether]

When running sandboxes I'll have some local stuff semi-prepped, say within a day's travel of the PCs' current location. I'll have generated points of interest for another day or two outside of that. And I'll have a couple of locations roughly prepped in the wider world.

When it comes to more focused campaigns, something with a theme and PCs with goals, it's so much easier to just prep NPCs and obstacles for their goals. Most everything else can be handwaved or improvised. Just follow the PCs around and throw complications in their way. It's so, so much easier.

The Game Master's Handbook of Proactive Roleplaying is great for that. Some of that I was doing already, but there's a lot of great advice in there.

I'm very much a VTT as digital white board referee so that 10-minute break would be plenty. If I already have the perfect map, I'll use it. If I have time to find a good map, I will. But it's not awesome map or nothing. A janky drawing is always good enough.

I ended up having some generic locations prepped with maps and then a section of generic NPCs to quickly pull out. I could have quickly created a random encounter chart if I needed it.

The big thing is that I didn't do it all at once. Just a map here or there over the weeks.

 

[Eyes of Nine]

Setting Creation/World Building is fun - I loved the charts in the back of the DMG 1e and in my Traveller sets, and as a youth with more time than sense spent many a dreamy hour imaging the world.

But for play purposes, I focus on the here and now like someone said up thread - a town, some npcs, an adventure (5 room dungeon probs), not much more, and make everything else up as I go along.

That said, there's the third option after Top Down, Bottom Up - that's Steal Away. That's what I do these days, as it seems that my sense (hopefully) to time ratio has invererted. So Forgotten Realms, using Phandelver as an initial area; and moving out from there. Working for me!

 

[CF07]

The Game Master's Handbook of Proactive Roleplaying is great for that. Some of that I was doing already, but there's a lot of great advice in there.

I've never seen that book. Will check it out.

I'm very much a VTT as digital white board referee so that 10-minute break would be plenty. If I already have the perfect map, I'll use it. If I have time to find a good map, I will. But it's not awesome map or nothing. A janky drawing is always good enough.

Yeah I usually want a bit more than that, though I do reuse cool maps, rotate them, drop obstacles, terrain features, etc. A big dungeon can be really great in this context, too, because they're essentially a limited area sandbox. I also don't mind just switching from mapped to unmapped combat, at least in the game I run that's more spontaneous and player-driven.