Gimme Details, Baby…

Posted by Rampant Coyote on August 1, 2011

I enjoyed Craig Stern’s article last week entitled, “Using Details to Craft a Coherent Game World.” Now, on a surface level, Craig sounds like he is mounts the lofty heights of Mount Obvious almost as badly as I do. Details make the world come alive? Gosh! But… if it’s so obvious, why do so many developers ignore ’em?

Craig brings up some nice specific examples that are often lacking, and suggests some ways developers – particularly indies, though this article isn’t indie-specific – can make their game worlds come alive. He includes puzzle design, which some readers may consider an odd choice, but I was thrilled to see it added. As Craig suggests, a puzzle well-integrated into the world can provide a good deal of exposition.

Departing from Craig’s actual article, I had a few additional thoughts while reading it:

These details don’t need to be front-and-center, requiring a page of exposition text or half-hour cut-scenes to explain. This is a sin committed not quite as often as the lack-of-detail, but it is in many ways worse. The mistake many designers – particularly new designers – make is in assuming that all this background detail that helps make sense of the world is needed up front. Especially at the beginning of the game – when what players really want to know is, “who am I supposed to be, and what am I supposed to be doing here?” Enduring opening exposition explaining that in the Third Epoch of the Reign of the Foo-Gods the Drizzle-Fobbers waged war upon the Unicycle Wizards of Gagadius is not only useless, it’s downright detrimental.

Background information should be in the background. In my first five minutes of gameplay, I want to know nothing more than what I’m doing there and what I can do to start having fun. The only details I’m really interested in are about what’s in it for me. Later, once I’m emotionally and intellectually invested in the world, I may find myself craving more information.

While not all of the details are spelled out to the player, the designer(s) should be very familiar with them. It can and should permeate the writing, the level design, tucked away in the background where it belongs. But as long as the designers are intimately familiar with the details, and are reasonably competent in their craft, the world will reflect that in its coherency. And that foundational consistency and predictability will make players more comfortable in the world, and more apt to appreciate the story set before them. “Show, don’t tell” is still a key principle here. A series of articles on “Mega-dungeon” design illustrates how this attention to detail in level design is important even in pen-and-paper RPGs. Even without a word of explanation, these details make the world feel “right.”

If a detail is somehow important to the game / setting that the player must know, one rule of thumb (I’m discovering the hard way) is that it should be presented to the player three times in three different ways. Otherwise there’s a good chance that particular datum will be lost in the noise. Maybe it is mentioned once in a key conversation, hinted at in a monster encounter, and then suggested again by an unusual layout of a village. The point isn’t just repetition on the theme, but having the player experience it in different ways.

The main bad guy (if there is one) and his plot – and how it involves the player – should really be well fleshed out by the designer. This is perhaps the most important place where details matter. Even in indie RPGs, we’ve really gotten past the point where a one-dimensional “Foozle” villain who is tormenting the world just for the sake of being evil is sufficient. Designers should really understand where the villain is coming from, what their goals are, and understand how the villains see themselves. THOSE details should be reflected in the plots and subplots the player finds him or herself embroiled within.

Above all, the designer should always be asking “why” and coming up with answers to that question in the game world context, even if not all of those answers are given to the player. Why is the bad guy doing this? Why is this dungeon here? Why is the village still here when they are completely surrounded by powerful monsters that could probably wipe the whole place out before breakfast? Why are these monsters working together? Why are they working for the main villain? Why are there still armies of footmen soldiers when the world is full of wizards who can wipe them all out with a single fireball?

These sorts of answers are what makes worlds and stories compelling.

Filed Under: Design - Comments: 9 Comments to Read

LateWhiteRabbit said,

A.S.P.I.R.E.

Arts/Aesthetics
Social Structure
Politics
Intellectual Achievements/Science
Religion
Economics

A college professor taught me to break down all societies and civilizations using this model for studying actual history, and it works remarkably well for creating worlds from scratch.

If you cover all those areas and think them out well – how they interact and influence each other and why they exist that way, making them all fit together – you will have a detailed and believable culture that is self-consistent.
Rampant Coyote said,

Wow. What a great break-down. I’m going to use that.
Maklak said,

This is very loosely related to the topic of above post, but I couldn’t resist. What do you think of procedural lanscape / content / monsters / whatever generation? The very idea of procedural content seems counter to designing a world with attention to detail, exposition, main storyline, etc. On the other hand it seems logical, and intuitively appealing.

Bethseda games have proceduraly generated landscape, that was later enchanced by level designers. It worked out well, but some people still complained, that it was plain. Same with Darkfall, I think.

Anarchy Online (and Diablo) had proceduraly generated mission / level layouts. There was a pre-generated set of rooms and corridors and rules to connect them. The result was that size and layout of missions wasn’t repetitive. Even with hundreds of human-designed levels, it would have to be otherwise.
In Mass Effect 1 every sidequest base had exactly the same layout. With some procedural content they could be all diferent, improving the game. If they added another level of procedures for placing furiture, even better.

Dwarf Fortress guy is experimenting with modelling urban sprawl to have cities with sewers and catacombs in his game. They look quite plausible yet very confusing. His major cities would be very difficult to navigate in 3D view. Amazing thing about it is that he models flow of trade goods, so if there is a dwarven-made cabinet in some house in a human town, it has a reason to be there.

It seems, that everybody is using scripted events, lots of “action” and big choices rather than a big open world with some seemingly random quests and non-intrusive storyline. Perhabs they are right, because in an RPG it is the story, and its emotional appeal that seems to matter the most. And yet with the right algorithms it would seem possible to spend much less design time on landscape, level, encounter, furniture layout, loot distribution (one that makes sense, not something scaling with player level), and other things. I also spent months exploring Morrowind, while ignoring the main quest. I greately appreciated not being funneled in that game.
Rampant Coyote said,

There’s a place for procedurally generated content. I mean, I played Daggerfall to completion, after all. 🙂 And that was even more procedurally generated and much larger than any later Elder Scrolls game. 🙂

And I have of course spoken of my enjoyment of Soldak’s games.

I’ve had something kicking around in my head for a few years I’d LIKE to do with that, but it’s still a ways away (I have three FK games to finish…) But something along the lines of a Dwarf Fortress type idea with real RPG gameplay is sort of the idea.

And there’s a lot of crappy human-generated content out there, too, which isn’t really much better than something generated algorithmically.

I think the “perfect” (hah, hah) RPG would contain a mix of both. There should be enough that is procedurally generated *as a result* of player actions – stuff that is “organic” and not just triggered by designers – to make sure that every game is really unique and responsive to each player. But the core content would be human-generated, very carefully, handling things in a way a computer just can’t. Giving the world a real feeling of purpose and story.
LateWhiteRabbit said,

It is hard to do generated content that is as good as content that is human created and placed.

It is like setting a stage for a play or a scene in a movie. You want to emphasize certain details and subtle hints toward the story, including foreshadowing. Just like the composition of a painting uses certain principles to guide the viewer’s eye and evoke certain feelings, so too should a game designer carefully craft his or her game world to build to the perfect pace and emotional pitch, like a composer placing notes on a sheet of music.

Procedurally generated content can be fun and exciting, but it is really impossible for it to have the same emotional impact or as strong a story as a hand-crafted game.
Menigal said,

I’d be perfectly willing to have a less developed story, which is really just an excuse for playing the game, in favor of a well done procedural world and variety in replays. Something like Dwarf Fortress, like the Coyote said, would be a good start. Yeah, you’d have to build in some set cultural, etc. data to give the place some flavor, but it’s not like everything needs to be proceudral. :p

But it is the details that help draw you into a world. I’m sure quite a few people who played Morrowind never bothered learning about all of the background, but the fact that it’s there really shows through. Not as much with the generified Oblivion. I kept looking for an explanation of how a ship arrived in the Imperial City, considering that there’s no access to the sea from that body of water. Really not that much attention to detail in that one.
Maklak said,

I for one like finding details and info about game world in an RPG. I read most books in Morrowind, except I skimmed sermons ov Vivek. They were to hard to make sense of.

There is a gradation from DF-like all procedural to all man-made. There are also a few possibilities to use random generation for some things, but not others.

Landscape can be proceduraly generated, and there are quite good algorithms to do that. Game designers can later review a few proceduraly generated landscapes, choose the most appealing to them, correct some slopes, add dungeons, cities and other detail. Used by Bethseda.

Randomized dungeons – Anarchy Online, Diablo, Din’s Curse, like I said above.

“Dynamic content” – people consume and produce goods, economy fluctuates, there are reasons for wars, etc. This would be the pinnacle of non-lineraity, but seems to be very hard to do in a way fun to the player. DF adventure mode is going to be this. Detail can be added by pre-existing locations, items, lore, books, etc. Din’s Curse also has something like this, with dynamic quest system. Some aspects of this can be “faked” by scripting, like people having daily routines.

Procedural loot generation is used in pretty much all RPGs to fill in all those treasure chests. My problem with how it is done is that it is too often “standard treasure for level X” DnD style. I really liked how Morrowind handled this. You would usually find crappy loot, but something that made sense, like food and stolen goods in a thieves den, gears and scrap metal in Dwemer ruins, etc. I also like it when encounters use their resources and drop everything on kill. (In Morrowind everybody dropped clothes they were wearing and weapon they were using). If those goblins have a +1 sword, 2 potions of healing and a wand of lightning, their champion and shaman should be using them, not let them rot in a poorly-locked treasure chest. (Read goblin comic for a satire on this). In Mass Effect there was a mission on the moon. Player could find remains of a soviet probe (moon rover). What was inside? Upgrades for 23 century weapons. That somehow killed the mood.
jaes said,

It’s good to see more developers talk about puzzles – If done right, puzzles (and mini-games) can be good storytelling tools.
Hajo said,

“A.S.P.I.R.E.

Arts/Aesthetics
Social Structure
Politics
Intellectual Achievements/Science
Religion
Economics”

It seems I did some things right with the racial descriptions for my never-to-be project Solarex, when I wrote the race descriptions, and planned the society structures. But having it spelled out, I see what I was missing and what to improve. Thanks for this valuable piece of knowledge 🙂