Wednesday, February 06, 2008

Giant Worlds. Boring Worlds

One of the big innovations that's come with radically increased memory has been big, detailed game worlds. There's always been a wish for them, ever since Zork or even before, but these days they're pretty commonplace.

I guess I like big worlds, but it feels really empty and wasteful. Most of the world is created using some kind of automation - fractal landscapes, NPC generators, building brushes - but these things do not add any real uniqueness to the world. Most of the uniqueness comes from the painstaking hand-scripting that the designers do.

For example, there are dozens of planets you can land on in Mass Effect. Leaving out the big plot planets, the rest of the planets do have interesting things on them. Unfortunately, 90% of those interesting things are TACOs that mean nothing (oh, look, ANOTHER downed space probe...) and the other 10% is some short little mission.

There's nothing inherently wrong with that, except for the fact that planets are big. I mean, really big.

I mean, look at Oblivion, which takes up only a relatively small piece of one continent. It's got at least as many unique (and non-unique) bits as Mass Effect. Of course, their cities are really tiny. Even their capital is basically the size of a small village. Cities are big. I mean, really big.

Look at Crackdown. It takes place in one city, and it has quite a lot of unique and not-so-unique content! Vastly more than any city in Oblivion, vastly more than any star cluster in Mass Effect...

Because it doesn't really matter how physically large your gameworld is. In truth, your universe will only have as much content as you manually program in. Whether your game takes place spanning thousands of galaxies or entirely on one small plate of pasta, the same amount of content can be crammed into both.

A lot of people think it's possible to generate content using an algorithm. They (we) spend an awful lot of time trying to find a tool that will let a designer use a "drama brush" or something, something that will create unique situations without being so carefully scripted. I suppose it's a fine goal, but, again, you only get out of it what you put into it.

Even that's kind of missing the point. A giant world full of little adventures is always going to feel... sparse. No matter how much content you cram in. Because in the end, there is no attachment to any of it.

I visit Homeburg, I fix their spider infestation, and I move on. Homeburg never matters ever again. Even if I let a Homeburgerian join my party, it's not like they are Homeburg. They're just a member of my party with no intrinsic connection to any given place.

Giving the player a lot of places to visit means that none of those places is going to be very important. Therefore, even if you generate awesome content for every single nook and cranny, players will still not care (or will never reach it because they're busy being interested with their starting city).

To me, the solution lies not in bigger worlds, and not even in more content, but in changing the things the game focuses on.

Right now, games are extremely avatar-centric. Your whole drive is to get better stuff for your avatar (and his buddies, who are basically extended avatars). In some games, there is a drive to make your avatar a certain kind of person - good, bad, ugly - but even in those cases, anything you encounter is a tool to express your avatar, not something that has value on its own. I can't tell you how many times I just selected "the right dialog" without even bothering to read it, because I knew which dialog makes my character more good or evil.

This is handy, I suppose, because it allows the designers to simplify things. The only changing factor is the tiny point of light we call the player. Everything else is laid down in linear script, nice and easy.

Don't you think we're past that?

Don't you think it would be okay to make a game which revolves around the places, rather than the avatar?

I don't mean something like The Sims or Dungeon Keeper or Dwarf Fortress. They revolve around places, yes, but their gameplay does not vary and they have no real plot.

I guess I'm looking for something kind of halfway between Dungeon Keeper and Mass Effect...


Darius Kazemi said...

The closest example I can think of is actually from your essay about home and Chrono Trigger. Of course, the game follows the avatars, but it's about their relations to their home time/place, and the big events usually involve, "Oh, we changed history, now home is TOTALLY DIFFERENT."

Craig Perko said...

Yeah, Chronotrigger feels pretty close.

System Shock II is full of different levels, but it has a very strong sense that this is a specific place you're wandering around in, so that might be generally the right direction, too.

Joe Osborn said...

So this, too, is a trend towards "worldier" content, right? In the sense that player actions have a measurable impact on and consequences for the game world. I think Raph Koster is doing some work in this area, if I understand you correctly.

To go back to your Homeburgerian example, how would you rephrase it to be place-centric? Do they get homesick? Do they get "home(burg) team advantage" when they're closer to Homeburg? Is their favorite food only available in Homeburg?

Or is this still too avatar-centric?

What does place-centric design really mean? A place is nothing without the people inhabiting it. You can only consider a place in the context of the people who have a connection to it, even if that connection is tourism - otherwise, it's just scenery and you can pass it by.

The game series that I think best expresses the sense of place is Suikoden. Each character has a history, including a hometown - and on top of that, there's the base of operations that you establish and work out of. It's unusual for an RPG to have a permanent "home base" like that(Chrono Trigger does it with the End of Time), and it really lends some weight to Suikoden's fiction of leading an army. I don't think there are many games that are -more- focused on decking out a massive number of characters in equipment, but the significance of the base and the importance of revisiting and understanding prior towns really makes the world seem more interesting. So, I think I agree with you - we need more significance placed on the world, and less on the numerical combat values. But if I could get some clarification on what exactly you mean by place-centrism, I'd appreciate it.

Craig Perko said...

"In the sense that player actions have a measurable impact on and consequences for the game world."

Well, no, not really. This could be done by a linear story just as easily. In fact, the nonlinear examples (Dungeon Keeper, The Sims) are precisely NOT what I'm looking for, so unless there's been a magic breakthrough, a largely scripted plot would be required.

"Or is this still too avatar-centric?"


I'm not saying that people are passe! I'm saying that we're centering our games around a single point - the avatar - and we might want to think about focusing a few of our games on a plane instead of a point.

Suikoden - depending on exactly which one - is a kind of middling example. It's very good in that the world is very immersive, but the gameplay is the same as usual: solve the town's problems, move along, that's it. Your castle was a central focus in some respect, and it was a lot of fun to see (and hear!) it grow... but the castle never really does anything. It never matters. It's only very rarely a stage for any actual game event.

I think what I'm thinking of, primarily, is that most games focus on a single "point". You have an avatar or a team. The complexity you can derive from that is fairly minimal - levels and equipment and spells.

If you allow your game focus to sit on something that is not simply flat axes and some equipment slots, you allow the player to explore more. You give the game more options as to how it can interact with the player without being schizophrenic.

So, we think about a game with two levels. The "top" level is the normal one, with a few avatars and some numbers and equipment slots. The "bottom" level is focused on the places in the game. The avatars in this level are places rather than people.

So, while you're playing the "top" game, you can actually run around inside the bottom-level avatars. Meanwhile, on the bottom level, the places are treated more like points rather than fields, and they interact with other places and things in various ways.

The interplay between the two levels is what I'm looking for...

I don't know if that was at all clear.

Joe Osborn said...

Okay, I understand a bit better now. So in the short term, player choices affect the party - which weapons to wield, which monsters to fight - but in the long term, the consequences of their choices change the nature of the world they're fighting in. This makes the Chrono Trigger example make a lot more sense to me.

Suikoden 5 was the Suikoden I had freshest in my mind - the continuing importance of Raftfleet throughout the game, the need to trade extensively between towns for various reasons, and the purpose of returning to previously visited towns to seek out more sympathizers was how I felt it in particular fit into the discussion.

I have an example of exactly the type of two-layered game you describe right on the tip of my tongue, but I can't quite locate it - Actraiser, where you build your city, then fight around it? No, the battle scenes are irrelevant to your town - SimCopter, where you can fly around the cities you built in SimCity? Closer, but the divide is still too strong - Legend of Mana, where the placement of the various regions on a map also effects their nature?

I'm sure there was a building-game where you could insert avatars and play around in the built world, but I can't for the life of me remember what it was right now.

How much player control would you like to have on the "place" avatars? Do you want to build them and decide their foreign policy? Or would you rather their interplay be another plot line or simulated situation, outside of direct player influence? I think that interesting games could exist on both ends of the spectrum.

Craig Perko said...

It sounds like you understand it very well.

I can't say how much control I'd like the player to have, since I can't say I've ever played anything like it. Computer games have the unique capability to allow something like this - it's not something that you could really do even in a long-running tabletop.

David said...

So something like having your RPG inside a Civ game?

Joe Osborn said...

In a sense, but I think Perko is talking about a potentially very controlled environment. If the player were a diplomat or a general for such a Civilization, I think that could express the place-centrism pretty well. The basic idea is to tie the significance of minute-to-minute quantities like strength, speed, inventory to the presence of hour-to-hour quantities like city populations, factions, and particular places.

In other words, games currently simplify a civilization by reducing it to a number of problems to be solved. What Perko would like to see is a rejection of that simplifying assumption - that a civilization is more than just a bunch of ducks that must be lined up in a row, and instead represents a complex landscape for play, with local color, local customs, distinctive issues, and a life of its own.

It's like playing Hide and Seek - if you run faster, or see better, or even have x-ray vision, the terrain of the game and the context are so much more important than your abilities or your enemies' abilities. The home team advantage always wins there. So, it should be possible to make a game that rewards learning all the intricacies of a place, as if it were learning how to sweet-talk an NPC.

Having the terrain shift under your feet in terms of simulation or script is another thing, but I think that even static terrains, if they are complex enough, can be engaging. In either event, bringing the location to the forefront is a big step that only a few games have taken. Some Zelda temples are like this, and some Shadow of the Colossus boss areas. Unfortunately, neither of those places wants you to go back afterwards to keep finding new secrets.

To answer your question, I think that having your RPG inside a civ game would accomplish the goal, but it's just a single point along the axis of "place determinism", degree to which places decide their own destiny, in terms of place-centric games.

And it sounds like it would be fun (:

Craig Perko said...

Ooh, I like that.

Olick said...

You retouch on this issue that you want a world that is, in a sense, more dynamic than the game worlds we have right now. Personally though, I enjoy the avatar/group centric designs, for the simple reason that I am interested and care about the characters, where games fail to make me care a lot about the places.

But on inspection this could be a symptom, not a preference on my part. The worlds seem less 'real' than the characters that you play with, and because of that I am not interested in preserving, changing, or effecting the world.

I think that culturally the focus on the place you live is downplayed too. I rarely see people being portrayed as having a high amount of interaction with the local community. But I mean.. we're interacting in a community right now, even though its over the internet, and not a local community. Maybe most people think about it as a "I interact with a community" and not a "I am part of the community".

I have wandered a little, but the point is, I think that part of this is that, especially in the US, we don't have a really strong connection with the world around us, or prefer to downplay it in favor of things we have direct control over. And I think this is reflected in the games we make and enjoy.

Craig Perko said...

I think you're right, but I'd like to see that change.

David said...

This is kinda stream of concsiousness but on-topic :

As much as I love the Zelda games, it did always make me sad that the towns weren't bigger. And this kinda goes back to your RPG post, but I nearly always count the houses. Or in the SNES rpgs, look at the number of beds in a castle versus the number of people living there.

In Final Fantasy 6 (3 US) after the world broke, it was interesting to go back to the old places to see how they'd changed. Some chests simply didn't exist anymore afterward... which was interesting to me. It gave the places an expiration date. Going to a place you "knew" and seing it so transformed was a neat experience. The people moved around too.

As I reread the "being a general or ambassador" example in Joe's post, I thought about a game where, as a general, you had to make difficult choices, such as favoring a place or the people. You might have to move people from town to town, based on resource needs, or overpopulation, or an invasion. The people would hate it, but it would be necessary for the place itself. Maybe someone you cared about had to be moved as well. In older times, it wouldn't be as simple as writing a letter to someone or calling them to let them know you were at the train station... if the party moving from town to town were somehow stranded... it could go alot of different ways. You could use your knowledge of nearby towns to see where they were likely to be waylaid or stuck at a ford. You would know which towns were likely to help refugees or criminals... interesting thoughts that give the places personalities that percolate into the avatar thread (or supercede it depending) that don't necessarily have to be the same from game to game.


Craig Perko said...

Yeah, that's pretty much exactly what I was thinking.

It seems that it's a pretty clear thought, after half a dozen people have polished it for me.