Thursday, July 27, 2006

Star Trader: Socialization

This is absurdly long.

I keep saying that you can't have a story - or a socializing experience - without a game to back it up. I'd like to talk about why, and some potential solutions.

Most games use somewhat emergent rules. They use simple rules, allow the player to choose from a wide variety of similar options, and propagate the situation based on that.

For example, the difference between moving to an overhang or going for the tunnel in a Quake map isn't choosing between two scripted events: it's a mass of choices which leads to you emergently moving towards one or the other. If you aim five degrees off, or leap a second too early, you might not make it. You might end up on a slightly different part of your target. These things change the game state emergently, and exactly where you want to put yourself also varies based on the game situation..

Similarly, where you put your warrior on the battle map is a choice between a bunch of similar options. The best choice emerges out of the surrounding situation: how damaged your warrior is, how important it is for him to attack someone, how far the healer can heal from, whether he needs to block someone... all of these things add together to make an exciting experience where the best option is rarely the same option.

For these kinds of emergent rules, the important things are change and interaction. How many enemies are there? Where? How well equipped are you in comparison? Is your skill high enough? When you act, the situation changes: the enemies are now in a different (but predictably different) situation, maybe one of them is in your sights now.

All good games use this basic method. All the pieces touch, and the pieces change based both on their actions and the actions of those around them.

Now, a social game.

Most attempts at social games use isolated social interactions. You flirt with A, A likes you. You punch B, B hates you. Cut and dry.

Boring as hell.

There's no emergent behavior, no interaction between the pieces, and no change worth mentioning.

This is why so many games which attempt to bring in socializing really do it poorly. For example, Oblivion, king of the misguided attempts to socialize.

The requirement to make this sort of thing "good" is to include more emergent behavior.

This is why I'm always saying that just socializing, or just adaptive drama, is not a good game. In order to make a good game which uses these things, you have to have an emergent ruleset under the hood.

Now, by its nature, an emergent ruleset requires that you have "organic" options - choices which vary smoothly, rather than iron-set choices.

For example, whether you plan on moving someone five spaces to the right and one space up, or four and two... that supports emergent rules. Simply clicking "move" does not. At least, not one-one hundredth as much.

But almost all social games thus far use iron-set choices: "Joke", "Flirt", "Be nice". This is a primitive method that was proved a poor idea with games like Dragon's Lair. When this kind of rule set is applied to non-social games, the game sucks. So... why would it miraculously work for social games?

Now, Santiago Siri has an interface. It consists of a 2D field, and you click somewhere in the field to determine the type of response you make. Instead of "joke" or "flirt", you have "+45 analysis, -30 compassion". Or whatever your axes are today.

This is a good start, but there are a lot of problems with it. First, his back end evidently converts these beautiful numbers over into a standard tree branch navigation system, so you might as well be clicking on dialogue options in the dark. But that's not the fault of the interface.

Second, the interface offers no feedback. You're operating blind. Therefore, you cannot make complex decisions very easily.

A game with a 2D grid shows enemies, houses, your people, and so forth. Without being able to see these things, you cannot make meaningful tactical selections. There's no emergent behavior, because you don't get any feedback so you can't alter your own actions.

Sure, you can give gross feedback in other ways, such as dialogue. But this is very inefficient, to the point where you won't be able to communicate at a "wide enough bandwidth" to allow for emergent behavior.

Okay, enough complaining. What are some solutions?

To make a more emergent system, you need to do three things: let pieces change, make pieces interact, and communicate to the player.

Let pieces change: There's a million ways of doing this. Although it's not being done efficiently, we'll pass on it for a moment and come back later.

Make pieces interact: This is almost never done in a social game. The closest you get is doing guild stuff, so if guild member A likes you, guild member B will also like you. That's not interaction - it's an excuse to make your characters brutally uninteresting.

In most non-social games, interacting is done via actions in the game. The healer looks around, sees his buddies, and positions himself safely to heal.

Social games focus on word interactions rather than actual interactions. So you get "be friendly" and so on as "actions". First, those aren't emergent actions, they're iron-hard actions. That's a bad idea. Second, they don't change the actual game world very much.

To interact, there are a few simple rules. First, pieces have a proximity to each other. Usually, this is a two dimensional map. Occasionally, it is a three dimensional map. It should NEVER be a one-dimensional map or, even worse, arbitrary. That doesn't provide for placing your own pieces on the map and herding enemies in interesting ways. (This map doesn't have to be a physical map.)

Second, pieces have a relationship to each other. A tank knows it should be between the weaklings and the enemies. A weakling is careful to stay out of bow range while still remaining within range to help with his skills. These relationships allow characters to compute how they should move and what actions they should take on the map.

Okay, so, interaction is clear.

Last thing: Telling the player.

Social games REALLY fall short on this. They try to communicate game state via text, or via statistics. Both of these options are pathetic. Most games communicate more by map position and velocity, adding in statistics to provide texture to the map, rather than to be a primary information source.

Sure, games (especially RPGs) use dialogue trees. But those aren't the primary game.

Now, some games (especially RPGs) use a very simplified rule set which essentially abandons "map battles" in favor of a slick, fast battle system.

This is not suitable for a social game for one, obvious reason: a social game is about building, not destroying.

Run into zombie #94942, you're eager to kill it and get it over with. So the battle moves fast. But run into Kate, you'll want to manipulate the situation so as to socialize effectively. And you'll want Kate to be a permanent part of the world - she'll have to crop up again and again. "Fast dispatch" is a term to avoid: you don't want to dismiss your characters quickly. Instead you want to aim for a more strategic style game, which means fewer, longer battles with very detailed, emergent play.

Okay...

Hopefully, you're getting this. If you got this far, you're very, very durable.

Now, here's the issue at hand:

How do we make a game which is suitably emergent while still being suitably social?

Well, my preference is to include more than simply social stuff. I want my social stuff to affect a real game, and visa-versa.

Let's make it something similar to a real-time strategy. We'll set it in the far future, where humans have become largely solo creatures, expanding to the stars individually and claiming big hunks of rock for themselves, nanobots doing all the work that we previously required groups for. Humans only socialize via on-line VR.

Our map isn't a star map: it's a map of "meta space". You start out with some miniscule amount of resources and can build a building. Then the idea is to trade whatever you're producing (iron? movies? guidance chips? comics?) with other "nearby" people. These people are people in your general social circle, so they're mostly newbs like you.

The trade is mutually beneficial. The idea is that through trade and negotiation, you can convince characters to move their buildings (or build new buildings) closer to your buildings, resulting in a more efficient trade. And, coincidentally, a tighter social relationship.

Of course, you can only build on your "land", and they can only build on theirs. Most land is "neutral zones", where nobody can build. Neutral zones are generally conquered by the people trading across/along them. Changing the boundaries, narrowing the neutral zones, and acquiring more land is a big part of the negotiation part of the game. As you expand, you can choose to take your friends with you, splitting any land you take. Or not, if it's too much trouble.

Your buddies are also talking to everyone near them, as well. You can give anyone any comm code you own, and visa-versa, allowing people to talk to people and set up trades through neutral territory. This will generally establish both of you in that neutral territory, and put you both close to the people between the two of you. It may also extend or cut off trade between people on either side of your expanding domain, which might piss people off.

You can also form pacts - such as marriage, or weird future-human pacts that do other things. These are buildings which must remain connected, and rapidly expand owned land towards each other, seeking an equilibrium between the two.

Wars, alien invasions, and so on are also quite possible. This would give you the kick in the pants you need in order to socialize "efficiently". It would be the threat, marching across the landscape, blowing up buildings and eating land. Or even absorbing people by force.

If you want to save your old friends, you've got to hustle. Infrastructure is expensive and difficult to move, so plan ahead! Or, you can simply bore a hole away from the enemy, trading with new friends, to give yourself time to develop countermeasures.

Anyhow, it isn't specifically "realistic", but it is complex and emergent. Trade will allow you to develop new buildings and new resources, and the characters should be written vibrantly, so you get a real feel for being "close" to people.

The tighter the two of you trade, the more your friend should be around, commenting in various ways. And, of course, close friends are more willing to ask and give help.

How about any of you? Did you understand my explanation? Do you have an idea for a game?

4 comments:

Patrick Dugan said...

A substantial, if not entirely reader-friendly, explanation of the emergent school of dramatic play; the reasons you cite here are precisely why I'm dissappointed with Storytron.

Is it nessecary to use material resources as a grounding for the social transcations? I suppose if games are material and formal constraints, and the social rules form the latter, then the answer is yes, but the nature of what we think of as resources should be reconsidered. Gold, mutton or whatever typically aren't enough to get the tight kind of thinking that touches on the social stuff, instead maybe it'd be better to move into some abstraction, things like attention, love, respect and all that, or at a low level of resolution things like politcal opinion. I'm going to do another draft of the Fianna play spec over the week while I'm away from a net connection (family vacation, paying dues to the genes) and I'll consider this in lue of the infratstructure notes I made last week.

Craig Perko said...

You're welcome to try, but I've had much, much more luck with nesting social games inside "real" games.

There's just not much meat to straight social gaming...

Chill said...

well pat, i think using resources like gold or mutton allows the player to build a model easily in their head, even if its wrong at first, which I believe Will Wright talked about before. Now of course what you do to add that extra element of depth is then design the resource as a metaphor for something else. If instead of using gold just as money, design the infrastructure so that gold is a metaphor for greed or some such.

Craig Perko said...

Just having characters which remember when you helped and hindered them and a few rules to determine what they think of it is enough. There's no need to think in terms of metaphors or other such academic things.