Wednesday, May 16, 2007

Not Good vs Evil, Please!

A lot of game designers these days use some kind of reputation/karma system based on good and evil. The idea is a simple one: easy to conceive, easy to program, and strongly rooted in the basic actions that players tend to take. Rescuing princesses and shooting civilians, the basic choices of players.

I don't really like the idea.

Some people don't like it because it is a shallow representation of something rather complex. It's hard to explore the theme of good and evil on a slider, you know?

I don't like it because it's a slider.

To me, the game dynamic is more important than the morale it represents. So, I hold all one-axis sliders in contempt. Good and evil is simply the most common. Other ones I've seen include law and chaos, emotional and reserved, even soft and hard. Sometimes, they are even combined for multiple sliders in one game, presumably to give it more "depth" - as if you won't notice the limits if there are more of them.

But a game is interesting mostly because of the way things are placed in "space". Mario is only interesting because his running and jumping navigates a carefully designed space full of fun obstacles. GTAIII is only interesting because roaming around the city in various ways lets you interact with the various things that are in the city (roads, buildings, people, power-ups) in various ways. The leveling system in an RPG is only interesting because of the "bumps" in the way it progresses - aiming for that next power-up, that new sword, whatever.

A good and evil slider (sliders of any type) don't really seem to fit the bill, not quite. They're not really a very good platform to make things bumpy. There's not really much "bumpiness" in your path to the light or dark, and even if there were, one-dimensional paths don't exactly have any easy ways to detour around bumps. Because they are not intrinsically linked to a space, sliders are kind of... background noise, like being able to choose your radio station in GTAIII. Sure, it adds spice, but it doesn't add gameplay.

Factions are the same way: factions are essentially a crap-ton of sliders, some of which always move in opposite directions. This doesn't lead to agency or choice, it just leads to being swamped with over-simplified "choices" that are never difficult to make. You know the old saying, "if you're a Jet, you're a Jet all the way"... it's never a choice as to whether to help the Sharks or the Jets, although theoretically you could choose your enemies. It may add some replay value, but I don't like sacrificing play for replay, you know?

It might be possible to add bumpiness to sliders, but how about instead we talk about a different method of doing the same kind of thing?

First, let's limit the world. A lot of games these days pride themselves on huge worlds teeming with life. Of course, the difference between any given person or place is nonexistent, which is actually a very depressing thing to suggest. Instead, lets imagine a world like an old-school RPG, where the world has only maybe ten towns, and perhaps twenty people in each town. On such a scale, each person can be a specific, unique person.

I think there's a big future in these "small worlds". I think we'll see a massive increase in the number of games that pack a thousand times the experiences in one-one-thousandth the space.

One of the big things these little worlds can accomplish is a sense of community. A real sense that the individuals matter - both to you and to the world. The feeling that you know specific people and that they know you. And I think that is a much stronger place to come from if you want to explore morality and character. The whole exploration of good and evil often comes down to who you love, who you serve, and who serves you. That's not something that's easy to do when everyone and every place is identical.

The dense, interconnected web of people are the bumpy terrain for the player. Each person is an individual, and can be dealt with as such. Helping or hindering a person will obviously make them tend to like or hate you, with caveats. Also, people feel specific ways about specific other people and things, and this also guides the way they feel about you. And, of course, individuals act on their emotions differently.

To some extent, this can be thought of as a slider for each person, but the key here is that each person is interconnected... and those connections can change as the world and the player interfere. I think that's a good level of bumpiness and complexity.

Now, I'm not talking about complex AI. We are talking about a lot of dialog, at least the normal way you'd do it, but we're not talking about complex AI. Just simple affinities for given people, actions, and things.

This isn't a perfect solution, but I think it walks in the right direction.

Other ideas? Comments?


Anonymous said...

Out of curiosity, did you read my post about our implementation of heroism/villainy and how it's actually a measure of interactions across up to nine levels of social involvement?

Craig Perko said...

I haven't read anyone's blog in about a month. Got a new job that sucks up most of my time. :P

I'll go look for it now, but this post certainly wasn't aimed at any specific theories of yours. It's aimed at KOTOR and Jade Empire and so forth.

Craig Perko said...

Erm... I can't find your post on that subject.

Anonymous said...

No,I didn't get the feeling you were talking about me specifically, as your objections to the Bioware approach mirror mine. I figured your objections to my approach would be a bit more in depth. ;)

Here's the link.

Craig Perko said...

Well, I think your solution is kind of tangential to the problems I see. I don't see it as a bad solution in theory (it's very similar to mine), but you seem to be adding a lot of complexity for reasons I can't see.

I think it's because you're thinking more in terms of philosophy and such. To me, framing it in any kind of judgment other than "Anne likes you, Bob hates you" seems... kind of on-the-nose, you know? Limiting.

Anonymous said...

I can see that from a social modeling perspective. The structure I propose in that post is geared to better integrate individual user experiences into a participatory culture which we can draw on for extended and meaningful plot arcs. It allows inter-tribal mythology to develop dynamically be drawn upon for textural reference later.

Craig Perko said...

As I said, more from a philosophical standpoint.

Patrick said...

Well, what you're proposing here is a re-distribution of content demands from animators to writers, which you can imagine I'm amenable to. But couldn't complex AI be the social algorithm that makes that kind of content more procedural? I mean, I'm not saying you need both, but the two conceptions are not mutually exclusive eh?

Craig Perko said...

Complex AI would be another post, but of course they aren't exclusive.