GDC 2007, Day 3

Some academics showed off what they considered to be the top 10 findings from game research this year. One of them involved the "playing to win" type player, and how even that type of player seeks to even the playfield by self handicapping or teaching the opponent. Another involved a bunch of data showing that a huge percentage of players spend a huge percentage of time playing World of Warcraft alone. It even used the phrase "together alone" as opposed to the phrase "alone together" that I used in my infamous article. A third finding had to do with ethical and moral exploration in games being a big, fertile, and unexplored area in games.

So, uh, I guess game researchers tend to talk a hell of a lot about what I write.

A panel moderated by David Edery (Microsoft) which included Raph Koster, someone from The Sims, and someone else from Neverwinter Nights talked about facilitating user generated content. Raph, always amusing, challenged the title of the panel "sharing control" saying that the users have almost all the control anyway, and that we're mostly along for the ride. The Neverwinter Nights guy agreed saying that maybe the players will share some control with us game developers. There were really interesting examples of players using content in crazy ways that were never remotely considered by the developers. As one example, Raph talked about how little development time was spent putting in dancing animations in Star Wars Galaxies, yet players made endless dancing videos on youtube and even orchestrated 150 person synchronized dancing scenes. He pondered "why didn't we just put in more dancing stuff and ship that? It would have been cheaper and we would have been on MTV." ha.

The creator of Castlevania talked about the advantages and disadvantages of 2D games. He said that 2D games are really good at capturing: 1) distance, 2) timing, 3) position, and 4) direction. In 2D, distance between objects and their facing directions are very, very clear. It's also pretty easy for the player to understand where a good position is in a 2D game and how to get into it. Finally, because those other things are easy, 2D games are able to focus on timing, rather than fumbling around in 3D space.

He also talked about how 2D, in some ways, is a great help to the team making the game. One person can be in charge of all the background in a level, such as "foggy village." In a 3D game, you'd have one person doing textures, one modelling just one room of that village, another fog programmer, and so on and so on. The fragmentation of the 3D team means each person feels like a cog in a machine, while the team member on a 2D project is responsible for a big chunk and feels more ownership, so he tries harder. 2D teams are also generally able to be smaller, which helps greatly with management and communication in the team. On the flipside, so much emphasis is put on 3D games that some team members feel they have no career advancement opportunities if they work on 2D games.

He also offered the interesting opinion that because of all the advantages of 2D listed above, that it's easier to create a 2D game that has the features you want and delivers the experience you want to the player. BUT, it's much easier to create a 3D game that has a presentation that impresses the player and gets him excited, as opposed to a 2D game where that is very hard. He thinks 2D games are unfortunately mostly for hardcore players who can appreciate the advantages, but that 3D games are inherently better at presentation because of camera movement, so they will remain the dominant form of game. That said, he also thinks 2D games will never die and that nintendo DS, Virtual Console, XBLA, cell phone games, etc all show many opportunities for 2D.

Chaim from Maxis gave an excellent presentation on the design of the editors in the upcoming game Spore. He talked about the difference between tools that let professionals create content and tools that seem to magically create awesome stuff when you hardly do anything. Photoshop can create great stuff is you know exactly what you're doing, but even a child can create interesting stuff with finger paints. Photoshop requires all sorts of technical knowledge to use to even a medium extent, but if you just put your hand in a paint bucket, then drag your hand over some paper, you get all sorts of interesting forms and shapes.

Spore wants to be more like finger paints. They want it to give you disproportionately great creatures/items/whatever for how much effort you put in. This way, your grandmother and other non-gamers can see what it's like to CREATE something, and once they do that with some success, they will be excited to try a little more complicated tools.

Imagine a large circle representing the set of all possible things a tool could make. The 3D program Maya, for example, has an enormous circle of possible things to make, as it could make any object/character/environment in any currently existing video game. Now imagine a much, much smaller circle representing all the "good" things one could make in 3D. Pretty much all of those are inside the first circle, meaning pretty much all the good stuff you could ever want to make, could be made in Maya. Too bad that the "good stuff" circle is damn tiny compared to all the really bad stuff you could possibly make in Maya. Even worse, imagine a third circle representing the content that an average users is *likely* to make. Unfortunately, there is zero overlap between what a new Maya user is likely to make with the set of "good" things that could be made. You are about 100% likely to make crap.

Spore wants align these three conceptual circles. They want the set of all likely things you'll make to be smack in the middle of all possible awesome stuff that you might want to make. Furthermore, they want as much of the awesome stuff you can think of to be inside the "possible to make" circle. At any given stage of their progress, they could look at the catalog of all the items made by various average people who get to play with Spore and see how "awesome" the resulting content is. It took a lot of iteration on the tools to get where they want to be.

One of the examples shown was the character creator tool. It was a hard problem because if you give the user the ability to make, say, *any* body for the creature, then players will tend to make very terrible bodies because the space of all possible bodies is so large. Furthermore, if the players could somehow assemble a bunch of polygons into some type of creature body, the animation system would have no clue what to do with it. So this very open system would be confusing to both players and the animation system.

Chaim (the prototyping master) asked one of the artists for help. This artist had a lot of drawings of creatures that Maxis hoped could be made in Spore (so they represent that circle of "good" things that we hope are all possible to create and even likely to create). Anyway, the prototyper knew that the artist had some kind of pattern he followed that let him always make good creature body shapes, but he didn't know what the secret was. The artist explained that all his bodies start out as a bean shape, and are then modified in only three different ways of extruding or bending or whatever.

The next creature editor prototype gave the player a 3D "bean" and a few controls to modify it in exactly the way the artist described. This structure, though very limiting relative to all the things you could make in Maya, turns out to pretty much always make good stuff. It is also very clear to the player what to do, and it's clear to the animation system how to animate anything that comes out of this structured system.

He also gave many other examples along these same lines. Quick notes are like, if you want to add legs to the creature, then allowing the set of all possible ways to put legs on would be difficult (how to place them in 3D space using a 2D screen and 2D mouse), it would lead to mostly bad placement of legs (the set of all bad places to put legs is way bigger than the set of all reasonable places) and furthermore the animation system would have too much trouble dealing with these wacky legs. So, what really happens is that all legs have feet that touch the floor. If you try to add a leg, the editor automatically puts the foot on the floor, and you move the leg around on that plane, which is very easy with a 2D mouse. It's fortunate that this gets rid of tons of bad places to put legs by not even bothering you with them, and the animation system is very happy too. They applied this same principle to many, many aspects of the editor.

I heard some people muse that creating doesn't mean anything if every choice is "right," but I think the overall approach is very good. It really will lead to empowering people like grandmothers who don't know they can create things at all, and will lure them into the experience. If they want more power, there are a couple levels of extra layers in Spore with more advanced features. If they want TONS of power, they can use Maya.

Next up, Ernest Adams talked about how he sucks at games and he wants more games for him and other people who suck but have money and want to play anyway. He talked about how a goal-oriented game can still allow diversions and sandbox stuff that is fun. Yeah we all know that but he's saying designers can take the sandbox activities more seriously and embrace the idea that it's perfectly valid to play around without a particular goal and not treat the notion as a second-class citizen.

Adams talked about FarCry takes place on this beautiful island with sandy beaches and blue water with fish and how he'd like to explore the game. But FarCry is, he said, allows you to explore that island if-and-only-if you want to be in a world entirely based on quickly shooting people before they shoot you. Of course, FarCry is simply not the game for him, but his point is that apparently MOST games aren't for him, which is a narrow state of affairs.

A game that offers a series of moral choices was an example of giving the player meaningful choice, but not requiring "skill" or challenge obstacles. I happen to be very interested in this exact type of game, but I guess that's for another time.

Oh, Ernest had a good line when he talked about how first person shooters have some of the most beautiful environments in the game industry, so "we have awesome nouns...and yet we have hardly any verbs." Rather than just shoot, he wants to ride a horse, climb a mountain, scuba dive, explore caves, go fishing, and other various activities involving tourism and exploring. He's saying that this style of play--that is play without gameplay--is way too uncommon. The reason, he says, is obviously because game developers are obsessed with games having to be hard challenges, which is less and less true as the market expands.

Everyone I've mentioned said a lot more than what was noted here, but I think I'll call it a day and get some rest. Game Developer's Conference 2007 is now over.

--Sirlin