All For Games: An Interview With Warren Spector

Gamasutra: I was reading your mission statement, which says you want to create the narrative with the players and not for them. How do you go about doing that without becoming a choose-your-own adventure or a Spore?

WS: Wow, this is the topic of my GDC talk this year. How much should I reveal?

Gamasutra: Sorry, that was just me thinking.

WS: No, that's a great question. In terms of game narrative, there's a broad spectrum of possibilities and implementation styles. On the one hand, there's Will Wright who's not even interested in telling stories to or with players, he wants to provide them with the tools to create their own stories. That's perfectly fine. I love Will and the fact that he exists. I love playing his games. On the flip side there are the roller coaster rides like Half-Life. I'm loving Twilight Princess right now. The Square-Enix games. Those are games that put you on rails, they're roller coaster rides. They're exhilarating, exciting, fun and challenging... all the things that games should be.

The Legend of Zelda: Twilight Princess by Nintendo

There's also a middle ground, and I don't think that it involves the choose-your-own adventure approach. There's a philosophy that I like to apply: as a developer I want to control the overall narrative arc. Using Deus Ex as an example, JC Denton has a brother, he works for this agency, and the agency turns out to be not what you thought it was and you have to switch sides because they turn on you. The terrorists are the good guys, and so on. All that stuff provides context and meaning for all of the minute-to-minute player choices. In that sense, I own all the acts and why you do things. Now, saying that, it's possible to own why you do things and leave how you do them in the players’ hands. The key for me is creating linked sandboxes and letting players explore those little narrative chunks on their own. I'll determine why it's important that you get through a door, but how you get through it, what happens and whether you kill, talk to or ignore everyone on the other side belongs to the player. That concept of sharing authorship is where the sweet spot of game narrative is. There are some things that I think we can do to take that to the next level, and things that can be done a couple of years from now that can take it to yet another level. The end goal for me now isn't for me to allow players to play a movie, ride a roller coaster ride or provide a sandbox so they can do what they want, but is to find the compromise where I can have a dialog with each player virtually. That's what's exciting to me.

Gamasutra: Do you think that's something that can be done independent of the designer? Games that have attempted anything like that in the past have struck me as only within the limits of what the designer can think of, but if I as a player can think of something that makes sense but I can't do it, that breaks the story for me.

WS: It's a very hard challenge, I'll give you that. The problem is that most people making games, especially story games, still rely very heavily on scripted sequences and interactions among objects. If the designer doesn't think of it, it can't happen. We started doing this at Origin back in the late ‘80s, but with Ion Storm and Looking Glass we tried to take it to new levels. We tried to use simulation a little more, or at least to script types of behaviors between types of objects and entities as opposed to specific interactions between players and things in the world. So, in that way, we can set it up so that players can exercise a little creativity. It's a little self-serving for me to say this, so I apologize, but in Deus Ex, things happened all the time that we hadn't expected or planned. The example that the team talked about most frequently, was that we had these objects called LAMs that were explosives that you could attach to walls. These were physical objects in the world that had size and mass, and what we figured was that you'd attach it to a wall and it would blow up when something comes near it. You detonate it or shoot it. What we didn't anticipate is that a player figured out that you could put one on the wall, jump up onto it because it was a physical object, and then put another one up and jump onto it, creating a ladder and letting him get to where we didn't intend. No one at Ion Storm thought of that, or planned it, but the fact that we simulated it deeply enough allowed players to create solutions to the problems we created.

There’s another one we were amazed by, but that we caught in testing so we can claim that we did it on purpose, even though we didn't. We had a character type called the man in black that exploded when you killed them. We wanted to create a challenge for the player who chose close combat instead of ranged. We gave the players a character development path that drove some difference in gameplay. If you were an up-close-and-personal fighter, killing a man in black was a risky proposition for you because he blows up and does a lot of damage to you. A ranged combat guy just stands back with a pistol and bang, bang, bang the man in black explodes, no big deal. What we discovered was that people were killing men in black while they were next to a locked door or locked chest, thus opening the locked door or chest without having to expend any lock picks, multi-tools or resources of any kind, which we never anticipated. That's another way that players made the story their own.

Players were getting online talking about how it was cool when this cool thing happened, but that “cool thing” was different for every player. That's harnessing the real power of gaming. That's different than “wasn't it cool when you killed the giant tentacle monster when you hit it 3 times in the big toe and then knocking it on its back with the sword of doom,” but if everyone does that then who cares? For me, games are all about unique experience. I did this and this and this, which is different from what you did, but we both get the experience of being a hero through the narrative context that we provide.

Gamasutra: There's still an importance on checkpoints there, though, in order to advance the narrative arc in a way that you've set out.

WS: Certainly, and I think there are ways around that too, which we'll explore over the next several projects. You have to know the narrative entry-point and exit-point of each major part of the game for that to work. Absolutely. That's why it's a hybrid of linear string of pearls game structure and sandbox approach, like Will and other folks use.