Warren Spector is well known for focusing on creating games where players can make meaningful choices. At a talk at the UC Santa Cruz's Inventing the Future of Gaming symposium, he set out to ask an important question, even perhaps the ventral question that defines his work: Can we level up game stories without compromising gameplay? But first, a caveat: "The games we're making you play now obviously aren't bad, they're tons of fun. I'm not saying any kind of game should go away, not be played, not be made, not be liked. All I'm saying is there's more we can do and that refining what we already do, while nice, will get us a whole lot further." He also made this stark observation: While there is plenty of "terrific" work being done in the indie space, and with games made specifically as art, "they're not changing the mainstream of gaming and I'm not sure they will." So he urged game creators who make mainstream games to "level up" their storytelling, and he had some concrete suggestions on how to approach that.

"All other media are made up of that single moment, the non-repeated moment. We are different."

The first important thing to keep in mind, he said, is that unlike in other media, "your story is exposed through the exploration of space." He noted four key things games can do:

And unlike other media, games must give each player a unique experience, he argued. Period. He would tell his designers: "If, at the end of this game, every player has the same experiences, go back and think." Importantly, he said, "this does not mean [players are] telling their own story soup-to-nuts. Players are exceptionally good at 'I did this, and then this, and that happened,' but they're not real good at narrative arc: Here's why this and this are important." His key piece of advice when constructing a story that allows room for the player is this: "In the dialogue between creators and audiences, you need to know what parts you are going to own and what part you need players to own."

"We need to build worlds, not sets."

He also railed against games that build movie sets -- with doors that don't work and empty rooms, facades. "We need to build worlds, not sets." If games are "grounded in exploration, how can you build a movie set?" he asked. "Sets are really easy to build. They limit player interaction. They're much more predictable than an open-ended world would be -- worlds are hard, worlds are tough. They're unpredictable." However, he said, "unpredictability is better for player-driven stories." He also argued for the development of the "virtual dungeon master" -- along the line of Left 4 Dead's AI Director, but even more advanced. "Good stories are constructed, not found," he said. "We need someone to create a story, but then we need systems, or some way for the game to respond to player input, to dynamically modify things and accommodate those unexpected choices we want players to make." And there is one more secret to unlocking the potential to game stories: "Every other every medium answers questions," he argued -- "here is what the director thinks about torture," for example. "There is room for interpretation, but they have to answer questions or they're unsatisfying." On the other hand, Spector suggested, "games ask questions, and players can answer them. I think that's key to the future of video game stories."