I think my last post deserves a little more time.
When I attended this year's GDC, I made sure to see Clint Hocking's presentation, "Exploration: From Systems to Spaces to Self." At present, he's one of the guys most publicly engaged with progressive single-player narrative game theory, emergent systems, and the whole Looking Glass legacy. His talk was, well, an exploration of the ways that players explore games, and thereby explore their own character through their in-game actions.
One example in his talk was the game Spider-Man 2. As I understood it, he argued that while the game hands the player the palette of Spider-Man's physical abilities to explore the physical space of New York City, the player is given no tools to explore the character of what makes Spider-Man a hero instead of just "an asshole in red tights." He noted the backstory of Spider-Man: that when Peter Parker initially gained his super powers, he just used them to win wrestling matches, until the day that he witnessed a robbery and didn't bother using his powers to stop it. That robber went on to kill Parker's beloved Uncle Ben, teaching Parker that "with great power comes great responsibility." The lesson is imparted by a loved one's being lost forever due to Parker's own decisions, his own inaction. From that point forward, Parker would be motivated to use his powers to prevent further tragedies.
The player of Spider-Man 2 has experienced no such personal tragedy, and has no motivation to much more than dicking around with Spider-Man's powers, beating up random criminals and returning errant balloons to gain points. It's the "what" of Spider-Man without the "why."
As I understood his presentation, Hocking went on to consider how one might build the intended characterizations and emotional responses in the player through the play of the game itself-- if you were going to build a game about Muhammad Ali, how to convey the conflict between his public persona and physical power, and how it would affect the following fights. And presumably, if you were going to build a Spider-Man game, how to convey the sense of responsibility through the play's mechanics and dynamics, though he didn't expand on how that might be done.
Immediately after the presentation ended, I began talking it over with a couple colleagues of mine. The stance I took was that, for the player to be legitimately invested in the theme of responsibility, the events that imparted the message would have to be personally meaningful to the individual player. They couldn't be concretely authored by the designer-- if Uncle Ben were going to be killed every single time you played this theoretical Spider-Man game, the event would be just as inevitable, and therefore just as meaningless, as if it happened in a cutscene. To truly affect the player, the designer can make his presence felt no further than creating and to some degree encouraging the possibility of Uncle Ben dying due to the player's own inaction.
Perhaps in this theoretical Spider-Man game, there are constants: the player is Peter Parker, post-bite, with his powers. He lives at the home of his Uncle Ben and Aunt Mae, and must return there frequently to sleep, eat, etc., and meanwhile become attached to his Aunt and Uncle through their interactions at the house. Among other things in the world, there are Robber NPCs with the goalset to rob other NPCs or banks, armored cars, etc. The Robber NPC type might also kill innocent civilians.
Now with just this base set of actors, we have the very barest possibility of the player witnessing a robber rob someone, fail to stop him, and that robber going on to kill the player's Uncle Ben. The possibility is there but extremely slight; only the rarest player would cease being an asshole in red tights through this series of events.
This is where tracking metrics come in. For one, Robber NPCs might tend to be much more likely to perform their robbery action when the player is nearby. Furthermore, perhaps each spawned Robber NPC maintains a record of whether or not he has entered the line of sight of the player while he's performing his robbery action, and from what distance and for how long. The Robber also records whether or not the player has deployed an attack at him at any point during or following the robbery action. Hereby, the system can fairly well confirm whether the player has duly witnessed a particular Robber perform his robbery action. The system can also record whether the player attempted to stop the Robber after witnessing his robbery action.
When it's confirmed that the player has witnessed a robbery and intentionally let the Robber go, that particular Robber NPC has a high probability of receiving the goal to kill the player's Uncle Ben. This wouldn't have to be as transparent as it seems-- the Robber wouldn't necessarily make a bee line for the player's house and shoot his uncle for no reason. When the Robber receives the goal to kill Uncle Ben, Uncle Ben might receive the goal to go into town to buy something, causing the Robber and Uncle Ben to meet in a plausible location for the act to occur.
Hereby, the game would set the possibilities in motion for a figure the player had grown attached to being lost through the player's own inaction, motivated by the dynamic events ingame instead of a pre-plotted series of scripted events. The metrics and goal systems would greatly raise the probability of the intended sequence of events occuring, but they would only occur when the player's actions spurred them to do so.
This would also bring up the interesting possibility of the game finding a player that had come to the table with the lesson of personal responsibility already learned, and acted with great diligence from square one without being prompted so by the events of the game. A player who stopped every robbery he witnessed would never lose his Uncle Ben because, for the purposes of the game's theme, he wouldn't need to. He would already be acting like Spider-Man.
I guess in large part the point of this post is to argue that worlds and narrative driven by dynamic AI actors needn't entirely preclude the idea of designer-intended events or story arcs, and could in fact make the authored elements more meaningful when received through the player's decisions and actions.
5.03.2007
Group 2
Labels:
game design,
game theory
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment