Part Four: An offer that can't be refused
Thought Experiment One: imagine Pong, tweaked to be more exciting. Instead of showing the AI paddle on the other side of the screen, we hide from the player everything that happens over the center line. An elegant pacing algorithm tracks the players returns and the score of the game, and fakes everything that happens over the line in order to constantly escalate the intensity. Pong matches would be full of ever more spectacular volleys reaching increasingly extreme climaxes. Every game would be more incredible than the last - and yet, very obviously, something would be lost, and this game would suck.
Thought Experiment Two: imagine a Battlefield 2 match where both team Commanders colluded to make the match as dramatically exciting as possible for the unwitting players on the ground. The Commanders would share information to guide their respective squads into interestingly asymmetric skirmishes. They would track the score and collaborate to keep things close, ensuring things like supply drops, artillery strikes and air support always arrived at the most desperate hour. Suddenly every Battlefield 2 match would be as good as the best one you ever had (unless you discovered the collusion, in which case it would all suddenly seem very pointless).
The principle challenge in designing and building a dynamic narrative that responds to player action is not in the building of it (which is, itself, a wicked problem), but rather in the designing of it. It is, in fact, the same problem as designing Extreme Pong: that in handing over drama management to an invisible algorithm, the player has no way to know his actions are driving the story and the dynamic narrative becomes indistinguishable from an authored one - and likely not as good.
With Commander mode, Battlefield 2 hints at a way to cut the Gordian Knot: it suggests the possibility that we can outsource drama management to other players. On the one hand, playing the Commander gives strong feelings of agency and on the other, players on the ground know the difference between a high level experience that is pre-packed by a designer and one that is managed on-the-fly by a player with an agenda.
The main problem with Commander mode is that it needs to be synchronized with the main game at runtime, making it a pretty niche kind of fun. But I believe this is easily fixed by decoupling the 'game' of player controlled drama management from the low-level action game being managed. If the management game is designed to be fun as a management game instead of as a part of a low level combat game, then one player's management game becomes another player's drama manager. This is where intramedia convergence comes in.
Thought Experiment Three: imagine Mafia 2, the open world game, shipped with Mafia Wars, the mobile application as its player narrative driver. Instead of spamming all of Facebook with requests to join a mob, Mafia Wars players would try to recruit the best players of Mafia 2 to join their mobs. Players of Mafia 2 would subscribe to a handful of Mafia Wars players, who would become the warring Dons of their game world, and the missions they were sent on would have meaningful repercussions in their worlds. As one Mafia Wars Don rises to power, his influence cascades throughout the game worlds of his subscribers. As other Dons rise to challenge him, Mafia 2 players find themselves embroiled in the familial intrigue of the best mafia fiction - contracted to protect, bribed to betray, or tasked to assassinate based on the real needs of real players (who would in fact be playing a different, but connected, game).
Unlike the Extreme Pong example, the beauty of this intramedia link between two different games connected via the cloud is that players on both sides of the line know there are other players out there returning their volley. At the same time, because the players operating at the Mafia Wars level are not tied to the minute to minute play the way Commanders are in Battlefield 2, they can appreciate the game at the pace of a social, bite-sized, mobile game.
Even more important is the bridge-building between two divergent audiences. The overlap in the Venn Diagram between Mafia Wars players and Mafia 2 players is probably not great, but then neither is the overlap being proposed between their games. Your Uncle Bob from Wisconsin who plays Wii Sports with you every Thanksgiving at Grandma's probably loves Mario Puzo books and clicks away at Mafia Wars three or four times a day between meetings, and would love nothing more than to have you as his man on the ground in his ongoing feud with Mike from Accounting. And with every click the decisions he makes in his game world could be changing the landscape in yours, and vice versa.