Unable to get back to sleep this morning, I had an idea. And I wanted to note it down before I manage another hour of sleep and forget it. So I apologize if it's a bit incoherent or difficult to follow; this is mostly for my own benefit.
One of the flaws of Department Nine, my 2007 Game Chef game, was that the initial setup phase took too long. Everyone sits down and comes up with ideas and draws up a complicated relationship web. Everyone needs to define several cause scenes for their event, and needs to relate these to other people's causes. And coming up with this stuff takes time. Then you spend a few very fun hours destroying the web, but that first hour of setup is too long. (It also can feel like you tell the story twice, which is thematic but not entirely desirable either.)
So I'm thinking that the initial setup phase can be subsumed into regular play pretty easily. Instead of doing all this prep stuff before the game really starts, we'll define each cause and other's relationships to it as we go. So each player still defines an event that they wish to prevent. But for the cause scenes, we have a stack of cards, each with a PC's name on it. You draw one at the start of the scene, and it's that PC's turn to deal with an event that leads up to their disastrous situation. Everyone else draws cards that define how they'll relate to the scene: maybe one card says that they oppose the primary PC for the scene, or that they roleplay NPCs in the scene, or that they're in charge of scene setting and such.
But the thing here is that the causes can be defined as we go. Which cuts down on prep and gives a player more material to work from creatively. They don't have to deal with a blank slate when coming u with events. They will already have all these things that other PCs did that they now can tie back to their event that they wish to avoid.
Well, it sounds worth playtesting anyway.