The process of generating situations in the way that I described yesterday is much, much easier for me when the style of the game is a well-established sort of beast. It's stupidly easy in a bizarre mega-dungeon, and it's fairly easy when I'm imitating a game that I've played in myself. Right now, for instance, I'm running a game for Dangerfox that's based on the game I've been playing with Trollsmyth for years. The setting is based on the one that Trollsmyth uses, the situation is very similar to the one I've been playing in for years, and I know the kind of scenes that I want to achieve. Most importantly, I know the kinds of goals that Dangerfox's character needs to have to make those scenes interesting, and I have a rough idea of how to give him the opportunity to develop those goals. If he doesn't, I have some ideas for how to adapt the game to accommodate other sorts of motivations.
That's really what's important here: Player and character goals. I've been making the mistake, in a lot of these games, of trying to get my players to do my work for me -- of looking to them to define the game, on the thought that they'll enjoy it more and it'll be easier for me if it's based on "what they want." Which is absolutely and endlessly true, but it doesn't do me any good if I don't know what kinds of situations to present at the table. I just sit there throwing either random situations based on my notes (or thin air) at them, or "logical" (ish) responses to their own actions, and without any yardstick for what makes a "good" scene I just keep getting more and more nervous, without any idea of what's "right" or any foundation for moving my notes or ideas or whatever else it is I'm bringing to the table into the game. It doesn't really matter if those notes came from me or the players: they're not the point.
Goals, though. Goals are something that only the players can come up with, and that can provide me a firm foundation on which to build something that I know is interesting, and fun, and that I can properly referee. A situation can always be built on the foundation of "here is something the player's want, and here is an obstacle in their way." A more interesting scene can similarly be built out of "here are two things the players want, set up in some way that they can only have one of them." (Unless, of course, they are very clever.) If I know the players (and characters) goals, then I have the game. A lot of my communication with players has historically been about determining what their aims are in-game; lately, I've been thinking as well about how to give them the information that they need to devise interesting goals.
That's another problem with these recent games: I've just been throwing "stuff" at the players in the early sessions, and hoping that they come up with interesting ideas about what to do with it. It'd be much easier, for everyone involved, to come up with a proper adventure, with some pre-loaded goals, at the very beginning, and allow players to develop their own ideas from there, once we're properly into the game. Much less flailing about for everyone.
Probably the most important thought I've had along these lines is that there are certain situations that I just can't render with any fidelity. There are a lot of situations that I either don't know enough about, don't have enough interest in, or can't picture clearly enough in my mind to be able to describe well enough that the players have enough information to make decisions about, in a granular way. In response, I've been attempting to get more comfortable with and confident at abstracting these situations.
For instance: I know nothing about wilderness survival. My spatial imagination is similarly underdeveloped. (Seriously -- my attempts to draw maps of places I spend every day in go hilariously awry because I just can't picture them properly, never mind imaginary places I've never been.) So it's difficult for me to handle a party running around the woods from simply a map and a key. Considering that this is what Trollsmyth, in the Pathfinder game I've been running for him lately, is doing, this has been a bit of a problem. I've discovered, though, that a general idea of the terrain, a random encounter chart, and the Survival skill have been good enough for me to fake it. I can take a point on the map and a few rolls and say, "Okay, here's where & how you found that thing you were trying to get," or "Here's what the area immediately around you looks like, and here's a Problem." I don't have a damn clue what any of the stuff inbetween these little interludes looks like, really, ("You walk for 2 miles through...") but it's enough to run a game on.