It seems that alot of the disagreement that has arisen here is due to different definitions of the word 'narrative'.
I am using the word in it's literary sense. Jeff, you seem to have assigned the term a looser definition, a series of causally linked events which when looked at as a whole 'happen' to make sense as a story.
My objections are kind of linked to the feeling I get when playing a strategy game against a computer opponent. I start thinking what's the point of this - I'm playing an unthinking machine - it cares not whether I win or loose. Hence I prefer face to face games, or computer moderated games with human opponents.
Your randomly arranged events would not feel like a narrative (using my definition) to me, since I as a player know they have been unthinkingly arranged by the 'machine' which is the game engine. On the first play of 221 baker street for example, you get a different feeling - that you are particpating in a sequence of events which has been arranged in very specific order, by a human adversary (the author or game designer).
Having said that, I think that what you are aiming for is sill unique and quite a challenge.
On the subject of computers, I think you need two. Firstly, you need a 'paper computer, something like the 4-way page flipper I suggested that is responsible for keeping track of game state. (you might be better off actually with some scheme of interlocking paper dials which you turn until they show the right values for the game variables, which then gives you a lookup reference).
Secondly, I think you as the game designer will need a piece of software to generate a large number of random arrangements of events which fit the causal constraints you require. This would be an easy brute force programming task. Simply generate a huge number completely random sequences then reject all of them that do not fit the causal constraints. There would be more intelligent implementations (particularly a genetic algorithm), but you could probably get away with something this simple.
This program you write 'compiles' the 'paper computer program' many instances of it - by auto-generating say a large number of the disks referred to above (each disk can be thought of like a program, a set of instructions executed by the paper computer). Then to play a game players select at random some of the disks, pop them into the paper computer and off they go.
I hope I've been pretty precise about what it is that I'm aiming for, but if my terminology has been a source of confusion, then I apologize. Maybe I should try to write a "sample story" that could unfold during the course of the game to bring something more concrete to the discussion.
Here's an example, in broad strokes: you show up at a house at the invitation of a mysterious host. Weird things start happening, and as you investigate the occurences, you learn things -- about the house, about the master, about why you have been brought to the house, about what you're supposed to be doing. As the game proceeds, you shift from making these kind of discoveries to attempting to fulfill your goal, leading up a climactic showdown.
The whole point is that I don't want the events to "look like" a story in retrospect. I'm trying to come up with a way to have the occurences "fit together" to form a story. To do that, I'm trying to develop ways of causing the story to "lock in" along certain story elements, and to "propel forward" at certain key points.
But yes, it's not a "story" in the literal sense; you're not reading a book or even playing a choose-your-own adventure gamebook. The point is to create a board game that offers a player experience in which players feel like they are participating in a rich story that unfolds before them as they play the game. I can't think of any games that I've played in which this happens. I haven't played "Tales of the Arabian Nights" but it sounds like it might be the closest thing.
That doesn't seem like a legitimate objection to me; it sounds as though you'd be making a prior committment not to like the game, rather than just evaluating it on its merits. It's like saying "I will never read a story written by a computer, since I know the computer can't write a story", without actually reading the story and finding out. If the story holds together and works, I don't see what the problem would be.
Have you played 221B Baker St? I don't get that feeling when I play the game, and it's exactly the problem this is looking to address. You do feel that there's a coherent story that you are trying to unravel, but you don't feel like a part of that story. The story is a static thing that you look at pieces of, and try to see if you can assemble the whole story from seeing only parts of it. I'm looking for something that actually gives the feeling of participating in a story; it would unfold as the game progress, it would respond to the choices you make. I don't want to knock the game, because it's a great achievement, but I think there's more that can be done.
Yes, this is something like the direction my other game project is taking. In that game, it's very important that everything be arranged in a very specific way, but the computer "story generator" still injects quite a bit of flexibility.
Hmm, this is worth thinking about some more; at a minimum, these "disks" (or whatever) might at least offer a way to keep the story line straight once it "locks in" along a particular path. That, I think, is what I really need; a component that gets introduced once players find out some basic facts about their story, to forbid contradictory revelations from subsequently being made.
-Jeff