Notion: Story Graphs in Theory and Practice

So, after a little contemplation it seems that a GM guides the story with a mental model of possibilities and potential stories. This model forms a directed graph of story nodes (that is, encounters). The game then traverses this graph of possibilities to produce a linear actual story. This linear story should then have certain characteristics of all good stories--coherence, rising and falling action (conflict), interesting characters and events, etc.

Here's a depiction of this: a graph of dark blue nodes with all possible connections (dashed arrows) and the path actually taken by the story (dark arrows). Some nodes have time requirements--they will happen at (or cannot happen until) a certain time. Similarly, some nodes might be dynamic, having contents that vary depending on which nodes were visited previously.


I must immediately note that most real GMs do not work this way! This is more of a brute force IN/computer game approach, where everything must be completely written before play even starts. As GMs, we don't plot out all possibilities in detail. For one, it's incredibly hard to think of all possible player actions that would necessitate a response. Secondly, it's a waste of time to do all this plotting for events that are unlikely to ever make it into the game.

Instead, I believe GMs work more like this:


That is, we have the main story line mostly planned, though there may still be a couple encounters (shown as outlined light blue nodes) we don't know all the details for yet. We may have a contingency plan for some of the obvious player choices that would take them off the main story line, and hopefully have a vague plan for how to work this back into the story. We might also have some world/content ideas with no idea of how these will affect the story (non-outlined light blue nodes). But these will be handy if we suddenly need to generate more of the graph to support a story that veers off of our planned outline.

In contrast, here's an approximation of the mental "graph" I was working from during our last Huffy session:


There's a clear starting point and a relatively good idea of where to end. There's a few ideas of the kind of action we might see--such as a rooftop chase or a rescue--but without any content--such as the characters or setting necessary for this action. These are shown as empty white nodes. And then there's a few other vague ideas that have no obvious story role yet, with a couple notions about how some of them might go together.

So, even if we do not need to generate a complete graph of all possible stories, how to we quickly generate a partial graph containing at least one complete possible storyline?