In order to perform story analysis, you must first create a model (i.e., specify rules and initial conditions on the model page). Once you've constructed your model, a model panel appears in the center of the model page.  The second tab on the model panel puts the panel in story mode. Here, you can examine the detailed causal behavior of the model using a novel simulation method called story analysis. Essentially, story analysis tracks individual agents over the course of a stochastic simulation in order to decipher the exact set of events that lead to the application of a particular rule - a so-called story observable. 

Unlike standard dynamic simulations, story analysis doesn't focus on a set of observables.  Story analysis, also called causal analysis, examines the specific molecular events that lead to the execution of one or more rules of interest.  Thus, story analysis gives the user an insight into an exact sequence of events at the level of individual agents.  Specifically, for story analysis the user must select one or more rules of interest, which are called story observables.  An agent-based stochastic simulation is then initiated, and the simulator keeps track of all agents and events that occur until one of the story observable rules is applied.  The simulation is then halted, and the history of agents and their interactions is analyzed to extract the list of individual agents and the rules that were causally implicated in the path to the occurrence of the story observable.  Story analysis discards superfluous information by only considering agents and rules that were causally implicated, only keeping rules and agents that were absolutely required to reach the story observable rule.  For example, story analysis will ignore repeated binding and unbinding of a key agent in the story to another agent that is not involved in any other way, but keep a binding event if it was necessary for the eventual execution of the observable rule.

Rather than generating individual stories, this procedure can be repeated multiple times (iterations) within a single story analysis, and the results are automatically amalgamated to compare the prevalence of different causal histories that lead to the execution of the story observables.

The setup to run story analysis is relatively straightforward.  The model rules and initial conditions are defined below the model panel on the model page.  Additionally, you must specify the reaction volume on the simulation tab of the model page.  On the story tab, you need to specify which rules to track as story observables (via the rule picker), the number of iterations to run (how many stories to collect), and an upper time or event limit on how long to run each simulation while looking for the occurrence of a story observable rule.  Once the story analysis has been run, the results appear in the story viewer on the story tab.