# How Results Files Work

Because results files are, in a sense, “abridged” save files, there are some commonalities with save files in their behavior.

• they represent the model state at a specific duration of cycles, albeit with significantly less model data
• when restored, a result file is the current model state (note the model restore command works on both save and results files)
• results can be stored in binary or text formats; the latter are larger but more amenable to data extraction operations with other tools

On the other hand, where the model save command creates a save file, its parallel model results command is not used to create a results file but to specify the criteria for creating a set of them.

Save files are created explicitly and singly (in most cases — saving in a FISH loop would be an exception). Results files are created automatically and implicitly according the interval value specified when the request for results output was constructed.

Perhaps most importantly, though a restored results file is treated by the program as the current model state, the model cannot be cycled from that state.

Results files never become the active model state on creation. They may only be made active via the i Project pane or using the model restore command.

Results File Anatomy

In their makeup, results files are comparable to save files, with the following differences.

• The text header only contains the program data; no commands are included.
• The data contained in the binary portion of the file is limited to the data requested by the user for the results files.