I chose Mörk Borg as the partner to my non-authoring experiments because I find the basic content of the game to be very rich in the lore of the setting.
I also find the game quite limited, and best suited to one shot games. You are lucky if a campaign lasts for even a month of your character’s life, so there is no point in getting attached to them.
Character creation is almost entirely random so here is no input from you.
This is the ‘authored’ play game, called that because using a Mörk Borg-esque oracle system, based on d20 and the variable DR, most of the details of what happens have to be made up by the player, as an author would have to create the action for a novel. Almost all solo tools say that they can be used as a writing tool, thus ‘authoring’.
Non-authoring, in this case using a concordance, the detail comes from outside of the player, from the solo system itself. The player then has to interact with that content, either selecting the most appropriate suggestion and/or bending the meaning slightly to make it fit the current scene.
In practice, this seems to lend itself to shorter, very specific campaigns. If the content comes from a closed system before long you will see the same things coming around again and again.
What I want to do is feed a concordance tool all the text from Mörk Borg and then see how I get on. Admittedly this is a limited text, but it is also a limited test. If it doesn’t work I will give it an entire novel to chew on and see if that makes a difference.
In the longer term I would like to marry up a true machine learning AI, all of the free works by Robert E Howard, and actual play transcripts from people playing Conan RPGs. With the intention of trying to create a very specific Hyperborea GM. If that works then it should be possible to make AI GMs for any genre of popular game.