I have always liked gamebooks with interesting game mechanics (not so much the traditional CYOA or Fighting Fantasy) and one great source for that in the past was the annual Windhammer Prize for Short Gamebook Fiction that ran from 2008-2015. I believe all the books are still available for download. Many interesting experiments there and I really enjoyed some books (but it was a while ago and I can't name any of my favorites now).
The Lindenbaum Compatition is a newer attempt to do something similar and it has also resulted in some books that I enjoyed reading/playing several of the books from the first year (have not taken the time to look into the entries this year):
As much as I dislike AI slop, one of the first things I did when I first saw Chat GPT was to generate some gamebooks to use as test input-data for my gamebook-generator script. Description with a graph showing one of the stories:
https://intfiction.org/t/pangamebook/52856/17
It was useful for making those test gamebooks. I also thought (too much) about how it would be possible to use a LLM to generate gamebooks, but probably best to first randomly generate some kind of structure (directed graph) for the story and then make many smaller prompts to ask for the book to be written one branch at a time. However even if I ever get around to experiment with that I will certainly not release any code (or slop) because AI-generated gamebooks seems like the last thing the world needs.
Regarding the Cycle Removal Problem. This problem is actually called Feedback Arc Set (it is a set that "hits" all "feedbacks" (i.e. cycles) with arcs, i.e., edges). Its undirected variant is called Feedback Edge Set and is actually simply Minimum Spanning Tree, i.e., a problem solvable in O(m log n) time, compared to the NP-completeness for the directed case.
While the cycle removal algorithm is a fine heuristic, it can perform arbitrarily bad.
Sorry I only know the spanish community. But the last years, the gamebooks (spanish) have been revival with a lot complex and interesting things.
For example loops, like as Groundhog Day (film), the book is https://www.puntodeheroe.com/elbucle.html
Or open worlds, like as modern RPG or GTA, the book is https://jose-tamayo.itch.io/la-leyenda-okiri .
Or Elite like gamebooks, the book https://sites.google.com/view/spacegom/inicio .
Plus, there are other new mechanics in other gamebooks.
I have always liked gamebooks with interesting game mechanics (not so much the traditional CYOA or Fighting Fantasy) and one great source for that in the past was the annual Windhammer Prize for Short Gamebook Fiction that ran from 2008-2015. I believe all the books are still available for download. Many interesting experiments there and I really enjoyed some books (but it was a while ago and I can't name any of my favorites now).
https://www.arborell.com/windhammer_prize.html
The Lindenbaum Compatition is a newer attempt to do something similar and it has also resulted in some books that I enjoyed reading/playing several of the books from the first year (have not taken the time to look into the entries this year):
https://www.lloydofgamebooks.com/p/voting-is-open-for-202420...
I am curious about other mechanics. Can you tell us more about what kind ? Could you point us towards some resource on this ?
These books appear to make heavy/exclusive use of AI for art. Are they also written by AI?
As much as I dislike AI slop, one of the first things I did when I first saw Chat GPT was to generate some gamebooks to use as test input-data for my gamebook-generator script. Description with a graph showing one of the stories: https://intfiction.org/t/pangamebook/52856/17
It was useful for making those test gamebooks. I also thought (too much) about how it would be possible to use a LLM to generate gamebooks, but probably best to first randomly generate some kind of structure (directed graph) for the story and then make many smaller prompts to ask for the book to be written one branch at a time. However even if I ever get around to experiment with that I will certainly not release any code (or slop) because AI-generated gamebooks seems like the last thing the world needs.
Regarding the Cycle Removal Problem. This problem is actually called Feedback Arc Set (it is a set that "hits" all "feedbacks" (i.e. cycles) with arcs, i.e., edges). Its undirected variant is called Feedback Edge Set and is actually simply Minimum Spanning Tree, i.e., a problem solvable in O(m log n) time, compared to the NP-completeness for the directed case.
While the cycle removal algorithm is a fine heuristic, it can perform arbitrarily bad.
Talk (with rubbish sound sorry) on how Ian Livingstone plotteed fighting fantasy gamebooks.
https://www.youtube.com/watch?v=JNSL7zTAy44