Development Blog 4 - Appendix A - Procedural Storytelling and Defining Good Procedural Content
This week's research has been considered beyond the scope of the project. It should be considered supplementary and informative to the procedural research, but not strictly applied to the goals of the project.
Procedural Storytelling
On this entry, the topic chosen for research was procedural storytelling. Procedural storytelling is a wide-ranging subject and covers a wide range of procedural implementations; from procedural characters, mission structures, generative text and procedural worlds, it cannot be pinned down to singular sub-category of algorithm and the scope of discussion is too wide for the subject of this projects scope. Its complexity can range from a game simply presenting multiple pre-made choices and tracking these, as seen in this authors prior made example of a simple text adventure, to the use of Language Learning Models with key inputs and memory regarding key narrative events, such as Hidden Door which generates interactive stories based on user input.1 2
Ultimately, this author does not foresee a narrative element being built into the projects scope. This is due to mesh terrain generator being chosen as the primary procedural content, and whilst this is not explicitly at odds with narrative implementation, it would require further time beyond the scope of this project due to the systems required to integrate procedural narrative and world generation. This is certainly possible, with research being done with around these subjects – for example, Balint and Bidarra (2023) describe the creation of procedural world spaces based upon inputs of inputted narrative and knowledge base. 3
Principles of Defining Good Procedural Content
Short (2019) raises several important principles for producing procedural storytelling that have been discerned, which can to an extent be applied to Procedural Content in general. 4
Simple procedural content.
Perceptual outcomes vs technical outcomes.
Tying generated content to mechanics.
Introduce constraints and parameters.
Simple procedural content can still result in meaningful results. This can be seen in X-Com: Enemy Unknown, in which rather than wholly newly generated environments, small hand-made sections are placed in different areas with different pathing and enemy spawns. 5
In a narrative perspective, this can be seen to influence the player in how they choose to explore the game space.
The importance of perceptual outcome verses technical and mathematically outcome can be evidenced in the game Hitman: Blood Money (2006). 6
Players interactions within a level are tracked, such as weapons used, enemies killed and player detection. Upon conclusion of the level, a procedural newspaper is presented detailing the exploits of the player in a diegetic context.
The complexity is as simple as tracking player actions, yet the outcome presents a responsive world detailing the known actions of the player providing a narrative outcome on player actions.
Tying generated content / characters to mechanics can help to avoid repetitiveness in the player experience. This can be seen in Middle-earth: Shadow of Mordor (2014) Nemesis System in which the procedural generated enemy characters have personalities, voice lines, abilities, skills, roles and even occasional unique traits.
This provides room for player engagement and relationship building with the enemy characters. 7
Furthermore, Short (2019) highlights the importance of constraints and parameters within procedural content to provide results that remains contained within the expected gameplay of the given application.
This can be extended in concept to the projects goal of generating a terrain via Mesh – a completely random terrain mesh would not fulfil the idea of realistic terrain.
Therefore, introducing constraints in the terrain modifiers, such as height, depth, water level and colour will help constrain the generator into producing realistic terrain. This could be extended to biome building, with limits to ensure terrain is generated within a desired landscape choice.
These points contribute to quantifying what could be considered good Procedural Content in that the user does not experience too similar and repetitive gameplay while attempting to balance out the randomness from being too varied, preventing ruining of player immersion (narrative) and player experience (gameplay).
References
DodginJam (2024) Adventure Novel. GitHub. Available at: https://github.com/DodginJam/adventureNovel (Accessed: 7 December 2024).
Hidden Door (n.d.) Hidden Door. Available at: https://www.hiddendoor.co (Accessed: 7 December 2024)
J. T. Balint and R. Bidarra, "Procedural Generation of Narrative Worlds," in IEEE Transactions on Games, vol. 15, no. 2, pp. 262-272, June 2023, doi: 10.1109/TG.2022.3216582.
Short, T.X., 2019. Procedural Storytelling in Game Design. Boca Raton: CRC Press.
XCOM: Enemy Unknown, 2012. [PC, PlayStation 3, Xbox 360]. Firaxis Games, 2K Games
Hitman: Blood Money, 2006. [PC, PlayStation 2, Xbox]. IO Interactive, Eidos Interactive.
Middle-earth: Shadow of Mordor, 2014. [PlayStation 4, Xbox One, PC]. Monolith Productions, Warner Bros. Interactive Entertainment