I’ve written a few times before about handling stats in ChoiceScript games, and making particular choices available. But in writing my own WIP, I also wanted to make sure that the story felt distinctly different if the player gave the protagonist a different personality — not just in terms of which choices they were able to make (or make successfully), but also in terms of the inner narrative.
With that in mind, I set for myself the following (silly) goal: when I ran randomtest, after the very first segment of play, none of the narrative output should be repeated across more than 4000 of the randomized playthroughs. That means that
- many plot beats are reached only on 1/3 playthroughs (or fewer)
- those plot beats that do occur every playthrough are narrated in at least three different ways, depending on the player’s stats and relationships
This speaks more to the fiction than to the mechanics, but the aim was to make the moment-to-moment texture of the story feel malleable, not just the plot structure.
This was also a good time to do more with the extreme ends of my choice spectrum: as discussed previously, I wanted to give some acknowledgement to players who managed to work their way into the top (or bottom) 10-15% of particular stat ranges, because that demonstrated a commitment to playing a particular way and should probably be understood as representing more deliberate agency than other approaches. So a lot of my alternate narration is designed to capture those high-end or low-end variations in how people view the world.
As I’ve often found before, it often enriches an interactive fiction to approach that story with some mechanical disciplines in mind.
One of the stats in my WIP represents how well the protagonist deals with pressure and whether they’re able to remain unflustered — and thinking about how a low-cool, highly anxious character would view events gave me a new angle on communicating the stakes of the story.
The influence went the other way as well. Anxious/calm characterization was easy to convey in a way that fit the existing stat. Another stat tracked social perceptiveness, and again, it was easy to write some high-end narration in which some players will notice more about the NPCs’ emotional state.
For some of the others, though, it was less obvious how that stat would change the protagonist’s way of seeing, so I had to put more consideration into what that stat represented.
For instance, a stat that represents the protagonist’s reputation is less obviously linked to perception per se, but it would change the amount of deference they were shown by other characters, and the way status exchanges played out. It might also mean that that character was more sensitive to implied slights; they would be accustomed to being shown respect, so any lack of respect would feel a little more marked.
In the end, there will likely still be some exceptions to my “no line appears across all 10K playthroughs” dictate: for instance, some lines of NPC dialogue are likely to stay the same no matter what the protagonist has done up to that point. But I liked the enrichment to the story that came from taking that approach.
4 thoughts on “Stats and Narrator Viewpoint”
It seems to me that a lot of the methodology you talk about here for Choice of Games can be used by parser-based game authors as well. Personally, I like the concept of a evolving game world that is shaped by player decisions and play style, and I have tried to do it in a few experimental parser-based games. However, I found that it is a daunting task for the author to pull off, and none of those experiments have yet seen the light of day. My current project does give a nod in that direction, but compared to what you are doing, it is a pretty feeble attempt. Thanks for an inspiring article. You have given me a lot to think about.
“[…] In the end, there will likely still be some exceptions to my “no line appears across all 10K playthroughs” […]”
I understand the intention, but i wonder if it is really worthy. Most of the time i see interactive readers to explore very few times the story, even only once; and then to ask for more of their particular branch of the story.
That sounds incredible. I can’t wait to read it. As a novelist-turned-game-dev, I salute your narrative priorities. And I shall definitely steal some of your techniques for my own current story—it’s easy enough to alter sections of text based on stats in ChoiceScript—but on a muuuuch less impressive scale.
Technically I already have altered my text a little based on personality. I’ve only written my first chapter, but there’s an opposed stat of flexibility vs focus (which changes based on whether you dawdle on the way to meeting a friend), and there’s a bit of dialogue in which your friend says something about how you’re “Always late” or “Never late”. I think it’s great to give the reader textual feedback on their personality choices right from the start.