Murderworld
Announcing my text adventure about the X-Men!
My newly released game started off as a joke, a bit of fan fiction in reaction to a chain of events that led to me becoming the world’s leading expert on the 1989 video game, X-Men: Madness in Murderworld. Author and game developer Chris Baker reached out to interview me about that game for his book, Wrong! Retro Games, You Messed Up Our Comic Book Heroes, and in that conversation he sensed my conflicted feelings. X-Men: Madness in Murderworld is not a good game, but it’s got some good ideas and he joked that I should remake it.
That stuck with me.
In parallel, I had begun to explore the world of interactive fiction. Interactive fiction goes way back to the early days of computing and initially homogenized around a particular style of gameplay, but these days interactive fiction is a vibrant tapestry of storytelling methods. Broadly, interactive fiction is based on the principle of a player reading through a piece of writing and getting to exert some measure of control over the piece. I had always wanted to make a graphical adventure video game (a genre descendant of interactive fiction), but art and animation are not things I’m good at creating with any great speed or skill. Graphics-free interactive fiction (IF) offered the core of what I liked about graphical adventure games: gameplay driven by narrative.
In 2018 an ambitious collaborative project named Cragne Manor was announced and it seemed to me an interesting way to learn the IF development engine, Inform. Cragne Manor is an “exquisite corpse” style project where each section of the game is designed by a different author, with two (amazing) managers of the project working to tie everything together to make the game playable and semi-coherent. The game turned out spectacularly and my contribution to it was mediocre. It remains ignorant hubris that I chose that project to learn Inform while working alongside some of IF’s greatest authors.
Still, through that process I realized that the IF engine, Inform, offered a path to that thing living in the back of my brain ever since Chris Baker made that one joke.
I’ve been working on my game in fits and starts since 2018, and with no particular deadline I kept getting distracted by other projects. However, occasionally I’d peek at the writing I had done and didn’t totally hate it. (Which felt like a sign that it was still worth pursuing.) With the start of 2025 I decided that the international Interactive Fiction Competition (IFComp), held annually in the fall, would be my deadline to finally finish my work.
My game, Murderworld, is not a remake of that game from 1989, but is inspired by one of its core tenets: that in order for video games about comic book superheroes to be true to their source material, they shouldn’t be entirely about punching people. Superheroes like the X-Men are not interesting because they use their superpowers to go around getting in fights, but because they’re people burdened by the responsibility of wielding world-changing power.
As I worked on my game, it stopped being a joke (I hope) and turned into a “real” game, something I’m very proud of and something I hope you’ll enjoy playing.
***
From here I get into major spoilers and the nitty-gritty of the game’s development, so if you haven’t yet played the game and plan to, you’re free to go!
Ultimately the problem with working on a game for so long and with lengthy breaks in between development cycles is that me–the writer and designer–changed over those years. I’ve gotten better at my craft and often when reviewing parts of the game I made years earlier, I was left thinking of that work as trite and clumsy. Which then led to redoing those sections to try and bring them in line with my current sensibilities, creating more work and a longer timeline. Further, Murderworld definitely suffered from scope creep as again, it turned from a joke project to a full-fledged game. Originally I planned just for the Blackbird “tutorial” segment and then some puzzle rooms in Arcade’s lair.
But I realized that I was being a hypocrite. I criticized X-Men: Madness in Murderworld (1989) and other X-Men video game adaptations for not understanding the source material, but I myself was creating a game where the only verbs were action-oriented. One of my favorite podcasts, Jay and Miles X-Plain the X-Men, describes the X-Men as their “favorite superhero soap opera” and that tagline reminded me of what I was missing. I described it above, but the “real” people that make up the X-Men superhero team are what’s interesting about these stories. I added the Mansion chapter to the game.
With the Mansion chapter I wanted to show the characters’ relationships with one another and their relationship with their superhero jobs. Instead of their superhero codenames, in that section of the game everyone is called by their first names. The player’s mission for their inhabited character is personal and deeply selfish (go call your friend; go find something to eat; etc). Largely, these people are burnt out and bored of the constant threats to their lives. They all just wanted a nice weekend.
“If this was the first time an attack like this happened, I think we’d all be a little more focused on the task at hand,” Kurt says with a shrug. “Another day, another violation of our superhero hideout.”
The game then transitions to the death traps of Murderworld (the place) punishing those characters for their selfishness and lack of teamwork. The game finishes with the shining examples of the young kids who use teamwork to save the day.
With the final product, I’m not totally sure I met that design decision as well as some players would have liked. A few players only wanted to save the world and solve all its problems, and seemed to chafe against the apathy of these adult superheroes. The characters are focused on petty issues while their home is literally on fire. Some players wanted to put out that fire first and foremost despite what the characters wanted to do and what the game was telling them to do.
One thing I did to soften the problem was a total perspective change late in the game’s development. (Ugh. It was painful going through the entire code, line by line, looking for pronouns at that point and I’ve almost certainly missed a couple.) I shifted everything from a second person perspective to third person. No longer was it “you” who were trying to get food from the snack machine, but Nightcrawler. “You” weren’t arguing over the phone with Storm, it was Cyclops.
Also, a lot of the character motivation and exposition is kept in dialogue with non-player-characters (NPCs) and I didn’t love the default text parser format for that script delivery. Every NPC has large tables storing their responses to a variety of dialogue topics, and initially there was nothing at all to prompt the player into querying NPCs. Early testers were completely missing those conversations. I later added topic suggestions to help spark and guide conversations, which seemed to do ok, but I was never entirely happy with it. Reading player transcripts from the competition, I could see players were still skipping most dialogue. Text parser games just don’t do conversation very well. It’s tedious to type “verb + name + topic” over and over again. One day if I do another parser game, I’ll go find a dialogue extension for Inform or try to implement some other mechanism that might bring me closer to the dynamic conversations I had in my head.
A Lack of Planning
There was a long period when I wasn’t confident about the game’s structure and it cost me significant development time. I was panicking about the game’s growing length and for that period I thought I’d make it so the player only played through one Murderworld section. The player would only see the section for the character they inhabited in the Mansion chapter. If you wanted to play the other characters’ Murderworld sections, you’d need separate playthroughs. I tossed that idea, realizing that while the Mansion chapter has definite, separate per-character threads (e.g. the Colossus version of the Mansion is notably different than Storm’s), they weren’t *that* different. Plus, I realized that I needed to stop apologizing for making a big game.
However, that waffling caused me to really screw myself when I eventually arrived at developing Nightcrawler’s section of the Mansion chapter. See, I had been developing each Mansion section by character, in alphabetical order and hadn’t given proper thought to the systems governing that whole level. I was still thinking of each player character as an individual slice of the whole game despite them sharing a lot of the same systems. In particular, NPCs moved around on timetables and in response to player actions. What then about a character like Nightcrawler who can easily break sequence by teleporting to any room at any time? I ended up spending way more development time trying to reroute all the systems I had built for Colossus, Cyclops, and Dazzler, and making them fit a character with a verb set like Nightcrawler.
A Fan Fiction Game
It was a fun and challenging exercise working with characters who have had 50+ years of extensive history. Also, it was amusing to find myself encountering many of the exact same pitfalls that writers of these characters have faced for decades. I’ve touched on the game design nightmare that is the teleporting Nightcrawler, but all six characters had their own design challenges.
Wolverine, for instance, is undoubtedly the most popular X-Men character of all time and has been comprehensively written a thousand different ways. I struggled to pick “which” Wolverine I wanted to write and in the end boiled him down to two archetypes: the “berserker” and the “detective”. His Murderworld section has two paths based on how the player behaves and while it ended up being far more bloated than the other characters’ sections, I was generally happy with the result.
The most difficult character to plan for, however, was Storm. Storm is my favorite X-Person to be sure, but how on earth do you design a game around her? Her weather-controlling powers are too versatile and all powerful. When trying to plot out her Murderworld section I immediately understood why decades of writers have given her debilitating claustrophobia and leaned on it. I immediately understood why writers have invented scenarios to eliminate her mutant powers completely, rendering her a normal (but still bad ass) human. How does one make a threat that’s scary to Storm when she’s so much more powerful than the rest of her team?
It’s a design challenge that needs an analysis and a hearty discussion that won’t fit this post. I don’t think I accomplished it as cleanly as I would have liked but I’m happy that I at least avoided those two narrative crutches, especially the claustrophobia.
Finding the End
The Mansion chapter “macguffins” turned full characters later in the game–the six children–are original characters and their ultimate significance in Murderworld can be attributed in part to a 2023 Narrascope talk by Aster Fialla about thematic puzzle design.
I was struggling to figure out how to end Murderworld and it was that talk that clarified my path. I realized that I needed to return to the game’s principle theme (teamwork) and use it to close out the story. I was tired of the six adult X-Men by that point and frankly, I was worried about my ability to generate even more puzzles using their set of superpowers. On the narrative side I didn’t feel that the characters had done anything to earn their freedom from Arcade’s lair. They hadn’t worked as a team. They hadn’t learned their lesson.
Separately, in the back of my mind, I had formed little backstories about the six kids you rescue in the Mansion chapter and it occurred to me that they’d be a great answer to the weary adults’ predicament. They were fun to write and I could even see them in their own game someday, so I planted some narrative seeds should I ever want to work on that spin-off…
A random easter egg: the kid, Henry, is named after my son. (My) Henry and I were talking about superheroes and superpowers when he was about four years old, and he said that if he had powers he wanted the ability to pull out other people’s thoughts and print them on the palm of his hand so he could secretly read them. It was such an unexpected answer from a four-year-old. What was in other people’s thoughts that he needed to know? Why choose that power over, say, the ability to fly or super strength? Doubly hilarious because he couldn’t even read yet!
Inform Ranting
Inform 7 isn’t well suited for a game of this size (I don’t know if there are *any* parser development tools that would be!) I absolutely could have been more efficient with my code, but the 600,000+ word project made scrolling through the IDE unbearable by the end. Using headings as bookmarks to jump via the Index tab can only carry you so far (plus the Index tab has a fun quirk of jumping back up to the top of the list whenever you run the game, losing your position). Also, two editing windows in the IDE were not enough.
And, my god, I am so angry about how Inform injects line breaks into everything (or suddenly, doesn’t). For a game with a lot of variable text based on player characters and a whole bunch of other factors, I eventually gave up fighting with it. If you saw an errant line break, just know I likely wasted too much time trying to fix it. In the end, it was just too frustrating to try and hunt down every possible branch of a situation, with some paths causing a line break in the text and some not. I wish Inform didn’t include breaks at all. Leave it entirely up to the author to manage! Please!
I am entirely to blame for one thing, however, which was a fear of Inform’s “skein” feature. I spent so much time using “test me” commands to test different paths and mechanics, when I could have been using the skein. Thank goodness I eventually explored it because I don’t think I would have survived developing the final part of the game or the final round of QA testing without it. I do wish, however, the skein was more dynamic with drag and drop abilities, copying and joining threads, and other such things, but most projects are not as big as Murderworld so maybe that’s not something everyone would care about.
Post-Comp Release
If you play certain IFComp games online during the competition, the IFComp website keeps transcripts for authors to read and it took a while to muster the courage, but I did eventually dive into every transcript. Thankfully there were only a couple big bugs to fix (shout-out to the game’s testers for setting me up for success), but there were plenty of new synonyms to implement, a few puzzle clarifications, and some quality-of-life features I wanted to add. No matter how much testing you do, a bunch of players banging on your game in different ways will always surprise you.
The biggest new thing is the addition of “achievements” which you can review at the end of the game. The problem with a game that offers different paths, optional solutions, and hidden interactions is that players don’t know they’re experiencing something different based on a choice they’ve made. To them, the game is just running as normal. For instance, as mentioned above, Wolverine’s Murderworld section has two completely different paths. That was a lot of work most players will never know about, let alone see!
Letting players know that a decision they made has significantly altered the course of a game is something that has plagued designers for years and I’m certainly not the one to fix it today, but maybe these achievements will give players a peek into what might have been.
(Also, with this new version I added a little secret bonus for interactive fiction nerds.)
Thank you for playing and for reading!
