Some of you may not have gathered this from my last two posts, but I, Rutskarn, was hired to write the award-winning* script of Good Robot. What you might also have missed (because I never actually brought it up) was that I was hired to write the script by Shamus in 2013. Which was a very different time for gaming; the Oculus Rift was a potentially overhyped newcomer and had yet to revolutionize videogames forever, The Last Guardian was still in development, and gaming channels on YouTube grappled with the thankfully short-lived “Wild West Apocalypse” model of copyright claims and takedowns. Heady times.
But at least one thing in this crazy world has remained constant, and that’s my willingness to (under reasonable duress) actually write Good Robot. Which I did–all the way back in 2013. Now three years later one thing in gaming really has changed, and that’s Good Robot, because the old script is now completely useless. As in, nothing could be saved–not even the model of storytelling. The changes we’d made to the game meant a total rewrite was necessary.
Let’s talk about why.
This is actually the first of a two-part series, but you don’t need to read either one to understand the other. That’s important, because this post–and only this post–is absolutely one hundred percent farm-certified Spoiler Free. The next one will hint at the context of the new game without, I think, spoiling anything, but consider yourself warned. Now–let’s hop right in.
*Shamus has already guaranteed me a lock on this year’s Goldun Riter Awward, but I’m also gunning for the Horace Peasmasher Wrote Man’s Pennant, a prestigious prize granted thus far only to Wizards of Wor, CD Boggle (Sponsored by Kellog Cereals!), and Seasons of Mystery: The Cherry Blossom Murders. And even–dare I say it–a Writer’s Guild award, that I might join The Force Unleashed and Assassin’s Creed: Brotherhood as the premier storytelling achievements of the modern age.
The story of a game isn’t always important. How the story is told, however, is absolutely crucial. A game with good mechanics and no story can still get by; a game with a good story can be appealing; a game with a good story that doesn’t jive with the mechanics is awful, and I’m not just talking about your ludonarrative dissonances–I’m talking about cases where getting a narrative across crudely or inexpertly steps on players enjoying themselves. Let’s give an example.
I like Morrowind. I’m fascinated by Dunmer mysticism and politics at the game’s heart, and learning about them makes exploring and dungeon-crawling that much more fun. But at the end of the day, exploring and dungeon-crawling are what I’m there to do, and if getting that story meant the freedom to do those things was taken away from me, I’d have resented the intrusion. Morrowind’s story and gameplay work because they’re both based on freedom to do what you want; you pick what topics you explore with NPCs, you pick what books you read, you pick what quests you do. Having the story happen in pre-scripted cutscenes or infolink conversations or what have you would break the flow of the gameplay, take away the player’s freedom, and do more to frustrate than entertain.
This brings us almost directly to Good Robot.
Even back then, approaching the game from Shamus’ rough story outline was a dicey proposition. The idea was to include a rough arc to the story–nothing too fancy, but something that would noticeably change how you see your character and the setting over time. It would introduce a conflict, introduce another perspective on that conflict, then resolve it via the final boss battle. So how do you tell that story through a twin-sticky shmup?
More than it could be about any idea or story, Good Robot is about:
- Exploring, but only to gain enough power to proceed. You’re very much driven by the need to acquire power-ups and experience points rather than the urge to see new scenery.
- Levels that are procedurally generated nonlinear mazes.
- Blowing up robots. Something that requires a pitch-perfect awareness of all the chaotic action around you and no distractions.
So how to tell that story? Cutscenes are right out. We can’t use audio logs, because those will either distract players from robot sounds or–worse–will force the player to stop blowing up robots. That’s not really what they need from this game, and having large text dumps would make things even worse.
Furthermore, the only location where we know that a player is going to be on a linear straight-and-narrow channel is the exit and entrance areas of levels, right after the bossfights, and constraining all of the game’s story to those small areas (to make sure nothing important was missed or experienced out-of-order) would leave the game feeling a little lean story-wise–even though, in retrospect, that might have been the best way to approach that build of the game.
The answer we ended up with was a compromise. All story would be delivered by floating nodes that can be willingly accessed, by a player, to deliver a short story punch. Since the levels are nonlinear, all the nodes within a level are nonlinear. The “real” story stuff appeared between levels and only rarely; if you read everything else, you’d ideally gain a little more perspective on what was going on, but it wasn’t required. These nodes would offer small opportunities for players to stop and recuperate not predicated on leveling up their character (which would have been a bad idea to tie to story progress for a whole bunch of other reasons). It wasn’t perfect, but it worked.
Time passed, Pyrodactyl came aboard, and the game changed tremendously. Now the player’s movement is even more liberated, the levels are even more chaotic, and the player’s need to stop and take an occasional breather is taken up by little terminals where you can modify your build–and the game is explicitly styled as a roguelike. The question was raised even more portentiously than before: how do you tell a story in a game this nonlinear?
You don’t. You absolutely don’t. The roguelike thing alone sinks it. If you make engaging with your story mandatory (cutscenes, dialogues, audio cues), players who die over and over will hate it because it’s an arc they’ve seen the first few feet of a dozen times and the first few inches of a hundred times, but never anything past that. Imagine it took skill to watch The Dark Knight and slipping up meant starting over from the beginning. Would people still like the clown robbery scene, or would they still vomit reflexively after hearing the first line of dialogue?
If you don’t make the roguelike’s story mandatory to engage, players will tend to seek it out on their first run through an area and then only very rarely after that–because doing so every time would be a willful hassle. So then when they do break through, they’re going to be frustrated because they don’t remember anything that happened in the first half (because a few text nodes don’t stand out too strongly in the memory against a million robot fights) and now they have no context for understanding the second. It’s a nightmare to write around.
All of that’s assuming players find the story as compelling as their own story of struggle and survival against the cold, unflinching AI. Which they won’t. Nobody talks about the Orb of Zot from Stone Soup Dungeon Crawl except when they’re talking about finding it.
All of this led to my decision to eschew any linear narrative in the game at all. When players visit terminals, they’re greeted with headlines from the colony’s news cycle. Taken together these paint a picture of what the colony was and suggest, rather than narrate, how it got to be what it is: an arena for endless robot duels. It’ll hopefully provide an interesting sense of texture and exploration without distracting the player, forcing them to revisit content, or relying on them to remember specifically what they’ve read before. Furthermore, it takes a backseat in every mechanical and narrative sense to the player’s own struggle. This might not have been the only way to do it, but I’m happy with how it turned out.
Next time: we’ll put on our Very Mild Spoiler Hats and talk about the plot of the last Good Robot build–and by ‘talk about,’ I mean ‘share completely.’
Quakecon Keynote 2013 Annotated
An interesting but technically dense talk about gaming technology. I translate it for the non-coders.
What did web browsers look like 20 years ago, and what kind of crazy features did they have?
What is Vulkan?
There's a new graphics API in town. What does that mean, and why do we need it?
The Game That Ruined Me
Be careful what you learn with your muscle-memory, because it will be very hard to un-learn it.
What Does a Robot Want?
No, self-aware robots aren't going to turn on us, Skynet-style. Not unless we designed them to.