The monsoon came early this year. A storm swept through town in June and besieged the walls of my house, all day and throughout the night. Lashing gales and crashing rain assaulted the storm shutters. The clamor heightened around two in the morning; I was jolted from sleep then, disoriented and tangled in covers, not certain in which bed I lay, halfway between waking and dreaming. The rain smashed against the window just above my head. The gales screamed around me.

That was when I heard the woman. Through the window across the room, calling from the house’s back door:

“Please! Let me in! Let me in!”

In my dazed state I dismissed the voice as a hallucination. The hourglass of my consciousness drained down into reality, but for a long time I thought, surely, no one could be there.

But I heard the security gate bang. I heard the woman scream again, voice less distinct. I stirred, realizing that someone was in the yard, and looked out the window to see.

It was too dark. The rain came down hard and fogged the glass. I went around to the front door and flicked on the lights—to let the woman in, to check if she was real.

No one was outside. The woman was gone.

In that state of half-slumber, roused by the storm, I really had heard a woman begging to be let in. I remember her voice. The sound was “real,” in the sense that any noise descrambled by our minds is “real.” But it is obvious in retrospect that it was an auditory hallucination, caused by a tired mind’s misinterpretation of the keening wind outside my windows.

It was a very dark and very stormy night. Without powerful lights mounted outside the door, I never would have been able to see, without venturing into the rain, that there was no one outside. Perhaps I would not have been brave enough to open the door without the assurance that it was safe to do so. Then how could I have been certain the woman wasn’t real?

Consider this same encounter in the year 1800. In the dead of night, during a long storm, a woman screams for help beyond my window. At first I dismiss it for a trick of the mind, but when the screams persist, I investigate.

I see no woman through the glass. But it’s challenging to tell. At most I might have an oil lantern to light my way, but even so much as a candle would be difficult to ignite in the dark without a pre-existing flame. The shadows around me are viscous at night. I am keenly aware that I am a diurnal creature, acting in a time God did not intend for me.

The woman’s screams have stopped. I hear only the wind. But I still cannot be certain she is gone. I do not know that the woman was a hallucination. I am too afraid to open the door, and so return, warily, to bed.

When I relate this story to my friends the next day, each has had similar experiences. They have heard the woman calling out for help during storms. Like me they have never seen her, but neither have they seen in clear light that she is false. Over time, as we tell these stories, we propose explanations. Perhaps there really is a woman whose spirit haunts our village. We call her Screaming Susan and speculate that she was a young girl thrown out from her father’s home during a terrible storm. No stranger let her in, and she drowned in the rain. Now she returns at night to relive those final moments. One man says he thinks he caught a glimpse of her once, through his window. New details are added over time.

In this way, a ghost story is born. At its heart is an element of truth, and the need to understand.

This story seems ridiculous in the modern day. Ghosts are not real. They are not a very good explanation for anything. But for almost all of human history, belief in the supernatural has been completely ubiquitous. Today we stand atop a corpus of enlightenment that seeks to deny the existence of the otherworldly, but it is only because I have the language to understand what an “auditory hallucination” is that I can explain away an obvious apparition, rather than to say it was a ghost or a monster or a message from the gods as my ancestors might have.

But I don’t think the Enlightenment is what killed the visceral belief in the paranormal that pervaded the past. Reading folklore today, when I consider what it must have been like to live in the stories of monsters and demons, I think it was humanity’s literal enlightenment that spelled the end for a world of magic and monsters.

For many months, rats infested my ceiling. You would not believe how loud these tiny aberrations were. Anyone who heard them would have sworn homeless men were holding a fiesta in the walls. For weeks I had no clue what they were, until I set up a trail cam and caught the fuckers running around my kitchen late at night.

Without that trail cam, I wouldn’t have thought they were rats. I would have thought my house was haunted.

This ancient house I live in is becursed by strange sounds. A rapping, tapping noise still sometimes comes from the kitchen. Noises bang out from the yard and basement. The walls groan like old men rising from their chairs. More than once I have sworn that a burglar has broken in, because some strange noise pounded from one room or another.

But always I turn the light on, and I see that no one is there. No one and nothing. If I go outside with a flashlight I might realize that the banging from outside has, somehow, despite its volume, been caused by a rabbit stuck in my fence.

The revolution in technology since the end of the 1800s has completely changed how we live. We all know that the Internet didn’t exist before the 1980s, or that people rode horses and drove carriages before 1900. No one forgets that airplanes didn’t exist before the last century. As a Type 1 diabetic, I am keenly aware that I would be dead many times over had I lived any time before 1922.

But how often do you consider that light was a luxury before the invention of matches? It often comes up in my novels. In the book I’m currently writing, the most useful spell any character can know is one of flame or illumination—to have a flashlight powered by your voice, when contesting with the darkness. Yet this is no magic in the modern day. It’s something we take entirely for granted.

But before friction matches, there was no easy way to light your surroundings, even with a candle. A hearth had to be tended constantly. Shadow was the endless enemy. Any task at all at night was challenging. The lightbulb fundamentally changed humanity’s relationship with day and night. In so doing, it also changed our relationship with the supernatural.

We have all had cameras and flashlights in our pockets for twenty years. In that time we have learned that there is nothing hidden in the shadows of Earth. No UFOs, no monsters beyond men, and no ghosts. When something happens in the dark, we can shine light on it. Each ray from a flashlight is a tiny, instantaneous exorcism. We are so numb to this miracle that it has become part of our instinct; when I worked in a haunted house last year, I realized that nearly every visitor would pull out their flashlight the moment they entered complete darkness—even voluntarily.

Of course, using your flashlight in a haunted house ruins the experience, as well as blinding the actors. But spoiling the horror is just the point. As light tears the horror from the carnival ride, so too does it tear the horror from life. An illuminated world is one where ghosts and monsters cannot exist.

I don’t believe in ghosts. There are no cryptids I think might be real. But the past was a dark and scary place, full of strange sounds and unknowable mysteries. If I’d had to live there, by the light of candles and lanterns, even in the Age of Enlightenment, I think I would have been just as superstitious as anyone else. And I think you would have been, too.

Suno is an AI tool that lets users generate music in two-minute chunks. Describe the kind of song you want to hear and receive it within seconds. Like most AI services, it can be extensively trialed for free: enter any prompt and make the music happen. I’ve never come close to exhausting my free credits, though they appear to have recently locked the more advanced models behind a paywall.

I enjoy music. I listen to a lot of it. But I don’t know anything about how it works. I can’t play any instruments, can’t read sheet music, and can barely sing. I could never write a song.

This makes Suno an arcane and mysterious platform to me. It does something that I cannot do myself. Though AI image generation is functionally similar, I can conceive of myself drawing a picture; I cannot even imagine what it would be like to write a song.

Now I can. It turns out it’s as easy as writing, “A 1980s New Wave hit about buying a houseboat.” Or, “A 1920s musical number, low bass, rhyming chorus, angelic choir background.”

If you are curious what Suno might generate as a response to the latter prompt, you can give it a listen here:

Suno is extraordinary. It seems to understand rhythm as well as most human musicians. What it produces is often legitimately catchy, and it does this in seconds. Perhaps all the world can be reduced to mathematics; and what is music but a mathematical expression of noise? Maybe it’s no surprise that computers can quantify melodies so easily.

I spent the better part of a day turning Edgar Allan Poe’s poem “The Bells” into a song. It struck me in college that the poem had an intensely lyrical quality, and I’d wanted a suitably gothic musical arrangement for it ever since. I even considered composing one myself at one point (though I was swiftly defeated).

No longer. Here, in just a few hours, I had something that matched what I had imagined those years ago. It is rhythmically varied, sonically interesting, and fits my taste in music precisely. Though most of what I’ve generated on Suno isn’t very compelling, I’ve had “The Bells” on repeat for the last three days. I like it. It’s a song tailored specifically for me. I like it not just for AI, but as a real song. It could be better—but it’s the closest thing that exists to what I want.

I won’t share “The Bells” here. AI “art” works because it can be tailored for the individual, generated in seconds or hours. My AI “art” isn’t tailored for you, and you probably won’t appreciate it. This is why posting your Midjourney images to Pinterest is such a waste of time. These things work because they are bespoke, not because they are inherently high quality. Were I to upload “The Bells” here, it would surely be nothing more than cheaply generated slop to most strangers. But to me, it’s the song I’ve always wanted.

The End of Culture

Now imagine a future where everything is “The Bells.” You have an idea for a book you’d like to read. You ask your preferred LLM to generate it. Within moments you have a novel tailored to your every preference, and though it might not be flawless, might not be Literature, it is good enough to read like any book off the shelf at Barnes & Noble. Except it is your favorite book, because it has everything you know you like.

You’re feeling like listening to something new. You could open up Spotify. Or you can have Suno take a look at your preferences and generate a few bangers that match what you’ve had on repeat all month.

You have a game you’d like to play. GameGPT makes it for you instantly. It has read your angry Internet blog and knows exactly what you do and don’t like. It is the best game you have ever played. Its small deficiencies are compensated for by the sheer fun you have seeing your old ideas come to life.

This future has no vision. It is a world in which no one is ever challenged or confronted by anything uncomfortable. Ideas are no longer transmitted between humans, and no one ever experiences anything beyond the filters of their AI tools. You will never be surprised. You will never encounter anything new.

In this world, nothing ties groups together. You have nothing to talk to your coworkers about. There are no metaphors that make sense to your whole family at the dinner table. The bonds of society have broken down. Culture itself has ceased to exist. There is only the Machine, and the Individual.

This is the future of AI entertainment.

“The Bells” is a great song. I legitimately enjoy listening to it. But I can’t help but see the far-reaching consequences of a technology that creates culture fundamentally tailored to people, rather than populations. What will the civilization enamored with these tools produce? What will it be remembered for? And what will happen to the people within it?

For those who insist AI will never rise to a level at which it competes with human-produced art, because the AI lacks a soul or some similar nonsense, I would say that Suno has already, in its infancy, proved you resoundingly wrong. Computers can and do understand what makes music work. They will only get better at it. They do not need a “soul” for this. And if music can be written by the machine, so can books, games, and movies. It’s just a matter of time.

Maybe I’m being melodramatic. But I believe this truly could be the end of culture as we know it. We have already experienced the breakdown of society with the decline of broadcast entertainment, the rise of the Internet, and the digital age. AI will kick the process into superoverdrive. Yet what is to be done?

AI is here to stay. There’s no stopping it. The only question is who gets access to it. For my part, I think that is better you and me than someone else.

When I was young, I divided the games I played into two categories: the ones where I could kill everyone (including my friends) like Halo: Combat Evolved and Fallout; and the ones where I could only kill the people I was allowed to kill (including my enemies), like Knights of the Old Republic and Metal Gear Solid. My preference was for the former. To test the boundaries of these categories, I always tried to murder everyone I met in any game that I played. This tradition continues to this day.

Today, I view these two categories as the primary types of narrative philosophies in game design. I call them Constrained and Unconstrained narrative systems. In a game with a Constrained narrative, there is only one path to the conclusion. In an Unconstrained narrative, player agency determines outcomes.

Constraint comes on a spectrum. Halo is obviously not an Unconstrained game overall, and every game is Constrained by the limitations of the designers. Still, I use this distinction to better understand the design intent of a given title, especially a roleplaying game, and to determine if it is or is not doing what it set out to do.

I’ll be comparing roleplaying games going forward. Fundamentally, especially when I was younger, I understood this distinction in games with branching narratives as the separation between BioWare’s games and Obsidian’s. Let us go to them for our first examples.

A Case Study

Every campaign of Obsidian’s Fallout: New Vegas ends with the Battle of Hoover Dam. Likewise, every campaign of BioWare’s Dragon Age: Origins ends with the slaying of the Archdemon. The end point is fixed. These games are both cRPGs and generally considered to be of the same genre. They are both highly regarded, and they both respect player agency.

But they take incredibly different approaches to how the player should be allowed to reach the climax, and both present the climax in very different ways.

BioWare

No matter what decisions the player makes throughout the course of Origins, the overall journey will remain the same. After the prologue is concluded, every playthrough will follow the player to Redcliffe, Orzammar, the Circle of Magi, and the Dalish. The order can vary, and the actual content of the beats will change based on player decisions, but each event must occur. This is the Constrained philosophy of narrative design.

The Constrained philosophy is plot-driven and character-focused. It is about telling a specific story about specific people, rather than allowing the player to tell his own story.

Origins is not Constrained in the sense that the player’s choices don’t matter, or that agency is not respected. What makes it Constrained is that the end can only be reached by following the plot the designers intended. There is no way through Origins that does not follow the main storyline. Every faction must be recruited first.

Obsidian

In New Vegas, the player’s initial goal is to find Benny and recover the Platinum Chip. The developers do not care how he does this. He can follow the plot as it appears in his quest log. He can use other clues to work it out for himself. He can march to New Vegas immediately in a straight line and take vengeance before doing anything else at all. When the story’s focus shifts and the decision must be made of what to do with the Platinum Chip, the philosophy does not change. The player’s goal becomes larger than revenge. He is presented with factions to side with or against, and how he goes about bringing forth the climax is up to him.

The player can make alliances with factions, or he can skip them. He can kill everyone in the world, or he can kill no one at all. He can do what Caesar or Mr. House want, or he can do it for himself.

No matter which option he chooses, the game still ends at the Hoover Dam. That is still some kind of Constraint. But nearly everything in the game is optional. New Vegas can be finished extremely quickly if the player knows exactly what to do and where to go. The game does not care. It is a sandbox. What he does with the sand is up to him.

The Unconstrained philosophy is player-driven and setting-focused. It does not care about plot or pacing, and it is less interested in characters than games that are Constrained. Fallout epitomizes the Unconstrained philosophy: the individual quests are Constrained, but players are given huge varieties of options for solving problems, from pacifism to genocide.

As a child, I was right to divide games along the line of mass murder. Dragon Age: Origins does not bother explaining why the player isn’t allowed to kill all of the dwarves in Orzammar. Any killing spree in that game is performed with the express permission of the writers. The implicit reason is because rolling up and butchering Lord Harrowmont and all of his friends would not be a valid solution to the succession crisis, and it is not something that your character, the last Gray Warden, would do.

The writers of New Vegas have no interest in telling the player what his character would or would not do. The systems only tell us what the character can do. The rest is up to him.

Both philosophies have their places. Neither is inherently superior to the other. What is vital is that developers understand the kind of games that they are making. If they don’t, disaster will strike. A Constrained game that permits too little behavior can start to seem frustratingly arbitrary. An Unconstrained game that allows too much without reckoning with the consequences will seem broken and bizarre. Before designing systems, stop to ask: what kind of a game is this? Is it sandbox or story-driven? Can the player kill anyone he sees, or will some be untouchable? Can the player rush to the end of the story straight away, or does he have to follow a plot?

The answers to each question do not matter. All that matters is the intent behind the design at large. Constrained and Unconstrained RPGs can both be excellent for different reasons. Only a game that doesn’t know what it wants to be will end up bad.

Edit: For an example of a game that has no idea whether it is Constrained or Unconstrained, look no further than Starfield.

October, 2010. My parents are away for the weekend. My eldest brother has come home from college for babysitting duty. He shows me Evil Dead for the first time at midnight on Saturday. After a long discussion about the film, I'm ready to head to bed.

"Wait," he says, "it's not that late. Let's play a game."

He installs a client called BYOND on my PC and sets up his laptop on a nearby table. BYOND has a blue UI that looked ancient even then. Apparently this is a free platform for creating multiplayer games, but alongside every game--and there are thousands--is a playercount that reads 0.

"What are all these?" I ask.

"They're games," he says. "Scroll down to Space Station 13."

I find it in the list. Unlike the other games, it has sixty players online.

Stay alive inside Space Station 13

The story goes like this:

In 2003, a grad student named Exadv1 created a multiplayer atmospherics simulation. This was its sole purpose. You played one of a number of different roles onboard a space station, and as holes were blasted into walls, compartments were depressurized, and the oxygen in the station was replaced by methane or helium, the air composition at any point would be simulated.

Depending on whether or not the air was breathable, you'd start to choke. Depending on whether or not the air was flammable, a spark might start fires. And if there was no air at all, you'd choke and then freeze (unless you had a space suit).

Over time the simulation became more and more robust. By the time I started to play in late 2010, the game had diverged into a number of distinct servers, each with separate code and lore. All shared a few common features:

Medical Simulation

How much blood is within any player at a given time? Are any of his limbs broken? Have his organs sustained damage? Is he blind, and can he still breathe?

All of these questions are answered in SS13. The game individually simulates any aspect of your character's medical condition that you could ever want to know, from how well your liver is functioning to whether or not you've fractured your toes.

Hands Simulation

The primary gameplay in SS13 is acting as a puppeteer for your "2D spessman," as players like to put it. The core of this is the management of your hands.

You have two of them, ideally, although it's possible to lose one or the other (or both on a bad day). With your hands you will pick up objects and interact with everything else in the world. How good you are at manually swapping between your hands via X or MMB and then using them via Z is fundamental to determining how "robust" you are (the game’s term for how close a player is to reaching its incredibly high skill ceiling).

Hands and medical simulation are two of the main things that make SS13 so special. Although not exactly realistic, these allow server developers to implement systems that normal, more intuitively designed games could never hope to design. They permit the mechanization of things like reloading, which, in most games, are completely token. Not so here. Here, how good you are at reloading determines how good you are at the game as a whole.

Roleplaying Elements

Space Station 13 is a roleplaying game. Upon entry you create your character: give him a name, choose preferred roles, and select a haircut (this is vital, or else you will be called Bald).

Different servers have different levels of seriousness expected for roleplay, from "low RP" to "high RP." In general, no one tolerates "icky ooc"--in character knowledge being shared through out-of-character chat--and "oocky icky"--out of character knowledge being explained in a non-diegetic way through in-character chat.

Even high RP on SS13 tends to amount to little more than writing your dialogue in a funny accent, but this culture of roleplay does provide a framework for enabling truly gripping emergent narrative.

Don't Throw Away Your Shot

You have one life. If you die, you're out for the round. You'll have to wait for the next to join in once again. This has huge ramifications for player behavior.

Mayhem

This is what makes Space Station 13 worth discussing. Without central conflict, it's an interesting RP anomaly on a vast internet, fun to dick around with now and then but lacking anything worth holding onto.

But then come the traitors.

And the wizards.

And xenos.

And changelings.

And Syndicate.

On a traditional SS13 server, each round has some kind of hazard. A number of players are selected to be traitors, for example, and must work in secret to assassinate random targets, corrupt the AI, or destroy the station entirely via a nuclear device.

I stopped playing SS13 in 2013 because these elements weren't important enough. You'd be a traitor one out of every twenty or thirty rounds, and your odds of even interacting with a hazard before round-end were negligible.

That meant not enough conflict, and not enough conflict meant boredom.

I filed BYOND away and went on to other things.

Colonial Marines 13

In 2017, I started feeling nostalgic. So I redownloaded BYOND, expecting only to play a few rounds of Goonstation before going back to something else.

I was expecting a population of ~100. Instead, I saw 800. I recognized most of the servers, but the one at the top of the list stood out: it was called "Colonial Marines."

I love Aliens. I like it so much that I even beat Aliens: Colonial Marines. I clicked that server’s name very quickly. My life would never be the same again.

CM13: An Overview

In Colonial Marines 13, most players assume the role of marines investigating a distress beacon on a Weyland-Yutani colony, while the rest play either the xenomorph hordes or any remaining survivors. This playmode for SS13 preserves the game’s complexity while introducing formal deathmatch and competitive elements in a semi-structured “medium RP” environment.

This format gives SS13 what it always lacked: reliable conflict. That conflict facilitates endlessly engaging gameplay.

The Best Military Sim Ever Made

CM13 would not work if it was simply a deathmatch between marines and aliens. Instead, it is a combined arms military simulation, with a complex division of roles.

On the ship, players can control Command, Medical, Requisition, or Engineering staff, as well as the Military Police. Command helps coordinate marines; Medical fixes up their injuries; Requisition manages logistics, with Command’s help; Engineering keeps the USS Almayer operational; and the MPs grief players.

These roles together make sure the marines can do what they do. Any round of CM13 is a complex logistical dance between the soldiers on the ground and the officers in the sky.

But these are only support. The real gameplay occurs on the ground. The rest of the players assume the roles of either xenos, or the members of one of the four main squads. Squadmembers can be Leaders, Weapon Specialists, Smartgunners, Engineers, Medics, or PFCs. Each is self-explanatory. There are also two Tank Crewmen who drive the APC (formerly, a tank: RIP).

I’ve never been very Robust, and I usually preferred Medic and Squad Leader. I have played thousands of hours of Medic over the years. I find patching up my friends more enjoyable than shooting aliens. But Squad Leader gets the best vantage over what makes CM13 most fascinating. It is his job to coordinate Close-Air-Support and Orbital Bombardments with command and relay information to the CIC. As an SL, you will truly feel the combined arms splendor of this game.

Combined Arms

A marine in CM13 feels like a soldier on a futuristic battlefield.

A gunship rushes overhead to strafe the enemy, as requested by you, the Squad Leader.

A tank rolls up to give backup.

A Scout Specialist lazes coordinates and sends them to the CO, requesting a cluster strike on the planet.

Unseen logisticians work to keep everyone supplied.

It's the versatility of the above mentioned systems that enables this sort of gameplay. Of course, these complex systems, all held together by little more than the duct tape of player engagement, can often backfire, like when your orbital bombardment is a few tiles off-center and hits the evacuating dropship instead of the overwhelmed landing zone:

But moments like these are exactly what players remember most. As Squad Leader I have fired Orbital Bombardments that won the game many times. I have also fucked up my coordinates and obliterated dozens of marines more than once. If it was not for the ability to fail, the success would not feel so sweet.

Emergent Narrative

I have only explained the game’s systems so far, because these systems work together to create emergent narrative. When we sit down to play CM13, there is no story. All we know is that we are soldiers in the employ of the USCMC in the year 2200; we are expected to act accordingly, but the plot that follows is determined by gameplay. Our only “script” is our objective.

But a three-act structure shall always emerge from each round.

The first takes place on the Almayer, as the marines get ready, the threat is assessed, and a plan is made.

The second takes place on the colony, as any survivors are rescued and the war against the xenos rages.

The third, given a xenomorph victory, takes place on the Almayer, as the hive attempts to board and hijack the ship.

Throughout these three acts, your friends will fall. You will make miraculous escapes through deadly terrain, evading xenos at every turn, only to survive because of one lucky twist. And if you make it to the end, you will have one hell of a story to tell.

You probably won't make it to the end.

That's the price you pay. Most rounds will end when you get shot in the back by the new PFC standing right behind you with his shotgun out, or because you walked onto a hidden facehugger trap, or because the CAS pilot accidentally nailed you with a Fatty bomb during his strafing run and blew you into gibs fifteen seconds after deployment. Like the Orbital Bombardment onto the Alamo dropship, the potential for disaster is what makes victory sweet. You cannot have the good in this kind of simulation without the horribly bad.

Your heart would not pound as you made your run for the escape pod if you knew it did not matter.

An Example

We're pinned down at the Forward Operating Base. I'm playing as Alpha Squad Smartgunner, which means that my bullets pass through friendlies. It's been a long round; I've participated in four or five xeno kills, which is a decent haul.

Our SL didn't make it. It's taken some coercing, but I've persuaded our lieutenant on overwatch duty to promote me to assigned Squad Leader (aSL). This gives me access to the Command radio channel and makes it my job to keep the Combat Information Center up-to-date with battlefield intelligence. The most important thing is to make sure we're ready to evacuate should the need arise.

I check all of the vulnerable spots over and over again. We don't want to get caught off guard. The xenos have back off on their assault, but I don't know if that's because we're winning or because we’re getting flanked.

I check telecomms. Tcomms is connected to caves in the back-end, which makes it a probable spot for a xeno flank. My motion detector scans the area.

Nothing.

I go to check the north caves next, and it's manned and secured. Just as I'm preparing to move back to tcomms, I see another aSL shouting over comms:

"QUEEN TCOMMS!!!11!!"

I issue a Move Order and run back ASAP. Through the barricades, past an autoturret, and where 30 seconds ago there were no xenos and no holes in the wall, the Queen stands with five Tier Three xenomorphs.

There is no winning this battle. They are slashing our barricades, and most of the marines are already dead. We are being overrun.

I don’t bother to stay and fight. I turn, and I run.

Our FOB is built at LZ1. But if I can make it to LZ2 via the south, I have a chance of being extracted.

But the front entrance to the FOB is overrun. I have no way out. I’m fucked.

"Hop in!" a voice calls over local chat.

The APC pulls up beside me. They had the same idea, and they’re about to make a break for LZ2.

I leap inside, and we tear out through the FOB. The last thing we hear before comms go down is that the Normandy is coming down to pick us up. A few xenos chase us, but they back off as we reach the LZ.

The Normandy touches down. The APC rolls inside, and all of the marines pile out.

"LAUNCH!" I cry. "GET US THE FUCK OUT OF HERE!"

My motion detector pings xenos all around us. At least a dozen of them, just outside of the walls.

"LAUNCH! THEY'RE ALMOST THROUGH!"

I hear a creak. The west doors are being pried open.

"We have fifteen seconds of cooldown,” the Pilot Officer says. “Can’t launch. Hold on.”

I have made it so far. I cannot stand the thought of dying here. I climb back inside the APC. Just then the dropship engines start to whirr. Looking through the porthole, I see a lone Lurker xeno leap through the doors, just as they close again. Its scythe arms decapitate a single marine as we take off.

Then we’re in the air. But the Lurker is now on-board the dropship with us, and it’s going for the Pilot.

I have to do something. I rush out of the APC and try to stop it. But it slips into the cockpit, and the door slams shut behind it.

I’m a Smartgunner. That means I don’t have DS cockpit access. On the other side I hear gunshots and slashing. Screaming. Death. I can't do anything to help. Then comes the dying roar of a xenomorph, and all is silent.

The ship lands itself.

The CO rushes to us and opens the cockpit door. The Pilot killed the Lurker at the cost of his life. He can still be revived; but the Alamo, our primary DS, has been captured. Now the xenos are on their way, 40+ of them, and we stand no chance. Only five or six marines are left.

Command decides it’s time for us to call evac.

I run as fast as I can toward the north pods. The door slides open just as the floor beneath me jolts. I trip inside, landing atop a cryopod, as I hear motion down the hall.

The lights have all shattered, but I have NVGs as a Smartgunner.

A xeno is running toward me down the hall.

I slam my fist against the launch button, and the pod takes off.

You have successfully left the USS Almayer. You may now ghost and observe the rest of the round.

Writers Leave

Sometimes the admins and developers of CM13 will run roleplay events, with scripts and pre-determined stories. These range from months in planning to a few fun adlibs to engage the shipside crew during marine victory.

It all sucks.

It doesn't matter how brilliant your script is. Nothing will compare to playing to find out what happens. Constructed narrative is never as compelling as the narrative that emerges naturally from systematic forces. I have, literally, thousands of stories like this one. I could fill a book with them. They are stories so exciting and improbable that no author would ever think to write them.

There is story in CM13. It is important, and it acts as a facilitator. Writing and systems should always work together: to give context to the world, to let us know how to talk and act. But the systems should tell the story.

A narrative has its own benefits, of course. Themes and characters and dynamic relationships are pretty cool. These things are important, and they are not likely to be found in CM13. But if what you want are thrilling stories, CM13 can generate them endlessly.

Takeaways

Perhaps I have made CM13 sound like the most amazing game ever. In some ways it is. Let me now lower your expectations and explain why you shouldn’t play it.

It is horribly laggy. When server pop is too high, it is unplayable. This is all the time.

The servers themselves are miserable to play on. CM13 has had tyrannical, insane, and anti-social moderators for generations at this point. It was better for a time, and now it is much worse. If you think that the server admins obey their own rules against metagaming, I have been in their servers, and I can tell you that they are all on Discord together while they play.

The server’s development is totally incoherent. Occasionally a brilliant and revolutionary update will drop with fantastic new content. More usually a baffling and untested change will be merged into the live server, and something no one wanted will be changed. The balance is a mess, and developers and moderators are rarely on the same page. Developers, moderators, and players make up the three estates of the Ancient regime. They are approximately just as dysfunctional.

Then there is the fact that everything good about CM13 is also something bad. Every round is a gigantic timesink. Most end in anticlimax. The playerbase is hostile and antisocial. Nearly every encounter you have with xenomorphs will make you want to vent on Github. Probably worst of all, the game is aggressively addictive. I played way too much CM13 in college; it is the only time I have been addicted to a video game, and it adversely affected my social life in a way that I am not proud of.

I often want to play CM13 again. I never will. Part of that is because we have outgrown each other. Part of that is because I cannot afford to get hooked again.

There is still a tremendously important lesson to learn from my time playing it. That is the primacy of systems. Games are games, and even narrative games must be systems-driven. The token roleplaying elements in SS13 sometimes create memorable moments, but they do so haphazardly and unreliably. The systems in CM13 create great stories again and again and again.