Another reason LLMs are not the future of dialog

Executives have been trying to get rid of writers ever since moving pictures were a thing. Samuel Goldwyn allegedly threw a writer out of his office, and told his assistant, "I never want to see him again! Until we need him."

Many years ago, a Canadian network, probably the CBC, decided it could save a lot of money if it made a series about a bunch of commuters taking the train home. The actors would improvise their dialogue based on the day's headlines and whatever else came into their heads.

The show was a ratings disaster, but it was cheap to produce, so it stayed on the air longer than it should have.

The latest idea for getting rid of writers is large language models, aka generative AI. So far, the demos have not been promising. AI-scripted characters turn out to be boring and creepy. 

Well, think about it. If a roomful of live actors can't keep your attention, how can autocomplete software do it, when it has no concept of what it's saying?

Think about improv. The audience for live improv is tiny. If The Groundlings weren't a farm team for Saturday Night Live, who knows if it would exist as a business? Why? Most improv is a stunt. It is like the dog who walks on his hind legs:  it does not walk very well, but it is fascinating to watch it succeed at all. 

It is hilarious to watch the actors interpolate truth based on "yes, and." But you can only watch so much of it. No one is binge-watching improv. 

I don't doubt that there are sublime moments in improv here and there, but you don't see much live improv on TV. Saturday Night Live runs on cue cards. Paul Feig lets his actors improv, but he does dozens of takes and edits a performance together in post.

So if people are not, by and large, watching live human beings making stuff up on the spot, why would we watch autocomplete software do it? 

It remains to be seen how long it takes Hasbro to realize that no one wants D&D scenarios based on autocomplete, either, but I figure we've got another year of this before LLMs implode like NFTs and Beanie Babies.

Delivering the Goods

I'm watching GODZILLA MINUS ONE. You might think it is a movie about a giant undersea monster with nuclear abilities.

But it is actually a movie about survivor guilt in post-WW2 Japan. The hero is a kamikaze pilot who abandoned his suicide mission. He considers himself a coward. Other characters are struggling with having survived when so many of their family died. 

If you deliver the goods in a movie or game, you can do whatever the hell else you want. THE L WORD had sexy lesbians. Lesbians, and lesbian-curious folk, were watching to see sexy lesbians. The writers got away with making their characters really flawed people, because they delivered on the sexy lesbians. Without the lesbian sexiness, network executives would probably have said, "These characters aren't likable, make them nicer." But no one was watching to see nice lesbians, they were watching to see sexy lesbians. So the execs left the writers alone. 

Similarly, in GODZILLA MINUS ONE, there is all the nuclear monster spectacle you could ask for. Which means the filmmakers were able to tell a story about survivor guilt. I mean, maybe they started with Godzilla and thought "How can we make Godzilla fresh." But I suspect someone wanted to make a piece about survivor guilt, and then realized they could bring a bigger audience to it by doing it in a monster movie. 

Survivor guilt, by itself, is pretty heavy. Who wants to see that? Some people, but maybe not many. But if there's a nuclear monster, then sure! Bring it. The best horror stories are really proxies for stories about theme that are too bitter to take on directly. 

We're working on a war story right now. But we're turning it into a ghost story, because who wants to play a game about how war is horrible? 

 Deliver the goods, and you are free to do what you want!

Listen to the Fans!

I’ve talked about how to use feedback: listen to the criticism, be wary of solutions other people offer. Until your game comes out, fan feedback is often particularly dubious. They only know the games they’ve already played. They’re making guesses about your game. If they criticize it, they’re only criticizing the version of your game that they have in their head.

(This is a little like when your fellow devs criticize an idea you have. Often they are criticizing what they think you want, when what you want is something else.)

At a certain point, though, your game gets an announcement trailer, and now they have some data to react to. Do they like what they see? And is what they saw the same as the game you’re making? Then you’re in great shape. You’re selling what they’re buying.

On Fragpunk, we did our best to make the game wackier than competing hero shooters; and we gave the game a novel mechanic, allowing players to change the rules of the game before each round. (For example, give their opponents Very Big Heads.) To our delight, when our announcement trailer came out during the Xbox Showcase, social media and critics were all talking about how our game was a “wacky answer to Valorant.”

What if they love what they see and it’s not the game you’re making?

In my book on writing movies, Crafty Screenwriting, I suggest pitching your screenplay before you write it. Rather than writing a whole script and then trying to sell it, pitch a bunch of ideas out to buyers and see which spark interest. Then write the script that got the best reaction.

Samuel Z. Arkoff used to take posters for movies to potential buyers. The movies that buyers wanted, he commissioned scripts for those movies and then shot them.

What if you try to make the game they want? Don’t do it if it will hurt the game, obviously. But maybe the game the fans want is a better game. Maybe lean into that.

We essentially did that on We Happy Few. It wasn’t an announcement trailer, it was demoing the game at PAX. As I discuss elsewhere, we were working on a procedurally generated game, but what fans most liked in our demo was the hand-crafted individual encounters. So we went back home and pivoted the game to focus on hand-crafted encounters.

Listen to the fans.

Naming Names

At some point, you may be called upon to propose a title for a game. This generally won’t happen on a AAA game, where the title is the province of the marketing department, and is something like Assassin’s Creed: More Templar Shenanigans or Splinter Cell: Tracklist. But on smaller teams, writers are often involved. I’ve been part of the team coming up with names for We Happy Few, Stories: The Path of Destinies, South of Midnight, Biomorph and, as of yesterday, the company I work for, Netease, announced the game I'm working on, Fragpunk.

Developers usually give their game a working title or a code name to begin with. We Happy Few started as Glimpse. South of Midnight was once just Midnight. Some games only have a code number. The video game industry is secretive; most companies don’t like anyone to know what they’re working on until they're ready to start building awareness.

A working title can inspire people, and give them a sense of what sort of game they’re working on. Midnight is the witching hour, and the game is about a girl with witchy powers; if it had been a more humorous game set in the South, we might have codenamed it Moonshine. Glimpse referred to an early game mechanic where the procedurally generated world would regenerate whenever you weren’t looking. (We quickly realized that would just be annoying.)

The title needs to be something that players feel good about playing. I'm not sure I'd want "Alex is playing Shower With Your Dad Simulator" to come up on my friends' Steam feed, though obviously there are people who do.

It also can’t be too hard to type. I pushed for our game to be called I’m Afraid We’ve Come to the End of Our Time, but that was perceived as too long, in spite of What Remains of Edith Finch and Everyone's Gone to the Rapture. (We did eventually make a little spinoff VR game called We’ve Come to the End of Our Time.) Even Call of Duty gets abbreviated to COD because twelve letters are just too darn many to type.

But the main purpose of the title is to get people interested in finding out more about the game. It can do that in a few ways.

It can, first of all, just tell you what the game is about. Thief is about a thief. Portal is about making portals. Civilization is about building your civilization. Where in the World is Carmen Sandiego is about figuring out where Carmen Sandiego is. Unpacking is about unpacking. Guess what I Was a Teenage Exocolonist is about?

More often, the title hints at what the game is about without stating it so baldly. The Deus Ex games are about technologically enhanced human beings. Deus ex machina is a familiar Latin phrase meaning “god out of the machine.” The main character, Adam Jensen, is part man (made in God’s image) and part machine.

(In classic plays, the writer would sometimes get his characters in such a pickle that the only way he could bring it home was to have an actor fly in, supported by a crane (a machine), playing a god (deus, who would then sort things out. Deus ex machina refers to the writer resolving the plot arbitrarily rather than through the actions of the characters themselves. It's an implicit criticism, like "hat on a hat," although H.G. Wells got away with it in War of the Worlds.)

Kentucky Route Zero is about a road trip. But routes are never numbered zero; and are you really still in Kentucky? Mysterioso.

Call of Duty is a game about war. “Call of duty” is an old phrase referring to serving as a soldier.

All of these titles suggest rather than saying. A playing seeing “Kentucky Route Zero” will hopefully think, “Huh. What’s that about?”

How can you be South of a time of day? The rule in marketing is “sell the sizzle, not the steak.”

We Happy Few suggested that our few townspeople were happy, which indeed they are, but only because they're on happy drugs all the time. There are also fewer and fewer of them. Players could guess that the title was ironic. But how?

Biomorph is about a critter (a biological) who takes the shape (morph) of other critters.

What sort of a game are you trying to sell people? Is it quirky? Is it a survival game? Call it Don’t Starve. Is it about an octopus masquerading as a suburban dad? Octodad. Is it a dungeon crawler which is also a dating sim? Boyfriend Dungeon. A bureaucrat in a depressing Soviet-style transit office? Papers, Please. 

Of course, a game title can just be plain mysterious. The Return of the Obra Dinn is pleasantly ominous.

But a completely obscure title may not help with marketing. Disco Elysium was a hit, but probably not for its title. Undertale? Sigma Theory? Umurangi Generation? Engare? Goragoa? Zoombinis? These titles are distinctive, and shoot right to the top of the Google search standings. So that's good. But they tell you very little about the game. It’s probably best when the title doesn’t just stand out, it gives you a hint at least of the tone of the game.

So, why is it called Fragpunk?

Play the game when it comes out, and find out!

Senior or Principal Writer gig!

We're looking for a principal writer for an online shooter. You *must* know from online PvP games. Also, be legally able to work in Montreal (i.e. you are Canadian, or a permanent resident, or have an active work visa).

A Game Writing Glossary

Here are some terms I've heard thrown around while we were talking about game narrative. Some are entirely idiosyncratic to me and my crowd. Others are common. 

What are yours?

Artifact

Something left over from a previous draft that no longer makes sense or serves its purpose. Narrative cruft.

Asset

A single object that exists in the game, such as a gun, a tree, a beer can. An asset can be easily cloned. Environmental artists make assets. Narrative folk may write text for assets that have labels. What’s on the labels of beer cans in this world?

As You Know, Bob

Dialogue that recounts what both characters in the scene already know, so the player can know it, too.

Bible

A compendium of all the information developers need to make the game. Out of date moments after it is written. The “narrative bible” might have information about the world story as well as the characters’ backstories. Once production begins, developers tend to stop updating the bible, so double check any information you read in it.

It’s helpful if writing a bible to focus on things that can actually appear in the game rather than Tolkien-stye lore about the Three Silver Trees of the Greensong. Also, remember that a bible is not player-facing. Information you wrote in the bible is not in the game until you, or a level artist, put it in the game using a narrative delivery system.

Break Story

Laying out the beats of the story, usually in a room (or a virtual room).

Button

A punchy way to end a scene, providing the energy to jump back into gameplay. “Needs a button.”

Cinematic

A short movie that interrupts gameplay. Level designers often hate them because they don’t like anything interrupting gameplay. May be pre-rendered, in which case it is literally a movie playing in the middle of the game. Or it may be animated via the game engine using the character models and environments of the game. Some studios use “cutscene” to distinguish in-game cinematics from rendered cinematics, but the distinction is not universal.

Clam

An overused phrase or gimmick. “He’s standing right behind me, isn’t he?”

Consumables

Anything the player character can use up, such as health potions, grenades, or food. Usually found in the environment or bought in stores in-game. If you find yourself in a large room with health potions and ammo scattered around, expect a boss battle.

Couplet

Two lines, usually brief, one answering the other. Often used to button a scene. “I’m a man!” “Nobody’s perfect.” John Rogers calls this “the basic molecule of script dialogue.”

Dependencies

Ways in which narrative delivery systems interact with other disciplines. Readables, for example, have almost no dependencies. They are assets that can be strewn about the environment by the most junior of level designer, or a narrative designer, or even a writer. (Gasp.) They don’t require animation. They may not require special art. If they convey lore about the world, they don’t affect the game’s story. By contrast, a cinematic has lots of dependencies. It can’t be written until we know what the story is. It will require animation and possibly mo-capping. An actor has to record lines, which have to be mixed. Putting cinematics into gameplay seamlessly can be a person’s whole job on a AAA game.

Downloadable Content (DLC)

Additional content for the game. May have new maps and tell a completely new story. “DLC” is also what your producer tells you when they cut the levels you’ve been working on for a year.

Encounter

What happens when the player character runs into someone or something. Level designers create and script encounters. If it’s a dialogue encounter, narrative folks will write the dialogue, Lord willing.

Environmental Artist

An artist who creates assets that will appear in the game. They may put them in the game map, if they don’t affect gameplay; level designers may put them in the map if they do. Narrative folk work with artists to make sure the environment tells the world story. For example:

This was once a grand palace, but now it is covered in weeds. Sprawled here and there on the floors and benches, servants, courtiers, knights, and ladies sleep dreamlessly.

Environmental Narrative

Narrative descriptions of the game environment, creating a world story, a mystery, a mood, or anything else that attracts us to play in this world and makes us care. Each major location in a game might have its own narrative description, in addition to whatever notes the art director is giving her artists.

Environmental scenario

A few items laid out in the environment to tell a story. For example,

A table covered with playing cards. Two chairs, one fallen backwards. Dried blood spattered on the cards. A bottle shattered by a bullet. Under the table, a scrap of paper – on interacting with it, it turns out to be an IOU.

Narrative folk write these descriptions to convey the world story. Ideally, an environmental scenario uses assets that already exist, or are already planned for the game. Creating a single asset for an environmental scenario can take an artist a week, which is expensive. The sooner you can get these to the environmental artists, the better.

The difference between an environmental scenario and environmental narrative is just scope. A scenario is a single collection of assets intended to convey a single idea. Environmental narrative encompasses everything that conveys what the world is.

Exotic

Something that only happens once or twice in the game, or in only one level. Exotic gameplay might be when a level introduces tightrope-walking as a verb, but there are no tightropes later on in the game. 

First Person

A way of showing the player the world in three dimensions as if through the player character’s eyes. The player character is usually invisible, though sometimes you can see your hands. First person games are more likely to have a blank main character, so the player feels like it is them in the game world. 

First Person Shooter, or FPS

A first-person 3D game in which you shoot people in the face.

Fridge Logic

A logic problem in your story that no one is going to notice until they get up and go the fridge. Often not worth fixing.

Gameplay designer

Someone who designs gameplay features, such as verbs. Features are active on many or all maps, unless they’re exotic.

Greybox

An early state of a level, before the environmental art is in. Looks like a bunch of grey boxes. No point arting up a level before you know if it works.

Hang a Lantern on It

To draw attention to something. If there’s something you really want the player to know, you hang a lantern on it somehow.

You can sometimes get away with inconsistencies by hanging a lantern on them. “How come he can fly and we can’t?” So long as the players know that you know it’s inconsistent, they are more likely to forgive you than if they think you’re being sloppy or think they’re dumb.

Hat on a Hat

Gilding the lily. Trying to make something stronger or funnier by adding something else strong or funny, thereby paradoxically weakening it. A hat does not need a hat on it. A man in a gorilla suit is funny. Give him a clown nose as well, and you have a hat on a hat. Also known as “gilding the matzah.”

Idiot Ball

When a character makes an out-of-character and obviously dumb decision, or acting on misinformation they could clear up in a text message, they are carrying the idiot ball.

Killing your Darlings

A darling is a bit of narrative you particularly like and are particularly proud of – but it isn’t actually good for the story as a whole. You must kill your darlings, as Eudora Welty said.

Laying Pipe

Setting something up early that will pay off later. Akin to installing pipes so you can use the sink later.

Level

A hunk of the game map that is loaded into the consoles or computer’s memory at one time. In Unreal Engine, a level is usually on a single map. Levels may or may not be apparent to the player. If levels are separated by loading screens, they will be pretty obvious, but next generation consoles don’t necessarily need loading screens. Often a single level designer “owns” or is responsible for a level.

Level Designer

Someone who designs the levels and what happens in them. Their job is to make gameplay fun. They have to know how the features of the game – especially the verbs – can work together. They are known for having a god complex. They create worlds, after all.

Loading screen

What you see on the screen while the console or computer is putting the level in memory. Usually a still image with some text. Can be used to remind the player what’s happened so far in the game, or to reveal some aspect of the world story, or give gameplay hints. The loading screens for Spec Ops: The Line mock the player, “Do you feel like a hero yet?”

Magic Hat

A Magic Hat is a special power that you don't explain because the explanation wouldn't actually resolve anything. The point of magic is that it's magic. You can create lore behind a magic hat (e.g. the One Ring, Excalibur). What you don't need is a scientific explanation:

Midi-chlorians were intelligent microscopic life forms that lived symbiotically inside the cells of all living things. In sufficient numbers, they could allow their symbiont to detect the pervasive energy field known as the Force. Midi-chlorian counts were blah bitty blah blah blah. 

And what gives midi-chlorians this power? Their magic hats? You're just kicking the can down the road.

Narrative Designer

Someone on the narrative team who focuses on getting the content into the game rather than creating the content. They might place assets into the level map. They might do some scripting. They work with the level designers to make sure that the gameplay is, as much as possible, also telling the story of the main character and of the world.

A narrative systems designer defines narrative delivery systems for the game. In a smaller team, that job might be folded into narrative design, or even writing.

The dividing line between writers and narrative designers is blurry, though. Some narrative designers write player-facing narrative content. Others don’t. Some writers do narrative design, whether they call it that or not.

Platformer

A game focused on how the player character gets from point A to point B.

Player-facing

Something players will experience in the game. Lore is not player-facing until it appears in a readable, a journal entry, or an environmental asset such as tavern sign.

Readable

An asset that exists to be read. Usually when you interact with it, you can read some text on a separate screen, for example as an entry in the player’s journal or inventory. A poster for shaving lotion that you can read as you ride into town is not considered a readable; it’s a “decal.”  Readables are by far the cheapest way to get lore into the game. Most developers will not put anything critical into a readable, since God forbid the player should have to read something. Baldur’s Gate 3 bravely puts information useful to the player in readables. That’s a game design decision.

Run and Jump

Describes a section of traversal in a game. “And here we have a little run and jump until we get into our first combat encounter.”

Second Draft

Something meaningful in screenwriting contracts and which producers like to put in their narrative pipeline chart, but isn’t really a thing. Writers may write a vomit draft, a rough draft, and a first draft. You may get your draft approved, and eventually recorded. But all the drafts between first draft and approved draft – why count them? How many changes do you need to make for it to be a second draft? Every draft should be considered a first draft until it’s recorded.

Shmup

A shoot’em up. A shooter. As opposed to a beat’em-up, a game focused on melee combat.

Shoe Leather

The narrative hoops you have to jump through to set something up dramatically -- e.g. establish that the hero is an orphan, remind the audience of their psychokinetic abilities, etc. all so it pays off later. Not that far off from laying pipe, but more of a criticism. “That’s all just shoe leather.”

Shooter

Any game focused on shooting NPCs. Game developers add “in the face” almost instinctually.

Snappy Banter

As a putdown, dialogue that sounds clever and punchy but has little emotional content.

Table Read

A readthrough of all the dialogue scenes. On the first day of rehearsals for a play, the playwright, director and all the actors would get together and read the play, sitting down. Our version might be on Zoom, to make it easier to get actors to participate. Pay the actors if you possibly can. You’re getting paid, aren’t you?

Taking the Curse Off

When something is cliché or dumb, you may not be able to get rid of it for reasons. But you can sometimes take the curse off it by twisting it in some way.

Telegraphing

Foreshadowing. Giving the players too heavy a hint where the story will go. You don’t want to telegraph the punchline to a joke. Comes from sending a telegraph saying you’re coming, rather than just showing up.

Traversal

Gameplay focused on getting from here to there. Narrative audio is traditionally constrained by the assumption that the player is traversing the level as fast as the game allows them, even though the player may not be. If the narrative is longer than the minimum traversal time, you need a game mechanic to prevent narrative audio from stacking up.

Unobtainium

A substance that everybody in the science fiction game world wants but is hard to get. It might be the thing you need to make the best sword, or a spaceship. In Avatar it is literally called unobtainium. Very common in fantasy too (e.g. mithril).

Up and Back

When the characters or plot go through action sequences, dialogue scenes, complications, etc.,, but the story does not progress. You could cut out an up-and-back and you wouldn’t feel anything is missing.

Verb

Any action the player can take in the gameplay, for example, “walk,” “jump,” or “dodge.” A shooter has specialized verbs like “shoot,” “take cover” (if it’s a cover shooter), “stealth takedown” (if stealth is a mechanic), and “throw” (if you can throw grenades or other consumables.) A platformer might have verbs like “climb,” “mantle” and “use zipline.” A walking simulator might have “examine.”

World story

The history of the world, as distinct from what happens during the game. What sort of world we’re in can be as, or more, important than the player character story. There may be a bible that defines the world story. It is already hilariously out of date.

Writer

A person with a socially acceptable and occasionally paying addiction. There is no known cure. Often an introvert.

The recipe of a work of art, according to Rothko

 We saw the Rothko retrospective at the Fondation Louis Vuitton in Paris. Fascinating to watch how his style changed -- how he found it, and then sort of lost it. This was interesting:

<blockquote>The recipe of a work of art -- its ingredients -- how to make it--the formula.

1. There must be a clear preoccupation with death--intimations of mortality... Tragic art, romantic art, etc. deals with the knowledge of death.
2. Sensuality. Our basis of being concrete about the world. It is a lustful relationship to things that exist.
3. Tension. Either conflict or curbed desire.
4. Irony. This is a modern ingredient--the self effacement and examination by which a man for an instant can go on to something else.
5. Wit and Play...for the human element.
6. Hope. 10% to make the tragic concept more endurable.

I measure these ingredients very carefully when I paint a picture. It is always the form that follows these elements and the picture results from the proportions of these elements. 

I still am most fond of the canvas he's got up at the Museum of Modern Art in New York. Like any Rothko, a photo doesn't begin to do it justice -- it is not one single color of blue, it is many blues. Up close and personal, it's mesmerizing. You could fall into it. 

How to Take Creative Criticism

(This is an adaptation of a talk I've given a few times, at companies I've worked for, and at the East Coast Game Convention. I wrote it about video game writing, but it is almost entirely applicable to film and television writing, and mostly applicable to a lot of other creative endeavors.)

In this business, you’ll get creative criticism from all sorts of people. Your boss, probably. Sometimes by co-workers. If you’re smart, you’ll ask for criticism from people who report to you. Whenever the game comes out, you may get criticism from random strangers on the Internet, although that is often not very helpful.

Sometimes you will get notes that solve problems you are having. Other times you’ll get notes you disagree with furiously. Sometimes they will be the same notes.

None of us comes to creativity without an ego. To be a professional creative, by definition, is to be someone who wants other people to have in their heads what you have in your head. If that’s not important to you, you don’t care. If you don’t care, you won’t do great work.

Ego regularly throws up obstacles to making creative criticism useful. How do you outwit your own ego?

Don’t take criticism personally

First of all, try not to take the feedback personally. This is hard to do. Someone is criticizing your baby! Obviously, they hate you.

When you take criticism personally, it becomes an attack. If you’re attacked, you defend yourself. If you defend yourself, you shut out the criticism.

(Note that feedback, especially public feedback, sometimes is a personal attack. The bigger the company, the more likely it has politics. Sometimes people will have it in for you. But you’re still better off treating an attack as if it is sincere, constructive criticism, if only to judo your opponent into loo
king like a jerk.)

All feedback is legitimate

If someone says they don’t like something, they don’t like it. You can’t tell yourself they ought to like it. If someone doesn’t laugh at your joke, it’s not funny, at least to them. If someone says they’re confused, then your writing confused them. If someone says they hate a character, then they hate them. It’s on you to figure out why, and what to do about it.

You may even be getting criticism from someone who does not like the genre you’re working in. What you choose to do with that information – whether you feel it is not relevant, or you embrace it – is up to you. But their criticism is still valid in the sense that that is their experience.

Distance yourself

It helps not to think of the work as your baby. Think of it as “the” baby. It’s not “my” scene, it’s “the current draft of the scene.” This also gives other people permission to critique it more freely. You can even criticize the work yourself. Maybe you have an idea of something that’s wrong with it, that you haven’t had time to fix, or don’t know how to fix.

Critiquing your own work, incidentally, makes anything positive that you might say about it more believable.

The problem may not be where they think it is

Just because all feedback is legitimate doesn’t mean that your note-giver is right about what the problem is. They may be picking at a symptom, not the cause.

I studied computer science in university. In programming, if something goes wrong three quarters of the way through, it’s often because you didn’t set something up properly at the beginning. In narrative, someone will say a scene is too long. But shortening the scene will not make the scene better. That’s because the real problem is earlier, where the game story failed to make you care about the characters.

Bear in mind, if you show someone a scene, they won’t have in their head all the scenes that went before it. You’ll know who the characters are, but they may not. It’s up to you to interpret the feedback you get.

Assume they are not an idiot

However, before you go off and decide that they’re wrong about the problem, allow for the possibility that the person giving you a note is not an idiot. There are almost no stupid people in video games. Right?

What often happens is, when someone gives a suggestion, what pops into our mind is an idiotic version of it. You may think they want a sweeping change. You may feel like they’re trying to cut the heart out of a scene or a sequence. Why would they want that? That’s just dumb!

Before you reject their identification of a problem, try to figure out what a non-idiotic version of what they’re saying might be. “If this wasn’t dumb, what would it be? Then ask if the non-idiotic version is what they meant. If it is, bravo! Now you have a good problem to solve instead of a bad one. Even if they did mean something dumb, at least they are clearer on what you’re thinking, and you’re clearer on what they’re thinking; and you have identified a non-idiotic problem your work might have.

Find the Truth

Back in the previous millennium, my acting teacher used to say, “Find the truth.” It’s easy to find something in feedback that might give you an excuse to dismiss all of it. “They are definitely wrong about x, so they are probably also wrong about y, z, a, b and c.”

Search for what is true rather than what is false.  Figure out how the critique is true. Your goal isn’t to defend your work. Your goal is to find how it is weak. Actively seek out the criticism, even if it is buried in confusion, contradiction or fluff.

The hard part of science is not finding evidence that corroborates your theory. The hard part is thinking up ways to disprove your theory. If you try to disprove your theory every which way, and it holds up, it’s a solid theory. But if you only try to support it, you’re leaving it to other people to point out its flaws.

Actively dig for the truth. Sometimes people will point out something that, you agree, is a problem.

But you also get the feeling that something else is bugging them. They just can’t crystallize what it is. So they grumble at it, and move on.

This is often a much more structural, serious problem than the thing they were able to criticize. It will require more work to fix. The temptation is to avoid the hint and focus on the surface flaw.

Don’t ignore it. Interrogate them (nicely). Try to work with them to crystallize what they’re having problems with. When someone tells you, “Nah, forget it, it’s fine,” you say, “No, seriously. I can tell you’re bumping on something. Let’s talk about it.”

If you don’t, they may come back with the same criticism in three months. Or, play-testers will have the same problem. Or, God forbid, players.

Don’t wait for the snippy Let’s Play video. Dig.

What they want you to do about it may be wrong

All feedback is legitimate. Most criticisms have some truth to them.

Many solutions proposed to you will be bad. Hopefully a more experienced writer can give you an approach that will solve the problem. Other people, particularly non-writing managers and people from other disciplines, will give you suggestions that fix the problem in front of you and create many more problems elsewhere. Even your boss may give you a bad idea, if you spent a month thinking through what you wrote, and they’re skimming it during an online conference call about optimizing frame rates.

Feel free to take the good bits of the criticism and ignore the bad solutions. Unless, of course, where you are, congeniality is prized over creative results. (That’s called “toxic positivity,” and most senior developers have experienced it one time or another.)

Find out what problem they are trying to solve

Surprisingly often, people forget to give you their critique, or gloss over it. They jump right to their solution.

It’s wise to find out what they’re trying to fix before addressing – or rejecting – their solution. They may be proposing a big, involved solution to a problem you can fix surgically, in a line or two.

Or, you may realize that the problem they brought up is actually bigger than their solution, and is going to require more work. That’s useful to know, too, even if it is not as cheery news.

“We’ll take a look at that.”

So, you really worked at understanding the criticism, and you know what their problem really is. But you’re still convinced that they’re wrong. They’re honing in on something that’s not the real problem, they want something fixed that isn’t broken, they want a change that will break things, they didn’t game it out.

You may find yourself from time to time in a situation where you can already see their idea won’t work. You can work it out in your head. Or maybe you already tried it out, like, three months ago. It didn’t work then, and it’s not gonna work now.

Imagining how unwritten things are going to turn out, whether they’re going to play nice together, what they’re going to break – this is a writer’s superpower. Surprisingly, most non-writers cannot do the math in their heads. They can’t imagine doing the math in their heads. So when you tell them you already did it, and that’s not going to work, they won’t believe you. They’ll think you are being precious. They may feel disrespected.

I have got in hot water for telling folks about problems I could foresee in a story. Two years later, sure enough, the story had those problems. (Didn’t help me any, but I was already at another company.)

Say “We’ll take a look at that.”

Then really do take a look at it. Actually. Really truly.

That gives you time to remember that the person who gave you the note is probably not an idiot, to find the truth, to find the deeper problem, etc. Time for your hackles to come down. Time to try out the suggestion, so you can say, “I tried it, and I couldn’t make it work.” Or “that doesn’t work but how about this.” Or “it works if we do it this way.”

Or, “I tried your idea and it worked. Thank you!” That generally goes down pretty well.

(You can even say this when they didn’t propose anything, but their notes triggered a good idea on your part. Or when they proposed something only tangential to what you did. They will often agree that it was their idea.

When I was a kid, we were told that we should be truthful about everything. But we were also told that avocado green was a good idea for your kitchen, so we should have known better.)

Is it a hill to die on?

Sometimes you have to execute a note you think is wrong. Videogames are a collaborative art. If you don’t execute the note the way higher-ups want, at some point they will find someone who will.

Play the long game. If you were right, at some point it will become clear that the note did not work. If there’s time and resources, perhaps it will get fixed. And, you know, maybe you were wrong. There have been times I thought I was right, and I was wrong, you know?

Is it a hill to die on? Probably not. I mean, if someone asks you to write something legit evil, then yeah, kick up a fuss. But if they’re just asking you to write something dumb, well, it’s a video game. Video games are sometimes dumb. Your career is more important than a video game.