Wednesday, June 27, 2012

Aji Again

I use a concept called "aji" when thinking about game narrative design. Here's a good explanation: Unlimited Hyperbole. That's ten minutes long, but the explanation is about seven minutes in.

I'm going to describe my version, but you can listen to his version if I bungle it. I think they're the same basic concept.

"Aji" is a term from the board game "go" (or wei'qi, baduk, etc). Very roughly, it translates to "potential". The idea is that you put down a few stones in a loose configuration. The idea isn't to finalize the capture of the area. You don't build a solid wall. You just put a few stones down in a good configuration. Then, when the enemy approaches, you can use those stones to support your response. That may end up being a solid wall, or a tentacle monster than destroys the region for everyone, or even an inversion, where you force them to waste turns taking those initial stones while you play lightly and aggressively elsewhere.

To me, plot in games is the same.

You put down a few elements in a configuration. You want the configuration to have good aji - good potential. As the players act, you want the initial elements to flex and bend but support the new elements that arise.

It is similar to go because the play style is similar. The player makes a move, the game makes a move. If the game makes too many moves, then the player feels either lost or stifled, depending on the situation. If the player makes too many moves, they feel like they are boxing air rather than fighting against an evil mastermind.

Now, Pinchbeck is probably focusing more on the idea of vagueness, of leaving things open to interpretation. I tend to focus more on the idea of pacing and player agency. But the two are, in my opinion, very tightly linked.

Establishing too much, too fast is a great way to lock the player into a boring, trite plot they feel nothing about. On the other hand, establishing just enough to make the player go "and then?" will keep the players interested and invested.

In games with a GM (tabletop RPGs), the lack of detail is a huge asset because it allows you to flex and turn when you need to. If the players aren't sure whether Darth Vader is a good guy or a bad guy, you can establish him as one or the other when play demands.

In theory you could do something similar in a computer game, but it would probably require a lot of scripting. But the vagueness and openness still pays off, because the player spends a lot of time thinking about the various possibilities.

That's my philosophy on narrative design in games. Put down pieces. Places, people, concepts - just a few. Allow them to be established as the player moves through your world, rather than trying to force them on the player. You only rarely need to spell things out - that's usually too slow, boring, and laborious.

It's really not that hard: it's mostly just a matter of self-restraint. I know you want to show off all the details of the cool bad-ass you invented. But don't. Or do it later. Let the player breath.

Saturday, June 23, 2012

I've been thinking about Hollywood and the music industry.

We talk about "remix culture". It's commonly accepted that people like to remix things. We hear a song, we want to sing our version of it. We see a super hero, we want to put him through adventures we think up. We play a game, we want to make a version of it that is "just like that, but with more of MY ideas in it."

Certain things, such as dubstep and sweding, are flat-out remix-only. Of course, they are done on the cheap, largely without any sort of publicity, because they know that at any time a lawyer could come and tell them that remixing their culture is illegal unless they buy permission.

Hollywood, the music industry, and even the games industry to a lesser extent are all desperately trying to control remix permissions.

Because they are a remix culture.

They produce almost nothing new. Everything is recycled and spit-shined.

Used to be that you did remixes until you wanted to prove you had the chops to do original stuff, and then you took the big risk of creating something truly original.

Nowadays it's the opposite! You have to do original stuff or lawyers show up at your door. You have to work your way up to doing remixes. You have to earn the right to participate in your shared culture.

The "culture engine" has stopped. The big boys that drive the culture have put the car in neutral, and the little guys who used to refine and consume the culture have gotten out and tried to push. But it's an awfully heavy car.

Well, that's been my Saturday Morning Thoughts.

Friday, June 15, 2012

It's a-me, Bowser

I guess it's cliche to talk about how Bowser isn't really the villain, but think about it. Bowser never kills anybody from the Mushroom Kingdom. Mario kills thousands of Koopa civilians in every game.

Bowser's villainous plans involve nonlethal surgical strikes against a known aggressive military power. His demands and further plans always boil down to "stop hoarding all the resources in the universe, you dicks!"

For example, here's one of Bowser's evil military plans:

"President Bowser! We've finally completed our armada of flying battleships! We can finally get vengeance for the tens of thousands of innocents murdered by those Toadies!"

"Yeah, sir, let's glass 'em!"

"No! If we kill them, we become them! Instead, we will perform a nonlethal surgical strike to capture their main military fortress and their leaders. Then we will negotiate for reparations and peace!"

"Snf ... you're my hero, sir!"

"Thank you! Now, man the nonlethal stasis beam cannons!"

Then, of course,the interdimensional horror that is Mario comes along and murders everyone. And the Mushroom kingdom releases another piece of propaganda about the evil "King" Bowser.

This propaganda isn't even vaguely believable. Princess Toadstool is some kind of willowy blonde lady? She's a mushroom. She's princess of her people. She looks like them. And Bowser? Thirty feet tall and breathing fire? Yaright, we believe that, Mario. He's just a turtle with a skin condition.

You're the only one that changes size and breathes fire, you pandimensional horror. The nuclear weapons program of the Mushroom Kingdom.

The only real question is why the people of Koopa Democracy keep electing Bowser if he's always losing to the Mushroom Kingdom. The answer is actually that Bowser is a really great leader. Think about it.

1) He always loses to the Mushroom Kingdom, but when he does so he lures the mass-murdering Mushroom Kingdom strike teams into some crazy death trap far away from Koopa civilians. And then he faces Mario head-on so no more of his men have to die.

2) The times someone else comes to power it tends to be an interdimensional godling with an indecipherable accent and a lust for murder. Sounds like they all come from the same place - Mario's world.

3) He's a really great peacetime ruler. First, he rebuilds everything between wars. Not just military armadas, but theaters and factories and playgrounds. Second, his people are so wealthy that even the random soldiers can't be bothered to grab the money, clothes, and food that litter the landscape.

It's a laugh seeing Mario smash what we loosely refer to as its "head" into brick walls just on the off chance we accidentally dropped a penny inside. Well, until it suddenly starts flashing and killing us while moving at mach speed.

Hell, in a world where that sort of creature exists and is on the leash of your worst enemy, wouldn't you keep electing the guy that keeps it out of your towns and cities? The guy who is willing to give up on military objectives just to protect his own civilians?

This is a real stand-up turtle.

Villain my ass.

Thursday, June 14, 2012

Social NPCs

Man, what a depressing essay to rewrite. The first version was deleted by a mysterious act of god - by which I mean Google saved over it with an empty body of text. So I'm going to write it again. But I'm going to do it while sighing heavily and looking very put-out.

Social NPCs are kind of a last-decade topic, but I don't care, I'm going to talk about them anyway.

Most games where you can socialize with the NPCs, the NPCs boil down to "I like or don't like you". This is "one axis" socialization. It's boring and easily gamed.

You can expand the number of axes, but you start to run into expression problems. How can the player tell that an NPC trusts them but doesn't like them, and why does it matter? What effect upon the game world will the NPC have if they like you and trust you, instead?

A lot of people have posited a lot of different models. Here's the simplest one I could think up - the simplest one that resulted in complex, fun, adaptive social behavior.

I call it the "Fuck Google for Making Me Write this Twice" model, or the FG algorithm for short. (Actually "face & gain" algorithm, but whatever.)

Every NPC has a confidence level. This determines how likely they are to come up with or participate in risky plans. Someone with a high confidence level will want to go out and kill monsters, or build a space ship, or talk to the weird tentacle alien, or whatever the game world's plans are like. Someone with a low confidence level will generally resist such plans, and tend to want to take the lowest-risk action. Which is frequently "stay home and prune the hedges".

There are two kinds of social events. One is the gain event, one is the face event.

Gain events happen whenever an NPC succeeds or benefits. Gain events affect an NPC's confidence, so they may grow more or less confident depending on your (and their) actions. Gain events happen naturally during a plan's execution: if it works out, the participants all get a positive gain event. If it fails, a negative one. You can also have gain events in response to things you didn't do, such as being given gifts or having continuous bad luck.

Face events happen whenever an NPC seems particularly strong or weak to those around them. These are not linked to anyone's confidence levels, but they will change an NPC's mood for some length of time. Face events most commonly happen at the same time as gain events: plans succeeding or failing (in front of people). However, they can also happen when events cause someone to look cool or foolish even if they took no action to do so.

These events are saved. Instead of saying English words such as "trust", we would think of it as "they have seen you gain confidence and face many times". Instead of "like", we might say "you have helped them gain confidence and face many times."

In addition, depending on how you craft your event structure, a lot of really organic NPC behavior can arise.

For example, you might do an event in front of as many people as possible, so the confidence gain and face events are as large as possible. Or, if you're uncertain of success, you might do the plan on the sly, hiding it from people, so it affects your confidence but not your face.

You might execute a plan to gain confidence or face, while the end result of that plan might cause someone else to have their confidence or face affected. Bullying would be to push their face down, or perhaps you're a carebear who helps them gain confidence. Depends on the personality written into the NPC and the structure of the events.

Similarly, when the player enters the stage, they will find it somewhat complex to simply optimize their friendships.

For example, if they help their new friend gain as much face and confidence as possible, they will be very well liked by their new friend... but their new friend will basically lead them around by the nose and tell them what to do. This is quite separate from what the in-world results of your plans did.

So maybe you farm face without confidence... except then you have an issue where not everything is determined by the total of both kinds of events. If you've helped someone gain face but not confidence, their plans (confidence-based) are not going to include you, even though they will be happy to hang out with you.

Or maybe there's a struggle to keep an overconfident friend from self-destructing due to his doomed, aggressive planning. Well, maybe the best way to do that is to simply make sure nobody notices him when he's doing something, so that when he fails he loses confidence but not face.

And, of course, NPCs can vary. This one starts nervous, that one starts overconfident. That one considers a high charisma to be equivalent to you gaining face events, while the one over there considers a high intellect to be equivalent to you gaining confidence events.

Everything ends up depending on the scaffold: there are definite limitations. And, of course, depending on how much smoke and how many mirrors you use, the lack of an intelligent social system may be quite obvious. It's not terribly realistic, but if we wanted to be realistic then socializing would be a giant pile of obvious optimizations.

All this requires to be interesting is a wide variety of plans that alter the shared world, and a predictable scaffold of alone/with friends/in a crowd schedules to give you a variety of face options. I've tested it with some prototypes in different kinds of games, and it seems to work okay in all of them, whether we're talking about plans being very rare and largely preprogrammed or common and done on the fly.

The key is that their opinion of you is based on how much confidence they've seen you gain (respectish), how much face they've seen you gain (leadershippish), how much confidence you've helped them gain (mentorish), and how much face you've helped them gain (lackeyish).

And now I'm done. Again.

Tuesday, June 12, 2012

Time Travel Games

For a long time I've been trying to come up with a good conceit for a free-travel time travel RPG. That is, a game where the players can roam to any time period they please freely, rather than being shuttled through time on a plot ferry.

The problem is that time travel pretty much takes care of all the human concerns that tend to drive us. Even the biggest challenges can be solved by simply popping back to before they started and changing a small variable. A time traveler who wants to keep his family from dying? Huge plot line! But the only way to make it difficult to do that is to add arbitrary levels of complexity and turn it into a fetch-quest thing where you have to first go back in time to change the things that are stopping you from changing the thing you want to change. It's not very compelling! It's just complicated and pointless.

So you introduce overarching time demon plot lines where in the end you have to fight against time itself or your older selves or some other stupid thing that is forced down your throat. Your personal goals have nothing to do with it - you have no interest in this plot except in that you can't escape it!

How do you introduce human interest into a story inherently about someone who has no human concerns? Even if they have human impulses - say, they love somebody - they can still just pop off to fix whatever might go wrong. The only game I can build out of that is time-traveling whack-a-mole.

Which, actually, sounds kind of interesting. BUT! Let's presume we want something with a bit more depth.

How about this? Time travel as self-improvement.

There are a lot of time travel games about restoring the timeline to your normal timeline.


Shouldn't it be the opposite?

Wouldn't a time travel go back in time and tweak it so his family was rich? So he was born with superpowers? So he was elected president?

You could even ally with other time travelers, cooperate to improve eachother's time lines, to use less of your life tweaking your timeline so you have more of your life to spend on other stuff...

"Hey, let's all just have the same parents. And back in the fifties, we can give them the mutant gene..."

Thursday, June 07, 2012


So, I read another person irritated by the way that the town blacksmith refuses to give a discount to the world-saving heroes. And, you know, they're right. It's almost like game designers don't really know what it's like to be world-famous, or to have a writ of unlimited authority.

Take Commander Shepard. He or she has saved the galaxy twice, and was famous even before that. He's like the Beatles, if they always drove around in a technicolor van labeled "Beatles".

And yet you can just pop off to see the refugee camp? Hell, Jackie Chan once had to walk down the middle of a busy highway to get away from a crowd of fans, and he didn't even save the galaxy!

When you enter the station, there'd be a whole squad of riot cops there to escort you, and a concierge with no particular scruples to get you anything you need.

Shep: "I'd like to go see the refugee camp."

Conc: "... ha ha ha, let's just pretend you didn't say that."

Shep: "Well, at least I'll pop off to the zocalo for some shopping."

Conc: "Hmm, can't advise you do that, either. Listen, why don't you just tell me what you need and I'll fetch it for you."

Shep: "Nonsense, it's just right through this door..." [opens door]

Voices: "Is. that. Shepard?" "SHEPARD!" "AAAAH! IT'S SHEPARD!" "WE LOVE YOUUUUU!" "SIGN MY FACE!"

Shep: [closes door]

Shep: "I need a type four barrel extender and all the fish."

It'd actually be kind of interesting to play a game where you really are that famous. I can tell you one thing: you wouldn't be randomly walking up to people and chatting with them. Not without seeing the look of confused panic in their eyes, at least.

Tuesday, June 05, 2012

Large but not titanic

I really do like analyzing the games I've played and looking for inspiration in the mechanics. Most of the time, the insights are fairly minor - for example, the only thing I'm gaining from the RPG I'm playing on the DS at the moment is that it might be fun to equip catchphrases rather than weapons and armor.

But sometimes you unearth a vast labyrinth of unexplored gameplay caves.

I've more or less finished with Dragon's Dogma. A variety of reasons mean I didn't get more than 20 hours in, but I enjoyed those 20 hours. And there was one part that I didn't enjoy, but I really wanted to.

In Dogma, many of the enemies you fight are quite large. Not truly huge - these monsters aren't walking landscapes like in Shadow of the Colossus - but between five and ten meters.

These monsters are large enough that they have multiple hit locations. You can climb on them, shoot projectiles at them, and so on.

This isn't quite the same as the "climb to strike its weak point" that I'd seen before. It's much more fluid. Let me see if I can give an example:

I was fighting a cyclops, the first tier of giant creature. I was trying to hit its head, thinking "yeah, that's gonna be the weak spot." Fireballs weren't doing much - I was having a hard time hitting the head, since the aiming controls use the same thumb as the firing controls. So I switched over to using lightning, which comes from the sky. But I missed with that and struck its arm - and it dropped its club. Cool!

Later on, with my melee character, I began fighting ogres, the second tier of giant creatures (the tier where the whole thing falls apart, but more on that later). There are a bunch of nice details here: ogres can grab characters and run off to finish them solo, but you can interfere by blasting their arms. You can climb on them, but they do this cool back-flop where you have to jump off quick, but then you can jump right on their heads. Moreover, you can use terrain to your advantage, standing on a small cliff such that their head is at the height of your basic swing and the hit zones for their attack are all below your feet (unrealistic!).

Now, this whole thing was a really interesting idea, and much more entertaining than simply killing orcs. But I think there's a lot more potential hiding underneath.

In the past, I've been thinking mostly of "terrain bosses" - that is, monsters so large that you have to navigate them in order to fight them. But I think that it may be worth considering monsters of the middle size: monsters that you can climb on, or whack at their knees if you don't want to climb on them, or shoot fireballs at their various limbs for different effects. The maximum size here is such that an archer firing at them would not automatically target a limb. IE, a giant dragon where the archer would default to the huge head coming at him is too big.

Having large but not titanic enemies allows you to suddenly have a lot more combat options than you used to have. Do you target their legs? Arms? Head? Do you smash their toes while staying away from their swinging hands and biting head? Do you leap on and climb so that you can get a good hit against a specific part of them? Do you use the terrain to hide, trip them up, or get above them?

You can even bring in team tactics: can you lure it to chase you into a doorway, getting it stuck and letting your teammates jump off the roof of the building onto its back?

The wafting scent of this kind of combat filled my nose as I played Dragon's Dogma, but it never materialized. There were a few reasons for this.

1) The NPC allies are too incompetent. They don't understand to hit the arms of the ogre to get it to drop whoever it is clutching, so if it is clutching you, you might as well reset the game. They don't understand to aim for the head. And they don't move very aggressively even at the best of times, and prefer to try to go toe-to-toe with the enemy. Why?

2) The enemy hit locations are not reactive enough. The cyclops was great: it had tusks you could knock off, an eye you could blind, a club you could knock away. But that was the apex. Every other enemy made little distinction between you attacking it in the leg, the arm, the head. Now, they did have reactive behavior - getting scared and cowering, or falling over stunned, or going berserk. Those were great, but I really wanted to smash its kneecap to slow it down

2a) Note: glowy weak spots are not reactive enough. There is a rock golem where you have to hit glowy weak spots. It's interesting for about five seconds. It would have been fine if those weak spots took less than five seconds to destroy each, but instead they took thirty seconds of continuous onslaught. Boring.

2b) Having limbs you can hack off, like the manticore's heads, is fun... but the manticore spams the attacks with all those heads, making it impossible to tell what's going on. Generally, I recommend that dealing damage to critical locations should reduce the capabilities of the monster and change its attack pattern, but the monster shouldn't have so many chaotic attacks that you can't tell what's going on.

3) The enemies are too tough. Jumping onto its head and stabbing it in the face accomplishes almost nothing. You have to just cling there and hit it over and over. Literally for a full minute, just cling there and smash smash smash smash smash smash... it'd be much cooler if there were more of them, but they were frailer, so that smashing them in the head really had an effect. Stunning them, bringing them down - then the question is whether you can finish them off before the two others get to you...

4) The enemies are too brutal. Even as a melee class, there are times when a single hit will do more than 70% of my max health in damage. There is no "retry this battle": if you lose, you go back to the last save point, which is a huge burden. If the enemies were less lethal or the game punished losing less, it'd be a lot more fun.

5) The climbing mechanic is too clumsy. Dozens of times it would malfunction. For example, grabbing a thigh and pressing "up" on the joystick. I want to climb. But instead, the hero simply crawls in circles around the knee until the giant kills me. Fantastic. Or another time, when my hero inexplicably decided to grab onto the end of the fist, just to make sure any given hit would kill me instantly.

6) Terrain was not used purposefully. I didn't even realize how cool terrain could make these fights until I fought an ogre up and down an ogre-height broken cliff. The poor programming of the ogre made this interesting - it couldn't simply punch me against the cliff I was standing amongst, because it only understood a few moves, and they were all either "attack the ground" or "grab something that's climbing me". The idea of "that guy is standing at shoulder height" confused the hell out of it.

A smarter ogre would have made that battle significantly more brutal, but the same basic idea could have made these combats more interesting in general. Lure the ogre into a ravine. Dance around a crowded forest with him having to uproot trees to get to you. Jump from crumbling fortifications onto its back...


So... could you make a game out of this?

The game would need:

0) The ability to move, and to attack different parts of different monsters.

1) Competent NPCs or multiplayer.

2) Monsters that change their attack pattern and stats based on what parts of them you destroy/injure.

3) Parts of them you can destroy/injure without having to spend all day whacking at it. Instead, allow for multiple monsters.

4) Don't punish getting hit too severely, because it makes the game un-fun.

5) A graceful climbing mechanic that supports easily moving from limb to limb, part to part. Also supports leaping on from other places. The monster may, of course, decide to snatch you up...

6) Terrain which can be used to screw with the relative height of the monster and its mobility, and perhaps even used to get it stuck, cause damage, etc.


Monday, June 04, 2012


I've been thinking about the way that most modern RPGs scale the enemies to your level.

This is more obvious in some games, like the Elder Scrolls games, where you'll just randomly start running into bandits wearing diamond plate mail and demons summoned from... well, it's unclear how they got here, but they want to kill you.

It's super prevalent. For example, in Dragon's Dogma, the enemies remain the same category, but scale by level. So if there were wolves in this clearing at level 2, there are a crap-ton of dire wolves in the clearing at level 22.

I would like to argue that scaling up the enemies by the player's level is a bad idea.

The argument in favor of scaling is that it keeps the pressure on just the right amount all the time. The designers can say "this area should be kinda hard", and it'll always be kinda hard, assuming your scaling method is decent. Without it, a player may overlevel and get bored of combat.

Old RPGs used a gating system, where any given area was a specific level. You could (and sometimes had to) level there, but the games had a diminishing returns situation: after you got a few levels ahead, the enemies simply didn't give you enough XP to get to any higher levels. And there's no real way to get the next grade of equipment, either: it's just not in the shops.

In this way, the old games could cap your grind ability, allowing people who want to grind to get some advantage while letting people who don't want to grind squeak through and catch up in the next area, where the old tier is achieved and surpassed in a flash.

But open-world games are not able to do that. If you can walk anywhere, then having areas which are higher- or lower-level effectively act as really irritating blockades: "Oh, you want to walk to the capital? Well, there's a dragon in the way! Nya-ha-ha, it's just a random encounter, too!"

So, instead, everywhere you go you will be faced by enemies that are just the right amount of challenge! GEEEEENIUS!

...Except, no, it isn't. It makes playing the game pointless. Why bother to play any of the sidequests when grinding and leveling up just makes the game harder?

For example, in Dragon's Dogma (which I'm specifically talking about because I recently played it), your character and your job class gain levels separately. I decided I wanted to try out a new job class - the mystic knight. Fine, right?

Except that the enemies are scaled by character level. So even though I'm this total wussy level 1 magic knight, I'm facing down multiple ogres at one time and other bullshit. The other party members are useless no matter what their level, of course.

This is the core of the issue. Not only is leveling up pointless because the enemies keep up, but if your character is not built to be minmaxed at combat, the combat rapidly becomes harder and harder and harder.

The other classic example is trying to play an Elder Scroll game as an herbalist or some other largely noncombat class: you're gloating over the way your speech skill is now 150, but then a dragon comes out of the sky and eats you. Since your "kill everything" skill is still at 3, where it started, you die.

It's crap. It's bad. It doesn't let me play the game the way I want to play it, instead it forces me to play the game like you want me to play it. And if I try to explore other kinds of play, it teases me by initially letting me, but then making the game too hard to continue.

That's crappy. Bad.

But... making zones where the levels are ten times your level is just as bad. It's not really open world at all. It's closed-world with really irritating walls.

My recommendation is to make the world open, but make it clear what combat level the zones are. Then make it possible to skip those zones for a small fee.

For example, you could walk to the capital, but there's level 100 monsters on that path! Well, people go to the capital all the time, so you could also fork over a few gold coins and take the train.

Of course, if there's a mission out in the wilderness outside the capital, you won't be able to accomplish it by taking the train... maybe you'll have to come back later, when your level is high enough.

This has to be combined with A) a speedbump rule, because fighting radically uneven battles is boring, and B) nonstatic level progression. That is, a level 100 baker can't get 0 XP for fighting a desperate battle against a level 10 goblin. Otherwise, a character not optimized for combat will never be able to achieve a level high enough to take on the enemies in a given area.