Thursday, January 31, 2013

Third Person Mouse Camera

Recently I've learned to create human(oid) models and animations well enough that I can actually start creating characters rather than dummy bodies. One of the things I've become addicted to is the idea that the character animations need to "understand" the world.

For example, I built a cartoon rig that can walk or run around, but when they get near something, they automatically interact with it in some kind of passive way. Running their hand along a wall that's too close, or putting their back to it if in stealth mode. Automatically looking at and reaching out for grabbable items. Ducking under or sidling through gaps that are technically large enough for them to just go through without such actions.

I really feel that these activities ground the player in the world. If the avatar is interacting with the world, then the player is getting that feedback and, in turn, the player is more "in" the world.

But I've run into an interesting problem: the third person mouse cam.

Not the programming of it. The "shoulder cam" is actually pretty easy to create in Unity. The problem is the gameplay it supports. Or doesn't support.

Shoulder-cam is not really very good for precise shooters. Instead, you get "action" shooters like Tomb Raider, where you just fire on the enemy in general, not with the level of precision you might get with first person cameras. This can be fun, especially if you configure the mouselook such that up/down doesn't change the vertical position you're aiming at, but instead whether the camera is squatting on your shoulder or flying far above. These settings can produce intense shooter action games.

You can do platforming, Prince-of-Persia style. Tomb Raider style, again. I think this mode of gameplay is overly linear, though: I don't really feel like creating such a linear design, and would prefer something more open-world for my immersion-centric approach.

The core difficulty, to me, is that the shoulder cam specializes at the middle distance. It's really good at targeting things that are three or four meters away. Action shooting happens at around that range, as does most platforming. At longer range you're better off with a first-person camera or, alternately, a proper birds-eye cam. Both allow you to finely target distant things. If you're up closer than three meters, your orientation and position begin to get muddy in respect to where the targets are, and things feel mushy and confusing. So you'd normally lock down the axes and go with something like a side view, where your movements can't accidentally make you face in awkward directions.

There are ways around these problems, but fundamentally I'm interested in trying to use the middle distance well. Let's assume that I'm vetoing action shooters and plaformers - perhaps those elements will exist in some amount, but they won't be the core gameplay. What else is left?

Map configuration/construction is one possibility. While Minecraft and its clones tend to be first person, all the action actually happens at the middle range. So a third-person world configuration/construction game should be possible, and even provide for more entertaining platforming.

Another thing that happens at the middle distance is socializing, so it's possible you might be able to create a game around the idea of interacting with people. Not sure how to make that a proper game, although I'll think about it.

Monster/robot training is another one you could do, where you feed/train/program your pets at middle distances, then watch them go off and fight or whatever.

Anything else?

Wednesday, January 30, 2013

Snap-Together Colonies

I've been playing some shooter-adventure-survival-horrory games recently. Not my favorite genre, not since it peaked with System Shock 2, but it still contains a lot of goodies.

One of the things I really like about them is that you feel like you're part of the world. Most of the plot points involve getting somewhere, fixing something, hacking something - the monsters are literally just in the way. Even when it does come time to exterminate the monsters, you do it by building or hacking some part of the world. Sure, there's generally also a climactic battle, but you win by using the world against them.

There are literally as many examples as there are shooter-adventure-survival-horrory games. Venting the queen out an airlock, lighting sacred candles, putting poison in the ventilation, setting the self-destruct, using vaccine syringes after stunning the boss, jury-rigging the solar array to fry it... it's so central to the concept, I think it's why most shooter-adventure-survival-horrory games... let's just call them SASH games... it's why most SASH games are science fiction, and most of the rest are in a modern world where you have magic rituals. These settings give you a world where there's lots of stuff you can do to use the world.

The settings kind of remind me of lego - the pieces snap together like modules. The residential section has a solar tower "snapped" onto the top of it, so when you need to restore power you can use that rather than trying to reconnect or restart the main fusion reactor. Or maybe the residential section has a commercial section snapped on, in which case there's no natural light and no easy backup power source... but you can evacuate into the commercial district and hope to survive. Evacuating into a solar tower is like evacuating to the attic of a burning house.

I began to wonder: can you build a game where you can construct facilities, but then you can also run through the facilities as they are invaded by aliens?

In this situation, each "module" you can snap into your little colony has a purpose in terms of the basic function of your colony, yes. But it also has a purpose in terms of how it helps or hinders the player when they have to play through it live. The alien attack never plays out the same way twice, so it's not guaranteed that things will unfold as you expect. Maybe this time, the aliens actually came in through the solar tower. Or cultists took over the residential zone, and you have to fight through them in order to configure the tower and get power flowing to other sectors through the residential zone.

A basic combination of statistical use, resource chains, and downfall variations can be used to make all this happen.

The solar tower is the start of a power resource chain, supplying power to whatever module is beneath it and, from there, to other modules connected to that module. However, although cheap to build, it is an ongoing expense, so it is generally mothballed once the proper reactors come on-line. Downfall variations are slim: as a resource source, its primary downfall purpose is to be turned on again. However, it can also be used to shine burning bright light onto other surface modules, or can be collapsed to deal catastrophic damage to the module below it and in any one direction.

On the other hand, a bio lab has massive downfall variants - everything from evil monster experiments run wild to 95% completed vaccines to cultists who want to destroy it. Some downfall variants require you to establish a supply chain (of power, life support, or transport), while other downfall variants challenge you to fight off enemies.


Taking into account how things fall apart when aliens invade could make an otherwise pretty samey "build a station" game rather interesting.

Tuesday, January 22, 2013

DLC and Saying Stuff

You know, I think I figured out why I hate DLC so much. Even free DLC.

It's because I think of game rules as having a lifetime. A game is simply the thing you do while exploring the rules. Eventually, the mechanics run out of things to say, and the game is over.

These days, most games don't have much to say. But the content goes on and on and on and on. I can't count the number of games where I say "well, I think I've done all there is to do in this game" and I'm only halfway through.

The game still has a plot, still has half a map left to explore... but I already know how it will play out. It has nothing meaningful left to discover. And, to be frank, there aren't any games with storylines and characters impressive enough to make me want to finish anyway.

So, when I see a game that has DLC, my first thought is "not only will the DLC fail to make the game any more interesting, it also strongly implies that the developers don't know what they are trying to say."


That was the important thing I wanted to say about DLC. Here's some unimportant stuff.

There are games where the primary draw of the game is socializing with your friends. In those games, even simple mechanics are supported by chatting with your friends. Is DLC okay in this kind of game?

Well, I don't play that sort of game, so I don't know the answer. If I want to play a game with my friends, I pick one that we haven't played to death. I pick one that still has something to say. Sometimes I pick one that I think will talk most loudly to my friends, sometimes I aim for one that I still want to see more of. But the choice is still about the game.

The idea of a giant virtual world is a fun one, but there are some fundamental problems. I think that today's virtual worlds are a big mess of different aims, and they have completely different things to say to different kinds of people. The cash shop in these games has goods aimed at each kind of player.

But in the end, if I'm going to be part of a virtual world, it'll be because I want to make things. Not simply live there in some vacuous virtual facsimile of pure consumerism. DLC and the ability for players to make things are diametrically opposed. So, for me, DLC in a virtual world says something a bit different.

In another game, it says "The developers don't know what they are trying to say". In a virtual world, it says "The developers don't care what you are trying to say".


The reason I bring this up is because I am creating diving game prototypes. The atmosphere of diving down into the depths is really compelling. But the things that experience has to say are slow things.

The experience of being in the dark, isolated, and probably a bit lost is not one which strikes instantly. It's something that grows in the back of your mind as you play. Moreover, although it is a slow thing, it is something that doesn't have infinite depth.

What I'm saying is that the diving game speaks about that kind of thing, but the sweet spot is probably 30 minutes in. Before that, it's mute. After that, you've heard it and are reaching out your hands in the darkness for something else.

So... the diving game has to have something faster to say, and something to say beyond the darkness.

And that's not so easy to plan out.

Sunday, January 20, 2013

Roguelike Rail Shooter (Paper Edition)

I built a very simple paper prototype to explore the concept of a rails shooter that gets generated on the fly.

The rules are simple: every node in the game is built out of a topology, gross motion, fine motion, transition, and one or more enemy presentations. I didn't even make unique enemies, by the way.

These can have slightly unique results/conditions.

So, for example, the first node in a random bit might be

TOPOLOGY: Large room full of shelves. High chance of enemies hiding in shelves.
GROSS MOTION: Run through and then turn around to face the enemies
FINE MOTION: Slow-motion opportunity
TRANSITION: Wall explodes (cannot leave room)
Wave 1: Dread enemies (corpses that get up)
Wave 2: Popup challenge (enemies that just pop up at close range)

As a result, this is what the player would experience with this node:

"Ah, it's a room full of shelves and corpses. We're sprinting through. Turning around and... oh, a slow-motion power up and the corpses are waking up! Ah, killed the zombies and AGH THERE'S ONE IN MY FACE! Okay, let's go through the doo- AGH THE WALL EXPLODED!"

Of course, this sort of very basic list system can't really reflect the difficulty of implementing this stuff actually happening in a 3D world in real time, but it does allow you to see how this stuff could be made to happen.

In this case, I rolled up about ten nodes, and after the wall explodes things go like this:

Still inside the shelving room. After the wall explodes, you turn and see the boss dropping a miniboss through the rubble. Fight!

You run to a maze of tight halls and circle through them until you kill the eight or so zombies within. Then you crawl into an air vent.

In the air vent, there's nothing. But peering out of the air vent is a cluster of enemies around an exploding barrel.

Jumping out of the air vent onto a straight stairwell, you first look down the staircase, then run down it. You're staring at the exit, but then you spin and look up the staircase - zombies are stumbling down it! Kill them! (Please note that "camera pointed uselessly at wall while enemies creep up from behind" is actually a fine motion event, not me interpreting an event to be like this.)

Through the door is the same shelving room you initially fled. You decide to stop and clear the room of the steady stream of zombies coming in from the broken wall. There is a quicktime event (shooting moving targets to avoid terrain damage is a common staple) where you dive through the broken wall

The medium-size windowed room through the broken wall has no enemies in it, but you can see enemies out of the window - and snipe at them, if you like. You pass through the door and seal it.

There's a large room with a small ramp on the other side. There's no enemies but, halfway through the room, there's a direction sting. Something like "Through there! We're getting close to the loading docks!" Then there's a sprint through a pack of slow, damaged enemies to crawl under a vertical door.

You end up in a hall packed with doors, and decide to clear it. A bystander is caught up, and you can rescue them by shooting a steady stream of incoming enemies and a miniboss. Unfortunately, the first door you try is locked.

You move around this large hall repeatedly, hopping around a bit, while shooting the various scattered zombies coming in through the doors. Then the boss arrives and carves away a big chunk of the level before winging off.

Still in the hall of doors, you stand your ground. There is a gore note as the boss drops a pack of corpses as it leaves. After falling wetly to the ground, the corpses stand up and attack, simultaneous with a few zombies coming in from another door. After killing them, you go through the damaged area...

This is the sort of progression you can easily get when you think of more than just map layout and enemies. Even with just generic enemies, you can see the framework for an interesting sequence.

However, the content required to build this is way out of my league, so I'll focus on maybe doing something similar on a much smaller scale.

Saturday, January 19, 2013

Roguelike Rail Shooter

I was thinking about whether you could make a roguelike rails shooter.

Obviously, you could create just a chain of rooms with pop-up challenges. But a rails shooter is a lot more than that. So let's consider it.

Let's split the content into four fundamental types.

1) Topology

2) Timing and motion

3) Enemies

4) Enemy presentation

We can separate the level out into premade topolgogical segments - random room-like objects. The rooms are not static - their elements can be permuted in various ways. A door could be locked, a roof collapsed, etc. This is important because the other three types need to be able to affect the topology to some extent.

The rooms get stitched together not just at the doors, but at the windows and overhangs and broken walls and collapsed roofs.

Since it's a rails shooter, how you move through the level and aim the camera is a major factor in the gameplay. This means you have to program in a lot of things like looking around, backing up, diving forward, and so on. These can be largely independent of the topology, except in that the topology has to allow it. Topology pieces come with several established possible paths, and you basically just stitch them together (including using multiple paths through the same room when possible). But you combine that path with the pacing and camera work to support your enemy presentation needs.

Enemy presentation is the key element of all rails shooters. The enemies aren't just arbitrary pop-up targets like a gallery shooter. They have an existence in the level. They associate with the topology: range, height, and obscuring stuff. They associate with the timing and motion, to: surprises, peeking, cover, parallax, spinning, glimpses. They combine, too: sniper opportunities, firing through windows, down from a walkway, up onto the roof...

Even further, enemy presentation also involves how many enemies, whether they are widely separated on the screen or clumped, and so on.

So, is it possible to make a game out of this?

I dunno. I think maybe I'll try a paper prototype.

Games Tied Down

I've been playing Farcry 3, and it reminds me of a really common flaw that AAA games have these days.

The core gameplay mechanic is often brilliantly polished and very interesting, but then the devs attach all these weights to it that just drag it down.

Farcry 3, for example, has three layers of metagame popup and story missions that are actually more boring than the random encounters. Seriously! A game where the random encounters are more interesting than the scripted ones!

This is hardly specific to Farcry. Nearly every AAA game these days features fantastic core gameplay that gets gummed up by something. Final Fantasy 498, Sleeping Dogs, Dragon Age - all had fun mechanics and good world design, but bogged down by various factors. (If you loved Sleeping Dogs, I bet you were logged in.)

At first I thought this bogging down was on purpose, to sell DLC. But so many of these games are bogged down by their characters, plot, or pacing - nothing to do with DLC.

No, I think the core problem is that game design teams are very big these days.

So the "A" team designs the core gameplay content, and then the "B" team does the 8 million pieces of scripted content the story needs. Or perhaps it's the "A" team, but under heavy time pressure.

Whatever it is, the core mechanic is typically one or two gameplay loops, and that can be iterated and tested on tissue-paper testers and just in general polished until it shines. But the scripted content is not so easy - each piece is an individual production, and while it can be tested by Q&A, it's hard to do proper tissue-paper tests on it.

So the things that can't be properly tested end up being significantly worse than the core gameplay loop. In the case of Farcry 3, I actually think the game is made worse by the mission plots. Even the cutscenes and dialog seems bad in comparison to the basic fun of exploring a world and getting jumped by a tiger.

I don't really have an answer for this, but I do have some possible suggestions.

1) Cinematic testing. Bring in tissue-paper testers to just watch the cinematics in the game, or at least the animated storyboards. Maybe read out a plot synopsis. Ask the viewers their thoughts on the various characters, what the things they liked and disliked were. Once their complaints are all about the stuff that they don't see, you're golden.

2) Understand what your core gameplay is, and that scripted missions need to embrace that. Your scripted missions shouldn't be Prince of Persia when the rest of your game is wandering around an open wilderness. Your scripted missions shouldn't simply be the normal incidental missions with stringent fail conditions.

Instead, make your scripted missions the same kind of gameplay, but in an amazing situation. If your gameplay is about wandering around the wilderness, your scripted missions should be about wandering around somewhere amazing.

3) Tissue-test your scripted missions, even though it's a pain in the ass.

Friday, January 18, 2013


I've been pondering more on the idea of using physical rules in a universe to drive not just the gameplay, but also the character design, plot, character arcs, level design...

This requires a good balance. The physics twist has to be interesting enough to drive all that, but also simple to grasp.

Two bad examples of what I'm talking about are Fez and BoomBlox. Both are good games, but they aren't good examples of what I'm talking about.

BoomBlox has a simple-to-grasp physics assumption, but the assumption is not deep enough to drive the story and character design. It's basically just a physics game where you blow stuff up, and build stuff to blow up.

Fez has a physics assumption which is deep enough to drive the plot and characters and setting - and, to a large extent, it does just that. However, it is difficult to grasp and brain-hurty. Even the most minor application of the physical assumptions is a puzzle.

Good examples include the recent Mario games. Paper Mario makes the simple assumption that everyone's made out of paper. This basic assumption allows for a huge variety of results. Characters are flat, so they stack. They can fold. They can be blown away, get wet, get glued down - everything you've seen paper do. You can also extrapolate to different kinds of paper, such as cardboard. Combined with the rich supply of content from existing Mario games (goombas and princesses and so on), the world feels super rich and easy to grasp.

Mario Galaxy is another example of how to do it right. The physics assumption is that the universe is made out of little planets, each with its own gravity. Because we're familiar with this kind of "Little Prince" assumption, we can really understand the results pretty easily, and it gives the game a lot of excuses for fun islands of uniqueness.

Mario Sunshine is anotherother example of how to do it right. The physics assumption is that water and paint are the primary concerns, so most everything that happens revolves around the nature of water and paint.

Neither are as sharp as Paper Mario, which pushes further, easier.

A wasted example is Epic Mickey, which was theoretically based on the nature of toons. In practice, however, it was mostly just an ordinary action platformer.

Kingdom Hearts is a pseudo-example, where the nature of the world is "all those movies and games you've played". This is a very rich vein to mine, which is why the games are so popular, but it's a bit too linked with pop culture for my interests (or budget). I would prefer to come up with a different physics assumption besides "self insert crossover fanfic".

I've been thinking about how to create these.

It's tempting to just flat-out steal one. As long as you don't steal the IP associated with it, that's legal and not immoral. I could make a game about paper cut-outs. I could make a game about paint, or toons. (Although I don't think I could call them toons, because that word is trademarked if I remember correctly.)

I could also take an assumption that had promise but didn't take it very far. For example, Portal. Portal's physics puzzles are excellent, but if you start over and assume that portals are common, you can start to build a world where the whole world - all the characters and settings and plots - involve the ability to pass through portals. As an easy example, snakelike creatures would be common, with the ability to pass their heads through portals while leaving the bulk of their body in their lair.

But instead of simply copying and enhancing, I'd like to think about the nature of these kinds of worlds.

Perhaps the core measure of this technique is how transparently the assumption becomes part of daily life in the game. The assumption has to be something that all the locals just take for granted.

But, on the other hand, it also has to be something that human players can easily comprehend. The idea isn't to make reality more complex, just to make it different.

A good example of that is Paper Mario - everything's made of paper. It's just taken for granted - toads hide by slipping under doormats or stacking on top of each other. People take care not to get crumpled or wet. It's part of everyday life. When a toad says something like "watch out! Don't let me blow away!" the player instantly understands the concern and what's going on.

On the other hand, if we were to ramp up Portal, we would need to make the ability to travel huge distances instantly part of everyday life. This would probably be too hard for the player to instantly understand, because the human mind is very much about object permanence. It takes a lot of thought for us to comprehend the idea that space folds like a crazy straw. We can't hold it in our heads very well, and there's not usually an intuitive leap. This is made worse because most of the time, the other end of the portal will be in an unclear location: not only is it unintuitive, but we're always working with hidden information.

I think the key is to pick something that players are familiar with, and apply it in a way that they are able to understand even though they've never thought of it before. Sort of like the movies "Toys" and "Cars" - everyone knows what cars and toys are, so the different personalities and plots and characters are really easy to understand.

Of course, easier said than done.

Wednesday, January 16, 2013

Action, Script, and Location

So, I was thinking about the content of games. In this case, we're talking about games where there are characters and a story - not completely abstract games.

In the past, I've mostly created non-action games. Turn-based or otherwise not having any time pressure. That was what the middleware I was using was best at.

When I moved to Unity, I was suddenly in an action-centric piece of middleware. It can do turn-based and so on, but it does action really, really well. And I was stuck picking my nose going, "how do you do action fun?"

Well, in the past two weeks I've create many action prototypes exploring different kinds of action games. I don't like creating "pure" games like crossword puzzles or Tetris, so all my prototypes are based on the idea that characters in the game world are doing something action-related. So space ship battles, beat-em-ups, rails shooters, whatever else.

But I noticed that most of my prototypes were about killing things.

In the past, most of my prototypes were about building things. Sure, sometimes the things you built then killing something, but by and large they were constructive games. When I switched over to action, it became a murderfest.

This got me thinking on the nature of action and violence.

So, after thinking for a while, I split games into two categories. Here I'm really talking about games with characters that go through plot lines, so that subset of games is split into two categories:

1) Games where you go out.
2) Games where you stay in.

The vast majority of nonviolent games are games where you stay in. You build your city, or tend your shop, or whatever. There's not really any travel involved, certainly not as the main mechanic.

The vast majority of violent games are games where you go out. You explore the world, or at least the region, and you face the various challenges that exist in that world. Nonviolent platformers fall into this category, too. They feature moving through the world as a major mechanic.

In terms of action games, most action games where you stay in are really boring to me. They are almost invariably "click on the things that pop up" sort of games, like the 10,000,003 restaurant manager games. I don't like them, and even if I did, it's really hard to create a plot or character arc when every level is basically the same.

(You can do really interesting stay-in construction games, but that's not what we're talking about at the moment.)

Go-out games are problematic as well. They have a level of complexity that is hard to generate. If you want to make the levels themselves fundamentally interesting, you typically have to build each level manually. That's a lot of content.

Most go-out games shore up their level content with other kinds of action challenges, most notably fighting monsters. Monsters are a plug-and-play challenge. It's very fast and easy to add monsters - you can even do it automatically. And it works well enough that even if your levels are crap, the game can still be fun. Roguelikes typically have very dull random levels spiced up by automatic monster additions. It works.

If you can't use combat as an added challenge, you need to either make your levels muuuuuch more involving, or have some other kind of plug-and-play challenge you can drop in.

Dunno what that would be in a nonviolent game. Math puzzles are always fun, I guess.

But fundamentally, the whole idea of an agent which challenges the player is somewhat oppressive. Whether its a monster trying to eat you or a bully who demands you solve math problems, it's the same category of challenge.

I wonder if there's a constructive, plug-and-play action challenge.


Thursday, January 10, 2013

More on Transitions

I wrote an essay (link) about generative content. I'd like to explain it a bit clearer.

Let's use the idea of a fantasy game where there's a hero guild, and you can build arbitrary heroes and take them out on quests. But we want the heroes to feel like they really have a personality and an effect on the world, rather than just being a bundle of combat stats.

We have generative content: the heroes. We want the content to mean more than just varying stats and classes, so we decide to use the idea of transitions. We build a random quest engine and a world... we put together a game. And this is what comes out:

The village chief asks you to go into a cave and kill off a small pack of goblins that's been marauding in the area. So you go into the cave.

Here's where the algorithmic generation of content begins. And here's where transitions begin.

The transition engine starts by asking "what do I want the player to feel, and about what?"

Let's say the player has been getting pretty relaxed recently, doing pretty well. So we want to give him a bit of a scare. The engine decides that the player should feel dread about these goblins. So the next piece of content has to be introduced in a way which causes dread - a very specific emotion.

So the cave opens up and becomes a catwalk above a vast cavern. And the cavern is filled with a raucous goblin army. Hundreds of goblins. They don't see you, but now you know that there is a goblin army. That's dread.

Lets say the player pushes on. Well, let's go ahead and punch things up. We'll give the player a taste of a similar emotion: dismay. The way we introduce the next piece of content should dismay the player. In this case, we'll say that the player stumbles across an alarm trap - and the sound of hundreds of angry goblins fills the air. They take shots at you from below, but the bridge is wide enough that they cannot hurt you. But they are coming for you. Stamp stamp stamp!

The algorithm is not terribly complex - it has a few concepts that a normal map generator doesn't have, such as the idea of specifically introducing rooms you can't yet reach, and introducing resources (a monster army) that aren't actively doing anything. But those aren't complicated concepts - other engines also do that, they just don't do it for the same exact reasons we want to.

Now, this doesn't rob the player of choice. They could have turned around, or rolled a boulder off instead of moving forward, or whatever. And it isn't a specific set of results you are aiming for. It's a heuristic that allows you to, in this case, make the player afraid of goblins.

Afraid of an enemy so dull and boring that they are used as the primary example of standard fantasy mooks.

If it can spice up a completely standard enemy, it can spice up the variations you create, too.

For example, let's say you have a rogue on your team. The rogue detects the alarm trap ahead of time. Yay! You now like your rogue a bit more. It's a transition - the transition that used to be a dismaying alarm trap is now an empowering discovery of that trap.

You can even give the characters some level of personality because you know what the transitions are. The characters can know that the alarm trap would have caused dismay. And they can react: "This is getting too dangerous, we should turn back" says A'li'ce. "I am not scared of goblins," B'-ob' replies in a harsh whisper.

In turn, these statements make the player feel emotionally invested in the characters. These random characters, these bundles of stats - suddenly they have a personality, they're interacting with the world.

This is much easier than trying to script them to do that in response to elements on a random map. If you stumble across an alarm trap on a normally-generated random map, there's a big chance it's out of position, not actually very dangerous. So the characters treating it as a threat is actually de-investing, because they come off as not actually connected to the situation at hand.

In this case, the transition is created very close to the time it comes into play, so the situation is well known to the engine, the emotion it is intended to cause is relatively likely to actually be caused, and the characters won't appear totally daft when they react to it.

Actually, the characters chatting about what might be ahead is, fundamentally, a kind of transition! As much as a door you have to kick down, or a dragon that flies down out of nowhere, or scattered cover that lets you sneak up on a patrol, or a villager crying out for help.

That's what I'm trying to talk about.

The Forever Game

So, like many would-be developers, I've played around a lot with games that build themselves. Let's talk about them a bit.

There are a lot of categories of algorithmic content. Some are very well tested and work great - for example, random map creation is very well done. There are many approaches and most of them work and have been put in games that are quite a bit of fun. Hell, "Roguelike" is a whole genre of game that relies on algorithmic map generation.

There are other kinds of algorithmic generative content. For example, the ever-popular Borderlands features randomized equipment.

So let's talk about how much self-referral the algorithmic content has.

In Borderlands you get random guns. The guns are just guns. Equip 'em, sell 'em, whatever. They don't really create any complexity, just variation.

On the other hand, a random map refers to itself. Each map segment is connected to other map segments. If you create random quests, each element of the plot connects to the next element of the plot. If you create random progressions, each phase connects to the next phase.

Typically, this is the weak part of algorithmic content. Any given chunk of content can be balanced and properly created, but when you start to tie them together you start to get a lot of complexity. In fact, in a game made by hand, the connections are often the most interesting part of the game.

It's not what the next room contains - it's how you're forced to face it, or how you have to work to get there. The transitions between the rooms.

Similarly, it's not that Alice is angry that is interesting in a plot point. It's why she's angry. What caused her to transition into that state.

Thinking of algorithmic content as simply balanced units of gameplay is a mistake. This is the core of the limitations the current algorithmic systems face: they think in terms of units of gameplay, because that's the obvious thing to think of.

If you think of a quest broker brokering random quests, you naturally start with the same assumptions: the broker will need to request a string of level-graded subcomponents. Kill ten rabbits in the D'e'tl'sle-fei'r' fields, then take a letter to H'eja'd'ifl's.

Instead, let's think of transitions. Each transition moves the player. It makes the player feel more emotionally engaged.

For example, the quest might start as "kill ten rabbits in X spot". The transition to that is player controlled, we'll just let that happen. However, we'll then have a transition where after you've killed eight rabbits, you are suddenly attacked by a vicious boar who chases you all around the level. The boar can only be killed by luring him into charging off a cliff.

See, this is a pretty basic pair of challenges: kill N enemies, beat boss. But the transition is what makes it suddenly interesting. There's a sudden jolt of panic when the boar blasts in from nowhere and completely shrugs off your arrows.

Surprise is low-hanging fruit. You can easily improve any algorithmic system by adding in elements of surprise. Collapsing floors, dogs jumping in through windows, whatever.

But there are lots of other kinds of progressions. A lot of them are simple pacing controls - a monster that stalks you from afar for ten minutes before it attacks. A treasure chest you can see but that it'll take some more rooms to get to. Secondary gameplay challenges mixed in with the primary gameplay challenges. Color notes mixed in with the gameplay.

A bit harder to tackle, though, is the player's emotional investment in specific characters, groups, or places. As an easy example: if a monster stalks you for ten minutes before attacking, it's going to seem a lot more personal and weighty than if you just get attacked out of the blue by the same monster.

NPCs, factions, and places don't have to vary wildly in terms of their mechanics or bonuses. You can build up loyalty and investment by having them act. They cause transitions which, in turn, causes the player to feel something about them.

If you are attacking a goblin fort and suddenly the elves stop by to rain arrows down on your targets, you'll probably feel something towards elves. If it was something you wanted, something that benefited you, you'll like them. On the other hand, if their arrows rob you of XP and light the whole place up as a giant fire hazard, you're going to feel that they are really annoying.

This doesn't require a delicate touch. There's no finesse required. Just understanding that it needs to happen, needs to exist as part of your random content is likely enough.


I guess I should probably go make a game like that, then.

Saturday, January 05, 2013

Star Ocean Design Lessons

So, I am once again trying to play Star Ocean: The Last Hope. Just bits at a time. And there are a lot of design lessons to take away from this game.

1) A decent combat system, decent crafting system, and decent visuals can make up for a totally shitty set of character designs and "plot" arcs.

2) If you're going to make a plot that has a lot of quiet side quests and pointless missions, fine. But make it somewhat open-world. Being railroaded into endless, dull worlds that have nothing to do with any part of any plot is no fun.

Even if, by sheer dumb chance, they end up having something to do with the plot in the end, that doesn't make up for feeling like you are really doing nothing important.

3) Don't change the personality of the avatar. The avatar is the means by which the player expresses themself. That's why most of them are so transparent. Even the ones that aren't transparent have a static personality, so the player can get used to this filter being put over all their activities.

But in The Last Hope, the avatar makes a dumb call without me, and I end up paying the price. For hours and hours and hours. He becomes a mopey, dull asshole.

I understand the idea. It's a character arc. In a book or movie, it'd work. But this is so unplayable that it's why I stopped playing last time. This time, I'm doing a little better because I'm skipping all the cutscenes.

In short, the avatar is working against me. If it was funny, then maybe it'd be acceptable. If he'd been a grump the whole game, maybe I could get used to it. But in this case, he was a fine mostly-transparent avatar until he suddenly became a real pain in the ass. For hours and hours and hours and hours hourshourshours...

4) Could we try to be a tiny bit less sexist?

I know that this game is not new - 2009. But it's still in full force. It's not even really limited to Japan: most games are depressing like this.

I could write pages about all the stupid, sexist things in this game. But, to put it bluntly, all the females are vapid, silly little things that like to dress up and obsess over men; all the men are powerful, upright galactic citizens fighting to save their planets and their universe.

Really, it's not even funny. Especially since the women are actually the better fighters. I haven't gotten far enough to play as the last two characters, but so far all the women are higher-tiered warriors than the men. They have higher damage output, better battle control capabilities, and usually have better specials as they jiggle around the battlefield.

So you start to like them. Even the mostly naked catgirl raises a few notches in your esteem when you realize she can kill off four enemies at once with a single combo chain, and her basic attack is a mob control technique combined with two finishers. Seriously? Oh, she's got low defense? I guess I'll just keep not getting hit.

But she's pretty worthless as a character. Every cut scene is cringeworthy as your most powerful warriors go "teehee I'm stupid! You men better take control, because I'm tooooooo stupid to do anything at all."

If you haven't played the game, you might think I'm exaggerating. I'm not. It's really that bad.

Ah, I've started to rant. Suffice it to say, the female characters are terrible.

They've got great stats, though. I can say that, right?

Thursday, January 03, 2013

Scared Avatars

I was thinking about horror games where having the player take actions similar to a scared person might make the game more scary.

For example, in some of Frictional Games' works, monsters would only see you if you looked at them. So you spent a lot of your time hiding in a corner, waiting for the monster to move on.

Also, there have been a few games by Quantic Dream where you manage your sanity by taking boring, day-to-day actions. The games weren't very good games, but it stuck with me, this idea that pouring yourself a glass of water and watching a bit of TV would calm your digital nerves. It's an immersive and powerful pacing tool.

I'm kind of interested in amping this up.

We can add in other things people do when they are scared and nervous, like hum to themselves, run wildly, look around a corner nervously, duck back, and then look again... a whole bunch of little behaviors. And then we can incentivise them gently, often transparently.

For example, maybe you will nervously run your finger on the wall if you walk along it at a specific distance. This creates noise, but also a small amount of san recovery. It's something the player will discover just in the course of playing the game, but they quickly learn to use it as a tool or avoid it as a hazard.

Maybe tidying up gives you san back. A book put back on the shelf? Plus one san. Knock a book off the shelf? Minus one san. Pick up broken glass. Straighten paintings. Scrub sinks... incentivise player behavior, rather than just making the avatar hear voices.

If you look around a corner and then duck back, there won't be a monster there. But if you just come waltzing around the corner, there's a monster in the nook. Don't even tell the players - just let their instincts figure it out. When they get scared and low on health and start acting cautious... that's when the pressure lets up. The moment they get cocky... the hammer comes down. It has nothing to do with how many resources they have, it's all about how confident they act.

Just kind of brainstorming. I think it would be fun to create a game where your avatar's connection to the world is more important than his stats. A game where the players learn to act nervous and scared.

I wonder whether it would make them nervous and scared.