Tuesday, January 31, 2006

Player Generated Content, Part I

Player Generated Content!

A quick essay by Craig Perko


Recently, there has been interest into the particulars of player generated content. Some people love this idea, some people hate it, but very few people actually understand it. I'm going to try to explain a bit about the complexities involved, and talk about benefits and problems. SecondLife, being the premiere game of player-generated content, will be mentioned quite a lot.

The very basic idea of player-generated content is that if your players generate content, you get free content. However, even more powerful than this idea is the fact that players who generate content are having fun on a number of levels that other games are hard-pressed to allow. A player who is creating content in a game is solving a puzzle, being creative, changing the world, competing, and usually performing for a crowd, all at the same time. And all in an almost zero-stress environment. That's pretty powerful, don't you think?

Some people say it's largely an excuse so that the developers don't have to create any content. I don't see why that's a bad thing, but even if it were, the actual content creation system is content, itself. Creating content is its own game. True, it may not appeal to all people, but made friendly enough, you might be surprised.

The biggest argument against player generated content is that most players will produce really crappy content. "Crappy" can be defined as whatever you want - unbalanced, offensive, artistically worthless. There is no doubt most players will tend to create things in one or all of these categories. People seem to equate this with other players having to live through this crappy content.

You're worried about someone building "the Tower of Wang", or just having a dungeon so tedius that it drives a new player off when they stumble across it and try to play it.

So you think, "well, we need moderaters to approve everything before it can get in the game world... we need a standard of quality... blah blah blah." Not only does this cost huge amounts of cash, it also cripples your game. Your game will only contain those things the moderators approve of. Unfortunately, game audiences grow and change. Moderators do too, but generally not in the same directions. Also, the lack of immediate feedback makes the content creation much less responsive and therefore much less fun.

This whole idea fundamentally misunderstands how "player content" works. This whole idea radically underestimates the power of the game's logic to shape the culture of the game.

There are a million ways to control content. However, there's a few fundamental drives and results to think about, and all content creation control systems stem from these same issues:

"Sharing" is a Four-Letter Word

The whole idea of player generated content is to share the content among the player base. If you create a giant monkey that sings Elvis songs, it's worthless unless everyone can comment on your giant monkey.

This means getting people to look at your giant monkey. Obviously, your content management system has to allow for that.

Here's the unfortunate thing: there exist the sort of people who would make a giant penis that sings Elvis songs, instead. Now, in some situations this content is fine, but in most situations, not really. These are griefers.

You see, there's an inverted bell curve here. The people who most want to use the distribution system are the people who have the worst content and people who have the best content. Most of the people inbetween the two are likely to restrict themselves, making content for only themselves and their friends: they have a real idea of what their content is worth.

Of course, it's much easier to create griefer content than good content, so given a "flat" distribution system, someone cruising looking for content is more likely to find Winky the Talking Penis, rather than the cool hovercar races.

This means you don't use a "flat" distribution system. This is a very complex subject, and there's an unlimited number of ways to weight the systems. We'll talk about some of them in later chapters, but a simple example would be a ratings system, where high-rated content is more likely to discovered than low-rated content.

The other solution is to somehow limit submissions by content, but this generally involves some people who stand there with big rubber stamps saying "okay" and "not okay". That's bad for a whole host of reasons, not least of which is the fact that it costs quite a lot of money. We'll talk more about that in a later chapter, as well.

Gotta Spend Money to Earn Money

Another fundamental issue is how content is created. This not only includes how much skill is required to manipulate the creation system, but also what kind of in-game requirements there are to create things.

For example, many old multiplayer text games (call them MUDs or whatever you wish) allowed you to create any content you wished. The commands were usually relatively simple, especially for a programmer, and the descriptions were simple text. Skill-wise, there was no minimum skill for creating basic objects of any kind. It all fell to your creativity, your storytelling skills.

Contrast that with games such as SecondLife, where in order to create an object you have to build a complex three-dimensional figure and then paste pictures of your own creation all over it. At that point, you can start creating the scripts which determine how it runs. Only after you have enough skills to do all that can you begin to actually create with your creativity and your storytelling skills.

There's an obvious barrier, and this accounts for much of SecondLife's slow growth.

There's also in-game restrictions. For example, some MUDs allowed you to create content only if you had a specific rank, or if you spent "points" which could only be earned over time and from others using your already-created content. SecondLife restricts permanent creations to (A) your person or (B) land you own (which costs real-world money).

The in-game restrictions are an attempt to keep griefers and the totally inept from diluting the world with their content. There are good ways to do this and bad ways to do this. I'll talk about both in this series.

The skill barriers are a product of complexifying games. In general, the lower the skill barrier, the more people will create. Of course, you can never please the hardcore crowd, no matter how much control you give them: but the casual gamer just trying your content creation system doesn't want to be faced with a million indistinguishable options.

I will talk about how to create a fun, simple system that doesn't sacrifice hardcore control in later chapters.

Red, the Blood of Angry Men!

Most people are happy the first few times they create something beautiful. But as their skill increases, so do their goals. You get people who want their creations to do something.

This is a touchy subject. Just how much control over the world can you give your players? How can you balance it?

Speaking from experience, if you allow players significant control, they will quickly find the "optimal" system of control. In SecondLife, you can nuke an entire sector, killing everyone in it. You can also release what is essentially a fork bomb, and crash the server. This kind of makes a bunch of people with guns rather obsolete.

How can you allow players an outlet for their constructive urges while balancing their destructive urges? How can you let someone create a castle surrounded by a horde of werewolves without allowing for an army of self-replicating naughty bits? How can you make sure those werewolves don't cause grief to the players? Or crash your server?

We'll discuss that in a later chapter, as well.

Lastly, but certainly not leastly:

The Luxury Clause

It will seem, when you step into a permissive game full of player generated content, that 80% of the content is porn.

Actually, it's generally much lower than that. But porn catches your attention much better than pretty houses and snazzy watches.

The basic idea is that people will create where they perceive the level of simulation to be most fertile. This depends largely on what kind of content creation you allow. SecondLife has a huge number of vehicles, costumes, and architecture, because those are the three things that their content creation system is built to support. The level of simulation is deepest there: an almost unlimited number of options and a relatively easy interface.

SecondLife does have combat, but it's pretty poor. This is because SecondLife itself is not built with combat in mind. When simulating combat, the rules are too simplistic and unreactive. Hence the construction of nukes.

But a game's level of simulation isn't limited solely to what you program into the game. It's also what's been programmed into the player's minds. Social interactions of all kinds are deep simulations, even if they're not programmed into your game, because the players interacting are complex and varied in and of themselves.

Of course, the primary social interaction for most players seems to be sex.

This whole thing is due to "The Luxury Effect". The basic idea is simple: you take care of what you need before you take care of what you want. It's really unusual for people in the middle of a raid in WoW to suddenly dive into a deep social interaction. However, once the raid is over and you're back at base, the situation is reversed. There's no worries about getting killed, no need to be at tip-top tactical attention. So you can relax and let your hair down.

There's a couple of ways to use the luxury effect to your advantage. Eve Online uses it by simply making your characters faces. I'm sure there's plenty of sexual sessions in Eve, but I'm also quite sure it's rarer than in a game where your characters are clearly depicted, usually as sexual beings. What you want is shaped by what you see.

Similarly, Eve's environment is one of continuous unrest. You're never really "secure". It's a tense, worrisome environment. This means that most of the player's time is spent dealing with needs rather than wants.

SecondLife is quite the opposite, with absolutely zero needs and an extremely detailed character creation system which emphasizes your character's sexual nature.

Still, even in systems of zero luxury and zero humanity, if you allow your players to create things, they will often create porn. This is not a solution to that: it will decrease the amount of porn, but will not drop it to zero.

What it will do is get more players involved in the content creation system. If creating content makes it significantly easier to survive, then most players will give it a more persistant try than if it's just a luxury item.

I will discuss, in a later chapter, some options you have for shaping the culture of your game by setting the difficulty, display, and the sexuality of your game world. We'll also talk about how Eve Online and SecondLife actually enhance immersion, while WoW and other such games actually cripple immersion. This is an interesting topic because Eve Online and SecondLife are largely polar opposites, yet they both know this secret.

And, yes, it is related to player generated content. Tangentally. :)

Here you have what I consider to be the biggest issues with player generated content. In the near future, I'll discuss more of the specific issues and solutions.

More on Endings?

Those of you with good memories may remember my rather detailed post about endings.

The Escapist put out another issue yesterday, and one of the many rather long-winded essays they have is "The Contrarian: Masks is the Woods". Because it is quite long, here is the link to the relevant page.

John Tynes is one of the wordiest writers I've bothered reading in quite a while, but at least he is saying something. He has two comments of note, only one of which will be discussed here. It is to be found on that page: he read a drama between two guilds. He got into it. It was exciting. Then he found out that the guilds moved on, merged with other guilds, and vanished.

Now, the article is a bit schizophrenic about whether it's talking about making stories or role-playing, but my intent is to look at it specifically in terms of stories. In my opinion, there really is no such thing as "role playing" as we call it. There is only brutally limited storytelling.

And the storytelling is what is faulty in these on-line games. All the attempts of the more creative players to create a story of their own - these attempts fail. Not out of any problem with the players, not even out of any problem with the story.

They fail because of the game.

The game is intended to always keep going. The game is intended to never end.

That doesn't work, when you're telling a story. As I mentioned, stories have endings.

For example, the whole shebang in the Contrarian this week: it could have been ended. It should have ended - wham, the good guys lose, cry me a river of tears. That's that.

"What next?" is a brutal question to ask. It can be dealt with, but it has to be handled like a high explosive, because if you tell it wrong, you cancel the power of your ending. In this case, the power of the ending was destroyed by the sudden revealing that the whole thing was in vain - that the groups simply dwindled away and vanished. Brutal. What was the point?

In order to correctly handle someone asking "what next?" you need a deep understanding of what viewpoints your story is enabling. Presumably, the story of the triumph of evil of good is a dark and sinister story. The "what next" would have to keep this viewpoint in place, talking of dark and sinister acts. To keep the same kind of viewpoint tweaking as the original story had, after a few chunks of time dedicated to showing the utter darkness, you would then show tiny lights. People fighting against the darkness, for example. Then the second story could begin.

But it's a delicate situation. You can't cut away, you can't act outside the viewpoint, you can't break the "mental map" of the way the world is supposed to be.

Unfortunately, games like MMORPGs don't much care for the story you're trying to tell.

On every level, MMORPGs sabotage your storytelling. You fight against the MMORPG, which doesn't allow for meaningful communication and has extraordinarily strict and arbitrary limits on the kinds of events that can take place and what a character can do. Also, the lack of any meaningful death kills most stories flat.

If you do, by some miracle, get a decent ending, the MMORPG doesn't pause. It shouts, "and then? What then?"

Under the restrictions, it is almost impossible to handle the transition from ending to beginning properly. And these are just hobbyists: even experienced writers screw up with something this delicate. Fighting against the game mechanics and the power of player drift, these stories are universally doomed.

The question is whether a MMORPG could be created to enable storytelling rather than cripple it.

The idea of an MMORPG which enables storytelling well is a powerful one. There is no story more powerful than one which is actually about real people. The idea that what you are reading about really happened (even in a fake world) is a powerful one.


Woo! Ludic vs Narrative!

So, ludic vs narrative has been done into the ground so much that the only way to talk about it is like this.

Sunday, January 29, 2006

Optimal Play

This isn't intended to be harsh, but it may sound that way. You'll have to take it lightly, I've been thinking in terms of alien invaders and mad scientists all day.

I got into a bit of an argument (Greek-style, rather than Bronx-style) with Patrick.

The gist is this: he believes that with perfect information (meaning "complete visibility", in relative vision terms), there is an optimum play pattern. Even if the player can't really figure out what it is.

There's three problems with this. One: it's wrong. Two: it's incomplete. Three: it's inapplicable.

Let me address the second one first. Assuming that, given perfect information, there is an optimum play pattern, you also assume that a player has perfect skill at playing. I haven't shown the other two to be wrong yet, but the second is painfully, obviously wrong. Sure, in chess it's not hard to play a piece exactly where you want it. But how about Duck Hunt?

Yeah, it's awfully clear exactly where you should be shooting. Does that mean you can hit the ducks? That depends on your skill and how fast the ducks are moving, doesn't it?

This is a core factor in many (probably most) games - the skill it takes is not just one of determining the best play path, but also playing it.

So, the idea of "ideal play" is incomplete, because an ideal path cannot be followed in any game where physical skill is required.

Now, the idea of "ideal play" is also flawed because it is wrong. There is not always an ideal play path, even assuming perfect information.

Rather, there might be, but we have not proven that there is, because many games with perfect information appear to have no ideal play path. Go (wei'qi, baduk), for example. This may simply be because the complexity of the game is too high to compute the information we are given, but the complexity doesn't need to be very high before this lack of an ideal path hits. Which means that, functionally, games of perfect information quite often have no discernable ideal path.

In addition to being wrong and incomplete, the idea of "ideal play" is also completely inapplicable:

A slight deviation from ideal play often radically changes the rest of the path. And, as mentioned, there is usually no way to (A) determine and (B) stick to an ideal path. Given you generally want your game to be as vivid and replayable as possible, most designers choose to maximize this drift rather than minimize it, where possible. In the past, this was impossible. These days, it is becoming more and more possible.

For example, what is the "ideal play" in Gran Turismo? There is no "optimal" optimal play. You might be able to compute "personal" optimal plays for any given individual, but even those will spread erratically.

If I'm better at cornering but worse at handling the shifter, my ideal car would be far different from someone who doesn't have a feel for corners but knows his engine intimately. Actually, knowing your engine intimately sounds rather painful, but each to his own.

The point is, even if me and MisterSecondPlayerMan have the exact same skills, we probably won't have the same optimal path after the first hour or two, because we will have bought different cars, made different tweaks, won and lost different races. This is despite all the cars' info being clearly available, either in the game or online: if we make slightly different choices, these choices compound into radically different playstyles.

And that's good! That means replayability! That means vivid, reactive gameplay! That means solid game balance! It's a dream!

Perfect or imperfect information has no effect on the calculation of optimal path. The amount of information available does, as does the difficulty of computing said information, but you can easily have a simple game with perfect information or an imperfect information game with huge tons of complex information. Checkers vs "Romance of the Three Kindoms Nth".

Optimal path calculation is just another half-assed shortcut to game balancing.

Saturday, January 28, 2006

Combat Sucks.

Despite appearances, this isn't really a rant. It's my solution.

Have you ever played a role-playing game? Did you notice that combat sucked?

It doesn't matter what kind of RPG it is. Tabletop, computer, LARP. The combat sucked.

In some cases, they've tricked you into believing the combat doesn't suck by strictly limiting their combat. But it still sucks.

For example, you have a party of characters. They fight a group of monsters. By some magic, everyone strikes and then backs off, allowing others to strike, or heal, or whatever. If you've ever seen a fight by anyone really trying to win, that's not how it happens. One of the fighters will simply pound the snot out of the other. There's barely time for the other fighter to respond, let alone back up and drink a soda. (Evidently, one of the warrior's primary assets in a fantasy RPG is the ability to down potions between swings like a frat boy and beer.)

Not only is the idea fundamentally screwy, it also produces some limits that cripple the dramatic capability of the engine. There's no way to fight a hundred zombies in a turn-based game, even though it would be easy if you were actually the power level you're supposed to be. Assuming you could even fit all the zombies into the combat engine, they would each get to go in turn, and the "lucky twenty" system that is in every RPG known to man would let them annihilate you.

Some games tackle this situation. For example, you'll get games like FF Tactics, which have a strict combat map. However, even then, there's a very tiny unit cap, and it doesn't feel very "dramatic" - getting hit by a warrior? You magically have time to turn, run, drink a healing potion, and turn towards him again inbetween his sword swings. Drama is all about being trapped and beleaguered, and that certainly doesn't count. The only drama in a game like that is when you're trapped in a corner.

Other games use swamping rules, or limit the rules to prevent swamping, or any of a hundred other ideas that end up not working very well.

I've been mulling this over for quite a while, thinking of various solutions.

I'm running a little game now, just to get back into fighting trim, which uses some interesting rules. I decided to make a combat engine unlike anything I had ever seen, one which would solve these problems and produce dramatic combat.

So I decided, right up front, that there was never ever ever ever going to be a one-on-one fight. The rules would technically allow it, but it would simply never happen. I would balance the game for party-against-zounds!-enemies combat and party-vs-huge-ugly-boss combat.

To do this, I treat large numbers of enemies as if they were a single boss creature. Units come in several sizes (up to unit size five, which is somewhere around two tons of enemies), and increasing the unit sizes changes the statistics of the monster. More HP, easier to hit, does more damage, etc.

Moreover, the game uses a hit-and-damage engine which totally rocks. You see, you have a flat damage number. Like, say, six. You roll to hit the enemy. Lets say he has a minimum hit roll of ten.

If you roll from 10 to 19, you hit but do no damage. If you roll from 20-29, you do your damage minus their applicable armor. If you roll 30-39, you do twice that. 40-49, three times that. Etcetera.

That's not exactly nonstandard, but here's the key: when someone gets hit, they get discombobulated. That first zero damage tier is not actually a hit: it's the creatures flinching as bullets go by them, or as you brandish the weapon. For every tier of hit you do, the enemy's minimum hit roll decreases by one. (This rule does change a bit, depending on the enemy.)

This works really well. The party fought a huge dog-monster with a hit difficulty of 20 - which is extremely high. The sniper hit him first, dealing damage and dropping the hit minimum to 17. The next people all grazed him, dealing no damage but dropping the hit minimum to 11. The next round, they managed to get it all the way down to three and someone ended up giving it the final shot with his wee little pistol. During this time, however, it had pounced into the group and was busily chewing on a giant robot. The natural reaction was for several people to jump it with melee weapons - which have subtly different rules, although I doubt the players have figured that out yet.

Of course, not all monsters have the exact same response to getting shot at. Statistically, it depends on their reactions as to what their result is. The dog-boss fight was intended to be easy - just to give the players a taste of things to come. After all, he wasn't backed up by any other monster packs.

Thing is, the combat system feels different from other systems. You're not fighting six feral dogs, or twelve zombies. You're fighting a pack of dogs, or two lurches of zombies. The dynamics feel much different. It also makes grenade use work easily.

Instead of saying, "dog A jumps you, you take 3 damage, dog B jumps you and misses, dog C jumps you, you take 2 damage, blah blah blah", you say, "A pack of dogs jumps you and takes you down. They do 12 damage." Then, instead of aiming at a particular dog, the person on the ground says, "Kill them! With my sword!" and swings wildly at all of them.

The idea that you would strike at just one of the enemies threatening you always seemed pretty stupid to me, and I think this is going to work great. So far, so good. Next session on Monday.

Behold the True Power of Data!

I don't live near any kind of grocery store, and I don't have a car. So I decided to try one of those grocery delivery services. In this case, Peapod. I ordered more than a week's worth of food - more than $50 of groceries. When they delivered, they brought me: a gallon of milk.

Everything else I ordered was out of stock.

Now, admittedly, I order some pretty unusual things, so I can see them being out of stock. However, there's about fifteen things wrong with the whole situation. Not least being their bizarre assumption that I would pay a $10 delivery charge for a $4 gallon of milk.

To their credit, when I called them after turning the delivery away, they said there would be no charge. But the whole situation is ludicrous, and worse, there's no way to prevent such things from happening every time I order.

This is caused by a half-assed data center.

Why doesn't their database know what they have in stock - and what they are getting delivered before the loading time for the vans? If they don't have any, say, Balance Bars, why do they even let you order them? A decent database should return a "zero in stock" when you're out, and the UI should simply not allow you to purchase them, either by leaving them out entirely or by noting their out-of-stockiness.

And if you're getting in more tomorrow, you know how many you're getting. Or you're really, really inept. So, the database should find it simple to simply say, "this is the number we're going to have tomorrow when the vans load (A), and this is the number the vans are going to load (B). So we will have (A - B) in stock and available for purchase."

Furthermore, a good database can flag idiotic situations like pending delivery of 5% of your order. It should be able to tell someone to contact the orderer in situations like this and ask: "Do you want me to just deliver milk?"

This is basic database stuff. I mean, this is the sort of thing that I can do - and have done, and will do. It's not hard stuff.

Don't companies understand that data is their most important asset? They'll blow all their cash on some idiotic legacy database like FoxPro, or on some overblown "modern solution" like the various German "super-databases", and these things won't work right. Because you need to have a data specialist set it up for your company's needs. You think there's a turnkey solution for something like PeaPod? No!

If someone says there is, they are lying.

The funny thing about data specialists is that most of them get kickbacks for choosing to use a particular set of software - this guy is an MS guy, that guy is Oracle, so those are the solutions they will recommend. Those guys generally come expensive, but they're still better than actually walking over to MS or Oracle and asking for a solution. The software companies don't know anything about your needs, and never will. But they'll be happy to pretend they do!

Of course, my suggestion is to find a nice, independent data specialist like me. One who has no problem using MySQL, a stable server of virtually any platform, and whatever front end the job needs, whether it be MS or PHP. We can give you the right solution for your needs, as opposed to bludgeoning our solution into a squiggly shape that kinda-sorta fits. Ah, the power of knowing your options.

Actually, my suggestion is to find me, but I'll understand if you don't. Still: find a data specialist (or company of data specialists). Don't try to buy a "solution" straight from a software company. They will hose you. They don't care.

And once you get your data working for you, you won't have Peapod's trouble. You'll have a level of elegance and automatic customer satisfaction that will make Peapod turn green. Er. Greener.

Wednesday, January 25, 2006

Tricking Players into Having a Good Time

As I may have mentioned, one of my skills is running live games: tabletops, LARPs, etc. Certain kinds of players find my games exceptionally fun, and I take some pride in that. This essay will tell you just a little bit about how I try to run a game.

I'm not perfect, of course. But I definitely do better than the "average game master". In case you're thinking this doesn't have any application to whatever your hobbies are: you're probably right. But maybe it'll be an interesting read. You'll never know until you try it.

The thing about a tabletop game is that everyone really does have more fun when they role-play. By that, I don't mean the rigorous enforcement of speaking in character. I always found that idea a little irritating, and I don't think very many people want to play that way. What I mean is that most of the people I run want to be part of a story. They want their characters to live in the world, even if the players do not. It doesn't matter whether the player speaks in character all the time, so long as their character speaks in character all the time.

This sounds obvious - or perhaps confusing - but it's something that few game masters build their games to allow. I try to do so.

In fact, I very, very rarely run a ready-made game system, because they are not built to pull the characters into the game, they're built to push the players into the game. Let me give an example of what I'm talking about.

Let's take GURPS. If you're unfamiliar with GURPS, just pretend I'm talking D20.

In GURPS, you spend a great deal of time making your character and picking personality traits. These personality traits cost or provide a certain number of points, and are intended to be acted out in session. For example, if you take "honorable", you're expected to act honorably.

This is a "pushing" method. The player takes a trait, often because he wants the extra character points or the stat modification, after which there is no built-in reward or penalty for acting out these traits. It's up to your GM to keep prodding you along with threats and entreaties.

Sure, some players do fine in that situation. But most do not. So I use game systems which have a built-in pull to them which rewards a player for making his character act in-character.

To explain, I'll use Nobilis. Nobilis is a fun and deeply defective diceless game in which everyone plays gods. In addition to a few other cool features, one of the most notable features of Nobilis is the flaws system. When you take a flaw, it gives you no character points. Instead, when it comes up during game play, you get "miracle points", which you can spend to increase your power temporarily.

This means that players who would normally have "powergamed" find themselves suddenly acting in character for maximum payoff. If they take "scared of heights", it only pays off when they act scared of heights.

Powergaming by role-playing. This really does work, but Nobilis is only a primitive example. Largely because it's pretty rare to use up enough miracle points to need to earn more.

My greatest attempt to perfect this was in my "Bastard Jedi" game. I ran in a "Not Star Wars" universe which used a lot of the same basic tenets. There were Jedi, and Sith, and the Force, and light sabers, etc - but I threw away all the particulars.

Now, this game didn't run... uh, smoothly. At all. At the time, I had just finished a Nobilis game in which half my players dropped due to time constraints. So, when I started up the Jedi game, I decided to take all comers, expect half of them to drop, and run two sessions.

I got stuck with between sixteen and twenty players for the entire course of the game. Almost nobody dropped. So the game was a bit hectic. But, still, many people did enjoy it (and many people didn't), and the core tenet of the game system was surprisingly effective.

The core tenet, of course, was "light side vs dark side". Each character had a number of personality axes. I think there were five or seven. They were axes such as "humility vs arrogance" and so forth. Each axis was obviously one side light, one side dark.

The way the system ran was this: you rolled a number of dice depending on your skill in a particular Force power plus the absolute value of what was driving you. So, if you had an arrogance of three, you would get three extra dice for using a Force power out of arrogance, whether it was a healing touch or a lightning bolt. Moreover, I would sometimes increase (or further decrease) the axis if you used it, so someone who used Force lightning out of anger would find their anger stat increasing... making it more useful to act out of anger next time. It also worked the other way: acting out of compassion was likely to make acting out of compassion more powerful.

This led to a natural tendency for even stat-mongers to give their characters a strong personality. And it worked. Flawlessly. Every single character in that game had a distinct personality, because they were very much rewarded for doing so.

I'll run it again someday, I think, with fewer players. Or more GMs. But I just wanted to explain:

The game system needs to support the whole game experience. If you want your players to have deep emotional attachments to the game, you need to reward them for pretending to have deep emotional attachments. Automatically. Continuously. As the rule, rather than the exception.

And this works for all players. I had people who were absolute munchkins, and they added to the game for once, instead of detracting. I had people who liked to role-play finding that they weren't being left behind on the power curve. I had people who were total newbs role-playing better than people in other games who had a five-year history.

Unite your rules and your goals, and nothing can stop you.

Oh No! The End is Near!

The thing about good stories is that they end. A series can be popular even if it has no particular end, but it is almost never considered classic. As an example, things like Pokemon and CSI don't actually end. They merely continue on in infinite permutation.

I think this is largely because stories are mostly about changing your viewpoint. That is to say, a story starts with the audience seeing one thing, and it ends with the audience seeing quite another. A series can have dozens of "ends" - one or two for each episode - but it only has one final resting place for the viewpoint: the last episode crystalizes the shift.

I don't know if I'm being clear, but what I'm trying to say is this: over the course of a story, the audience is drawn into seeing things the story's way. This emotional and intellectual involvement in the characters, the situations, and the world itself lends great power to the story. An ending takes all that momentum and slams it into a final "view" of the world. Generally, ending with the good guys winning.

Even if an ending is poor, the last episode of a series you've watched is very potent. You build up an intellectual and emotional involvement in the world just by watching - even if you're not particularly enjoying it. For things such as movies, with their shorter time spans, you need to have extremely efficient draws to get the same level of involvement you can get from a series. It's just the fact that the series simply has more time to let the audience grow to like the world.

I want to talk about endings. I had to go talk about everything else, first, because I wanted to explain why endings are so critical.

I want to explain that it really isn't hard to have a good ending. In fact, it's not even hard to have a good ending that doesn't end your series. The problem is, your ending is only as good as the view it leaves the audience with, and that view has to be something spectacular in comparison to the normal view.

Some series who don't want to alter their "winning formula" try to achieve this with great fights, large explosions, and walk-ons dying. There is some catharsis to be found in that sort of thing, I suppose, but an ending is called an ending because the story is over. The story has accomplished what it set out to accomplish. It has shown the audience whatever it had to show. No worthwhile story has an unlimited number of episodes in it.

The ending is the most powerful part of the whole story, because it has the weight of the whole story behind it. You've probably noticed this. If you've ever watched four or five hours of a series all in a row, finishing up the last episode, you'll notice that the ending leaves you feeling very different for several hours - even if you were displeased or unsatisfied with it. The power of an ending simply cannot be overstated.

The problem is that much of the power of an ending comes from the brutal finality of the ending. Once a story has ended, it cannot be ressurected or continued without diminishing the impact of the ending. You put in a series that you like, you watch it, it ends with a potent and beautiful set of scenes. You are in awe for all of two minutes while you switch the disk out for the next plot arc. Then that beautiful final viewpoint is crushed under the heel of the new plot arc's viewpoint.

But fully utilizing the power of an ending is, perhaps, even worse. Because that means that your wonderful journey is over. Forever.

After all, how many people have burned to watch more of a series that just ended well? I know people who would happily watch another eighty seasons of Babylon Five or Rah Xephon. The endings were permanent in a way which was almost brutal to the audience.

In my opinion, endings do not need to be permanent. They do, however, need to be temporary. Meaning that an ending viewpoint has to be dwelled on for a fairly significant period of time before being replaced by the next story arc's preliminary viewpoint.

The way you do this is simple: encourage the audience to watch extras after the ending. Whether it's the ending of a plot arc or the final ending of the show, extras serve two purposes:

1) They give your story a feeling of being made by humans. Moreover, by humans the audience now knows. This doubles or more their chances of evangelizing and buying other products you've put out.

2) So long as they aren't radically inappropriate, they give the audience time to soak in the final viewpoint. Extras don't shift the audience's viewpoint like the next story would, because they are not intended to be stories.

For the sake of all that anyone has ever held sacred, put extras on that final disk, whether it's truly the last disk or simply the last disk of a plot arc. It doesn't matter what these extras are. One of your actors jabbering on about how he kept trying to date one of the actresses is fine. Production sketches. Fan art. Anything.

To be honest, I think that out-takes are probably the absolute best extra you can put on a disk, so long as your ending was reasonably upbeat. Any scriptwriter worth anything can write you some out-takes if you don't have anything suitable. Also, light personal humor about your various actors and difficulties is also fantastic.

If you are planning on releasing a story without extras, you are shooting yourself in the foot. The twenty-first century marks the beginning of the Extras Age. The audience wants to hear about all your stupid little troubles, all the ideas that didn't make it in, all the lame jokes your cast makes. And you want them to hear this stuff, because it doubles or triples the likelyhood that they will evangelize and buy whatever else you've put out in addition to radically enhancing the potency of your ending.

Extras, folks. Seriously.

Monday, January 23, 2006

A Matter of Size!

This post is a further exploration of the concepts found here. So you should read that first. Or at least look at the pretty diagram.

I've continued to think about how to represent games. Some people are down on the music analogy, but the more I've thought about it, the more I think something very similar to it is necessary. After all, each "instrument" we use, such as a close combat FPS system or a puzzle system, is more than simply "on" or "off".

Not only does each loop have a "difficulty", it also has a "current state" and a "type". For example, if you get injured during an FPS loop, this makes the next challenge more difficult. This is not just inside the game loop, though. For example, your exploration game loop uses the same hit points as your combat game loop. So, when you get injured exploring, it makes combat more difficult despite the lack of combat-related injuries.

In a diagram, this can be represented fairly easily with touching circles and a label. However, such a diagram doesn't tell you when you'll fight enemies, or how difficult the traps will be. You'll need another diagram for that, and both are missing the critical link between the two. We need a clearer representation.

Fortunately, we can use a "music-like" representation. Here:

I'm not suggesting this is in any way the BEST representation, but I think it does get the job done. As you can see, every play loop has challenges and upgrades. Challenges generally knock you down, upgrades generally make you stronger. These are represented as red or green circles.

The end state is represented by a "cross". This cross is not only within the active play loop, but can also propagate to other play loops. If the cross is missing, the end state is unmoved in other game loops, or is simply on the circle, in the primary game loop.

The position up and down the ladder represents the "power" of the challenge or upgrade. For example, a low-lying green dot might represent a clip of ammo, whereas a low-lying red dot might be a small monster. On the other hand, a boss is up on top, and tends to drop the player's state to quite low thanks to the hammering the player receives.

As you can see, some upgrades and challenges affect other game loops, others do not. This gives you a feeling as to what they are, and lets you fine-tune your player's power in any given game loop. In addition, one challenge or reward can be in multiple game loops, reflecting a complex situation such as dogs and snipers attacking simultaneously, or a boss which requires puzzle solving.

Now, this is awfully detailed. Generally, there is little need to be this detailed. Instead, we should create "chords", which are commonly used multiple-playloop relationships. For example:

Using a simple "outline the block" system, we can quickly identify our "chords" that we use. Then we can write them in simple progression, in plain English:

Discovery, long-range accent, short-range accent, empty beat. Discovery, trap, long-range combat, ammunition. Upgrade, short-range fight, short-range upgrade, double challenge. Secret, discovery, jumping puzzle, heavy double challenge.

You can even shorten them further, by category. For example, you could call it a "short range accent diminished" or simply "S-" for the quick little short-range fights. You could call the short-range only upgrade "S+".

Similarly, the difference between a secret and a discovery is a small one, and both are simply the opposite of a jumping puzzle.

The patterns are quite clear. You just have to write exactly what the pattern is, so that the game developers know what kind of progression to have. You can write several different progressions, in the same way a song has several different chunks of melody that have different feels. The choice of the player will usually determine which melodic pattern is played at this moment - whether they crawl through the ducts or cross the bridge makes for a very different experience.

But you don't need to specify which, or either. You simply specify the melodic progression(s), and let the developer use them as his artistry suggests.

I think there's got to be a way to simplify the final output - I'm just dashing this off in my last moments before bed, though, so I really can't think of anything. The point is: you need to see the effect of an event in one play loop in other play loops. It's critical. What that event actually is doesn't matter: only what effect it has. The actual event will be implemented by the developer. :)

Something Nicer...

To keep the rant from being my front post, here's a little something:

Seth Godin made a throw-away post about fear. Actually, several of his recent posts have been about fear, but this is the only one I'm gonna chime in on.

"Creativity" may not be an enemy of fear, but I'm pretty sure "adaptability" is. And adaptability is, after all, deeply linked to creativity. :)

The Science Section


I walked down to a Barnes & Noble yesterday - it was on the way to other errands. I decided to cruise through and find something fun to read. Now, because I am an exceptionally geeky person, "fun to read" usually implies things like "1000 opening baduk problems" or "chaos theory made easy". So, I went over to their science sections ("chemisty", "physics", "general science").

Dismay! Hate! A full half of their titles in the "general science" section were nonscientific and even antiscientific pieces of propaganda, such as books on ghosts and "living better". Stuck in-between texts on string theory and introductory game theory were books on... ghosts and new age religion.

Okay, now, I'm all for challenging the mainstream. Scientific theories need to be attacked, because that is what makes them stronger or replaces them with a stronger successor. But these things are not attacking: they are undermining. They are replacing true logical thought with something straight out of the fourteenth century!

Science arises from logic, and logic is not "belief". Logic never has been "belief". Logic is "disbelief". If you demand more evidence, if you disbelieve propaganda, you are being logical. Your priest tells you the world was created by an invisible man in seven days? He's full of shit until he can prove it. Your trailer-trash daddy told you blacks were inferior? He's full of shit, too.

And if a book tells you that you and a pack of friends meditating can change the crime rate of New York, guess what? The author's fucking full of shit. His "results" and "reports"? Made up. You know why? Because he's full of shit.

Listening to this sort of thing without any kind of critical thought means that you, too, become full of shit. It's the transitive shit-for-brains theorum. I could write a book about it. And put it in the "general science" section.

The result of believing whatever someone tells you may be benign this time. After all, there's nothing inherently sinister about believing ghost stories. But that same lack of critical thinking leads straight to Xenu.

There needs to be a new section. A section called something other than "general science". A section for titles that are obviously not science.

Nonscientific and even antiscientific books are not bad - so long as they are clearly represented as such. Look, "Live Long Enough to Live Forever" sounds like an interesting book, so does "Infrastructure: A Field Guide to the Industrial Landscape". But why are they in "general science"? It boggles the mind.
Well, it boggles the mind which is being used. Obviously, it doesn't boggle the "minds" of people who don't know science from an invisible man in the sky.

If this post was too bitter, I apologize. I'm just a huge fan of, you know, using your brain. I find it to be the most rewarding thing in existence. So, do me a teeny favor, and the next time someone says something kind of unlikely, think: "I bet he's full of shit."

Thursday, January 19, 2006


That's a bit off. The post I just made popped up beneath my silly little review post, instead of above it.

Go here for my latest post.

"Patterns in Game Design"?

Darius is asking me for a review (or at least comment on) this book. There's a sample of their first chapter up, which is evidently what he wants me to discuss, since I don't have the book itself.

Well, I'm a little short on posts at the moment (other things on my mind, but there's a doozy coming up about game notation sometime soonish), so this will make an okay post to keep the blog churning.

I hate it when people's "sample chapter" is their first chapter.

Nobody I've shopped with for books buys a book based on the first chapter. The first chapter is universally just an "orientation" chapter. And "orientation" is another word for "99% bullshit".

So, thanks to the sample chapter, we know the book is about applying various "patterns" to game design in various ways. After glancing briefly at the first chapter, you quickly flip to somewhere near the middle of the book, to see if there's any actual content to back up the lofty goals.

Flip, flip...

Wait, I don't have the book. All I can see are my fingers. Hmm. I could use a manicure.

I'm not trying to offend Darius - or anyone else. The idea of the book is fine, but I could write the same opening chapter for three book ideas I have. It's hardly indicative of the content. It's like wanting to see a trailer for a movie, but getting the FBI warning and the "rated R" logo. Why they put up the first chapter instead of chapter, say, 12? Who knows! Gross incompetance, I assume.

If anyone has the book, I'm perfectly willing to borrow it and read it. I can give an explicit, in-depth review of the book, because that's the sort of thing I love doing.

However, I do have a few things to say about their concept.

It's not my cup of tea. Their stated idea - which may be misleading, I haven't read the book - is that they offer a wide variety of design "patterns" both found in the wild and raised time right here at home. I don't know what they mean by "patterns", but I get the sneaking suspicion they mean "theories of game design". If that is true, then most of their book will be crap or obvious, because nearly all theories of game design are crap or obvious.

If they spoke more clearly and with less of a "this is my thesis" miasma, I would probably be more likely to assume they were doing something new and interesting, instead of what boils down to, in essence, assimilating other people's work.

The Other Kind of Score

(This is a second take on an earlier post. This one, hopefully, is a bit more useful. It contains a few ideas about how to make a notational system that works. No guarantees, of course.)

The basic idea behind any kind of notation is that you are assuming a standard knowledge, and then applying that knowledge. For example, musical notation assumes you know how to play the instruments. Scientific equations assume you know the math to manipulate them. This is, of course, not entirely true of works which are teaching you to play or manipulate - but even then, they assume you're only trying to learn the algorithm or instrument they are teaching, and already know the stuff they rely on.

The idea that the "notation" of a game needs to somehow be "complete" is ludicrous. In fact, the "artistry" of notations (from novels to sculpture and back around to screenplays) is that they are incomplete. They use what the audience and/or performer will know, and then reach heights unimagined a hundred years ago - because they are standing on a platform built by thousands of people all working in the same direction.

However, unlike music, or screenplays, games are interactive. If they weren't interactive, you could just write them like screenplays. Like most people do now. "The zing jing jumps out at you. (Fight ensues, stats are blah blah.) You continue on to the bibble bobble. (Acquire the Goingiboing Gnarkfark.)"

This isn't suitable. People seem to think games should be linear. Games should not be linear. At least, not all games. The idea that the choices you make choose between utter failure and standard advancement is bad for a lot of reasons, not least that it multiplies the content required to make a game that plays the same length of time as a nonlinear game. (A nonlinear game has much higher replay value.)

Nobody I've seen has figured out a good way to represent the interactivity of games.

This is because people are using the word "interactivity".

Calling games "interactive" is like calling music "sound". Yeah, it's true, but it doesn't mean anything useful. Clipping your toenails is interactive. That doesn't make it a game any more than dropping an egg counts as music.

This is because it is the pattern of interactivity and the pattern of sound that makes games and music games and music. How do you make a pattern of sound? You use an instrument that produces similar sounds but different tones whenever you manipulate it. Then, you can create a wide variety of patterns, all with the same base "noise". By simply writing down the variation in pitch and length, you can record much of the pattern - although the secondary characteristics get lost in the shuffle, because you are writing in a way all musicians, regardless of their instruments, are trained to understand.

How do you make a pattern of interactivity?

Well, one would assume you use something which creates similar interactive experiences when used, but can be manipulated to vary the tone. Like, say, a play loop. For example, the FPS play loop always produces the same basic "tone" of gameplay, but it can be manipulated to produce hard fights or easy fights, fast fights or slow fights, long pauses or difficult maneuvers. Teaching how to "play" these loops - even what loops exist in the first place - might be a bit difficult. The MDA framework provides a nice start for many play loops. It can't represent full games - it can only represent play loops. Hence the need for something which can go further, represent multiple interacting play loops.

There's nothing wrong with having one play loop any more than a song that has only one instrument. However, usually having only one play loop is used for emphasis, like a soloist in a song: patterns only really grow complex when you start mixing multiple play loops.

Like a song which has a guitar, a bass, a keyboard, and a drum, you're going to end up with a much more complex and usually more enjoyable game if you use a variety of game loops, like FPS, stat growth, and linear narrative. The pattern you are representing grows much more diverse and complex.

So, okay, our "instruments" are play loops. And you are expected to know how to "play" them. If you can't "play" an FPS "instrument" to program a game experience, stay away from scores that include FPS instruments - or get another person to join your band. Or, like someone who can't play guitar but can play a keyboard, get yourself something you understand that can fake the instrument passably well.

Now the critical part is representing the interactivity. To some extent, this is actually easier than it seems, because nearly all games use the standard "pattern repeat repeat repeat variance" method of play. The simplest example would be the steady increase in speed throughout Tetris: it's always the same pattern of play, but at an increased speed.

It's true in all games. Each character has a "refrain" they play whenever you interact with them. They are goodhearted scamps or foulhearted skanks. They are brutal warriors or terrified victims.

This extends throughout. The enemies follow the same refrains: the guys with shotguns always play this way, the guys with autocannons play that way. Even - and here's the hard part - even the level layouts and placement of supplies has repeating refrains. In Doom III, you could always tell exactly when you would be attacked, and you always knew exactly when you would get more health and ammo - except in a very few cases.

These patterns, these "refrains", are the game proper. But the exceptions are often what makes or breaks a game. Anyone who plays blues can tell you, the refrain doesn't have to change, because the part it supports - the variations - are endless.

Of course, you need a good refrain. That's why blues musicians tend to stick with the few standard refrains, and rock tends to have a few standard drum beats.

So there would be two "parts" to writing down a game design in this way. Writing the refrains, and then writing their exceptions/collusions. Like music, the refrain determines not only the feel of the game, but also its genre. Which play loop instruments you use in what ways... that's important.

How does this write interactivity?

Well, the refrain is something which progresses without exception. It always goes "buh-wheee-o-whee-wheat", then you can sing about how your woman left you. In the same way, games will always have the same "feel" to them, even if the exact path of the player is up to question. The player will always fight zombie-type monsters here, fire-type monsters here. The player power-ups will always be located after boss fights. The player will always have so many enemy encounters after powering up before enemies power up to match. Maybe fights during the refrain will always go "easy-easy-hard-medium". Or "long range, medium range, long range, short range". Lots of level designers do this by accident, never even realizing they are doing so. See Doom III for an example. :)

Refrains can be anything. They are one or more instruments that play a recognizable "tune" that sets the game's feel.

Now, I can't keep calling them "refrains", or at least I shouldn't, because "refrain" means something very specific in music terminology, and I'm not using it the same way. These patterns in the game - depending on the size of the game, there's often dozens or hundreds of "refrains". A game like Tetris has only one, but a game like, say, Halo, would have a symphony-sized ream of them.

This doesn't seem so different from writing little "pieces of level", does it? You say, "He'll run into fire guys here. They'll be mostly long-range, except a few panicky melee fights. There won't be much healing." That's a refrain. Except the refrain is a much more elegant way of jotting it down, because it refers to the intrinsic but unstated reward and challenge structure. For example:

This could represent the situation I just talked about, with the lines representing long range fighting, short range fighting, fire guys, healing, and ammo. "Played" in your mind (or even on a keyboard, if you like), there is a definite syncopation and rhythm which defines this part of the game.

This represents "first tier" interactivity. It represents play, but not as play actually would be. For example, what happens if the player decides to turn back and explore the level for fifteen minutes between some encounters? The rhythm is suddenly very different. This can actually make a huge difference if you're in a game with something time-related happening. For example, if you heal over time.

In short, it doesn't represent how play loops change, it represents how play loops DON'T change. It represents the part of the game which remains constant.

How can it represent true interactivity? How can it represent a live AI squad roaming the level looking for you? How can it represent the other player stealing health power-ups he doesn't need?

It can't. That's jazz. The designer knows his duty: his duty is to have the play loops form this refrain (and a long list of other refrains, presumably). How he handles the loose ends depends on how he "plays" the design of the game.

Can you codify this into something simple? Not that I can see. I can see putting in "bars" to represent where significant breaks in play can happen, or little slides for where things lead directly from one to the other, but I don't see how you can demonstrate what is functionally 7+ dimensional space on a 2D display.

Instead, you represent the important parts. Like the way you take a photograph instead of sculpting a full duplicate of the scene.

Do I think this is "the" way? No. But it can represent the primary play loop patterns, reward/risk situations, "mood", expectation... it's better than the others I've seen so far at large-grain game depiction. This score has the high score. :)

It's a start.

Tuesday, January 17, 2006

Languages Marching On!

(This is an extremely long post, and I'm afraid I wrote it at about 3 AM. I'll re-write a new post on the same subject while I'm more awake, but you'll have to forgive the way this one bounces around. Long story short: I tried the kind of notation Lost Garden is suggesting more than a year ago, and it doesn't work.)

I'm sure you've already heard about Lost Garden's notational essay, unless I'm your sole source of game design information.

In short, he or they or she or it is speccing out a bare-bones methodology for describing gameplay. There's a couple of severe flaws, aside from the irritating writing style, which includes, in all seriousness, phrases like "Let us praise the advent of music notation as a potent and world changing enabling technology."

Obviously, the author has a lot of good points. Right now, we're in a world of primitive designs, like cavemen scratching out stick figures. There's a couple of new artists coming along. Some are trying sculpture, some are trying paints. Inside those realms, each person is trying to codify the whats and hows, and we're ending up with calligraphy, impressionism, cubism, and so forth.

Which one is "best" will probably never be a question that can be asked, any more than what the "best" style of art is. Some people will go on liking cubism, even though it's the most wretched style of art ever born. It's just the way people are, and if people like it, it has value.

However, there is some scientific merit (or demerit?) to most game design theories that are coming out these days. They can be largely reviewed with logic and a bunch of statements with the words "is like" in them.

For example, the notation he is proposing "is like" musical notation. Long story very short, yeah, musical notation gives about 80% of the song for about 20% of the effort. The rest of the song is gained through the nuances you flesh out after playing through it a few times and understanding when to stick some verve in and shake up the rhythm just a touch. The stuff computers have a bitch of a time doing.

He wants to use notation for representing "mechanical" and "emotional" pieces of the game. Verbs/tokens/rules make up th e mechanical side and actions/rewards/risks make up the emotional side. This is, of course, brutally standard fare. In my opinion, it's also a false dichotomy. The very concept of separating "mechanical" from "emotional" is like saying you're going to separate "food" from "taste".

His first cue is "buzz notes", which are essentially rewards. That's fine. Rewards are great things. Of course, every player gets different amounts of buzz from different things. For example, lots of people found they loved playing dark side in Jedi games. I couldn't stand it - not because it was too evil, but because it was too childish. Therefore, those buzz notes didn't resonate with me. Does he address this? Well, sort of.

He suggests "buzz channels", which are essentially different instruments. For example, light side and dark side would be different instruments. Actually, they'd probably be several different instruments. By simply noting what the player prefers, you can weight given channels heavier. Although he doesn't mention this, it is inherent.

He moves on to actions. "Verbs", as people have taken to calling them. And totally falls apart. Verbs without a state are like people running around saying "love eat hate date". What are they saying? Who knows? Can you track the feedback? No, not really. It might be mildly useful for linear games, but god help the person who tries to track a combinatorial game using this system.

Even saving the state of the game into these things - which I think he is suggesting - is not going to help. You don't need the state of the game. You need the state of the feedback loops. You need to see the patterns that forge the present in order to see how they will forge the future.

Now, I'm not saying this is useless. I'm saying you can't design a game with it. You can tweak a game with it. You can identify weak points in your play. However, at this stage he's assuming that players can't look ahead. "Low points bad!" No, not at all. (He knows this, but he's going from a theory of "burnout", which is like talking about how big the ocean is without mentioning it is made of water.)

Low points are great, so long as the player knows there is a high point coming. Many players delay the rewards until he can slam them all down together. This ends up looking like "unrewarding" play, because there's a long dearth of rewards.

Trickles of rewards are not the answer. He knows this, but for the wrong reason. Let me sum up: you do not and never did want a steady stream of rewards. You use your rewards to build up anticipation of future rewards! That is the whole purpose of early game rewards!

Moreover, the system in no way allows for player-controlled rewards (such as delaying the lesser line vanishings to snag a Tetris). It does not allow for anticipation. It does not allow for player-chosen rewards, such as when a player decides that grenading the jeep is the fun part of the game.

He mentions "expected rewards", but he grossly (in both meanings of the word) simplifies this critical part of game design. In fact, if all a system does is manage expected rewards, it would be 80% of game design. You want that 20% work for 80% result? Expected rewards management.

He does have some nice resonance between suppression, burnout, and risk. But without expectation, it's like talking about a cake and getting only frosting.

He talks about game structure, talking about a "progression of notes" - he's really talking about the cycles of game play loops. Which loops are active. If you want to stretch the metaphor, he's talking about which parts of the orchestra are playing at any given time. But he doesn't say that.

Now, the attempt is good. Any attempt is good. It doesn't deserve to be attacked like this. But it's in my nature.

I just don't think it's much of a breakthrough. I tested reward waveforms and pulse modulation methods several years ago, and I found them wanting for several reasons, not least of which are the ones mentioned above. These systems are simply not adaptable enough to deal with the actions and reactions of a diverse set of gamers.

They are theoretically useful for tuning pacing issues in a linear-ish game. Half Life 2 could have used it, for example. But the method simply does not have the robustness needed.

In order to be robust, there are a couple rules that need to be acknowledged.

First, the execution of a game is, like the playing of a song, always going to change from developer to developer. Right now, we're a bit like jazz. There's nothing wrong with that, and jazz doesn't take much notation. A core beat, a basic melody, then let the musician play. Similarly, we don't need to write the whole game out. We talk about the core elements, and the developer is familiar with how to actually glue the core elements together, with his own special brand of blues and soul.

Those are two separate elements. First you have to learn your instrument, then you can learn to play songs. These are two entirely seperate pieces of the puzzle.

And they are addressed with two different but related languages.

The language of a particular game would be a language of rewards. When you offer them, what they are, how much the player can push them. The language of all games would be one of play loops. Each play loop is an instrument - or a few instruments.

So, like someone who scrawls "guitar" or "soprano" at the beginning of a staff, before writing the pattern of rewards, you specify a kind of loop. Is it a personal empathy loop? An engine breakthrough loop? A statistical improvement loop? Is it played "with verve", perhaps meaning the player can push it around? The developer is assumed to know how to "play" the play loop. Otherwise, he needs to go learn that, first, like you need to learn to play the guitar before you can play "Stairway to Heaven". Well, arguably.

Then you can talk about the rewards. Remember, the strength of a piece of music isn't in the notes it plays. It's in how the notes relate to the other notes. It's not exactly which rewards get pinged: it's when they get pinged, and what other cycles are going on to support them.

We're talking about harmony and confluence of rewards. It's not how a play loop runs. It's how a play loop connects and resonates with the other play loops.

The fundamental problem with this, of course, is that it is still very difficult to talk about interactivity.

Still haven't figured that part out, yet.

Monday, January 16, 2006

The Quick and the Dead

My favorite article of the day? This one. It's about how people make their decisions about whether they like a site within one twentieth of a second of seeing it.

Now, I'm not the sort of person who usually toots my own horn... uh, more than once or twice a day... but, you know what? I've been saying this for at least a year. I've said - from the beginning - that the first moments are the most important in any piece of media. What someone sees in that first instant.

Not minute, although the first minute is more important than the next five.
Not second, although the first second is more important than the next five.

People make their decisions in a flash of neural static. If you can make sense out of the patterns in the static, you can influence people radically.

Hence my commentary about pictures being more "catchy" than words a few days (weeks?) back.

Now, the funny thing about the article is that it is posted on a site which does not obey that precept. The top half of the page is freaking banners. Woo! The left quarter and right quarter are vertical lists, jammed full of text. The article title is tiny - almost unnoticable. And the article itself is jammed between a maze of cramped pictures and ads. Moreover, the color on the sidebars and main banner is red, whereas the article is a washed-out white. Note: Red is a terrible color to make secondary features, because it is so eyecatching. The red isn't even in a "container" shape, it's in a "shattering background" shape. It's a terrible layout!

I just find that, you know, a bit ironic.

Anyhow, the first moment of a game - or a book, or a movie, or a cartoon - isn't the first moment of the game. It's the first moment someone sees something about the game. It's the first moment someone sees an ad, or the front page, or a review.

Obviously, your game content matters. Largely so you can have snazzy ads, front pages, and reviews. The end result drives your sales far more than the actual content of the game, because people assume that other people know what they're talking about... even if they're not actually talking. Getting a good review will sell more of a shitty game than an excellent game with shitty reviews. Classic marketing shit.


Sunday, January 15, 2006

Plot Algorithms Revisited!

One of my abiding interests over the past years of my life has been in developing some kind of automatic plot generation algorithm. Something where you start up the game and it grinds a bit, then comes up with a new, interesting episode of a continuing game.

I've had some very minor successes on this front in the past. However, now I'm going to approach it from a new angle: perspective shifts!

As I stated at the end of this post, a story is all about a shifting perspective. Think about it. Think about your favorite stories. Their whole purpose is to shift your perspective, usually by shifting the perspective of the characters within the story.

Even the mostly-empty episodes of soap operas and children's shows base each episode around one or two perspective shifts. You get to see a new situation, a new character, a new reaction. Just pulling stuff at random off my local media:

Young Einstein: A distorted view over the world of science - the trials and tribulations of scientific thinkers, seen through funhouse mirrors. The perspective shifts from the highs to the lows with a manic intensity.

Hikaru no Go #2: The bizarre perspective of "what happens when a 1000-year-old ghost is forced to both attend grade school and meddle in the world of modern go?". Watch as his perspective slowly shifts to account for the modern world... and watch as your perspective shifts as you see it.

Teen Titans, "Employee of the Month": A highlight on Beast Boy which reveals his character in new depth and detail, giving him the opportunity to face his fears. Also very funny.

Paranioa Agent #10: Offers a unique (if somewhat confusing and highly unconventional) insight into death even before the final reveal.

You can go on pretty much forever, but I would hate to bore you too badly. If you can come up with an example which doesn't contain what the audience considers to be a shifting point of view, I'd be surprised. Some that don't appear to are actually contrasting their current POV with POVs from earlier episodes, so if taken as a whole, they are shifting.

Remember: it's just as often that the audience's perspective is the one that is shifted. New information that the characters never uncover can lead us to think wholly differently about a situation. Heck, just one origami unicorn can totally change our perspective on an entire movie, let alone a few tears in rain.

It's probably possible to encode this, don't you think? Some kind of meta-language to define what kinds of things people can see, and the directions they can see it from?

I'm thinking about a "spotlight". A spotlight shines a particular color of light through a particular chunk of fog to highlight a specific piece of reality with its warped, warbly light.

The light would be a source. A coloration. A filter. For example, "zombies", or "old money" or "getting married". It would determine what the gross shape of the plot would be.

The fog would be the view travelled through. The set of values used to warp the light. For example, "family duty", "cowardice", "war", and other such humanizing values. It would determine, largely, what the activities in the plot arc would be.

The land would, of course, be the thing highlighted through all this. Like "the insignificance of mankind" or "everything is funny if done wrong". This would determine what activities are taken out of the set of activities the fog allows for.

As you can see, just combining the example makes for some strange stories. How about cowardly zombies showing you that everything is funny if done wrong? It's a very different story from cowardly zombies showing you the insignificance of mankind. It's also very different from wealthy nobility caught in a war highlighting their insignificance.

The difficulties would be in getting the computer to translate the components into a coherent whole. What would it present to you if it was going to be a story about cowardly zombies? What would the progression be?

Anyhow, I'm just rambling. If I get any further on the idea, I'm sure I'll write it up here.

Actually, I want to use something similar for Her Majesty's Hackers. I want to use painfully cliched soap opera plots, and then totally wreck them with social ineptness and rampaging mutants.

"Can't you see what I'm trying to say, John?"

"Give me a hint. Is it about the Xerxes project?"

"No! It's about you and me!"

"Oh, so you finally noticed!"

"Noticed what?"

"That I'm actually your clone, of course. Isn't it obvious?"

"But I'm a clone of the director!"

"There's only one question left: is it unscientific to love yourself, my beautiful twin?"

"It will require some testing, I think."

"My lab or yours?"

Managing Time in Comics

Okay, I've got a question. When you look at this comic, do you skip to the punchline?

(Image is 9 Chickweed Lane, which is probably in the top five syndicated comic strips if you ignore the bizarre Sunday strips, like this one. I'm reprinting it here for educational purposes.)

You see, the way I read that image, I read the upper cells from left to right, but then - bam - I zzzzzip past the left cells on the bottom, drawn straight to the final frame's vibrant colors. This happens a lot, and I think it's largely because the perspective of these sorts of things usually urges you to shoot through them. To a lesser extent, maybe it's because there's no natural flow from one row to the next, so your eye just falls down the page, rather than skipping back to the left.

I find this is a common problem - not just in Chickweed Lane. There's something about a cell or page layout which can cause me to pause, to process it as a chronological moment. Other cells, like those above, just slip through.

So, any of you that know the magical art of cell framing can chime in on this. My feeling is as follows, as someone with marginal "armchair" knowledge of the subject:

All the images I look at which cause me to pause are "bumpy". They do one of three things, all of which are "bumpy":

1) They are shaped funny. Crooked, or single-cell-two-vertical-cells-single-cell, or shattered into pieces, or something else which makes the actual cell itself look funny. This can also be as simple as wide-cell-skinny-cell, although you obviously get lesser results with lesser attempts.

2) The content is ragged between it and the next cell. Most people tell you to guide the audience's eye through the page. This is the exact opposite. This is having the "momentum" point smack into something which gets in your way. For example, a white cell leading to a black cell. Or the line of action leading square into a dialogue box.

3) There's a large amount of white space. I used to think this was the most important, but I've seen examples with very little white space that were still clearly pauses.

All three of these are actually, I think, type 2. They are attempts to control the flow of the eye across the page. Unfortunately, Chickweed's attempts always encourage rapid flow, rather than breaking it up.

Interesting, I think.

What do you think? Do you see it the way I see it?

Friday, January 13, 2006

Somebody misses me!

I knew that somebody would miss me when I moved away from Seattle, and I was right. Since literally the day I moved out, Seattle itself has been crying non-stop. 26 days of rain, straight.

Buck up, Seattle. We had good times, but we knew it was never meant to be.

Scaling and Comparison

One of the things which fuels games is comparison. This is most easily explained through strong story games, such as an RPG. You start off fighting slimes, end up fighting dragons. The dragons are inherently more impressive.

By using what a player expects to give a sense of reality to what is otherwise just a bundle of numbers, games make that bundle have some emotional response. This lets them frame different parts of the game as having different emotional "ratings". For example, the goody-goody wood elves feel one way, the ancient necromancer's ruins feel quite another way. Going from fighting boars to fighting demons will make you, the player, feel like you have taken a dramatic step forward.

This is also true of gameplay comparisons, both gross and fine. For example, most games give you more and more choices as you go through the game. This is to let you know you have advanced. Fire 2 is better than Fire 1, but having access to both Fire 2 and Ice 2 is better yet. Higher complexity requires a wider perspective.

On a fine level, comparison is what really drives games. You game for the "cool" moments. The victories, the goals, the interesting bits. In Tetris, there's a rush when you succeed at the four-lines-vanishing trick. In Halo, there's a rush as you see a jeep flipping five times after ramming a pile of explosives.

A lot of games these days are "crossing genre barriers". People liked GTAIII and Halo because, to a large extent, they allowed for several genres all packaged into the same game. Why do people like these cross-genre games?

Comparison spikes.

You spend most of the game walking around. Your viewpoint is low to the ground, you move at a few miles per hour, and you're awfully vulnerable. Then you hop into a vehicle. Suddenly, your speed, durability, and view are all radically increased. It's a huge boost, a whole new perspective.

Similarly, interacting with things that have this huge boost is a heart-pounding experience. The foot soldier trying to deal with an incoming jet? Trying to capture a tank in GTA? These are not just "the next enemy", as in an RPG. They are associated with actual capabilities and difficulties because you have been in their shoes. You have (or will have) been on both sides of that equation.

Of course, I think this is mostly Relative Perception ("Player Vision") related, which means it's easiest and most powerful in physics-based games.

But look back over the popular games, and you'll see they almost always contain one or more of the following four elements (usually two or three):

1) A big license. (Generally associated with high-end graphics.)

2) PvP

3) Huge relative perception shifts.

4) Casual persistant play

Here's GameFaq's most popular pages list. This isn't necessarily a good representation, as people coming to GameFaq are looking for walkthroughs and cheats, but it is indicative.

01 Dragon Quest VIII - PS2

Casual persistant play on a huge license with massive perspective shifts.

02 Kingdom Hearts II - PS2


03 Animal Crossing WW - DS

Casual persistant play on a license. Are there perspective shifts here?

04 World of Warcraft - PC

All four.

05 WWE SD vs Raw 06 - PS2

Casual persistance on a license, probably with PvP.

06 Resident Evil 4 - PS2

Casual persistant play on a huge license with massive perspective shifts.

07 Dead or Alive 4 - X360

Casual persistance, license, PvP.

08 GTA: San Andreas - PS2

All four. I think this one has PvP, doesn't it?

09 Xbox 360 Hardware - X360


10 Mega Man X Collection - PS2

Big license, casual play, some perspective shift.

So, yes, you already knew about PvP and casual persistance being important, and the license is a big "duhhhh". But most successful games have big perspective variations throughout the course of play. This is important! Successful indie games all have perspective shifts. Unsuccessful ones generally don't. Alien Hominid - perspective shift with new weapons and radically unique boss structures. Katamari Damacy - all about perspective shift right from its core.

Perspective shifting is fun. Everyone likes to see things from new angles. Everyone likes to mow down hundreds of something they had a hard time with, or spend a weird and difficult level dealing with something they've never thought twice as they used it before.

Just don't let it distract you from the greatest perspective shift in your game: the story.

That is, after all, what a story really is. Didn't you know?

Wednesday, January 11, 2006

Time is Quality?

A thought came to me about the price of goods.

Presumably, you have at some point purchased a marginally expensive semipermanent product. For example, a car, or a cell phone, or a lamp, or a movie or something. Something you buy and plan on keeping.

Now, what happens when you see the product a few weeks later for half price? You think, "damn! I was cheated!"

The thing is, price reflects value. Always has, always will. If people are willing to pay $50 for a movie DVD, it must be a good movie! If they're only willing to pay $10, maybe not so much. This is usually pretty accurate, and most people think this way to some extent.

But in today's world of planned obsolescence, prices will always drop so long as a comparable product is being newly produced. Buying an old movie is cheaper than buying a new movie, even if the old movie is three times better. The fact that most such products - from cars to music - are price-fixed to a standard "release price" means that even crappy products have a high initial price. In this case, price does not reflect value.

What really reflects value is not price, but the price drop over time.

For example, "Bruce Gigolo" was originally sold for a price roughly the same as other release movies. Now you can easily find it for $5 - and it's been there for a while. It dropped like a rock, because the movie was worthless.

On the other hand, "Valkyrie Profile", a PS1 game, debuted at a comparable price to other games. Now, five years later, the price has doubled.

Sure enough, as expected, Valkyrie Profile is a top-notch game.

This is kind of obvious when you think about it. The market determines the real price of a product, regardless of your original price fixing, regardless of your planned obsolescence.

But it isn't the price that matters. It's not even the price relative to other goods of the same type: lots of companies get away with charging more for shitty products because of their position in the market and their billion-dollar advertising campaigns.

What matters is the price change over time.

(For free products, I would imagine it's churn: the less churn in percentage of total users, the higher quality the product.)

Evident Greed

I'm catching up on Seth Godin's blog, and he has a post which I found interesting. In it he debates himself as to what the "best" strategy is when it comes to (A) advertising others and (B) advertising yourself on a blog. He says he "doesn't know". Of course, he does know, he just hasn't really put it together.

The fact of the matter is that there's no permanent "right way", nor will there ever be. There are niches we fill, and these niches change over time and what passes for space here in the online world. The key is to know what niche you're in and how you can act while in that niche.

Seth isn't advertising his products enough, flatly speaking. He takes the "high road" and only mentions his products a few times ever - in a blog that generally posts twice a day. This is noble, but it totally misunderstands the nature of who he is, where he is, and what he is offering. It is the approach of a man who is used to writing books, and subconsciously expects his blog to be read in one chunk, rather than in tiny pieces over days and weeks as he writes it. It is the subconscious expectation that all posts will be given equal attention, rather than some posts being speedbumps and others being freeways.

What he sees is: "I just talked about that a few (10? 12?) posts ago. That's pretty recent. They'll think I'm greedy or self-absorbed if I post another post on the subject."

That's the standard line of thought: your blog is more important to you than to anyone else. You are hypersensitive and, more importantly, you don't have any sense of the amount of time and material passing through your audience between their visits to your blog.

What an audience sees is a tiny blip. They see a post that takes fifteen seconds to read on a blog which takes five to ten minutes a day. Aside from the fact that people need to be reminded for dozens of reasons, it's also virtually impossible that someone will be offended by a tiny blip of an ad that is utterly on-target. It's like going into an art store and seing a print of the Mona Lisa on the wall with a little price tag. Are you ever going to get upset by it? No! It's a good product in the right place, and it doesn't exactly get in your way.

As a matter of fact, it's more irritating when he works ads into the body of otherwise unrelated posts, like he did in the post linked above. It's out of place and detracts from the flow of the post.

All it comes down to is the difference between greed and eagerness. Greed is what comes out when you have 25 ads and pimp yourself every day. Eagerness is when you talk about products you think are great - whether they be yours or others - enough to make people remember that they are great.

To do less is to underestimate your worth and to serve your audience poorly.

Tuesday, January 10, 2006

Finding Finder

Holy shit! Darius was right! Carla Speed McNeil's putting Finder on the internet! For free, even.

Which is actually a bit strange, since it's better than every pay-to-read webcomic out there combined...

If you haven't read Finder, you can find my favorite story here - online free! Wow! As well as my second favorite story arc here.

Then, go buy the trades. You can buy them here. Honestly, this is a comic made to be held in your hands. If you like what you see, don't read it on-line. Buy the trades. Tell her I sent you. She'll reply, "who?"

If I had to come up with my top three favorite comic series ever, two of the slots I would argue with myself for hours, but the other slot belongs to Finder.

Seriously, read a bit of it, then go buy it.

If you have read the early stuff, you can find the much-awaited (by me, at least) "Five Crazy Women" online here.

The internet is richer than it was before.

A game...

Well, I suffered a harrowing setback on Machine City. So, I took an afternoon off to do a little relaxing.

You can find the result here. I call it "Rettamitna". It's free, of course.


As mentioned, I'm playing catch-up. As such, I just discovered a post called "Beauty is Pixel Deep" on Terra Nova. It's a little old, but it shows something interesting.

It shows how your assumptions can make you miss things which are obvious on the field.

Reading the relatively short essay, it is clear that Nick Yee is assuming "beautiful" is the same as "attractive". Which really isn't the case for two reasons. First, beauty is relative. Second, beauty is only one facet of attractiveness.

Now, what Yee is asking is if you can manufacture a social framework with the "memetic" choices you allow the player to use. Obviously, I believe the answer is "yes". Half my posts are on the subject. But in order to understand the complexity of the issue, you have to understand the complexity of what people consider "attractive". Interestingly, it can be extrapolated on for use in virtually any situation.

Being "attractive" doesn't necessarily mean being "sexy", although a huge number of the pubescent players on MMOGs will tend to be guided mostly by the various bits fastened to their crotches. For example, Gandhi was attractive. People followed him, loved him, starved for him, died for him. But Gandhi wasn't sexy.

On the other hand, being sexy usually makes you attractive. It's just not a two way street: you can be attractive without being sexy. In fact, you can be pretty ugly and still be attractive.

Because attractiveness is relative.

Attractiveness is based around the environment of the judge. Or, more classically, "Beauty is in the eye of the beholder".

So, Horde is uglier than Alliance. But the game actively pits the two sides against each other, often very successfully. Moreover, people who choose the Horde are fully aware - and usually eager - to join into the darker world those monstrocities inhabit.

Then the pressures hit those who hardcore play the horde:

The Horde people help you, the Alliance people hinder you.
The Horde is exalted, the Alliance insulted.
The disturbing Horde women are available, the pretty Alliance people aren't.

After a few weeks of these pressures building, people who play the Horde can undergo a severe aesthetic shift. When asked, they'll probably still say that Alliance women are beautiful, Horde women are disturbing. But they'll be flirting with Horde women, thinking about Horde women, and trying to cyber with Horde women.

A common alternative, for people who don't make quite such a severe shift, is to think like Conan. While playing the game, they idolize the dark culture of the Horde, revelling in the "might is right" ideal. These people are doing the same thing, just on a slightly less disturbing level.

Obviously, this takes several important assumptions. For example, it only applies to hardcore horde: people who play horde for a significant number of hours each week, and don't play Alliance. It also applies only to people who play only WoW - if you go off and play City of Heroes, the isolation is broken and your preferences are recalibrated.

Similarly, it assumes that the people who chose Horde aren't coming in with a preference for disturbing physical shapes. I'm sure that some Horde players come in with pre-existing fetishes.

This growing extremism in isolation is found in every situation. From cults to SecondLife, people will persue goals they can see, and the importance of goals outside their worldview diminish with every passing moment. Isolation, often self-enforced through the forming of private groups, keeps these people from recalibrating to an accepted norm. So they get weirder and weirder.

This isn't necessarily bad. In fact, you can use this to do some awfully interesting things. Some of my live games have been based around the concept. But you need to understand that it does happen. Preferences are not static. People will change shape to fit whatever container you provide them with.

Really, most games are about isolation: about the wall between what the player experiences and what the "real" world is. Games are about isolating particular components of culture or reality and letting the player live them without overt interference from the normal world.

Managing your walls - governing the isolation of your players - is a skill which will allow you to create a strong and vital community of enthusiasts. :)