Tuesday, May 24, 2005

A Taste of PAC!

Okay, I'm going to talk a little bit about Pattern Adaptation Control (PAC). I'm not going to talk about the nitty-gritties in this post, just the general idea. Nitty-gritties will come to those who wait.

When someone plays a game, they are toying around with a rigid rule set. The exact nature of the rule set doesn't matter to this post - it could be Tetris, it could be Final Fantasy 47. But it is important to remember that both the AUDIOVISUAL components and the PLOT/LEVEL PROGRESSION count as definite rules - an action causes a reaction. That is, of course, in addition to what you would normally consider game rules, such as 'four connected puyos vanish' and 'you get money for killing slimes'.

The thing about a rigid rule set is that it is predictable (even if it doesn't make any kind of sense), once you've been exposed to it. It forms a pattern, and the player is in there, learning the rules, mastering the pattern.

Obviously, there's some very definite twists on this. For example, the pattern you get from playing Tetris is a wholly different kind of emergent pattern than the one you get from advancing a narrative - a pattern that is only marginally interactive.

But the core idea is the same. Your job is to manage the player's understanding of the pattern. You want him to enjoy the pattern, revel in its permutations. And he wants to. Everyone who sits down to play a game is thinking, "golly-gee, I hope this is fun". All of them. "Golly-gee".

So why do we produce shitty games? We know what they want. Why don't we give it to them? Because we (and by "we", I mean "they") don't know how.

The art of Pattern Adaptation Control (PAC) is precisely that. It controls HOW, WHEN, and WHERE the player adapts to a pattern by controlling the methods with which the pattern is revealed and played. It also makes clear what elements the 'best' patterns for this purpose have. In theory, it works for highly emergent patterns - such as Tetris - and also for rigidly progressive patterns - such as a narrative.

In my game-du-jour I'll be using early PAC theories to attempt to lure the player into the pattern of the game. I'll tell you some of THOSE details later.

I like PAC. I am, as far as I know, the only person who knows anything about the subject. I'm the PAC-man. Goo-goo-ga-joob.


Justin Hawkins said...

Bet we could have one mean discussion about progression of gameplay through a storyline.

In paticular, the horrible use of todays goal oriented quest systems. I say this in relation to the masses trapped playing the online games with no real substance or plot. I will say that the massive online role playing games that let you be part of the game world and help shape the fututre of the game ar erather good approaches. But the lack of any consistency in the plot from these such games is horrible.
My solution to this delima is to write semi-artificial intelligent coding that bases its objective approaches off virtual-world aspects. This combines the best of both worlds with minimum object oriented coding.
What do you think?


Craig Perko said...

Well, I'll tell you, I've thought a LOT about that exact subject. There aren't a lot of options that are actually affordable. For example, your option requires a fairly serious breakthrough in algorithms and analysis before it's really feasible.

I still think the solution lies in letting players affect the game world, but you start to encounter some pretty serious problems pretty quick. How do you keep your game world from devolving into pornographic chaos or stagnating into 'those three guys run the place'?

I've come up with some ideas, but they've taken the back burner in comparison to the game I'm working on now. Most of my ideas involve restricting player vision... but that crops up all sorts of NEW problems.

It's a hairy problem.

Justin Hawkins said...

The game I am working on is a whole new genre,
(RTERS Genre= Real Time Empire Research & Strategy)
A whole new approach to the strategy of building an empire more realisticaly than civilization III or settlers can convey.
The backbone to the game is the super size research tree that is designed never to fully finish researching. Also because of the extent of this tree, almost everytime you play, it will be a virtualy different game.
Ontop of the research back bone lies a complex resource managment system and a advanced military abilites. The resource management system is designed that once you set it up it is virtualy maintenence free. The military aspects of the game are designed to grow and advance with the technology tree. Your whole empire is based on the technology and resources you have access to, I mean you will actually have to trade if you want to advance faster.
So if those give you an idea, without talking your ear off.


Craig Perko said...

Well, I would hesitate to throw my energy into the game you describe. It's not that any one piece of it is bad - after all, the game I plan to make in the future has most of the same features.

But there's two types of multiplayer gamers. The first likes SKILL. The second likes PERSISTANT WORLDS. Your game nimbly avoids being of either type. Presuming it takes several hours to play a game, it's too slow to be 'easy to learn, hard to master'. Simultaneously, it's simply too fast to be considered a persistant world.

So, I don't think I would back it. I don't think it would succeed.

But the IDEA isn't a bad one. It's just not something the market would like, I think.