Warning:  there might be a couple of light spoilers, but everything is in the trailers anyway so… nevermind that.

I hope what I’m going to say won’t sound too stupid, agreeing with everybody and not making a fool of myself is really hard. n_n

That being said:


  • DON’T duct-tape features together to make Frankenstein-style abominations.
  • DO find natural sources of gameplay/setting harmony and consistency.
  • DO identify the axes from which your game is built in order to know what you should and shouldn’t do (it also works for other forms of intellectual construction).


This has been my main design paradigm and method for the past year. I understood the base concept about a year ago, and spent a few months refining it into a model for quality creative direction. The main idea is that in quality artworks, everything just fits perfectly; you can call that harmony, purity, focus, or whatever.

There are 4 main axes that need to be taken into account when using this model: Perimeter, Exploitation, Justification, and Legitimacy. These axes are used in the following way: you define a perimeter, then exploit it, by justifying that your exploitation of the perimeter is valid, and making the legitimacy of your perimeter obvious.

Simply put: find a nice place to dig (like a goldmine), and DIG deep, don’t just scratch the surface.

I realized later that that described very precisely what QUALITY is about.

Here’s a definition of quality we were taught at university: “a product’s ability to satisfy known or unknown user needs”.

Making quality art is somewhat easy, you just define what you want to do, and do it.
If you define something worthwhile to make, and succeed making it, you have quality.

It’s a sort of exploration of the Ideosphere: define a need for exploration like “hey, we should totally make a game about …” and satisfy that need.

That’s what happens all the time: you publish a trailer for your game, and create a need for your players to put their hands on it.
If you satisfy that need, you win, because you only have to satisfy the need you’ve created, which you can probably quantify.
If you don’t, you frustrate your players and lose.

You need to define very precisely what you’re making and announcing, it’s very very important.

Let me explain.


I’d say, I usually start defining a game from an idea, but there’s often a couple of other questions I need to answer before I can do anything. “What am I going to do with this idea ?”, “How am I going to do it ?”, “What options do I have ?”, “What should I do ?”.

I’ll use examples:


So I guess the first idea was the portal thing right ?

What did they do with that mechanic ?

A Strategy game ? A fighting game ? nope, a Puzzle Game.
Third person ? RTS view ? nope, First person.
How are the portals used ? you have a gun.
What do you do ? You escape from a science facility.

What’s relevant in here ? I would say you can sum up the game by “first person puzzle game with portals”.
So we have “Portals”, “Puzzles”, and maybe “First person view” as keywords to describe the game, “Portals” being the main one.

Portal is what happens when you combine portals and puzzles like this:

When you think about it, almost every game, or artwork has a couple of base axes like this:

  •  Super Meat Boy: Platformer, Hardcore, OldSchool – Everything in the game just says “I’m the revival of the oldschool platformer games (but without everything that went wrong at the time)”
  • Braid: Time manipulation, Puzzles, Depth
  • Angry Birds: Demolition, Puzzles, Simple

The thing is, these axes I’ve just given aren’t necessarily valid, but that’s what I work with, and it works very nicely.
It even works for other domains than video games, music and film for instance: any artwork can be described as being made by combining a couple of axes.

You can see an intersection area on the Portal axes diagram.

That is what I call a perimeter.


There’s a couple of things to be said about perimeters.

What happens when you don’t actually have any intersections between your axes ?

Just take features and throw them together, you’ll find out that your features are interchangeable.
From what I know, that leads to hollow games, clones, and cookie-cutter production.

Take a look at this graph:

Games like this are just the sum of their parts, like dead bodies, and dead bodies aren’t exactly engaging by default.
It’s like making Frankenstein-style abominations: not games, but bits of games stitched together. See the difference between the two ?

So we have to find intersections, to find synergy between ideas, which creates new real ideas to work with, before we start making anything.


So what do we do to make games that are designed naturally, and not artificially ?

Well there’s one thing that happens sometimes, and that is really awesome: there is a strong link between mechanics, and setting.

How awesome is that ? well, it’s a start for NOT making duct-taped stuff.

I call that Hyperweaving.

What is Hyperweaving ? It’s when there’s an equivalence between your gameplay and setting: they imply each other.
Your mechanics and setting are basically the same: you could say that you have a consistent game world.

I’ll take Mighty Jill Off as an example: Jill is sent downstairs by her queen, and she has to climb back up.
Jill climbing up and you making her jump are the same thing: you can’t have one without the other.

That’s organic design, it’s consistent, natural, not stitched together… you get the idea.

Another example, Half-Life: you’re escaping from Black Mesa, which implies shooting stuff.
You escape from a science facility, and you need to shoot stuff.
You shoot stuff, to get out of a science facility.

See how hard it is to separate the setting and the mechanics ?

Again take Counter-Strike for instance (sorry for the Valve over-representation here, they just happen to make insanely high quality games from what I understand):

How could we get the same gameplay without the terrorists vs anti-terrorists setting ? (I mean, the bombing/hostage rescuing and team infiltration)
How could we have a terrorist vs anti-terrorist setting without some good old team infiltration and bombing ?
It would need some serious re-arranging for that to be possible.

In hyperweaved games, everything just fits so well together, you have a hard time isolating elements because of how tight and closely woven together things are. Organic design is when everything is just too intricate for you to see through.
And it’s really engaging to feel like you’re not presented with a mechanical system instead of good company to spend time with.

I feel I need to give an example of a non-hyperweaved game.

I hate saying nasty things about specific people. I wuv everyone :3. It’s just regretful that people make mistakes.
I won’t give names, but I’ll say I’ve played a AAA game recently that could have had the exact same story but with a totally different gameplay to it.
Or the exact same gameplay but with a completely different setting.
That specific game, didn’t have anything special gameplay wise, and the setting was just slapped as Art and story assets, without much to do with gameplay.
It was very conventional.

Conventions should never be where you stop, but only where you start.

A hyperweaved setting is what you’re looking for.


Well then, turns out this post ended up being longer than expected.
I guess I’ll have to divide it into several parts.
So for the moment, I’ve only talked about perimeters.
Next time: exploitation and actual design methods.

In the meantime, if any of this is either too obvious or too confusing for you, please tell me. n_n
Please don’t pay too much attention to the fancy words I use.
I don’t want to pretend I’m bringing anything new, everything is there under our nose, we just need to understand it.

I know this is stuff I’d like to have understood years ago, and I hope you can get something out of this post too :)

Have a nice day n_n