Coherency


Orienteering

One of the most challenging things in games and in any creative work is in figuring out what direction to go in.

The basic strategy most folks will use when they don’t have other process ideas is “copy-paste-modify” or “remix”: Take something that looks appealing and try to make your own spin on it, maybe add in a few other ideas. The majority of creative techniques are an enumeration of remix methods.

Mood Board

What I have found from trial and error is that the limitation of remixing mostly has to do with the “truthfulness” of the thing we end up saying through the resulting work. What we’re making is media, and media can’t help but communicate something.

If we only modify random elements of an existing work, then we also communicate random things. Sometimes this works, but more often it does not. Games are built up out of a lot of expectations and assumptions, and they can easily become confusing and arbitrary by changing even one small rule.

Communication

Note that the content of the message is not the issue. There are numerous “definitions of game” that seek a universal goal for games. This is not about the goal. It is about whether it is comprehensible at a basic level. Clear communication must precede profound statements. The statement can actually be very trivial: A lot of music, for example, isn’t aiming to communicate more than “this is a nice rhythm that you can dance to.”

Understanding is built on knowledge, and knowledge is in turn built on truth. What makes something a true statement, and not gibberish?

We can apply a bit of philosophical reasoning to assist us. If we can predict in advance which changes will read truthfully, we can be more confident about what happens when we go to build something!

Venn diagram

The tool we have for achieving this prediction of knowledge is called a “theory of truth”. We believe something is true because we have a theory that justifies it. 1

There are many theories of truth based on different premises and making slightly different arguments, and they often overlap, explaining the same statement in different ways. For this article, we’re focusing on the coherency theory of truth, and my own definition of coherency:

A proposition is coherent if it supports our existing knowledge.

That is, it isn’t contradicting propositions we already know. “the sky is blue” and “the sky is red” are propositions that do not cohere. Preferably, it follows from other propositions by logical entailment.

If I hear “the suspect is very tall” and “the suspect is a professional basketball player”, I would expect the two statements to be coherent, since even if I cannot know for sure that they are both true, it would make sense that either one follows from the other.

Square Wave Approximation - Credit Thenub314 - Own work, CC BY-SA 3.0

In communicating, perfect accuracy is only desired some of the time. Many situations prefer approximation and rhetorical value. Statements that build up coherent knowledge also build up a convincing piece of fiction. This is a power we can use for good or evil.

Let’s talk about games again. How do I make coherent statements about a game for a practical purpose like doing game design?

When I realized that coherency explained problems I had been having with my own design work, I started looking for a way to make a formal process. The result I eventually came up with is a form of brainstorming called the “coherency game”. I used the coherency game in this past Ludum Dare 42 to establish the basic game concept, and I’ll quote from the liveblog I made:

The way the coherency game works is to start with a quick brainstorm and make a list, including the theme or other requirements:

  • Running out of space
  • Hard drive
  • Oxygen supply
  • Space-time continuum

I don’t need a lot of ideas in this list. A rule of the coherency game is that I can only keep around as many ideas as I can remember. Four is fine to start.

Now, I start comparing every pair of ideas to see if they have a reasonably strong coherence, or whether they contradict each other. This is an informal process and relies on your good taste, which I’m sure you have. Coherency isn’t exactly consistency, but consistency is a good starting point for judging coherency. What we want is concepts that can support each other well.

  • Running out of space : Hard drive ✔️
  • Running out of space : Oxygen supply ✔️
  • Running out of space : Space-time continuum ❌ 2
  • Hard drive : Oxygen supply ❌
  • Hard drive : Space-time continuum ✔️
  • Oxygen supply : Space-time continuum ❌

Where the ideas don’t make sense together, I should start changing, adapting or combining them. This optimizes our idea set towards coherence. The reason why we want coherence is because it puts the whole concept on solid ground, where it’s easy to get our imagination going and think about which features are important and give them a hard priority. When the concept is incoherent, its implementation tends to expand in scope without ever resolving.

  • Running out of space
  • Hard drive
  • Vapor
  • Recycling

What I noticed in the last set was that while hard drives and oxygen didn’t spark any ideas, there is a gas seal in hard drives, which made me think of vapors, which made me think of vaporwave and how it recycles stuff. So I swapped in both vapor and recycling.

  • Running out of space : Hard drive ✔️
  • Running out of space : Vapor ✔️
  • Running out of space : Recycling ✔️
  • Hard drive : Vapor ✔️
  • Hard drive : Recycling ✔️
  • Vapor : Recycling ✔️

Vaporwave

OK. This is looking like something. A trash planet vaporwave aesthetic, where if you run out of bytes on your hard drive the game is over. 300 megabytes. Let’s add in some more concretely mechanical aspects so that I can start thinking about how to build it.

  • Running out of space
  • Hard drive
  • Vapor
  • Recycling
  • Driving
  • Sector editor
  • Running out of space : Driving ✔️
  • Running out of space : Sector editor ✔️
  • Hard drive : Driving ✔️
  • Hard drive : Sector editor ✔️
  • Vapor : Driving ❌
  • Vapor : Sector editor ❌
  • Recycling : Driving ❌
  • Recycling : Sector editor ✔️
  • Driving : Sector editor ✔️

Although driving fits with vaporwave and its love of cruising in old sports cars, it doesn’t fit with “vapor” - the gas - quite as well, and sector editing is a bridge too far. Recycling and driving are also kind of iffy.

  • Running out of space
  • Hard drive
  • Re-circulation
  • Driving
  • Sector editor

I decided to try putting vapor and recycling back together and focus on something that would give a sense of movement.

  • Running out of space : Hard drive ✔️
  • Running out of space : Re-circulation ✔️
  • Running out of space : Driving ✔️
  • Running out of space : Sector editor ✔️
  • Hard drive : Re-circulation ✔️
  • Hard drive : Driving ✔️
  • Hard drive : Sector editor ✔️
  • Re-circulation : Driving ✔️
  • Re-circulation : Sector editor ✔️
  • Driving : Sector editor ✔️

So: You’re driving through your hard drive’s sectors to edit them. The drive is running out of space. There’s a race track representing the drive that you “circulate” through. There are still a few things to work out about implementation, but I think I can run with this concept. Time to take a break.

In subsequent entries in the blog, I work out more details about the game’s scenario and play structure by referring to these ideas and reusing them in different ways, over and over.

When we have a coherent set of ideas, we have a powerful feedback tool for every other question about the game: “What coheres best with the main themes?” This tool alone is sufficient to power most of the development process, because a truthful answer to this question will expand the set of coherent thoughts further and further, producing unexpected good ideas through the process of “filling in” the ideas that already work.

World of Warcraft with HUD mods

In general, project scope is hostile to coherency. Why is this? It’s actually pretty straightforward: more scope means more unrelated ideas, and hence more combinations that can be incoherent when paired with each other.

Scope also compounds: If the response to “it’s not good enough yet” is “try adding more stuff”, you go even farther away from coherency. This is how you can end up with a large investment into a dead project. The common wisdom “less is more” isn’t about simple reduction, but about maximizing effectiveness.

Flea Jump Height

Projects running with big budgets and lengthy schedules tend to suffer the curse of being barely coherent enough to hang together, because they are straining to cram everything in and the different parts overlap and conflict, creating more work to do to resolve the mess, and a more difficult collaborative environment. Coherent projects have a lower actual scope, but they punch above their weight.

Zak McClendon’s “Big Board” design strategy is a great follow-up to the coherency game. Forged within the production cycles of very large, high-profile projects, he adds certain formal elements to the process of prioritizing ideas past a brainstorm, which helps in organizing a group discussion and generating buy-in. He discusses it in full in this PRACTICE 2018 talk.


  1. ( I am glossing over all the details and critiques of these theories. This is the secondhand version of how I learned it in the 101 course. ) [return]
  2. I rejected this one, but as someone pointed out, “there’s a time-travel or hyperspace version of FTL right there.” [return]