Abstract
Problem: Why do game development teams struggle to agree on even basic concepts, and what does this mean for how games should be led?
Approach: Tim Cain uses a memorable college AI class anecdote about defining "a cup" to illustrate how surprisingly difficult it is for people to agree on definitions, then extends this to game development.
Findings: Even highly educated people cannot agree on the definition of a simple everyday object. In game development, where concepts are far more subjective, disagreement is inevitable. Committees produce consensus — which means nobody hates the result, but nobody loves it either. A game director who can make final decisions is essential.
Key insight: If a room full of smart people can't agree on what a cup is, expecting a team to naturally align on game design concepts without a decisive leader is unrealistic.
1. The Cup Exercise
Tim opens with a story from his college AI class. On the very first day, the professor posed a seemingly simple challenge: define "a cup" — but in terms precise enough that a computer could follow the rule.
A student immediately offered: "A cup is a concave object that can hold liquid." The professor countered: "So, like a swimming pool?" The definition was refined — it must be something you can hold. Counter: "Like a bucket?" Further refinement — it can only hold a few ounces. Counter: "Like a plate?"
This cycle of definition → counter-example → refined definition continued for the entire 90-minute class. The discussion spiraled into minimum and maximum dimensions, acceptable slopes for concavity (not too shallow like a plate, not too steep like a test tube), and whether handles are required (what about teacups without handles?).
The point the professor was making: it is extraordinarily difficult to define even the most common, everyday objects in a way that includes all valid examples and excludes all invalid ones.
2. Definitions Get Harder With Subjectivity
Tim connects this to his earlier video on defining "RPG." Unlike a cup, which is at least a concrete physical object, genre definitions involve inherent subjectivity. His approach was to create a continuum — a checklist of RPG characteristics where the more items a game has, the more it qualifies as an RPG.
Even then, people disagree on where on the continuum something stops being an RPG. If the checklist has 10 items, does a game need four? Two? Eight? There is no objective cutoff point.
3. Why Committees Make Bad Games
This definitional problem scales with team size. Tim argues that committees produce consensus, not vision. Consensus means everyone settles on something nobody hates — but also nobody loves.
He illustrates this with how Carbine Studios got its name. The team submitted names they each loved, but every beloved name was hated by someone else. They ended up with "Carbine" — the name nobody hated, but also nobody liked. Tim's verdict: "Try making a video game with people like that. Kind of nightmarish."
4. The Role of a Game Director
This is why a game director is necessary. A game director says "the buck stops here" — they end debates, make final decisions, and keep the project moving. The conversation ends, a decision is made, and the team moves forward.
4.1. The Fallout Effect
Tim shares that on Fallout, simply having a project leader often resolved arguments before they reached him. Designers and artists knew that if they couldn't agree and brought the dispute to Tim, he'd resolve it unpredictably — he'd often just pick whichever option was easier for him to code. Since they couldn't predict his decision, team members were motivated to work things out themselves rather than risk getting an outcome neither of them wanted.
4.2. Co-Direction on The Outer Worlds
Tim acknowledges that multiple game directors can work, citing The Outer Worlds where he and Leonard Boyarsky shared the role. But this only succeeded because they had clearly divided responsibilities: Tim handled programming and system mechanics; Leonard handled art and narrative. They collaborated on the story early, then Leonard took the lead on it. The key was clearly defined roles with no stepping on each other's territory.
5. Takeaways
Tim's two key points:
Never assume people will agree — even on the most basic, fundamental terms of your game. If college students can't define "cup" in 90 minutes, your team won't naturally align on complex design concepts.
Always have someone who can make the final call — even with just two people. Tim notes this is why he liked having three co-founders at Troika: any disagreement could be resolved by vote, always producing a majority (3-0 or 2-1).
For solo developers, the disagreement problem doesn't apply during development — but it will surface the moment players get their hands on the game. No feature is universally loved or hated. That's just the nature of game design.
6. References
- Tim Cain. YouTube video. https://www.youtube.com/watch?v=zX7SZWTj5_w