How long should a game prototype take?

Days to a few weeks. A focused prototype that answers one question can be built and tested in about two weeks: a few days for the question and paper version, about a week for the greybox, and several days for playtesting and a decision.

What is the difference between a prototype and a vertical slice?

A prototype answers whether you should make a game, using rough disposable work. A vertical slice answers whether you can make it at shipping quality, with one near-final example of each feature, and usually takes months.

Should I reuse prototype code in the final game?

Usually not. Prototype code is built with shortcuts to answer a question quickly. Once the idea is proven, rebuild the foundation properly with clean architecture and performance in mind.

How much does it cost to have a game prototype built?

Hyper-casual prototypes commonly cost $1,000 to $4,000 and take one to three weeks. The studio's multiplayer FPS prototype starts from $2,499, and its current Spark package starts from $999.

How to Prototype a Game Idea in Two Weeks

Q: How many people should playtest a prototype?

Small rounds work best. Nielsen Norman Group's usability research suggests about five users uncover most usability problems, and recommends several small rounds rather than one large one. That finds problems; it does not prove market fit.

The short answer

To prototype a game in two weeks, write down the one question the prototype must answer, test the idea on paper, build only the core mechanic in a greybox with placeholder shapes, fake everything else, playtest with real people in small rounds, and end with a written decision: kill, pivot or continue. Two weeks is enough because a prototype is not a small game. It is an experiment, and the code is meant to be thrown away.

Prototype, vertical slice or MVP?

	Answers	Quality	Typical time
Prototype	Should we make this? Is it fun?	Rough, disposable	Days to a few weeks
Vertical slice	Can we make this at shipping quality?	Near final, "one of each thing"	Months
MVP	Will players keep playing and pay?	Shippable, small	Typically 4–8 weeks for a focused mobile game

Developer and educator Rami Ismail draws the line clearly: prototypes help you decide whether you should make a game, while a vertical slice shows whether you can. He recommends prototypes be fast, cheap and precise, limited to specific questions and treated as disposable.

Days 1–2: write the question, then play it on paper

Write one sentence: "This prototype will tell us whether ___." Good questions are specific: "Is swiping to build a bridge satisfying for two minutes?" or "Do players understand the merge rule without a tutorial?" Add what result counts as a pass. Then try the core rule with paper, cards or a whiteboard. Dan Cook has described making Triple Town's first prototypes on paper and whiteboards before a rough digital version. Paper finds broken rules in an afternoon.

Days 3–8: greybox the core mechanic

Build only what the question needs. Unity Learn describes greyboxing as blocking out a scene with primitive grey shapes so you can build movement, triggers and interactions before final art. Practical tools:

Primitives and ProBuilder for levels and objects.
Unity's free Starter Assets for ready-made first- and third-person controllers, including mobile touch controls.
Free or AI-generated placeholders where a shape is not enough. If anything AI-made ships later, Steam asks developers to disclose AI-generated content.

The Experimental Gameplay Project at Carnegie Mellon, where students made games in under seven days, put it simply: fake it liberally and build the toy first. Skip menus, saving, settings and polish. Game designer Jesper Juul proposed a two-week version: week one for mechanics and controls, week two for tuning and progression. That is the rhythm here.

Days 9–12: playtest and tune

Put it in front of people who have not seen it. Do not explain; watch. Nielsen Norman Group's research on usability testing found that about five users uncover most usability problems, and recommends several small rounds over one big one. That is guidance for finding problems, not proof that a game is fun or will sell, but it fits prototyping well: test with about five people, fix, test again.

What to record:

Did they understand the goal without help, and how long did it take?
Where did they get stuck or bored?
Did they want another go without being asked?
What did they say the game was, in their own words?

To reach testers on devices: Apple's TestFlight supports up to 10,000 external testers, Google Play has internal and closed testing tracks, and Steam Playtest lets PC players request access without affecting the main game's reviews.

Days 13–14: decide

Kill if the core is not fun after honest tuning. Cheap now, expensive later.
Pivot if one part sparked interest; write the next question around it.
Continue if it passed. Write a short design document (our GDD guide shows how) and plan the MVP or a market test.

For mobile ideas aimed at broad audiences, the next test is often marketability: short ads or playable ads to see what installs cost. Our hybrid-casual production guide covers that path.

Mistakes that waste the two weeks

Asking several questions at once, so no answer is clear.
Polishing art, which makes testers comment on visuals instead of play.
Explaining the game to testers, which hides exactly what you need to learn.
Keeping prototype code for production. Rebuild properly once the idea is proven.

If you want it built for you

We build prototypes to answer specific questions, typically in one to three weeks. Hyper-casual prototypes commonly cost $1,000 to $4,000, our multiplayer FPS prototype starts from $2,499, and our current Spark package starts from $999. See hyper-casual prototyping or our packages. AI tools can speed up placeholders too; our look at AI game generators explains where they help.

Sources: Rami Ismail: prototypes and vertical slices · Game Developer: prototype a game in under 7 days · Jesper Juul: the two-week variant · Unity Learn: greyboxing · Nielsen Norman Group: test with 5 users · Apple: TestFlight · Steamworks: Steam Playtest · Raph Koster blog: Daniel Cook on Triple Town prototypes. Checked October 14, 2026.

Prototype a game idea in two weeks