The vertical slice: what goes in it, and what to leave out

What it is for

• Proving the game is fun, to yourself first. A slice that is not fun is enormously valuable information delivered early.
• Proving the team can finish things, which is what publishers and funds are really assessing. Anyone can build systems; finishing is the rare skill.
• Establishing the quality bar that the rest of production will be measured against and costed from.
• Producing marketing material. Screenshots, a trailer and a store page all come out of a slice, which is why the interface in it matters more than teams expect.
• Making the remaining budget estimable, because now you know what one polished chunk actually cost.

Choosing the slice

The choice matters more than the execution. Pick badly and you polish something unrepresentative for three months.

1. Take the core loop, not a set piece. The thing players will do a thousand times, not the scripted moment they will see once.
2. Include one of everything that defines the game. One combat encounter, one traversal section, one system interaction, one progression beat.
3. Choose something scalable. If the slice only works because it was hand-built, you have proved nothing about the other twenty hours.
4. Avoid the tutorial, which is atypical, and avoid the finale, which is exceptional by design.
5. Ten to twenty minutes is the usual length. Long enough to show a loop repeating, short enough to finish properly.

What has to be final

The trap: a slice you cannot scale

The most expensive way to build a slice is to hand-craft it. It looks wonderful, it wins the pitch, and then production discovers that nothing in it was built in a repeatable way.

• Bespoke assets that will not become a library. If the slice's props, animations and interface were one-offs, the cost per minute of the rest of the game is unknown.
• Hard-coded systems. A slice held together with scripted triggers proves the moment, not the mechanic.
• A quality bar the budget cannot sustain. If the slice took three months for ten minutes, the twenty-hour game is a hundred and twenty months of content at that bar. Publishers do this arithmetic.
• No tooling. The slice is also the moment to prove that the team can build content efficiently, which is the actual question behind "can you finish this".
• The honest version: build the slice with the pipeline you intend to use, and let it be slightly less pretty than a hand-made one. It is a better argument.

How long a slice takes

• Typically 10 to 15 percent of the total project time, which for a two-year game is around three months.
• It always takes longer than planned, because it is the first time the team finishes anything, and finishing reveals every gap in the pipeline.
• That overrun is data. The ratio between your slice estimate and its actual cost is the multiplier to apply to the rest of the plan, per how long a game takes.
• Do not extend it indefinitely. A slice that has been three weeks from done for two months is a team that cannot finish, which is the thing being tested.

Slice, prototype and demo are different things

A public demo is a slice plus everything a stranger needs: an onboarding path, settings, a clean exit, and the first ten minutes designed rather than assumed. That is a real amount of extra work, covered in the first ten minutes.

Presenting it

1. Let it be played, unassisted, if you can bear it. Watching someone play your slice cold is the most information you will get all quarter.
2. Have a build that starts in the slice, with no menu diving and no setup.
3. Fix the crash-on-launch class of problem first. A slice that fails in the first minute is judged on that minute.
4. Bring the plan with it. The slice shows the quality; the plan shows how the remaining hours reach it, and the second half is what funding decisions turn on.
5. Do not narrate over it. If it needs explaining, that is the finding.

A vertical slice does not prove you can make a good game. It proves you can finish one thing, which is the harder claim and the one being tested.

A slice checklist

1. Is it the core loop rather than a set piece?
2. Does it contain one of everything that defines the game?
3. Could the rest of the game be built the same way, with the same tools?
4. Is the interface in it final, or is it placeholder in every screenshot?
5. Is there audio, mixed?
6. Does it hit target performance on the target hardware?
7. Does it start in the slice, with no setup?
8. Can a stranger play it without you talking?
9. Do you know what it cost, and what that implies for the full game?
10. Is there a plan attached showing how the rest gets to this bar?