Unity or Unreal? Choosing your game engine

The engine is the first technical decision in a game, and the hardest to reverse. Switching after six months means rebuilding most of what you paid for. Choose it the way you would choose a building site: before anyone pours concrete.

Both engines are good, and both have shipped successful games on every major platform. The question is which one fits your team, your platform, your art and your budget. We look at five things, in order.

1. The team you can hire

Unity is scripted in C#. Unreal Engine uses C++ and Blueprints, its visual scripting system. That one difference shapes your hiring.

• Unity: C# is easier to learn and quicker to iterate in. In most markets, mobile game developers with Unity experience are the larger pool.
• Unreal: Blueprints let designers and artists build game logic without code, a real advantage on a content-heavy game. Performance-critical systems are written in C++, and strong C++ game programmers are fewer and cost more.

Ask who exactly will work on your game, and what they have shipped in that engine. An engine the team knows beats a better-suited one they learn on your budget.

2. The platform you ship on first

Both engines build for iOS, Android, PC and consoles. They are not equally comfortable everywhere.

• Mobile first: Unity is the usual choice. Builds are smaller, the engine runs well on older and cheaper phones, and ad networks and analytics tools reliably offer a Unity plugin.
• PC and console first: Unreal is the usual choice when the game is 3D and visually ambitious. Its rendering, lighting and large-world tools are ready on day one.

Decide where the first release lands, then where the second might. A phone with 3 GB of memory is a harsher place than a console.

3. The art style

If the pitch says “realistic”, Unreal gives you more before you write a line of code. Nanite handles very dense geometry, Lumen handles dynamic lighting, and the character and cinematic tools are strong. Unity can reach high-end visuals too, but the team assembles more of the pipeline itself.

If the game is 2D, stylized or low-poly, Unity is lighter and faster to work in, and its 2D tools are mature. A 2D puzzle game in Unreal is possible, and rarely sensible.

4. The budget, to build and to ship

The license is the smallest line on the bill. Unity charges a subscription per developer seat once your company passes a revenue threshold. Unreal takes a royalty on a game’s revenue above a threshold. Both have changed their terms in recent years. Read the current terms on their own sites before you sign, not in an article. That includes this one.

What moves the budget is people and time. High-fidelity 3D art costs more than stylized art, in any engine. And on mobile, a larger download loses installs. Size is a business number before it is a technical one.

5. Life after launch

A live game is a calendar: events, new content, balance changes, fixes. Ask how each one reaches players.

• Can content update without a new store release? Both engines support it, if it is designed in from the start.
• Where do accounts, saves, leaderboards and remote settings live? Unity and Epic each offer their own services, and independent backends such as PlayFab work with both.
• Is it multiplayer? Unreal has a mature networking framework built in. Unity has its own package and is often paired with a service such as Photon.

A short rule

• Mobile, 2D or stylized, a wide range of devices, a small team: choose Unity.
• PC or console, realistic 3D, a larger team with C++ skills: choose Unreal.
• In between, such as a stylized 3D game for both mobile and PC, the team’s experience decides.

For the record: we build in both engines. SANAM, Tamr: The Date Farm and Kick Arena are mobile games, and all three run on Unity. That says something about phones, and nothing against Unreal.

Before you decide

The dabb picks hard ground before it digs. Do the same, and settle three things before the contract:

1. Ask why this engine and not the other. A partner who can argue the other side has thought about yours.
2. Ask for a prototype of the core loop running on your weakest target device. Measure frame rate, download size and load time.
3. Agree that the project files, the source code and the store accounts are yours.

Then commit.