[rybosworld]

One thing that stands out as odd to me is the small selection of really "good" game engines.

Unreal seems to have the most features and the steepest learning curve.

Unity is comparatively easier to learn. Fewer features but C# support is a big plus. The major downside is that the company is run by incompetent MBA brains.

Godot is open source (awesome!) but doesn't seem ready for prime time. It doesn't have as many features as the other two and the commitment to GDscript seems... odd.

If I were picking a game engine today, it'd be between Unreal and Godot, but I'm not sure I'd feel great about either choice.

[johnnyanmac]

>One thing that stands out as odd to me is the small selection of really "good" game engines.

not much odd about it. 3d game engines are a large endeavor, somewhere between making a Large desktop application and making a full blown OS. I'd say the largest ones are comparable to making a web browser engine or even a full blown IDE suite.

There's also some slight snide in that there is no truly "good" game engine. Just ones that teams put up with enough to get across the finish line. The sheer scale of the engines and the nature of them being developer suites means running into edge cases is inevitable.

There are several mature 3D solutions out there, but the ones mentioned have the 3 largest communities and are strongly supported. But to throw out a few others

- Open3dEngine (O3D3), and fork off Lumberyard, which is a fork off CryEngine, is probably the biggest off-the-shelf competitor to UE when it comes to delivering AAA level games.

- Stride is an engine mentioned often in discussions, and it has a similar feel to Unity. But it simply isn't as mature as Godot and lacks some platform support like Mac

- UPBGE is the spiritual successor to the defunct Blender Game Engine, with a similar pitch: create your game without ever leaving your modeling suite.

-Finally, while not a fully featured game engine, I do want to give some note to Ogre3D. It's one of the oldest and most battle-tested graphics libraries out there, is MIT open sources, and is there for the kinds of developers who rolled their own engine/framework and simply used Unity as a rendering backend. Ogre doesn't include physics nor input (nor an editor), but it's really good at throwing your hand rolled game logic and giving it something visual. I could make similar other recommendations for stuff like Raylib and BGFX, but Ogre still being supported after over 20 years is admirable.

There aren't tons of choices, but there are choices out there if you are willing to get your hands dirty.

[jay_kyburz]

I'm thinking about going back to 2d. If Factorio and Slay The Spire can be smash hits in 2d, perhaps I can as well :)

I'm a fan of Love but I have to admit the amount of Lua I wrote for just my prototypes did start to get a little scary.

I might have a play with raylib.

[mnk47]

If you like or don't mind using Go, check out Ebiten as well. I think it's similar to Raylib in spirit, but it also supports Nintendo Switch, iOS and Xbox. IIRC it uses native graphics APIs (Metal for MacOS, DirectX for Windows) while Raylib just uses OpenGL.

[cmovq]

Interestingly both of the games you mentioned use custom engines. I think people overestimate the difficulty of rolling out your own 2D engine.

[jayd16]

It really depends on what your game/team needs. If you're trying to make a match-3 then Unreal is maybe not the best choice. If you're trying for competitive cutting edge graphics in an engine that is mostly done so you can actually focus on the graphics, then Unreal is top of list.

I think its probably better to look at Unreal as state of the art but licensable. Other big game companies have cutting edge engines but they're proprietary and not available to license.

[rudedogg]

> focus on the graphics

Unreal isn't good just because of the graphics. It has loads of tools built in, and that's where most of the value is in my opinion.

This video will give you an idea: https://www.youtube.com/watch?v=k2IP5DYQ0-0

[zzo38computer]

I think that which game engine will be "good" depend on manyu things including what kind of game, what computer system, FOSS, licensing, other features, etc.