I’m pondering what pitfalls there might be in 2D development.

You perhaps could share a link or two and lead me near to some good articles?

Or perhaps share your own experience.

Juuso Hietalahti


  1. Call me old fashioned… but I like that old skool 2D style. :)

    Greg – yeh, it makes sense. I guess one question could be something I blog about in my next blog post.

  2. From a development point of view, the biggest problem I’ve seen is that 2D artwork is now a more rare skill. Most professional artists do 3D work. When we got an artist to add some more artwork to Meridian 59, we basically had someone create a 3D model of the art and turn it into 2D.

    I think you also need to keep in mind the type of artwork you need for your game. I think 2D art is a lot better for high detail work, especially on an indie level. But, as others have said, you get some good advantages with 3D in some situations like isometric views.

    Hope that helps.

  3. I agree with the previous posters. Even if you’re going to do a “2d game” there are many benefits to using a 3d engine. For example: A top-down view has benefits of being able to model all your game objects with a 3d modeling tool rather than drawing pixels — even if your model is ‘flat’ with a ‘pixel looking’ texture.

    Personally, I think you can really get a lot of mileage with a 3d engine in a single plane to reduce the complexity of physics and movement.

    Have fun.

  4. Dexter, yeh.

    @Lumooja: “You might lose all high quality customers and get lots of low quality customers instead.”
    -> That’s an interesting statement :D

    @Ronin: Yeh, if I’d wanna do isometric… I’d probably use 3D and then ensure that renderer handles the isometric rendering.

  5. 2D can be a great medium to work with if you use it for the right things. Don’t try and simulate 3D in 2D when you can just go 3D instead…
    I started developing my first game in 3D, but due to little funding and no time I decided to go for a 2D game first instead. But I’m not making that 3D game into 2D, I came up with a new game instead that works perfectly with 2D.
    Also, 2D is great if you want to develop in flash. Sure flash has some 3D engines out there too now, but it’s still primarily 2D.
    Pitfalls I have found so far is that you have to be careful which engine you choose for your game. Some engines only support tilemaps at this moment(like flixel) while others can be a bit complicated cause they’re in still in early versions(like PBEngine).
    And the usual saying goes here too. Don’t reinvent the wheel ;)

  6. I would be worried about the customer base. You might lose all high quality customers and get lots of low quality customers instead.

    From a gamer’s perspective I don’t see any benefit in 2D games, but rather I would like some fancy 3D special effects.

    The only real 2D game which I would play still today is Command & Conquer 1, but even that tries to simulate 3D graphics.

    Nowadays you don’t have to simulate 3D graphics anymore, but you can make a 2D oriented game with real 3D graphics. That’s what Westwood also thought, and made all the new Command & Conquers with a 3D engine (you can zoom down to an unit and see it like from FPS perspective).

  7. 1) There are no pitfalls. Only challenges. Period. This is something we try to hit on hard at Double Cluepon. We see a great many games, edging toward photorealism, pretty graphics, and zero gameplay. That is the suck. While pretty is good, it means nothing if the same, or more effort is put into the design of the gameplay itself. I still play horrid looking Atari 2600 games…it’s the gameplay!

    2) Ragnarok (which is an isometric has full camera movement, pan and zoom. Their rendering model is pretty effecient. It’s also an older game. Makes money. As does Trickster Online, etc. Isometric is vastly underrated, and great things can be done with it, if you’re willing to sit down and do some actual design. Not to mention, there are some awesome Flash iso engines out there, with good camera functions. =)

    3) Side scrollers are still a very viable genre. One only has to look at Kongregate, or any web game portal to realize this. Platformers are pretty much a win/win if, as I said before: you are willing to sit down and do some design. Chibi Knight is a good example of a recent sidescroller, with some simple but effective gameplay design. http://www.kongregate.com/games/BoMToons/chibi-knight

    4) As an indie house working with both the puzzle and the MMO genre, we look first to design, and what should be done before we get to the pretty aspect of things (stuff like how it will look). Even Raph Koster said this, i.e: get the basics down, then try some styles. We try to think of problems more like a challenge, than a show stopper. The only show stoppers we know of are undocumented API’s. =)

  8. I see 3 main pitfalls with 2D for isometric.

    1) You’re more lor ess stuck with only showing 1 viewpoint of a character or building unless you pre-render all views separately.

    2)It’s hard to rotate the game screen.

    3)Character can’t change their outfit because doing so will require re-rendering all animation with all outfits.
    The total number of animation frames and time required will go up quickly.

    To sum it all up, You can’t really change the shape of buildings or character in real-time.
    If I can help it, Prague will be my last 2D game.

  9. Tough question. Maybe my canceled project BIONIC can serve as an example (see http://bit.ly/c300my ).

    I wrote an isometric engine for this game alike the engine used in Baldur’s Gate. The pitfall here was my lack of experience in generating game content. So I tried to enabled very complex shapes (speak concave shapes). Take an isometric archway as example: When a character stands underneath the archway, what should be drawn first? The archway or the player?
    The only solution was to divide concave objects into multiple convex objects. This raised the complexity of content creation into unmanagable heights. Finally I quit the project because of this problem (and some minor other problems).

    Conclusion: Don’t use 2D for something 3D solves implicitely. Rendering 3D models to 2D sprites and arranging them in pseudo 3D (isometric view, 2.5D) is a dumb decision. It kicks out built in collision detection, visibility information and is a waste of time.

    @juuso: Is that what you meant by “pitfall”?

Leave a Reply

Your email address will not be published. Required fields are marked *

Pro-Human Quiz: