No announcement yet.

Game concept questions from a dedicated ameteur

  • Filter
  • Time
  • Show
Clear All
new posts

    Game concept questions from a dedicated ameteur

    Hello UDK Community!

    This is my first post, but I'll do my best to explain what we're trying to accomplish, and hopefully this wonderful community can look through it and let me know what will need to be custom coded and/or what roadblocks I might run into, as the game concept me and a few friends want to work on is definitely not an FPS.

    Technical setup:

    Level Design

    Our basic idea for level design is to have an auto-generated 3d grid of cube-like pieces to create a terrain, which is first step of playing the game (think Minecraft, but not procedural, and not endless). Basically, nothing will be a traditional level, it is a game that is generated and locked to the size you select. The user will then passively control a collective of units that are mostly autonomous, so the camera moves around like an RTS, and there is no 1st person perspectives.

    Just as an example, we would love the "grid" to be made of blocks roughly 8x8x8ft that are placed next to one another and will likely end up being in the millions (including vertical air space). Basically, the gamespace might be something like 256*256*64 worth of reused meshes of different kinds.


    Besides basic camera controls, digging/mining and constructions such as doors and walls need to be designated via Mouse controls. Basically, you designate areas to be mined, and miners will dig the area out when they aren't doing something else, like eating and sleeping. The same goes for walls and stairs and doors. Also, workshops can be built that allow for the creation of items, such as armor and weapons, crafts for trade, ingots made from ore, food, etc.

    Keyboard buttons will control Z-Level movement as well, since we really want a 3d-experience.


    At any one time, the user will not be looking at a render of the entire map, for obvious performance reasons, so there is definitely a need to only render portions at a time. Hopefully that won't be too difficult. To change what you are looking at, you move the camera around on the current level, and traverse Z-Levels using a keyboard command and/or a menu button most likely.


    We are planning on having a turn-based style of gameplay that appears more or less real-time. It's a 'tick' system that gives ever actor a 'turn' each tick. Many tasks will take longer than a single 'tick,' but I think you get the idea. If you have any questions, let me know and I'll try to explain better or provide a simple example.

    Screw it, here's an example:

    There are two unit on the gamespace. Unit 1 is an Orc and Unit 2 is a Dwarf. Based on some calculation, an order is determined when more than one actor is allowed to take a new action on the same tick, as it matters who goes in what order (Think of a Dungeons and Dragons-style ruleset). The Orc goes first and is one square away from the Dwarf, so he moves to the square adjacent to the Dwarf, which takes 3 ticks (The move is instantaneous, as are most actions, but the Orc is now 'delayed' 3 ticks). The Dwarf is free to do what he wants, so he attacks the Orc, who is now adjacent. The Dwarf attacks immediately, and is now 'delayed' 4 ticks. In three ticks, the Orc can now make a decision, and he decides to attack. The Orc attacks and kills the Dwarf, and the Orc is delayed 4 ticks. After 4 ticks, the Orc can now decide what to do next.

    This might change a bit if issues come up, but the basic principle is to have an action queue that delays the unit after an action is taken. I think this kind of system will be easier for us to manage and balance, as we are very much in a Dungeons and Dragons mindset when it comes to rules, especially combat rules.

    Thank you for your time, and if anyone has any ideas that are better and/or simpler to accomplish the same goal, feel free to point us in the right direction. We are really ready to learn the UDK and get something moving in the right direction.

    You're difficulty is going to be in having an auto generated level. That changes things a lot. And pretty much everything you have designed will require some special programming, which is also something that makes it difficult.


      I wasn't happy to hear that, although I kinda figured everything I posted so far would require that (I tried to post about only the elements I believed would be the most difficult). I am still pretty optimistic overall, especially after seeing a thread from December about spawning static meshes to create a dynamic level system such as mine. I'll just have to work on each problem as it comes up and see how far we can take it.

      Thanks for taking the time to read and post, instead of completely ignoring me. I am planning on getting the $99 indie license and really taking a serious crack at this. We already have about 1/3 of the design of the game decided upon and written down, we just need to start implementing as soon as we have the design completely figured out. I'm fairly certain that game engine limitations will force our hands from time to time to make changes, but the overall gameplay should be possible. Thanks again!


        You don't need the license until you're ready to sell.


          The only problem I see here are your autogenerated maps.

          Some RTS games have them, Soldier of Fortune 2 was an FPS game that had one, but to be honest I have yet to to see one auto generated map they can't live up to what a level designer can do in terms of visuals and game balance.

          It's pretty hard to teach a computer things like beauty, taste and even common sense (e.g. trees do not grow underwater) by algorithms. If you can get that working decently, there's still stuff like balancing (not putting one player at a gamebreaking disadvantage right from the start), pathing, visibilty and so on and so on.

          Static lighting and Global Iillumination won't be present in a satisfying form either, as this cannot be done on the fly.

          Are you planning to do this in UDK? If so, I doubt Kismet is powerful enough for that and I'm not sure UnrealScript is any better. you might just put years of effort into this and come up with nothing satisfying.

          My advice would be not to leave something that important up to chance. I'd rather create a multitude of maps and randomize where the player starts. That gives your level designers total control and the player an illusion of unpredictability.


            I'm sure it's possible with UDK, it's just going to take a lot of work.


              Originally posted by krillin6 View Post
              blocks roughly 8x8x8ft that are placed next to one another and will likely end up being in the millions
              I'm worried about performance...


                I also asked for a game concept... nobody answered