Last week I got blocking out some new multiplayer levels and in the process I’ve learnt something about how to develop games in general. From the outset I held back from jumping into producing a lot of maps before we had tweaked and tested the actually combat/gameplay. This is something I learned to be important after making Kingdoms Collide since that game featured many maps that didn’t suit or enhance the game-play they were simply made based on artistic merits. However, even after recognizing this problem with my processes I still wen’t ahead and made a nice fancy looking level “Ruins of the wolf king” for KR. But now since blocking out and testing these test levels I am finding them 1. A whole heap more fun and 2. A million times faster to modify or iterate to suite what the gameplay needs.
My advice to any new game developer from here would simply be, don’t art-up your levels until the end. Go ahead and model generic props etc. But don’t get decorating or beautifying your maps until they have been thoroughly tried, tested and adjusted while still in easy to manage block form. Why? Because once you place hundreds of grass meshes on your terrain how willing do you think you will be to change the position of a small hill? or how willing would you be to move a wall once it’s been decorated with bricks, moss, paintings, vines etc.
So yes, bad news from all of this, the current “pretty” map will likely also be the least fun. Good news is there will be many more fun and well balanced maps in it’s place.
Meanwhile Paul has been re-coding the player’s movement and rolling code to be more fluid, it’s disappointing having to pause work on exciting things to re-make such low level code, but I think once we show the results everyone will agree that it’s worth it.