Last night I attended the monthly Twin Cities XNA User Group meeting. Instead of having a single presenter we have six members of the group stand up and talk about some projects they’ve been working on for 15 minutes or so. I have a feeling that everyone could have just gone on and on about what they were doing but, alas, there was no time.
I gave a presentation of the XNA game I have been working on called Drop In. Drop In is a pretty simple game with 3 screens. Also all of the artwork was created with Paint.NET and the music as created by Zach Adams along with some of the sound effects being created using a free tool called sfxr by DrPetter. So you can say that Drop In was created entirely with free tools!
The source code is broken up into an engine and Drop In specific projects. Currently it will only run on a PC but with some minimal hacking you could probably get it to work with an Xbox. I think the main problem is loading up XML files for settings off of the disk which I’m not sure if it will work.
One of the most interesting parts of the game is the AI used for one player mode. I created a reusable MinMax evaluator in the game engine library. I looked at a great applet created by Yosen Lin at Berkley as a reference for improving my implementation of the MinMax algorithm with Alpha Beta Pruning. It actually seems to work pretty well too. There are some deficiencies with how I evaluate states I think and also there is a problem where the computer will evaluate all states and determine that every move it makes will result in a loss and therefore it ends up making an effectively random choice. It doesn’t account for human error and should still probably try to pick the “best” of the losing moves.
Anyway, I have posted the code in its full form so you can play Drop In and also as prebuilt binaries if you don’t care about the code. You’ll need the XNA runtime to play it or the XNA Game Studio 2.0 to build it. I have also posted just my game engine in case anyone else is interested in creating a 2D game or just taking a look at the MinMax algorithm I have created in C#.
Next month’s user group meeting might focus on AI if so then I can talk more about the MinMax algorithm I implemented for anyone interested. I might actually be looking to improve this AI a little by then too, I’m thinking of including some sort of mechanism for machine learning for example. I am also toying with some ideas for changing the goal of the game and adding in some features for complexity. For example perhaps the goal shouldn’t be to just get 5 in a row but instead to get as many in a row as you can when the board fills up. I’m also toying with ideas of special pucks that do more than just drop, such as pucks that push other pucks down and pucks that blow up and pucks that change the colors of other pucks. Etc.
I also want to make the AI evaluate the state continuously, not just on the computers turn and include the possibility of network play… which should be pretty easily actually since I have abstracted the players pretty well.
If anyone out there has any motivation to create a 3D Puck model for me I’d be pretty happy about that! I’m looking to take this game engine (and Drop In) to the next dimension and to do that I will need some good models.