Durandal

Background Information

What's special about it?: It uses a dynamic tree structure for storing guess factor data that passes existing information to new branches. It minimizes memory consumption by only expanding branches as needed (based on the collected data). Basically, it can use a greater number of segments (I don't know if that's a good thing, but it sounds cool) without a significant increase in memory usage. Worst case (rough estimate) memory consumption is 1.5 GB, and actual average consumption settles around 60-80 K.

How competitive is it?: Meh. It's okay. Top 100 (barely) in the general rumble.

How does the melee strategy differ from One-on-one strategy?: You could say that the melee strategy differs quite a bit (i.e. I don't have one).

What does it save between rounds and matches?: It stores all movement and targeting data between rounds and (currently) nothing between matches.

Where did you get the name?: In literature, Durandal (more commonly spelled Durendal) was the sword of the legendary Roland. However, my bot is a homage to the rampant AI from Bungie's Marathon games.

Can I use your code?: In a bit. I'm currently cleaning up my code/comments/making sure I give proper credit for the code that came from this wiki.

What's next for your robot?: Currently, I'm working on the tree structure I use for storing the guess factor info. I want to find the ideal ordering of segments (it currently segments on eight dimensions). Perhaps I could have the ordering dynamically determined for any given opponent?

Does it have any White Whales?: Dookious, the bot that inspired me to do this.

What other robot(s) is it based on?: It's not directly based on any other robot, though the foundations for the wave surfing/guess factor targeting came from the tutorials on this wiki.