A while ago my wife Yayoi and I saw a baby hedgehog in the park. It was small and pink and rather cute, but it still had spikes. So my Hedgehog bots are supposed to be cute and pink, but still sharp and spiky enough to win a few battles. Imagine my chagrin when HedgehogGF ran out of space and had to lose the pink.
I was keen to play with movement and targeting ideas now that I had a microbots 749 bytes to work with. HedgehogP(attern) was the result. I took a fairly standard stop and go / flat movement and added some stat collection to decide when to switch between movement modes and alter reverse percentages.
For firing I took the basic nano-pattern gun and added some correction for walls to make it more accurate. It seemed quite clever at the time, but in practice it failed to win that much.
So HedgehogGF was born. I used the fairly standard micro bot dual mode movement (stopNgo/flat) and added a standard guess factor gun. I did try lots of segmentation variations, but as I discoverd with the Neophyte series, the "usual way" of doing things is the way because it has been proven to work. Most of my ideas failed and in the final analysis all I could do was tune carefully and add in a bit of velocity randomisation to try and fool a few of the more basic pattern guns.
Actually the first version of HedgehogGF caused me a lot of pain as I had an error in my build script and the Wave class was omitted from the JAR. Not only did this mean it would not run when uploaded to Roborumble, it also meant that my size calculation was out by 108 bytes. I lost a lot of sleep the next two nights trying to trim 108 to make it fit! And after all that work I still can't beat Thorn and Waylander and WaveShark proves to be hard to beat heads up.
Note: the above is copied from the old wiki, so it rather dated.