Ugluk/Gemini

Version	Rating	Notes
1.1a	1772	Waves tracking limited to live rounds.	Performance varied greatly between the two for many bots, though the ratings averaged out to about the same.
1.1b	1770	Waves tracking not limited to live rounds.
1.2a	1780	Includes distance to target.	Another case of averaging out. I may have to develop a means of dynamically including or excluding layers.
1.2b	1782	Excludes distance to target.
1.3a	1773	Includes opponent acceleration.	Very few dramatic differences between performances.
1.3b	1768	Excludes opponent acceleration.
1.5a	1782	Distance to target increment: 40.0	Less finely grained distance to target is clearly superior, and also clear is that the segment is one of the more favored.
1.5b	1775	Distance to target increment: 11.0
1.6a	1818	Decay Rate: 95%	Not much to say. A very fast decay of visit count stats is not effective.
1.6b	1787	Decay Rate: 80%

---

Getting this error from Gemini 1.4a and 1.4b:

pedersen.Gemini 1.4b: Exception: java.lang.NullPointerException
java.lang.NullPointerException
    at pedersen.Gemini.getMovementMethodsMelee(Gemini.java:62)
    at pedersen.Gemini.init(Gemini.java:40)
    at pedersen.Gemini.run(Gemini.java:34)
    at robocode.peer.RobotPeer.run(Unknown Source)
    at java.lang.Thread.run(Thread.java:613)

I'm running Java 1.5 and Robocode 1.1.3. From the scores they've gotten so far (I haven't uploaded any battles), seems like it's not particular to my system... -- Voidious