Apoptygma 0.4 released. Bugfix, few minor changes, colours removed. Primarily a maintenance release for speed and stability. Should allow Apoptygma to operate nearer it's full potential. Sorry to everyone whose RR@H clients were ground to a halt by 0.3.
WOOHOO! I made it work. *dances like a maniac*
Apoptygma 0.3 features:
- Multi-variable non-symbolic pattern matcher
- Simply segmented absolute guess factor gun
- Virtual bullets based virtualguns system
- Pretty colours
- No compromises (unless you count movement)
On the competitiveness aspect... StatGun? is Apoptygma 0.3 without colours and with one more line of cleanup code.
StatGun? doesn't skip myriads of turns and occasionally freeze like Apoptygma 0.3 does.
StatGun? is codesize 771 (without colours) :(
1st: arthord.StatGun? 4317 1350 270 2292 404 0 0 27 8 0
2nd: arthord.micro.Apoptygma 0.3 1821 400 80 1267 74 0 0 10 27 0
Given another 22 bytes Apoptygma could be a competitor, but for now it stands as a successful experiment in miniaturization. As well as pretty. ;)
A new test robot by Kuuran, a first draft plus one minor revision (version 0.2) is ready and vaguely competitive. This robot tests how complex systems can fit in a micro, currently it boasts a VB based stat gun in a VB based virtual gun array along with one other gun. The original spec also called for a pattern matcher, which looked like it would fit, but in the last minute it had to be cut. This is a priority to get in before version 1.0.
To be uploaded and added to competition as soon as the repository is back up.
What's special about it?
I believe this is the most complex weapon system any micro to date has boasted, though that is somewhat debatable as the stat gun is still pretty raw.
Great, I want to try it. Where can I download it?
How competitive is it?
Top 30 micros. Not amazingly. The movement is poor.
How does it move?
The bot remains perpendicular to it's opponent at all times. The bot chooses a random distance in a random direction and moves it.
How does it fire?
It determines whether it's better to use pattern matching or stats to fire at an opponent, and does so.
How does the melee strategy differ from one-on-one strategy?
How does it select a target to attack/avoid in melee?
What does it save between rounds and matches?
Where did you get the name?
It came off my playlist, Apoptygma Berzerk is a band I am somewhat fond of.
Can I use your code?
Because of the nature of the pattern matcher I'm afraid I lied about open sourcing it. This matcher could allow for a re-engineering of the NanoSatan
line, so I'm going to sit on it for a bit longer.
What's next for your robot?
Fix the memory leak it seems to have.
Probably remove colours, fix an inherent instability in the matcher (you'll see it wig out occasionally), statistically segment the matcher ala Tron, add more stat segments.
Plenty, it's poor movement makes it's performance against simple guns with good dodging (generally poor scoring robots) quite weak.
What other robot(s) is it based on?
Not applicable. Though I suppose any Minibot statist owes some heritage to SandboxMini
Comments, questions, feedback:
Wow, looks like the number of Micro-statists has tripled in the last couple weeks :-p. It never occured to me to try and combine it with a PM, though (probably because of the complexity/number of dimensions of my stat gun). -- Kawigi
I guess knowing you'd need to do it better than DuelistMicro scared people off :p. As a pure stat gun goes FMicro is a far stronger - and thus far bigger - bot, this gun is running practically unsegmented. The big experiment I'm trying here is to squeeze virtualguns into a micro rather than to try and make a super-competitive bot, but right now I'm not having too much luck either way (hopefully it'll be both when I'm done ;) -- Kuuran
... and since the VirtualGuns minis that I know of can be counted on one finger... (of course, it might have been done before...) -- Kawigi