[Home]History of Kenneth

Robo Home | Changes | Preferences | AllPages

Revision 19 . . January 17, 2004 16:46 EST by Brainfade
Revision 18 . . January 2, 2004 0:04 EST by Brainfade

Difference (from prior major revision) (no other diffs)

Changed: 1c1
New bot 1v1 bot by Brainfade.
Kenneth's dead. Well at least for the moment. I originally released him just to see what kind of level i was at. Since then a heap of half-assed and badly thought out changes have significantly lowered his ranking.

Changed: 3,68c3

What's special about it?

It has a burning desire to be a giraffe...

Great, I want to try it. Where can I download it?


How competitive is it?

See for yourself
http://rumble.robowiki.net/servlet/RatingDetails?game=roborumble&name=brainfade.Kenneth%200.31. In it's current state not very but hopefully i can work a few of the bugs out and climb the table.

How does it move?

Its target-orbit motion with a half assed and badly implemented fluid wall adjustment.

How does it fire?

Statistical targetting using waves.

How does it dodge bullets?

It doesn't directly

What does it save between rounds and matches?

Between Round - gun stats
Between matches - gun stats

Where did you get the name?

It wasn't chosen by me, it may change in future...

Can I use your code?


What's next for your robot?

Debug my wall avoidnace and tweak a couple of bits and pieces. Then probably a new new gun segmentation when i can be bothered.

Does it have any WhiteWhales?

Fermat keeps comign back to sting me, but at the moment Timdog is whooping my ass

What other robot(s) is it based on?

None as far as i know.

Comments, questions, feedback:

Ok, so i put it to you that this bot get's the strangest set of results in roborumble. For instance it get's 54.8% against griffon and about 15% against spiral crash. The thing i'm not sure about is whether my problem is movement or gun related. My curretn thought is that the gun is weak against stop+starters, and also bots with long time periods between stopping (i have a segment for enemyTravelTime?). The other thing, is that i think the movement is weak against robots that approach or close in on me, as pre release i never tested this. There may very well be a new release before the weekend is out... --Brainfade

Well it turns out that while trying to clear out some of my debug print statements i mangled my targetting code. version 0.13 will hopefully not be so inept... -- Brainfade

YAY 52nd!! strangely enough fixing my gun bug has decreased my score against some bots. Although if i'm lucky i'll manage to keep my scalp of griffon, and add to it the one of VertiLeach. I'm moving back home tomorrow, so i doubt i'll change anything before then, but who knows... --Brainfade

VertiLeach isn't worried. Once you start working with your movement it will bite you right back. =) Congrats on your #52 rank! -- PEZ

New version - 0.13. Fixed a wall avoidance bug that wasn't too major, tweaked the gun a bit and tried to implement data saving. There may be a slight improvement, but nothing more... --Brainfade

Ok, Kenneth is sitting in 50th as we speak, which has put a reasonable size smile on my face. Unfortunately the increase in ranking seemed to coincide with my loss of big snipes, but i reccon i can cope with that. On my lunchbreak today i was manically scribbling down pseudo code for what will hopefully be a new kick ass movement system. I probably won't get round to actually coding for a couple of days (either that or i'll stay up all night). Before i do another release i want to clean up the code, as i'm beginning to lose track of what's going on. Looking down this page is should be called "Brainfade blather's on about very little to pass time while something else is running". Thankfully there is a post by PEZ in the middle so it doesn't look as much like i have a mental problem. Until next time... --Brainfade
* You should see Jekyl's page for mindless blather. You are not even close =^> -- jim

Nice to see another English Flag :) -- Paul

I have spent aaaaaaaaaaaages playing around with my movement trying to make it better, but have failed miserabley (in fact it's worse than it was). So in the hope that someone can give me some clues as to where i'm going wrong. the basic details are here: /Movement? please post comments, idea's, criticisms etc... --Brainfade

I hadn't randomly written here in a while, so i decided it was time for an update. I have been playing around with my movement trying to get it flatter. I have a new much simpler movement scheme that i am currently testing and tweaking. The release of Slain was triggered by finding a function that gave what seemed to be a good result. In practice it does better against the lower ranked bots, but does even worse against the top ranked bots. Hopefully the next release's movement will be all round better. Unfortunately i rashly dived in and altered my gun segmentation, and managed to break it. I hadn't kept a copy of the working code anywhere, so at the moment i'm stuck with head-on targetting. I'll probably just decompile an old version and steal it back. Or maybe this is just the inspiration i needed to make my gun suck slightly less (removing the takebots from my WhiteWhale list would probably come by doing this). My current target is to outrank Duelist, so i'm looking for enough improvement to jump 8 or 9 places. hopefully i can knock this all together before new year. Anyway, back to the tweaking... --Brainfade

Ok, so tweaking has been more successful than i thought it might be. After running loads of tests i have simplified my choice now to 1 of 3 values. One of these is better against statists, one is better against pattern matchers, and teh other is a compromise between the 2. I suppsoe at some point in the future i could allocate this dynamically, but for now i'm just trying to pick one of them. Chances are i'll make a release in the early hours of tomorrow morning, and thanks to my freshly working client (YAY!!) i can give Kenneth some cpu cycles to see where he's gonna be. I have a horrible feeling that my testing may have been slightly too specific and that the new version will around the same place (or god forbid) lower than slain, but oh well you can't win 'em all. If i place in the low 30's or better, then i'll be the one who's smiling like a cheshire cat... --Brainfade

Good luck dude! -- PEZ

So after the disappointment of my tweaked 0.2 movement, i have completely restructured the logic. I'm gonna release a fairly untweaked version because i want rid of the 0.21 movement. I reccon it'll place about 60, which i hope to raise using my complete new testbed of robots. Here's to hopefully forthcoming flatness... --Brainfade

So I've again indulged in my new favourite pasttime of mangling my movement code. While destroying it i noticed a few glaring errors that would have seen all previous versions be a bit better - but not to worry. My plans for a comprehensive testbed of robots also went out the window and i'm currently running 250 round battles against TheTestBot to tweak a parameter. The second i see a value i like i'll stick a new version out... --Brainfade
I'm currently pondering some ideas that will hopefully be more competitive, that may end up as a new version, but to be honest i want a name that sounds slightly more threatening :). Be back soon peeps... --Brainfade

Robo Home | Changes | Preferences | AllPages