[Home]GloomyDark/Comments20031213

Robo Home | GloomyDark | Changes | Preferences | AllPages

Nice bot! You are quiet close to DT:

 1st: pe.SandboxDT 2.11	        86696	34900	6980	38839	5976	1	0	699	302	0
 2nd: pez.GloomyDark 0.1	61457	15100	3020	39716	3552	4	64	302	698	0

GD beat DT on bullet damage, it's just the survival that's not too good. A little tweaking and you might force paul into an update. -- Tango

Is that from a 1000 round battle? I was going to say it was just luck, but not with that many rounds. Cool, but still not close enough. DT is a survivalist bot and thus sacrifices bullet damage for survival, and does it extremely well. Marshmallow used to always beat DT clearly in bullet damage, but over 1000 rounds always lost in total score. Of course that was earlier DTs. M doesn't have much of a say against DT 2.11. I'm glad GD looses with some face left. =) Thanks for posting those results. I have been running much shorter matches and only have tested long battles against Kawigi's and iiley's bots. DT feels a bit unreachable for me at the time. 2.11 is such a strong bot that I almost doubt I can do anything about it. Almost. I will try! -- PEZ

I wonder what drm.Magazine does that DT doesn't like... And GB, the strangest bot ever I think, seems to score better than Gloomy against DT too. I wonder if GB could beat DT if I give it a mirroring movement. Wait, I can try that. =) -- PEZ

Yes, it was 1000 rounds. I was also wondering what was happening with Magazine... -- Tango

I'm just running 1000 rounds of DT vs Magazine... it doesn't look like it was a fluke... DT doesn't seem to be able to hit it... -- Tango

 1st: pe.SandboxDT 2.11	42576	17400	3480	18630	3065	0	0	348	97	0
 2nd: drm.Magazine 0.39	24906	4850	970	17718	1347	20	0	98	348	0

Those are very SlowBots... that was less than 500 rounds and took ages... it looks like DT takes a long time to learn how to hit Magazine, but once it does, it is fine. That's the problem with small battles. -- Tango

That's interesting. In RoboRumble@Home teh chances that DT and Magazine meet twice on the same client are minimal. Which means, that with a movement like Magazine, which DT seems to have trouble figuring out quick enough you can beat DT. Magazine obviously does. =) My test with using MirrorMovement in GB didn't work out all that well though. It seems like its old movement was what DT has problems with, not its guns. Nevertheless I'll release this mirroring GB since it is a crap bot anyways. =) Funny that if you use a GoToBot mirroring is as easy as:

goTo(new Point2D.Double(centerX - (enemyLocation.getX() - centerX), centerY - (enemyLocation.getY() - centerY)));
I stole the code from Marshmallow's AntiMirroringTargeting? gun. (Which uses M's future position instead of enemyLocation of course). -- PEZ

Wow... Already someone using my graphing utility. :D - That's an excellent looking curve PEZ, it's very flat and spread out; curve flattening targeting bots (Fractal especially, quite obviously from the graph) will have a lot of trouble hitting it. Very nice. -- Vuen

Thanks for those words. I had actually hoped it would make a bigger difference with this tuned movement. But it hasn't exactly soared through the rankings yet. The automatic movement adapter (DT style) isn't in operation yet, since I can't get it to not wreak havoc. Maybe with that in place I can do some real kick-ass curve flattening. -- PEZ


Robo Home | GloomyDark | Changes | Preferences | AllPages
Edit text of this page | View other revisions
Last edited December 13, 2003 23:43 EST by PEZ (diff)
Search: