Robo Home | Changes | Preferences | AllPages

/Participants - /Rankings

Just an idea I had from my experimentation with teams and Droids. Basically this would be a OneOnOne competition on a small field (400x400). All robots would be Droids, meaning no Radar. Ive currently put together one which can beat sitting duck before time runs out, and about half the time will beat trackfire. Id definitely reccomend trying out an ImplementsDroid, just because it will have you thinking of robocode a bit differently. -- Jokester

Sounds like fun. I'd make an entry for it. --UnderDark

Nice Idea, but its harder than i thought! --Krabb

Ya, so far its been about firing a ton of low bullets until I find the target. Its going to take a lot of variable work and algorithms to hit something that moves... I would love to see a droid melee competition ;) -- Jokester

Shrubbery: Here's your chance to treat hitting a robot as a scan.

Jokester: Would a droid melee competition also be on a small battlefield?

-- Kawigi

Ok, I've got a little droid hack that can beat all the sample bots except Walls in one-on-one (400 x 400) and wins melee with all the sample bots (somehow) (barely on 400 x 400, but by a bit on larger battle fields, and this is without melee movement code, not that it would necessarily help). I figure to beat Walls one-on-one, I'd have to add some really Walls-specific code for the gun or use a more anti-walls sort of movement. -- Kawigi

I can beat every sample bot except Walls too! But i have still a bug.. Enough for today, i have to sleep :( But tomorrow ill upload my version :) --Krabb

Hmm, I havent done much with melee yet, Ill run some to see how it goes but I just assumed that smaller would be better. If larger seems to give you a higher margin, then we might consider standard field size. -- Jokester

400x400 is cool for 1v1 I think. I'd like to enter my DroidPoet of old. It's one of the very few bots in the Haiku Driod class. It's made for melee I guess, but I doubt it matters. =) I'll enter itin both 1v1 and melee. -- PEZ

OK. So DroidPoet wasn't very good at this from my tests. But I've made some small changes and the result is DroidPoet 1.0 and it does the following against DoBoh:

1st: pez.femto.DroidPoet 1.0	124196	31750	6350	66484	7263	8696	3652	668	339
2nd: Krabb.doboh.DoBoh 1.0	85924	16950	3390	47209	3611	12487	2276	370	635
Haven't tested with any other Droids yet. -- PEZ

I was so happy, that i could beat your bot :) But i made some small changes too, so that i get this result:

1st: Krabb.doboh.DoBoh	        37747	10300	2060	18778	1855    4355    396	220	182
2nd: pez.femto.DroidPoet 2.0	37425	9100	1820	17916	1675	5953	960	197	206
realy close --Krabb

Yeah! Do you have any EnergyManagement in DoBoh? I have a simple check I don't fire myself disabled, but I'm not sure if it's worthwhile. That last bullet could hit... -- PEZ

Up to now i have no EnergyManagement, and i think ill make the power only depending on the hit chance. May be its good to stop shooting at low energy, so that you win the round, if the opponent gets disabled. But who knows :P --Krabb

I suspect DoBoh would do much better if he didn't get cornered occasionally. -- Kawigi

Schhh! DroidPoet just loves opponents that gets cornered now and then. =) -- PEZ

I just love how I started the rules, and yet have been doing my tests on the wrong field size... -- Jokester

Has anyone of you already tested your droids against Dummy's BlindDroid? (part of his BlindFighters? team) ?? -- GrubbmGait

Yes, or at least against his perceptual droid NanoBot that's based on BlindFighters?. Someone should enter that one. And isn't BlindDroid? redundant? -- Kawigi

Well, you know who named it . . . -- GrubbmGait

Oh the BlindDroid? owned DoBoh hard :( But my new targeting system shuld beat him :D --Krabb

Adding dummy.PercNanoRobotDroidDummy? 1.0 as probably the original search-by-firing algorithm. -- Kawigi

If Perceptual, Nano and Robot are not a requirement, I'd much rather have the BlindDroid? participate, unless the perceptual-nano-robot-version somehow performs better ;-) . -- Dummy

I can't see why perceptual, nano or robot would be a requirement. That's probablt a separate competition. Nothing wrong with redundancy. There are very few bots that can lose half of their name and still be. =) -- PEZ

The only reason mine is an extends robot is because I havent had time to put in some movement, thus there isnt much need to do multiple things at a time. -- Jokester

Interesting, the only reason mine is a Robot is because then it doesn't beat itself senseless against the walls. =) -- PEZ

Well, make a release of just BlindDroid? then and enter it ;-) (or make some improvements to it) -- Kawigi

ImplementsDroidCompetition: nice idea. I have entered an old droid of mine: Hodur. Probably not much of a real opponent, but is has more chances of winning here then in the RoboRumble where it once was second to last...Were can i find the results of this competition? --Loki

Up to now you have to run the battles by yourself :) But if you want to know, who the current winner is :) It is the new DoBoh version! --Krabb

ok. RoboLeague is running... I will verify your claim ;-) --Loki

The new version, wait 5 min :) --Krabb

I added SoloDroid. I don't expect it to do well since it is a robot and droid, making it extremely hard to program with. However, it should be able to escape most of the other droids because it is a robot with a very simiple pattern. -- Kinsen

Wow, my tests have been way off. I was hoping to beat some one ;p well I guess its back to the drawing board. -- Jokester

My tests are off, too - I expected to not beat someone. I guess not losing to anyone by a huge margin pays off? -- Kawigi

Can someone post some results from their tests with all participants here? I'm a bit curious but still too lazy to run the competitioin myself. =) -- PEZ

Just click /Rankings :) --Krabb

Linking the full battles wouldn't be a bad thing, either, though ;-) -- Kawigi

Hmm, it turns out that I do beat the bots I expected to, its just that they are beaten less by the others, I'll have to start testing against some of these other ones... -- Jokester

Hmm, it turns out that I do just beat everyone - I don't remember ever beating BlindDroid? (but then again, I don't think I have the exact version Dummy uploaded). Also, I didn't test melee on 400 x 400, which may be where I was strong I guess. -- Kawigi

I thought melee would be on standard size (800 x 600). 400 x 400 seems very crowded with 8 (or more) bots. -- GrubbmGait

Yes, i think meele would be better on 800 x 600! --Krabb

we could also stick to the standard sizes: 800x600 for 1-vs-1 and 1000x1000 for melee. I tested the 8 bots on 400x400 and 1000x1000. The ranking was the same, even the battles lasted about the same. my 2 cnts. --Loki

I prefer 400x400 for 1on1 and 800x600 for meele --Krabb

Robo Home | Changes | Preferences | AllPages
Edit text of this page | View other revisions
Last edited June 24, 2005 22:02 EST by Krabb (diff)