Robo Home | Changes | Preferences | AllPages

Can you tell me why this bot totally and utterly eats Tityus 0.9.4 alive? -- PEZ

True , 59% against Tityus
How that works , that's what you want to know , isn't it :P Honesly i don't know.It's just a mirror bot that changes its movement every 5 (or was it 7) rounds to a modefied OldMan movement with a quite effective gun. It has good scores against most bots (around 50% although with top bots) but the micros and nanos with simple movements are worrying me.Maybe a movements score calculator that makes me move with the currently better movement will fix this --deathcon

I already had noticed that it's a MirrorMoving?. But Musashi (version 2.18 - with AntiMirrorSystem) fails to detect it's mirror-moving (at least by now). It seems not to be 100% mirror, introducing some random factor, am i correct? -- Axe

Read the description again. =) It alternates between mirroring and a flattener movement. -- PEZ

I thought Musashi re-detects at the start of each round to deal with this kind of alternation? This was my reasoning as to why he sometimes lost against mirrors when they nailed him before his high power bullets and perfect hitrate would kick in. -- Kuuran

I know, but i already saw in development that if i turn on my detection system to be less "sensible" in detecting that kind of moving, he dies so fast as any mirror. I am truly convinced that Musashi's AntiMirrorSystem is not 100% perfectly accurate, it can't reach OldmanXP? nor PrairyWolf?, Althoutgh(i never can say how u spell this) NaN, and HumblePieLite are vanished. PEZ, i am beeing slow in that AntiMirrorSystem page not because i ain't gonna share it, but because i'm not at least 90% satisfied with it (85% by now). What i can say is, if you know that they gonna mirror your movement (and that is a really important if), it's only a matter of you knowing where yourself will be, i dare that u can't find a way to fit this kind of system in your minis. Remember the mantra (a new form for it) : "Aim low, Hit hight!".

But OldManXP truly deserves my hail! A really good KISS bot. -- Axe

Aaach! sorry, a lot of sintax & meaning corrections above, i'm drunk. Yup! Soooryy. @) -- Axe

Well deathcon, it looks like this bot gave some people some extra work. You can count me to that list as well >-/ Locke is basically thrashed most of the time. Can you tell me a little bit more about its gun? --Vic

It's more because of it's movement than because of it's gun I think. If you're still using the Raiko movement and faces a smart mirrorer like OldManXP then you get into trouble if your gun is not as good as Oldman's to hit that movement. -- PEZ

XP is also a MAJOR problem for Raiko. Raiko juuuuust wins 52%, PB index -18.5. When observing a match I noticed that OldmanXP? forces R into a deadly GF1 streak a lot of times. These streaks are much more punished by XP than the other way around. So I was kinda hoping XP has a simple unsegmented GF gun, or even a LinearTargeting gun. I think its the first, because the DuelistMicro series are also giving me trouble, and I believe those are unsegmented GF guns...but I'm not sure. This would give me a vital clue about the weakness of Locke's gun. --Vic

Well, Raiko's gun might not be the best choice against Raiko's movement. I think many DuelistMicro's use pattern matcing guns actually. They sure are slow like PM bots at least. And they keep getting high PBI's against most surfers. Which often indicates PM gunnery. -- PEZ

Hmmm, so still no clues there... maybe deathcon can shed some light on the XP gun? --Vic p.s.: found it on the wiki: DuelistMicro uses VirtualBullets which in practise has the characteristics of an unsegmented GF gun I think.

Oh just detected that here is still traffic on this site. Of cause i will bring some light on XP, but as its a quite long time since i wrote that bot i will have to search for the source first ... --deathcon

Oh i finnaly found it - if uses a highly segmented guess factor gun with an unnormal segmentation. Many enemies get confused by the changing movement i guess. Need more information ?--deathcon

Well, of course I am curious what this unnormal segmentation is, because it is very effective against Raiko's movement :-) Could you tell us a little bit more about that? --Vic

I segment on the heading the enemy if going to move relative to my position. And i just detected the same bug that is in all of my old guns: i did e_accl (which is enemy accelation as you have already thought)

 = -e_speed+(e_speed=e.getVelocity) instead of = -Math.abs(e_speed)+Math.abs(e_speed=e.getvelocity())

ah, there's that RelativeHeading again. We've been discussing that a couple of weeks ago. I assume the other segmentation dimensions are more normal then....velocity, acceleration, distance.. right? --Vic

Depends on what you mean with other??? --deathcon

Are you saying that RelativeHeading is the only segmentation in XP's gun? That would explain a lot! --Vic

Oh no, its one of lots. As i said it is heavyly segmented. But additional to 'normal' segments i used to add one ore more 'unnormal' segments to my guns these days to find out which worked and which doesnt. --deathcon

Ah, ok I see. Thanx, maybe this will give me some clues why XP is so effective against Locke --Vic

Robo Home | Changes | Preferences | AllPages
Edit text of this page | View other revisions
Last edited May 5, 2006 4:20 EST by GrubbmGait (diff)