however, this paint is not as good, as pain can do.
If you know "not every pain hurts",you'll see...
* 1.40:I think the melee stragtegy start works from this version.
* 1.35:Added a melee stragtegy
* 1.32:Fixed a bug in movement.
* 1.31:Change the movement to Flattener Movement introduced by Lacrimas.
* added:you can set the behavior in the properties file.
* first released version
It sure is a pain for TityusMega! The development version beats Cigaret quite easily so I was hoping BlestPain wouldn't be so painful. But I was wrong. Blest eats TityusMega for lunch. Maybe it can exploit it's movement flaw. Would you care to analyse its movement some with Blest's stats and maybe you can see what's wrong with the movement? It's fun to read Blest's debug output though. Not as varied as TheArtOfWar maybe, but still cool. =) -- PEZ
Do use this method in TityusMega: translateInsideField?(Point2D point, double margin) ?<I think it is not a good method for movement,it can make you movement more predictable,because i used it once for my Spark,it did very worse than before,so i removed it,then it did better.The way to avoid it is just choose the point inside field to move.If it is a point outside,rechoose a new one.Maybe it not helpful on your bot,maybe have a try?. -- iiley
Hmmm, I don't think it's the wall movement. I could be wrong, but looking at the bot moving and checking the profile in the walls segment it looks better than other movements I have created. Yes, I do translate the destination inside the field, but only after some tweaking. I was thinking you could maybe see in BlestPain's stats where TityusMega is weak? But I'll continue to hunt for it. =) -- PEZ
I have no grapher for BlestPain,so i am hard to see something.But one thing,is TityusMega's movement is relative to enemy's firepower?It seems that BlestPain hits TityusMega with little power much.Then he choose the little power bullet to fight. -- iiley
Yes! I have tried to use the enemy fire power to create a flat profile over the full range. But I have only really tested 1.5 and 3.0 (using my patch of Fhqwhgads). When you say low power bullets, how low are we talking about? Thanks for alerting me to this. Now I have something to work on. -- PEZ
From the version info up there:
===== version 1.31(2003.12.15): ===== 1.31:Change the movement to Flattener Movement introduced by Lacrimas.
;) -- ABC
Je-je, I saw that! But I still wonder... That would make Lacrimas and Blest pretty much the same bot in 1-v-1 and I seem to have much bigger problems with Lacrimas than with Blest. -- PEZ
That's probably because Lacrimas was further improved. I suspect Iiley used Jamougha's Narcissus class in Lacrimas, since it saves .narc files. -- ABC
I have noticed those files too. Thanks! This is another piece of the complicated puzzle of sorting my problems bots. I would like to have a clue if they gun me down or outmove me. -- PEZ
I actually think that Lacrimas is the root of most surfing bots that exit today. As I recall, this bot remembers the last GuessFactor that it was hit at and avoids that factor. I can not remember where I read that but I know Paul mentioned it and iiley confirmed it. Thats why Lacrimas and Blest seem to move the same but Lacrimas can really put a hurt on you. -- jim
Yup. I think iiley might have been the first surfer. But I think he didn't dare go full monty with it and thus never unleashed the full potential. -- PEZ
I think ad.Neo/Quest? just may have been the first; I get the impression from old posts that Frakir bolted some sort of adaptivity onto the Cigaret movement and illey copied the idea back, and much improved on it, in Lacrimas. Any of those 'in the know' about the Quest code know if I'm talking rubbish or not? :) -- Jamougha
I'm a bit in the know of the Lacrimas code while it was still open source. I don't recall seeing any adaptivity until quite a while after iiley had let himself be influenced by Quest. But it was all a long while ago and iileys code has always been quite opaque to me so I might have missed it even so... -- PEZ
A long time has gone,when i back to this robocode world to have a seeing,i saw you guys are still hot loving robocode,i think i have to come back now (well i will have a job and a nice house so i must have a good humor to "play the stupid robot game") haha~~.~;] -- iiley
For 10 days rush about our studio,now i can sit down and surf the internet,maybe this weekend I can continue my robots' coding. Haha~~~ Nice to meet you PEZ and Vic, my robocode friends.;] -- iiley
We all miss u man! great u are back!!!! -- Axe
There is a problem on my new PC,robocode runs too slow on my PC,only 60fps with:
two Samples.Crazy<pre>800*600 field Visible Robot Energy Visible Robot Name not Visible Scan Arcs not Allow robots to change colors repeatedly maximize window (AthlonXP 2500+,512M DDR,Geforce FX5500 128M)
But on my friend's PC(Intel P4 2.4) it can up to about 500fps.Did all Althlon CPU runs robocode slowly?OR ... any one knows the reason? It will be very helpful for me,Thanx in advance. -- iiley
Maybe you have set the maximum fps to 60??? Or your computer doest find/can load your graphic card. --deathcon
Strange i have the same hardware and get 80 fps (1024*768 pixel ) --deathcon
Just tried it: AthlonXP? 2200+, 320 MB DDR and an onboard video card (no gaming is done on this computer) fluctuates between 110 and 120 fps for two sample.Crazy on 800x600 field. Using classic Robocode and not RobocodeGLV014. -- Kuuran
I set the maxinum fps 10000. And my computer know his graphic card.:)I think this is a strange problem,it views java draws graphics very much slowlier on AMD hardware than it on Inter hardware. Because i am coding some java games,it draws many graphics,and it become slowly too on my AMD hardware platform(still very fast on my friends' Intel hardware platform),and i dont know how to improve it. Sad~~;[ -- iiley
I've heard rumors that floating point math is faster on AMD hardware and integer math is faster on Intel hardware (of course integer math is faster than floating point math, too). -- Kawigi
I think they're both down to doing it in one cycle, so the higher Intel clocks would give them a slight advantage. Most of AMD's speed advantage and ability to go with lower clocks is due to their ability to do things more efficiently, but in integer math the floor is already reached. At least that's my understanding. -- Kuuran