2011-10-26T00:00:49 hey admins, there seems to be different types of battle resolution (attack,focus,closest,support). is it going to be stuck to "focus? 2011-10-26T00:01:47 focus is here to stay 2011-10-26T00:03:05 hrm.. what's focus? 2011-10-26T00:03:31 that's the name of the current battle resolution scheme 2011-10-26T00:03:41 i'm doing some test maps to see how formations will work in chokes 2011-10-26T00:03:41 :D 2011-10-26T00:03:59 ah, ok 2011-10-26T00:04:07 thought it was a certain stage of combat lol 2011-10-26T00:04:28 bobbydroptable: you can block a <3 wide area 2011-10-26T00:04:41 yes 2011-10-26T00:04:43 that's what i've realised 2011-10-26T00:04:52 anything wider than that and there will be exchanges 2011-10-26T00:05:29 *** xar0l has joined #aichallenge 2011-10-26T00:05:45 *** dvladim has joined #aichallenge 2011-10-26T00:06:11 *** Saulzar has joined #aichallenge 2011-10-26T00:06:29 *** canadiancow|work has quit IRC (Remote host closed the connection) 2011-10-26T00:07:08 janzert: hi, did you look at the color scheme on ant paste? 2011-10-26T00:07:17 working on a response :) 2011-10-26T00:07:32 McLeopold: Most of the new coloring looks better I think, but the white numbers on the bar graphs seem worse 2011-10-26T00:08:36 * janzert has been thinking about replacing the A/Z button with an options button that would pop up a options dialog of some sort 2011-10-26T00:09:00 hmm, there are people trying to clone with what I'm saying :-) 2011-10-26T00:09:13 *** TheDigitalNinja has quit IRC (Quit: Page closed) 2011-10-26T00:09:14 if I may assuming they heard it 2011-10-26T00:09:29 we could then add color themes in there and we wouldn't have to worry quite so much about the color blindness issue for the main theme 2011-10-26T00:09:51 also could have themes suitable for red-green color blindness and other types separately 2011-10-26T00:09:58 *** Sunshowers has left #aichallenge 2011-10-26T00:10:37 I agree, pushing these colors to vars was the first step 2011-10-26T00:10:55 I can make the bars use the black text, or sand color 2011-10-26T00:11:22 I think black will probably work best, but sand color might be interesting to see 2011-10-26T00:13:13 *** porco` has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2011-10-26T00:13:19 I've also been wondering if it would be more natural to have the speed increase and decrease buttons on the bottom, styled as double arrows >> and << basically? 2011-10-26T00:13:21 hold on... 2011-10-26T00:13:29 I think so 2011-10-26T00:13:50 it's the way the planetwars visualizer was anyway 2011-10-26T00:14:30 *** bmh has quit IRC (Quit: bmh) 2011-10-26T00:14:52 *** bmh has joined #aichallenge 2011-10-26T00:15:18 *** nishant has joined #aichallenge 2011-10-26T00:15:37 *** dvladim has quit IRC (Read error: Operation timed out) 2011-10-26T00:16:04 I'm using the java staretr package and get the following exception 2011-10-26T00:16:16 Exception in thread "main" java.lang.NoClassDefFoundError: AbstractSystemInputParser$SetupToken 2011-10-26T00:16:16 at AbstractSystemInputParser.parseSetup(AbstractSystemInputParser.java:86) 2011-10-26T00:16:16 at AbstractSystemInputParser.processLine(AbstractSystemInputParser.java:48) 2011-10-26T00:16:16 at AbstractSystemInputReader.readSystemInput(AbstractSystemInputReader.java:18) 2011-10-26T00:16:16 at MyBot.main(MyBot.java:15) 2011-10-26T00:16:17 Caused by: java.lang.ClassNotFoundException: AbstractSystemInputParser$SetupToken 2011-10-26T00:16:17 at java.net.URLClassLoader$1.run(URLClassLoader.java:217) 2011-10-26T00:16:19 at java.security.AccessController.doPrivileged(Native Method) 2011-10-26T00:16:21 at java.net.URLClassLoader.findClass(URLClassLoader.java:205) 2011-10-26T00:16:23 at java.lang.ClassLoader.loadClass(ClassLoader.java:321) 2011-10-26T00:16:25 at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:294) 2011-10-26T00:16:27 at java.lang.ClassLoader.loadClass(ClassLoader.java:266) 2011-10-26T00:16:29 can anyone tell me why this is happening? 2011-10-26T00:16:55 nishant: use http://pastebin.com/ for longer text 2011-10-26T00:16:55 pastebin.com is a much better way to post long tracebacks or code snippents 2011-10-26T00:17:12 alright 2011-10-26T00:17:17 but can you help me here? 2011-10-26T00:17:18 nishant: did you compile a jar file? 2011-10-26T00:17:29 to answer your question though, that looks like an incorrect classpath of some sort 2011-10-26T00:17:33 i used the make file provided 2011-10-26T00:17:43 and tried compiling all java files manually 2011-10-26T00:17:46 it was updated yesterday, I think, when did you get it last? 2011-10-26T00:17:58 and, windows or mac/linux? 2011-10-26T00:18:02 sometime yeaterday itself 2011-10-26T00:18:08 also what's the command line you're giving to playgame? 2011-10-26T00:18:09 im on a linus machine 2011-10-26T00:18:21 ./playgame.sh 2011-10-26T00:18:28 yes, the make file was bad, make sure you update 2011-10-26T00:18:46 i will try once again 2011-10-26T00:18:49 basically, it wasn't putting all the classes in the jar file 2011-10-26T00:18:59 the jar file is just a zip file, you can unzip it to check 2011-10-26T00:19:06 McLeopold: have you seen the new VERSION file? 2011-10-26T00:19:08 but what is the command in the .sh being feed to playgame.py? 2011-10-26T00:19:15 Is there any programmatic way to submit? 2011-10-26T00:19:26 amstan: no? 2011-10-26T00:19:28 bmh: you're welcome to make one 2011-10-26T00:19:33 McLeopold: download a starter package 2011-10-26T00:19:41 amstan: too busy making a bot 2011-10-26T00:19:57 bmh: it's probably just a wget 2011-10-26T00:20:11 amstan: it would have to get a login session first though 2011-10-26T00:20:20 2 wgets 2011-10-26T00:20:20 i tried the new version 2011-10-26T00:20:24 yeah -- wget posting the uname password and file 2011-10-26T00:20:26 version file is 2011-10-26T00:20:49 Really fixed java Makefile. Added jar to be a proper target in the makefile 2011-10-26T00:21:32 nishant: java -jar MyBot.jar gives you this? 2011-10-26T00:22:00 no 2011-10-26T00:22:06 can you try that? 2011-10-26T00:22:12 amstan: it just hangs 2011-10-26T00:22:22 that seems about right 2011-10-26T00:22:39 amstan: I need to find a teammate in Europe and another one in East Asia so we can follow the sun :) 2011-10-26T00:22:44 amstan: the program seems to be running but maybe the script is fawed 2011-10-26T00:23:36 it apparently cant find the class file "AbstractSystemInputParser$SetupToken.class" 2011-10-26T00:23:41 bmh: or you can do polyphasic sleep 2011-10-26T00:23:52 however, when i see the compiled files its there 2011-10-26T00:23:52 *** janzert has quit IRC (Read error: Connection reset by peer) 2011-10-26T00:23:59 amstan: I could also quit my job and work on this full time, but that would be a bad idea 2011-10-26T00:24:11 nishant: hmm, maybe i did break it in the last commit 2011-10-26T00:24:44 *** Relax has quit IRC (Quit: Leaving) 2011-10-26T00:24:45 why does java make up new .class files that are not based on the source files' name? 2011-10-26T00:24:49 amstan: can you please help me in fixing it for now, so i can try to advance 2011-10-26T00:25:00 nishant: makefile, line 13 2011-10-26T00:25:08 nishant: replace $(CLASSES) with *.class 2011-10-26T00:25:16 well apparently they are expectd to be in the code for abstractsysteminputreader 2011-10-26T00:25:41 hold on 2011-10-26T00:26:36 amstan: wonderful 2011-10-26T00:26:41 it did the trick for me 2011-10-26T00:26:47 thanks a lot dude. 2011-10-26T00:26:54 yeah, well, i broke it first 2011-10-26T00:26:56 please also update it to reflect this 2011-10-26T00:27:06 nonetheles...thanks 2011-10-26T00:28:36 antimatroid: re: tracking enemies in unseen territory. yes, that's pretty much what i was thinking... construct a rough probability distribution of where enemy ants might be given where/when you last saw them 2011-10-26T00:28:49 a1k0n: I hate you, that was my plan :) 2011-10-26T00:28:54 a1k0n: want to join my team…? 2011-10-26T00:28:55 heh. 2011-10-26T00:29:05 well ,this contest is mostly for individuals 2011-10-26T00:29:05 aichallenge: Alexandru Stan epsilon * r84a68ce / ants/dist/starter_bots/java/Makefile : java likes to make different class files - http://git.io/wEKgYg 2011-10-26T00:29:21 That's what you would say, you're an individual. 2011-10-26T00:29:28 a1k0n: yeah but i said perhaps you could just skip the probability aspect and track information about the most recent turn you saw an enemy ant that could reside in a square outside view 2011-10-26T00:29:33 a1k0n: i would do it, he has cool ideas 2011-10-26T00:29:36 the problem of profitting from that information decision-wise is still pretty open 2011-10-26T00:30:04 amstan: Of all people, I think a1k0n needs my ideas the least. 2011-10-26T00:30:04 antimatroid: i would argue that there's an implicit probability distribution in such an assumption which you can basically just ignore 2011-10-26T00:30:06 no probabilities involved there, just work out how to correctly propogate that info or whatever 2011-10-26T00:30:27 yeah, with probabilities i imagine it'd put more weight on squares with more access to them? 2011-10-26T00:30:46 whereas this would be more the ant has equal probability of being in any square reachable 2011-10-26T00:30:53 i guess. you could say there's a sort of diffusion coming from all anthills among the "routes" on the map 2011-10-26T00:31:24 *** Saulzar has quit IRC (*.net *.split) 2011-10-26T00:31:24 *** Israfel has quit IRC (*.net *.split) 2011-10-26T00:31:37 anyway a similar argument can be made for food, except it's uniformly distributed plus symmetry 2011-10-26T00:31:46 which can be used to motivate exploration 2011-10-26T00:31:46 Given that we have limited computational resources, I'm wondering if it's smarter to just devote the lion's share of the resource to foraging 2011-10-26T00:31:49 *** janzert has joined #aichallenge 2011-10-26T00:31:57 a1k0n: get out of my brain, kthx 2011-10-26T00:32:00 there's definitly a tradeoff. 2011-10-26T00:32:02 * bmh works on his symmetry detector 2011-10-26T00:32:08 food isn't as much of an issue, i'm more interested in hunting down enemy ants that don't want to be found 2011-10-26T00:32:09 heheh 2011-10-26T00:32:16 i don't even think it'd be that useful strategy wise, just awesome 2011-10-26T00:32:23 *** nishant has left #aichallenge ("Ex-Chat") 2011-10-26T00:32:25 i kinda want to make an FFT-based symmetry detector just to prove it can be done slash show off 2011-10-26T00:32:28 janzert: want to decide on the HeadTail issue? 2011-10-26T00:32:32 bmh: is that what you meant by unfair strategy? detecting symmetry is fine 2011-10-26T00:32:49 i'm a big fan of if it's possible then it's fair game, but i have done what i can to hide the symmetry from you :P 2011-10-26T00:33:04 too bad the code to produce maps is open 2011-10-26T00:33:11 why is that too bad? 2011-10-26T00:33:16 otherwise i couldn't compete 2011-10-26T00:33:20 a1k0n: Whatever, they can just swap out the map code in the end 2011-10-26T00:33:21 well, it makes it much easier to reverse-engineer 2011-10-26T00:33:26 but anyway yeah, i was gonna say 2011-10-26T00:33:32 you could just write a new mapgen for the tournament 2011-10-26T00:33:45 a1k0n: you can't do new symmetry times for the tournament though 2011-10-26T00:33:48 *** Israfel has joined #aichallenge 2011-10-26T00:33:51 *** Saulzar has joined #aichallenge 2011-10-26T00:33:53 If the bots could go to the wire, you could just has a cluster brute forcing maps until it finds the one you're on :) 2011-10-26T00:33:54 a1k0n: we pretty much implemented all of them 2011-10-26T00:34:03 types* 2011-10-26T00:34:04 amstan: true. 2011-10-26T00:34:30 plus, the food spawning is random on maps with more than on symmetry 2011-10-26T00:34:42 McLeopold: 'eh? 2011-10-26T00:35:14 parse error 2011-10-26T00:35:19 the tutorial map has 2 valid symmetries 2011-10-26T00:35:20 bmh: how are you planning to detect symmetry anyway? 2011-10-26T00:35:31 antimatroid: spatial hashing 2011-10-26T00:36:22 aichallenge: Scott Hamilton epsilon * r9029404 / ants/dist/starter_bots/javascript/Ants.js : 2011-10-26T00:36:22 aichallenge: Merge pull request #332 from teuneboon/epsilon 2011-10-26T00:36:22 aichallenge: Added check if tile is visible to the javascript starter bot - http://git.io/ov_hnQ 2011-10-26T00:36:27 you can work out possible player numbers from the dimensions of the grid, and finding another ant hill gives you a divisor of the number of players 2011-10-26T00:36:34 bmh: i can't fathom how that would apply 2011-10-26T00:36:43 try to find similar substructures on the map? 2011-10-26T00:36:59 it speeds up pathfinding 2011-10-26T00:37:14 knowing symmetry? 2011-10-26T00:37:18 *** delt0r_ has quit IRC (Ping timeout: 244 seconds) 2011-10-26T00:37:19 a1k0n: Exactly, find substructures 2011-10-26T00:37:22 i would love to see someone write code for that 2011-10-26T00:37:43 antimatroid: working on it 2011-10-26T00:37:50 i think the best approach is autocorrelation 2011-10-26T00:37:50 McLeopold: before deciding head/tail issue as presented I'm working on figuring out why we have 106 submissions who's errors already overflowed the current database field 2011-10-26T00:38:06 unicode? 2011-10-26T00:38:08 I think it has to do with unicode ascii conversions 2011-10-26T00:38:10 yep 2011-10-26T00:38:19 do we care? 2011-10-26T00:38:36 it breaks the submission error display 2011-10-26T00:38:59 janzert: I'm going to push the color changes, unless you object 2011-10-26T00:39:00 since the recorded data is no longer valid json 2011-10-26T00:39:22 isn't there like a mysql feature that can suggest optimizing a table based on its contest? 2011-10-26T00:39:31 with what's on ants paste now? 2011-10-26T00:39:33 oh, why the heck did I store a json x) 2011-10-26T00:39:36 yes 2011-10-26T00:39:40 how about getting a bigger data type for now then optimizing later? 2011-10-26T00:40:11 I'd prefer going to black text on the bar graphs but what's there would be ok I suppose 2011-10-26T00:40:17 a1k0n: That certainly looks more efficient 2011-10-26T00:40:26 amstan: the problem is we also allow for bot stderr, so I don't want to fill up the database with chatty bots 2011-10-26T00:40:39 janzert: it is black text 2011-10-26T00:40:44 bmh: it is if you take advantage of the fact that oyu can do it via FFTs, otherwise it's expensive 2011-10-26T00:40:57 a1k0n: i can't believe you didn't finish uni :P 2011-10-26T00:40:59 amstan: we should only be storing 1024 bytes of output now and it's overflowing a 4096 length field 2011-10-26T00:41:13 wtf? 2011-10-26T00:41:17 McLeopold: ahh, needed a browser refresh 2011-10-26T00:41:19 so like, water = 1, ground = -1, unexplored = 0, do a 2D FFT, multiply itself by its conjugate, find largest magnitudes 2011-10-26T00:41:21 amstan: right 2011-10-26T00:41:37 largest one should be at the translation offset 2011-10-26T00:41:41 McLeopold: looks good to me 2011-10-26T00:41:54 a1k0n: what if it's a mirror symmetry, or translational? 2011-10-26T00:42:02 err, rotational 2011-10-26T00:42:06 antimatroid: i'm a horrible student of subjects i don't want to take at the moment 2011-10-26T00:42:22 amstan: then you have to do something else with the boundary conditions; i haven't worked it out yet 2011-10-26T00:42:27 amstan: but it finds translational ones out of the box 2011-10-26T00:42:42 amstan: actually you'd do convolution instead of autocorrelation 2011-10-26T00:43:01 a1k0n: What should I put in my milk so I can grow up big and strong like you? 2011-10-26T00:43:02 and i managed to not get either of those terms, lol 2011-10-26T00:43:19 a1k0n: i'm in the same boat, i've almost failed units from a lack of effort :P 2011-10-26T00:43:40 bmh: haha. i was vegan for a while 2011-10-26T00:44:06 a1k0n: pescatarian here. 2011-10-26T00:44:07 bmh: but am too lazy to pass up things with milk and cheese 2011-10-26T00:44:10 aichallenge: McLeopold epsilon * rdbf52fc / ants/visualizer/js/CanvasElement.js : make bar text better - http://git.io/AsztSg 2011-10-26T00:44:12 aichallenge: McLeopold epsilon * r6cc4ba8 / (4 files in 3 dirs): Merge branch 'epsilon' of http://github.com/aichallenge/aichallenge into epsilon - http://git.io/O9SDSg 2011-10-26T00:44:23 i would go vegetarian but i think i'd run out of food i am willing to eat 2011-10-26T00:44:30 it's pretty easy. 2011-10-26T00:44:33 antimatroid: try Ethiopian. 2011-10-26T00:44:39 bmh: :P 2011-10-26T00:44:39 especially if there's mexican food on every corner where you live 2011-10-26T00:44:42 time to piss everyone off with another color change :) 2011-10-26T00:44:49 a1k0n: Do you think this symmetry detector is actually worthwhile? 2011-10-26T00:44:54 a1k0n: are you in the Mission? 2011-10-26T00:44:56 bmh: yes 2011-10-26T00:44:56 colour chane \o/ 2011-10-26T00:45:02 detecting symmetry gives you a lot of infor 2011-10-26T00:45:16 antimatroid: like where their hill is. 2011-10-26T00:45:23 and where water is 2011-10-26T00:45:35 and where food might be 2011-10-26T00:45:47 food is whereever not water is? 2011-10-26T00:45:49 bmh: heh, no, i live in sunnyvale 2011-10-26T00:45:53 McLeopold: is there a reason not to make the captured length configurable? 2011-10-26T00:45:57 food is added to the map symmetrically from ant hills 2011-10-26T00:46:11 a1k0n: as in sunnyvale goon? 2011-10-26T00:46:14 so it's a bit of an exaggeration 2011-10-26T00:46:14 a1k0n: Palo Alto here 2011-10-26T00:46:16 having a longer capture for compile errors than game errors might be nice in the future 2011-10-26T00:46:16 aka cheap wine? 2011-10-26T00:46:24 antimatroid: oh. That is a lot of info 2011-10-26T00:46:26 janzert: because we also capture bot stderr 2011-10-26T00:46:27 antimatroid: as in what? 2011-10-26T00:46:36 antimatroid: as in california 2011-10-26T00:46:50 yes? but that is the reason to make it configurable 2011-10-26T00:46:51 janzert: uh, if you can tell the difference, sure 2011-10-26T00:46:56 a1k0n: http://www.homeliquordelivery.com.au/images/sunnyvale%20fruity%20lexia.JPG 2011-10-26T00:47:01 ok 2011-10-26T00:47:07 that might be australian though 2011-10-26T00:47:19 antimatroid: what... is that? 2011-10-26T00:47:19 We can up the database and still not increase the limit of HeadTail 2011-10-26T00:47:33 a1k0n: it's a $10 cask of wine 2011-10-26T00:47:36 what I'm thinking is we would capture more when running the compiler, less when running the bot in a game 2011-10-26T00:47:49 Actually, those are in different database fields 2011-10-26T00:47:52 bmh: anyway, symmetry detector is not worthwhile until you have a basic probabilistic based bot.. which i don't have yet 2011-10-26T00:47:54 right 2011-10-26T00:48:16 a1k0n: what about water info and hill info? 2011-10-26T00:48:23 and number of players 2011-10-26T00:48:48 but let's not change any of the actual capture lengths yet, until I/we get the current issue with too long of captures sorted out 2011-10-26T00:48:49 antimatroid: well, once you're certain about the symmetry, then yeah, that would help a ton 2011-10-26T00:48:54 you can immediately start zerg rushing the other anthills 2011-10-26T00:49:11 have people done much on defending their hill strategies? 2011-10-26T00:49:14 a1k0n: sure you don't want to join team skynet? ;) 2011-10-26T00:49:22 janzert: did you already up it or did I/ 2011-10-26T00:49:27 we use haskell 2011-10-26T00:49:27 like making sure there's a permiter around it for which you know there's no enemies in there 2011-10-26T00:49:31 I'm having trouble even getting started :D 2011-10-26T00:49:32 I haven't touched it 2011-10-26T00:49:45 what is it actually? 2011-10-26T00:49:49 hmm 2011-10-26T00:49:50 game errors are at 1024, submission errors at 4096 2011-10-26T00:49:50 bmh: you're skynet :P 2011-10-26T00:49:56 you should be sky captain 2011-10-26T00:50:00 from lapd future cpp 2011-10-26T00:50:02 were the rules modified since the last contest? it used to explicitly forbid teams 2011-10-26T00:50:08 no it didn't 2011-10-26T00:50:12 teams were always allowed 2011-10-26T00:50:14 *** delt0r_ has joined #aichallenge 2011-10-26T00:50:16 ahh, I've only looked at submission error length 2011-10-26T00:50:16 not in tron 2011-10-26T00:50:17 a friend and i did our tron bot together 2011-10-26T00:50:20 it's been 4096 2011-10-26T00:50:21 really?! 2011-10-26T00:50:23 pairs were allowed in tron 2011-10-26T00:50:27 ah. 2011-10-26T00:50:30 they just said say so on your profile 2011-10-26T00:50:46 and now you can actually edit your profile 2011-10-26T00:50:50 haha yeah 2011-10-26T00:51:21 i can't wait to do tron again 2011-10-26T00:51:28 heh 2011-10-26T00:51:45 with ant maps... 2011-10-26T00:51:50 lol 2011-10-26T00:52:00 yes, i wanted bigger maps for tron anyway 2011-10-26T00:52:02 they were way too small 2011-10-26T00:52:18 bmh: it might be best, since i don't have a lot of time for the contest this time. i spend most of my time these days keeping my 3 year old from waking up the infant and vice versa 2011-10-26T00:52:44 a1k0n: Do you like Haskell? 2011-10-26T00:52:52 bmh: oh i thought you were skyramp? 2011-10-26T00:52:55 i do, but i don't know if i'd use it for this 2011-10-26T00:52:57 how many sky people are there! 2011-10-26T00:53:02 skynet 2011-10-26T00:53:10 yeah i know 2011-10-26T00:53:10 antimatroid: Like the guy who killed John Connor 2011-10-26T00:53:15 which was a starter bot last i checked, heh 2011-10-26T00:53:20 there is a skyramp on the leaderboard 2011-10-26T00:53:33 a1k0n: That is correct. It's been a busy two weeks. I recently moved London -> CA and started a new job 2011-10-26T00:53:45 bmh: what kind of work do you do? 2011-10-26T00:53:58 antimatroid: I'm a software engineer 2011-10-26T00:54:05 that's not very descriptive :P 2011-10-26T00:54:11 what kind of software do you make 2011-10-26T00:54:14 bmh: aren't you a googler? 2011-10-26T00:54:25 amstan: yeah. 2011-10-26T00:54:26 McLeopold: heh, and I just now realize that the compiler doesn't use HeadTail at all? 2011-10-26T00:54:26 bmh: heh indeed, i didn't know you were transplanted from london 2011-10-26T00:54:27 so many googlers! 2011-10-26T00:54:33 he's a noogler 2011-10-26T00:54:37 a1k0n: Boston -> London -> CA 2011-10-26T00:54:38 (so am i really) 2011-10-26T00:54:41 hmm, Did I break the viewer again? 2011-10-26T00:54:43 i'm not :( 2011-10-26T00:54:50 and sigh, zannick, anyone else? 2011-10-26T00:54:51 rejected my internship application. :P 2011-10-26T00:54:52 lol 2011-10-26T00:55:01 bobbydroptable: join the club 2011-10-26T00:55:09 McLeopold: seems to be updated and working here 2011-10-26T00:55:20 i didn't get the summer coding thing :( 2011-10-26T00:55:24 try when a user is highlighted 2011-10-26T00:55:29 summer of code that's it 2011-10-26T00:55:34 antimatroid: and i'm part of that club too 2011-10-26T00:55:38 i'm such a failure, lol 2011-10-26T00:55:44 yeh tell me about it *sigh* 2011-10-26T00:55:56 oops, yep. that's broken 2011-10-26T00:56:03 can't even figure out this ant shiz 2011-10-26T00:56:05 lol 2011-10-26T00:56:20 I haven't even gathered strength to apply to Google yet, haha 2011-10-26T00:57:17 one of hte maths phd students at my uni is interning in sydney this summer 2011-10-26T00:57:30 all you have to do is win this contest, write a blog post, and you'll get a referral 2011-10-26T00:57:42 oh yeh... win 2011-10-26T00:57:43 lol 2011-10-26T00:57:51 aichallenge: Janzert epsilon * r5d747eb / worker/engine.py : 2011-10-26T00:57:51 aichallenge: Merge pull request #325 from AndrewBC/epsilon 2011-10-26T00:57:51 aichallenge: Make bot output capture length easily configurable. - http://git.io/9zzS_Q 2011-10-26T00:57:54 ya, working on that. lol 2011-10-26T00:57:55 a1k0n: oh man, i should have wrote that blog post 2011-10-26T00:57:56 anecdotally it works 2011-10-26T00:57:58 *** cyphase_ has joined #aichallenge 2011-10-26T00:58:02 seriously! 2011-10-26T00:58:26 a1k0n: so winning tron helped land the job then? 2011-10-26T00:58:27 i wonder where mega1 is these days 2011-10-26T00:58:30 indeed 2011-10-26T00:58:39 *** cyphase has quit IRC (Ping timeout: 252 seconds) 2011-10-26T00:58:50 where he is job wise or ? 2011-10-26T00:59:03 *** cyphase_ has quit IRC (Read error: Connection reset by peer) 2011-10-26T00:59:04 i think he left finance and was just doing consulting for lisp 2011-10-26T00:59:06 whether he's doing this contest 2011-10-26T00:59:07 but not sure now 2011-10-26T00:59:21 *** Palmik has joined #aichallenge 2011-10-26T00:59:22 i doubt he'll do this contest, i don't think he had time for the last one, but i could be wrong 2011-10-26T00:59:23 *** cyphase has joined #aichallenge 2011-10-26T00:59:47 there's some big names from last contest around already though 2011-10-26T00:59:58 flag, greentea, i think i saw bix0r 2011-10-26T01:00:11 they're regulars huh 2011-10-26T01:00:19 and accoun is working on a bot i'm pretty sure 2011-10-26T01:00:54 they all wrote their bots during the beta and are resting on their lauerls 2011-10-26T01:01:13 yes, my point is they were all top 15 in planet wars i think 2011-10-26T01:01:21 so i expect decent bots from them if they continue 2011-10-26T01:01:28 mostly top 10 2011-10-26T01:01:51 aichallenge: McLeopold epsilon * rd087dce / ants/visualizer/js/Replay.js : fix highlight player - http://git.io/mri_Uw 2011-10-26T01:01:54 oh yeah, that's true 2011-10-26T01:02:14 so accoun made the new tcp server? 2011-10-26T01:02:23 or he's jsut hosting one? 2011-10-26T01:02:29 hosting 2011-10-26T01:02:47 He's getting practice at being an admin :) 2011-10-26T01:02:53 haha 2011-10-26T01:03:06 oh god, i'm still laughing 2011-10-26T01:03:14 we should complain when he changes his map 2011-10-26T01:03:31 call him a chitting russian 2011-10-26T01:03:49 great, now I'm highlighting the wrong player :( 2011-10-26T01:07:29 McLeopold: from what I can see we don't cut off compiler errors at all right now? 2011-10-26T01:07:34 or am I missing it somewhere 2011-10-26T01:07:51 if so I'm surprised more haven't overflowed the database field 2011-10-26T01:08:11 janzert: sry, can't remember 2011-10-26T01:08:56 ok, I'll add something to limit them 2011-10-26T01:09:37 Nice color change 2011-10-26T01:09:38 *** ChrisH_ has joined #aichallenge 2011-10-26T01:12:13 McLeopold: ping 2011-10-26T01:13:36 *** cyphase has quit IRC (Ping timeout: 240 seconds) 2011-10-26T01:13:51 @later tell McLeopold Those colors look fine to me. I can tell them all apart. 2011-10-26T01:13:51 ChrisH_: Ready to serve, my lord. 2011-10-26T01:15:04 lol, http://aichallenge.org/visualizer.php?game=14578&user=46 <-- wtf happened here? timeout protection kick in w/o scheduling any ants? 2011-10-26T01:16:09 *** Fixpoint has quit IRC (Ping timeout: 265 seconds) 2011-10-26T01:16:17 aichallenge: McLeopold epsilon * r369f6b8 / ants/visualizer/js/Replay.js : fix highlight player correctly - http://git.io/KVhA6Q 2011-10-26T01:16:21 hi 2011-10-26T01:16:38 woot for the orange ant: http://aichallenge.org/visualizer.php?game=14313 2011-10-26T01:18:06 lol sure sucks to be him 2011-10-26T01:18:39 he's a survivor 2011-10-26T01:18:44 did pretty well though, ranked #2 2011-10-26T01:18:46 he also destroys an anthill 2011-10-26T01:18:49 urgh seriously youtube? a 5 year old video and you want to impose copyright on the music? jeez ... 2011-10-26T01:19:06 blue got amazing combat tactic, and still ranked lower 2011-10-26T01:19:10 *** keithroe has joined #aichallenge 2011-10-26T01:20:10 well you gotta cap hills to win it 2011-10-26T01:21:19 *** keithroe has quit IRC (Client Quit) 2011-10-26T01:21:42 are we doing the log ant count on the graphs yet? 2011-10-26T01:21:47 sqrt 2011-10-26T01:22:37 aichallenge: McLeopold epsilon * r07041b5 / ants/visualizer/js/Replay.js : fix highlight player correctly, for reals - http://git.io/Mir_BA 2011-10-26T01:23:00 ok, that guy needs a medal for that game 2011-10-26T01:24:38 arrg! 2011-10-26T01:25:04 lol 2011-10-26T01:25:13 funny to watch 2011-10-26T01:25:36 lol that looks good 2011-10-26T01:25:38 I'm really, really, really sry 2011-10-26T01:27:01 aichallenge: McLeopold epsilon * re40e941 / ants/visualizer/js/Replay.js : X( - http://git.io/OSLaBA 2011-10-26T01:27:07 McLeopold: yeah, don't worry about it 2011-10-26T01:27:48 the ants are very electric 2011-10-26T01:27:55 I'm actually rotflmao over here 2011-10-26T01:27:56 McLeopold: i can do those "really fix" commits that break stuff too. 2011-10-26T01:28:24 have we considered decreasing the speed of games a bit? it's pretty hard to see what's going on when it moves so fast 2011-10-26T01:28:30 and we can't make it go any slower 2011-10-26T01:28:40 problem is that at this point you can't push -f stuff you ammended anymore since there's too many people watching 2011-10-26T01:28:44 wow, still broken :( 2011-10-26T01:28:52 antimatroid: the speed depends on the length of the game 2011-10-26T01:29:06 janzert: sure, but the current length and speed is quite fast 2011-10-26T01:29:16 even for short games? 2011-10-26T01:29:20 or just for long ones? 2011-10-26T01:29:26 just for longer i guess 2011-10-26T01:29:37 McLeopold: one more equal? 2011-10-26T01:29:37 wouldn't it be better to have a constant speed per turn as default? 2011-10-26T01:29:45 rather than constant time for the game to be visualised in 2011-10-26T01:29:48 antimatroid: +1 for that 2011-10-26T01:30:06 I think it's nice to try and go for constant time until that constant time gets too fast 2011-10-26T01:30:24 did planet wars do constant time? 2011-10-26T01:30:29 probably 2011-10-26T01:30:41 don't really remember 2011-10-26T01:30:44 amstan: actually, I don't think that's it 2011-10-26T01:30:46 at the very least it would be nice if we could slow the game down from the default 2011-10-26T01:30:50 and not just speed it up 2011-10-26T01:30:54 McLeopold: mostly kidding 2011-10-26T01:30:55 antimatroid: you can 2011-10-26T01:31:13 http://paste.aichallenge.org/0G5ci/?user=871 2011-10-26T01:31:14 you can slow it way down 2011-10-26T01:31:31 the second from bottom button on the right, with the snail in black 2011-10-26T01:31:36 last time i checked i don't think i could and i can't check atm :P 2011-10-26T01:31:50 yeah last time i tried it wouldn't let you slow it down any further than the default starting length 2011-10-26T01:31:51 I have it playing at about 1 move per second at the moment 2011-10-26T01:31:56 only speed it up then slow it back down to that 2011-10-26T01:32:01 hmm okay 2011-10-26T01:32:05 i'm fine then 2011-10-26T01:32:33 yeah, if you couldn't slow beyond the default speed it would be bad 2011-10-26T01:33:29 *** Areks has joined #aichallenge 2011-10-26T01:34:00 *** pedrosorio has joined #aichallenge 2011-10-26T01:35:27 *** the-mgt has quit IRC (Ping timeout: 260 seconds) 2011-10-26T01:35:38 *** pedrosorio has quit IRC (Client Quit) 2011-10-26T01:35:38 *** bmh has quit IRC (Quit: bmh) 2011-10-26T01:36:02 *** ashoka has joined #aichallenge 2011-10-26T01:37:26 *** Parasprites has joined #aichallenge 2011-10-26T01:39:37 *** ig has quit IRC (Ping timeout: 256 seconds) 2011-10-26T01:39:45 *** the-mgt has joined #aichallenge 2011-10-26T01:40:18 *** Parasprites has quit IRC (Client Quit) 2011-10-26T01:40:45 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-26T01:41:17 *** kfed1 has joined #aichallenge 2011-10-26T01:43:59 *** kfed1 has quit IRC (Client Quit) 2011-10-26T01:47:28 *** he_the_great has quit IRC (Ping timeout: 245 seconds) 2011-10-26T01:50:52 *** djr__ has quit IRC (Read error: Connection reset by peer) 2011-10-26T01:51:06 aichallenge: McLeopold epsilon * r6d66368 / ants/visualizer/js/Replay.js : fix visualizer sorted rank colors - http://git.io/k0atwQ 2011-10-26T01:52:55 *** djr_ has joined #aichallenge 2011-10-26T01:52:59 http://aichallenge.org/visualizer.php?game=14850 2011-10-26T01:53:02 McLeopold : visualizer is breaking in my Firefox and Opera. (both on Ubuntu) complaining against Application.js (Not enough arguments [nsIXMLHttpRequest.send]) 2011-10-26T01:53:13 ashoka: yep, that was me, just fixed it 2011-10-26T01:53:23 um... 2011-10-26T01:53:32 McLeopold: what did you do? 2011-10-26T01:53:56 what should i look for? 2011-10-26T01:54:17 ashoka: can you try again? 2011-10-26T01:55:08 *** ig has joined #aichallenge 2011-10-26T01:55:11 sure 2011-10-26T01:55:55 opera says this : Statement on line 166: Undefined variable: JSON Backtrace: Line 166 of linked script http://aichallenge.org/visualizer/js/Replay.js: In function Replay replay = JSON.parse(replay); Line 506 of linked script http://aichallenge.org/visualizer/js/Application.js vis.state.replay = new Replay(vis.replayStr, debug, user); Line 252 of linked script http://aichallenge.org/visualizer/js/Application.js func(); 2011-10-26T01:56:02 McLeopold: the visualizer changes look good, except one thing. 2011-10-26T01:56:37 while Firefox give : QueryInterface function QueryInterface() { [native code] } message Not enough arguments [nsIXMLHttpRequest.send] result 2153185281 filename http://aichallenge.org/visualizer/js/Application.js lineNumber 418 columnNumber 0 inner null data null initialize function initialize() { [native code] } 2011-10-26T01:57:25 McLeopold : my browser versions in this machine are bit older though. 2011-10-26T01:58:15 McLeopold: When watching a game for the "focus" player, and that player wins the match so that their score and ant count bars are on the far left, the white bar color blends into the white background and the score ant ant count numbers seem to be floating because there is nothing to separate them from the "scores" and "# of ants" labels 2011-10-26T01:58:56 McLeopold: A vertical black bar to separate the label from the numbers might help. 2011-10-26T01:59:02 ChrisH_: really? I guess I shouldn't have used red and green 2011-10-26T01:59:03 instead of pure white a light gray would probably fix it as well 2011-10-26T01:59:22 janzert: that's another good idea 2011-10-26T01:59:26 it'll still look whit against the brown 2011-10-26T01:59:28 mcleopold : ancient versions infact. firefox : 3.0.3 and Opera 9.64. may be I should just upgrade them :-) 2011-10-26T01:59:30 white even 2011-10-26T01:59:33 can anyone else get ashoka's errors? 2011-10-26T02:00:04 try FF 3.6 on ubuntu now 2011-10-26T02:00:31 can't get a reproduction here 2011-10-26T02:00:33 *** Vivitron has quit IRC (Remote host closed the connection) 2011-10-26T02:00:33 I didn't touch that code tonight, so I don't know, other than upgrade 2011-10-26T02:00:57 *** Vivitron has joined #aichallenge 2011-10-26T02:01:51 McLeopold : May be these were there for a long time and no one else is using such old browser versions :-) 2011-10-26T02:01:53 woot, just passed 15k games played 2011-10-26T02:02:57 ChrisH_: http://paste.aichallenge.org/0G5ci/?user=871 2011-10-26T02:03:15 nice! 2011-10-26T02:03:45 McLeopold: that looks good to me 2011-10-26T02:04:19 *** analyst74 has quit IRC (Read error: Connection reset by peer) 2011-10-26T02:05:03 *** analyst74 has joined #aichallenge 2011-10-26T02:06:18 @seen jmcarthur 2011-10-26T02:06:18 cirno_the_greate: jmcarthur was last seen in #aichallenge 18 hours, 33 minutes, and 41 seconds ago: @later tell amstan sorry about lambdabot. i was demonstrating a couple things and intended to kick her out when we were done, but forgot 2011-10-26T02:07:57 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 252 seconds) 2011-10-26T02:08:13 aichallenge: McLeopold epsilon * rf579c01 / ants/visualizer/js/Const.js : update highlighted player color - http://git.io/vFD9Ag 2011-10-26T02:08:13 aichallenge: McLeopold epsilon * r839f75b / ants/visualizer/js/Const.js : update highlighted player color - http://git.io/a8BiCg 2011-10-26T02:09:53 btw, if the "rainbow" of colors appears out of order, then somebody beat a higher ranked bot 2011-10-26T02:10:41 latest update live yet? 2011-10-26T02:10:45 yes 2011-10-26T02:11:04 hmmm... 2011-10-26T02:11:05 hmm, can't get my browser to refresh for grey highlight 2011-10-26T02:11:06 maybe not 2011-10-26T02:11:13 ah 2011-10-26T02:11:27 okay, forgot to git pull 2011-10-26T02:12:05 there it goes 2011-10-26T02:12:34 wtf.. ants.getIlk(loc).isPassable() is returning true for water tiles.. wtf!!! 2011-10-26T02:13:37 OMG.. found it. I overriding addWater but forgot to call the super.addWater 2011-10-26T02:13:58 grom358: have you tried lolcode? 2011-10-26T02:15:00 haha.. no 2011-10-26T02:17:47 *** porco has quit IRC (Remote host closed the connection) 2011-10-26T02:19:04 McLeopold: Having the almost white as the focused player color is SOOO much better than the almost black we had the last couple days. I'm really liking the color tweaks from today. 2011-10-26T02:19:27 yeah, that's what I wanted originally, but the graph didn't work with it 2011-10-26T02:19:35 I remember. 2011-10-26T02:19:49 Glad you found a solution. 2011-10-26T02:20:44 @rankings 2011-10-26T02:20:44 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-10-26T02:20:46 McLeopold: Top 10 players: GarySWest(80.7), MomoBot(80.2), sir_macelon(79.8), paratrechina(78.9), GreenTea(78.8), olex.s(78.1), Ender(77.7), ThisIsNotABug(77.4), xRaider(77.4), r-bit(76.3) 2011-10-26T02:21:00 woot, python on top! 2011-10-26T02:21:04 grrr, I'm just off the top ten right now. 2011-10-26T02:21:07 #11 2011-10-26T02:22:05 i don't even rank in the first page anymore :( 2011-10-26T02:22:25 ouch, you just dropped 3 2011-10-26T02:23:03 ummm, rankings page is blank for me right now 2011-10-26T02:23:18 well i'm 126 lol 2011-10-26T02:23:21 ChrisH_: you were all too crap :P 2011-10-26T02:23:38 huh yeh rankings page is broken 2011-10-26T02:23:49 stop breaking things mcleopold :D lol 2011-10-26T02:23:50 or rather, it has all the borders and column headers, but now player entries 2011-10-26T02:23:52 i'd be ranked higher if my bot played more games :-/ 2011-10-26T02:24:37 that was weird 2011-10-26T02:24:39 antimatroid: sorry, I'm not sure what you meant 2011-10-26T02:24:53 ChrisH_: ignore me :) 2011-10-26T02:24:57 *** vk has joined #aichallenge 2011-10-26T02:25:02 antimatroid: if you insist. ;) 2011-10-26T02:25:38 McLeopold: Did you get the pastebin I sent you a couple days ago with the page source with the bad navigation links? 2011-10-26T02:25:47 yes 2011-10-26T02:25:54 did it help? 2011-10-26T02:25:57 no 2011-10-26T02:27:00 bysin: I would suggest taking off your email address, or changing it to bysin at bla dot com 2011-10-26T02:27:19 its kinda my throw away email 2011-10-26T02:27:28 ok 2011-10-26T02:27:28 i'm not too bothered if i get spam 2011-10-26T02:27:32 thanks though 2011-10-26T02:28:00 did you see my swarm replay? 2011-10-26T02:28:50 i like the spread of your ants 2011-10-26T02:28:58 i'm trying to get the same thing ==" 2011-10-26T02:29:05 how did you do it if you don't mind me asking 2011-10-26T02:29:05 lol 2011-10-26T02:29:24 :-X 2011-10-26T02:29:37 hahaha fine ==" 2011-10-26T02:29:38 jeaous 2011-10-26T02:29:39 jealous 2011-10-26T02:29:40 :( 2011-10-26T02:32:58 its cpu intensive, which is why i chose C 2011-10-26T02:33:07 *** cybsy has joined #aichallenge 2011-10-26T02:33:43 *** nancy has joined #aichallenge 2011-10-26T02:34:05 hi 2011-10-26T02:34:09 *** nancy is now known as Guest29987 2011-10-26T02:34:35 does anyone get a threading exception when running the tools 2011-10-26T02:40:17 *** Guest29987 has quit IRC (Quit: Page closed) 2011-10-26T02:40:42 *** Puj has joined #aichallenge 2011-10-26T02:43:12 *** Larose has quit IRC (Ping timeout: 240 seconds) 2011-10-26T02:52:41 FYI: If anyone is using python 2.7- on cygwin and is having trouble with wierd unpredictable timeouts in deterministic code (or even "pass" in do_turn), don't run playgame through mintty, just use Eclipse and the program arguments in run configurations. 2011-10-26T02:53:49 *** theonewolf has quit IRC (Ping timeout: 240 seconds) 2011-10-26T02:54:53 *** theonewolf has joined #aichallenge 2011-10-26T02:55:16 *** cybsy has quit IRC (Ping timeout: 252 seconds) 2011-10-26T02:56:06 *** cirno_the_greate has quit IRC (Remote host closed the connection) 2011-10-26T02:58:44 *** xathis has joined #aichallenge 2011-10-26T03:00:50 *** jesionaj has joined #aichallenge 2011-10-26T03:01:34 *** nishant has joined #aichallenge 2011-10-26T03:02:24 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-10-26T03:02:57 getting a java.lang.CastException in java implmtation of code when i get to tutorial sep 2 2011-10-26T03:03:07 What line? 2011-10-26T03:03:10 anyone can help me with this? 2011-10-26T03:05:25 nishant, what line are your receiving the error on? 2011-10-26T03:05:55 line 58 of MyBt.java.... it deas with "Tile" 2011-10-26T03:06:11 What is that line in your code? 2011-10-26T03:07:01 hold on 2011-10-26T03:07:19 *** Nescio has quit IRC (Ping timeout: 265 seconds) 2011-10-26T03:07:21 Puj: TreeSet sortedFood = new TreeSet(ants.getFoodTiles()); 2011-10-26T03:10:51 *** onensora has joined #aichallenge 2011-10-26T03:11:24 *** Bluedgis has joined #aichallenge 2011-10-26T03:11:55 well i'm learning new things today. i might actually progress with the bot today woohoo 2011-10-26T03:12:20 whats your profile id? 2011-10-26T03:14:06 same 2011-10-26T03:14:09 don't even look 2011-10-26T03:14:11 its horrible 2011-10-26T03:14:15 lol 2011-10-26T03:16:17 *** nishant has quit IRC (Quit: Ex-Chat) 2011-10-26T03:17:59 *** grom358 has quit IRC (Quit: Ex-Chat) 2011-10-26T03:20:00 you're not bad at all 2011-10-26T03:20:03 what are you talking about 2011-10-26T03:21:04 its 126! ==" 2011-10-26T03:22:37 *** jesionaj has quit IRC (Quit: Leaving) 2011-10-26T03:22:38 and still too random (it wanders around quite abit figuring out what to do) 2011-10-26T03:23:13 My attack and defence code are done 2011-10-26T03:23:23 now to let it run on the tcp server all night to see how it does 2011-10-26T03:24:01 hehe nice 2011-10-26T03:25:24 *** zyberkiddy has joined #aichallenge 2011-10-26T03:28:55 *** barsandcat has joined #aichallenge 2011-10-26T03:29:56 wow my ant does really bad on low food maps 2011-10-26T03:30:00 Not sure how to do that 2011-10-26T03:30:17 open high food maps, my ant does well though 2011-10-26T03:30:47 Sounds like you may need better defense conditions? f4hy 2011-10-26T03:30:54 probably need to be more evasionary 2011-10-26T03:32:31 *** dmiles_afk has joined #aichallenge 2011-10-26T03:32:46 i dunno i think low food maps are a bit "luck of the draw" 2011-10-26T03:33:39 I'm still getting started. I'm digging around trying to see how the game interfaces with the bot still. 2011-10-26T03:33:59 Puj: ya, my defense normally kicks in after a certain threshold of ants, but on low food maps, I dont get enough to trigger it. gonna have to reconsider 2011-10-26T03:34:14 *** paulwal has joined #aichallenge 2011-10-26T03:34:18 hello 2011-10-26T03:34:49 *** brainfarm has joined #aichallenge 2011-10-26T03:35:22 My bot is moduspwned http://ants.fluxid.pl/replay.6614 you can see here my defenses is great but http://ants.fluxid.pl/replay.6617 it fails horribly 2011-10-26T03:35:38 *** aerique has joined #aichallenge 2011-10-26T03:36:15 f4hy, Why do they favor all going north? 2011-10-26T03:36:21 I guess your defense threshold is around 30 ants f4hy? 2011-10-26T03:36:41 because when you have directions = ['n','s','e','w'] they're going to do that 2011-10-26T03:36:42 lol 2011-10-26T03:36:55 When moving an ant, the documentation says to output "o row col direction". Are row and col the originating tile or the destination tile? To move an ant at 3,18 northward, would you output "o 3 18 N" or "o 3 17 N" ? 2011-10-26T03:37:08 origin tile 2011-10-26T03:37:18 Israfel: Lol, I have "if you dont know what to do yet go north" I have thought about what the default behavoir should be, but doesnt really matter much I dont think 2011-10-26T03:37:20 so 3 18 n 2011-10-26T03:37:23 *** mviel has joined #aichallenge 2011-10-26T03:37:27 which goes to 2 18 2011-10-26T03:37:37 ohhh 2011-10-26T03:37:41 of course 2011-10-26T03:37:54 f4hy, yeah, they're covering ground already covered by other ants. Spread them out. 2011-10-26T03:37:58 f4hy could take advantage of that... put lots of defenses facing south lol 2011-10-26T03:38:05 i was thinking in terms of x,y. As in x first. But.. row then col. Thanks 2011-10-26T03:38:08 f4hy, consider implementing a proportional calculation, distance ants can go from hill in direct relation to # of ants where when you have two ants, the second ant never leaves vision of the hill, then you can release this condition when you reach the threshold you currently have implemented? 2011-10-26T03:38:11 knowing that the bots will never come from the north XD 2011-10-26T03:38:38 Puj: but this reduces the speed at which you gather food, because you're now dependent on drops happening in that region. 2011-10-26T03:38:43 i think 2011-10-26T03:38:48 *** avdg1 has joined #aichallenge 2011-10-26T03:39:11 *** g0llum has joined #aichallenge 2011-10-26T03:39:14 Puj: that is too limiting early game 2011-10-26T03:39:18 the 1st ant can't defend anyway 2011-10-26T03:39:26 i say 2011-10-26T03:39:27 *** levis501 has quit IRC (Quit: levis501) 2011-10-26T03:39:33 track how long it takes to get your first food 2011-10-26T03:39:38 A game should have several phases. the first being fast expansion for resources. 2011-10-26T03:39:39 Puj: On low food maps you have to get as much food as posisble so all ants need to be gathering 2011-10-26T03:39:59 Israfel: obviously, but on low food maps, my ants never transition to phase 2 2011-10-26T03:40:03 Undoubtedly, but on low food maps, it is better to not lose, than to win. 2011-10-26T03:40:07 so I need to come up with better parameterss on when to switch 2011-10-26T03:40:21 f4hy maybe track how long it takes you to grab food between averges 2011-10-26T03:40:37 bobbydroptable, I think you receive the food_spawn_rate as a startup param 2011-10-26T03:40:38 if you know the food is coming in slowly, then kick in your defense 2011-10-26T03:40:45 Puj: I probably should just have something that checks for how much food I have gotten after the first 10 turns and retunes the thresholds 2011-10-26T03:40:50 burny: exactly 2011-10-26T03:40:53 bobbydroptable: 2011-10-26T03:40:59 burny: oopss ssorry 2011-10-26T03:41:18 "loadtime" # in milliseconds, time given for bot to start up after it is given "ready" (see below) "turntime" # in milliseconds, time given to the bot each turn "rows" # number of rows in the map "cols" # number of columns in the map "turns" # maximum number of turns in the game "viewradius2" # view radius squared "attackradius2" # battle radius squared "spawnradius2" # food 2011-10-26T03:41:19 f4hy, phase 2 should occur when they reach a certain area of exploration. If there's no food, then defending and outer perimeter to control territory or location of enemy hills. 2011-10-26T03:41:29 My defense is stupid, but damn powerful so far. 2011-10-26T03:41:35 nothing about food spawn rate here. 2011-10-26T03:41:40 jk 2011-10-26T03:42:00 bobbydroptable: your bot isn't told the rate 2011-10-26T03:42:07 *** xar0l has quit IRC (Ping timeout: 265 seconds) 2011-10-26T03:42:15 *** avdg has quit IRC (Ping timeout: 260 seconds) 2011-10-26T03:42:17 janzert yeh it was a respnse to puj 2011-10-26T03:42:17 That was my bad, I was looking at playgame. 2011-10-26T03:42:19 I've got great strategies, but still getting started on the programming in my spare time. ;/ 2011-10-26T03:42:27 people, throw your stuff up on the tcp sesrver. I want more feedback ;-) 2011-10-26T03:42:38 goodnight all 2011-10-26T03:42:42 mine doesn't even move 2011-10-26T03:42:42 :) 2011-10-26T03:42:43 gl f4hy 2011-10-26T03:42:44 cya 2011-10-26T03:43:12 Israfel: AI is mostly about the Aritificial part. We call have great strategies because we ARE intelegent. The trick is figuring out how to implement stuff. 2011-10-26T03:43:17 *** kaemo has joined #aichallenge 2011-10-26T03:43:26 Well, it's programmable. 2011-10-26T03:43:26 s/call/all/ 2011-10-26T03:44:00 I just hate that my rank resets on the main contest when I submit a new version 2011-10-26T03:44:07 its also getting the agent to have the same awareness of the environment as we do 2011-10-26T03:44:08 I cant show off to my friends being in the top 100 anymore 2011-10-26T03:44:11 f4hy, I break each one down into certain classes and the rest won't require any thought after that. 2011-10-26T03:44:40 gl all, I hope to crush your bots in battle. gnight gnight 2011-10-26T03:44:46 hehe night 2011-10-26T03:50:09 Hi there 2011-10-26T03:50:14 Hello 2011-10-26T03:50:31 *** vk has quit IRC (Quit: Page closed) 2011-10-26T03:50:42 does anybody succeed in running the coffescript or the js starter locally ? 2011-10-26T03:51:21 yes 2011-10-26T03:51:27 my bot works when I upload it but it crashes locally... 2011-10-26T03:51:32 *** adyxax has left #aichallenge ("WeeChat 0.3.5") 2011-10-26T03:51:55 It's usually path issues when running locally. 2011-10-26T03:52:52 *** avdg has joined #aichallenge 2011-10-26T03:52:59 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-10-26T03:53:17 *** WadimX has joined #aichallenge 2011-10-26T03:54:11 mviel wow . usually its the other way around isn't it? 2011-10-26T03:54:20 *** Puj has quit IRC (Ping timeout: 260 seconds) 2011-10-26T03:55:11 *** avdg1 has quit IRC (Ping timeout: 244 seconds) 2011-10-26T03:55:22 *** Egg5 has joined #aichallenge 2011-10-26T03:56:42 *** avdg1 has joined #aichallenge 2011-10-26T03:58:26 aichallenge: janzert epsilon * rc8ebaee / worker/compiler.py : Limit length of errors reported from compiler - http://git.io/x3xLuw 2011-10-26T03:58:26 aichallenge: janzert epsilon * r23d6663 / (ants/visualizer/js/Const.js ants/visualizer/js/Replay.js): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://git.io/Juihpw 2011-10-26T03:58:52 *** xathis has quit IRC (Quit: Page closed) 2011-10-26T03:59:11 help 2011-10-26T03:59:22 *** avdg2 has joined #aichallenge 2011-10-26T03:59:35 *** avdg has quit IRC (Ping timeout: 260 seconds) 2011-10-26T04:01:13 *** avdg1 has quit IRC (Ping timeout: 245 seconds) 2011-10-26T04:02:06 *** ikaros has joined #aichallenge 2011-10-26T04:03:44 Hello! Is there any admins here? 2011-10-26T04:06:39 amstan is here 2011-10-26T04:06:41 i think 2011-10-26T04:10:10 *** jezuz` has quit IRC (Ping timeout: 244 seconds) 2011-10-26T04:10:51 *** jezuz has joined #aichallenge 2011-10-26T04:11:34 wadimX what's up 2011-10-26T04:13:24 *** xathis has joined #aichallenge 2011-10-26T04:13:24 I've registered, but didn't recieve email. 2011-10-26T04:15:16 oh. sorry can't help then :) 2011-10-26T04:19:41 *** thestinger has quit IRC (Quit: WeeChat 0.3.6) 2011-10-26T04:19:59 i have to comment this line :"process.stdout.flush()" to make the coffee stater lacally 2011-10-26T04:20:18 if i didn't i get that: "TypeError: Object # has no method 'flush'" 2011-10-26T04:20:29 WadimX: Check your spam folder? We haven't gotten any error message back that the message couldn't be delivered. 2011-10-26T04:21:00 The subject should be "AI Challenge!" and it should be from staff@aichallenge.org 2011-10-26T04:22:52 janzert: Yes, of course, I checked all folders. 2011-10-26T04:23:01 *** master_ninja has joined #aichallenge 2011-10-26T04:25:29 and your username @ pisem.net is correct? 2011-10-26T04:27:56 http://aichallenge.org/visualizer.php?game=15274&user=871 2011-10-26T04:27:57 good game 2011-10-26T04:28:23 *** zyberkiddy has quit IRC (Quit: Page closed) 2011-10-26T04:31:20 I don't understand why olex.s didn't raze the final hill any time after move 924 though? 2011-10-26T04:32:04 surely leaving an open hill on your flank is a bad idea at any time 2011-10-26T04:33:09 neither do i 2011-10-26T04:33:22 janzert: i'd be curious to see some games with better bots and less food 2011-10-26T04:34:44 *** mehf has joined #aichallenge 2011-10-26T04:34:57 hey there 2011-10-26T04:35:11 http://aichallenge.org/visualizer.php?game=14837 has a pretty low rate although on a small map and rather high starting food 2011-10-26T04:37:29 personally i like that much more 2011-10-26T04:37:34 the other map just got a bit flooded 2011-10-26T04:37:57 Please help :) should playgame.py be runned with some key to use presetted ants and food on map? i created map to test bot, set there food and ants, but map starts only whith ants on it 2011-10-26T04:38:16 mehf: yes, I believe it is --scenario 2011-10-26T04:38:20 *only whith hives on it 2011-10-26T04:38:56 take a look at the options available with --help 2011-10-26T04:39:30 *** sigh has joined #aichallenge 2011-10-26T04:39:33 i know i'm biased, but i think the maze maps are really cool 2011-10-26T04:39:41 ok, thanx. bad for me, i have win2008 and some problems whith running original playgame.py, 2011-10-26T04:39:43 WadimX: can you confirm your email address is your username @pisem.net? 2011-10-26T04:39:47 *** Blkt has joined #aichallenge 2011-10-26T04:40:12 the maze maps are sweet, but I wouldn't want them exclusively either 2011-10-26T04:41:12 it is time to watch day9daily archive! 2011-10-26T04:41:52 hey fellas, i just finished writing a new starter package. what now? 2011-10-26T04:41:52 i.e. dinner time 2011-10-26T04:42:09 janzert: Yes, it's mine 2011-10-26T04:42:25 I'll try sending you another email 2011-10-26T04:42:49 this one will be from janzert@aichallenge.org 2011-10-26T04:43:02 *** Uhyz has joined #aichallenge 2011-10-26T04:43:21 janzert: Thank you 2011-10-26T04:44:23 *** Uhyz has left #aichallenge 2011-10-26T04:44:31 thanx, --scenario helped and it seem to work. could it run with visualizer? some way it dont want to run, getting out of range exeption in visualizer 2011-10-26T04:46:04 janzert, are you an AI Challenge admin? 2011-10-26T04:46:45 if it causes a visualizer problem that's certainly a bug, but I don't really know enough about the visualizer to help with that. McLeopold or mleise are the ones to talk to 2011-10-26T04:46:55 paulwal: yes 2011-10-26T04:46:57 *** ig has quit IRC (Read error: Connection reset by peer) 2011-10-26T04:47:44 janzert, I just wrote a starter package. Can we add it to the website? 2011-10-26T04:48:06 do we support the language? 2011-10-26T04:48:36 The language is Tcl 2011-10-26T04:48:53 it should be installed on your server, though it may need an ugrade to 8.5 2011-10-26T04:49:20 upgrade* 2011-10-26T04:49:33 we'll need to add support in the compiler infrastructure for it as well 2011-10-26T04:49:50 it's not particularly hard just will need to be done 2011-10-26T04:50:07 are you familiar with git and github? 2011-10-26T04:50:11 ok, how do we do that? I can help. It's an interpreted, non-compiled language. 2011-10-26T04:50:24 yep, somewhat familiar with Tcl 2011-10-26T04:50:25 *** mehf has quit IRC (Quit: Page closed) 2011-10-26T04:50:55 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-10-26T04:51:17 Sorta, should I send a pull request to a certain directory 2011-10-26T04:51:56 best way to procede would to fork the repo and add your starter then modify worker/compiler.py to add Tcl support 2011-10-26T04:52:09 basically copy and modify the Python section to fit 2011-10-26T04:52:15 the send a pull request 2011-10-26T04:52:15 ok 2011-10-26T04:52:45 is the http://packages.ubuntu.com/it/natty/tcl too old? 2011-10-26T04:52:45 janzert - seeing as you guys support C#, does that mean F# is a possibility? 2011-10-26T04:52:56 it would be really good if that package will work 2011-10-26T04:53:12 bobbydroptable: if mono supports it 2011-10-26T04:53:12 natty has a tcl8.5 package 2011-10-26T04:53:38 ahh http://packages.ubuntu.com/natty/tcl8.5 2011-10-26T04:53:44 and the pull request is basically uploading tcl_starter_package.zip to ants/dist/starter_bots? 2011-10-26T04:54:10 yep, 8.5 2011-10-26T04:54:26 janzert: I just received empty mail: Received: from [62.141.94.151] (HELO mx1.ks.pochta.ru) by node2-2.ks.pochta.ru with QIP.RU LMTP for ....@pisem.net; Wed, 26 Oct 2011 12:42:37 +0400 2011-10-26T04:54:42 janzert: Is it Yours? 2011-10-26T04:54:45 I used dictionaries, so 8.5 is required 2011-10-26T04:55:00 anyone know how to permantly shut up Firefox warnings for the replay? 2011-10-26T04:55:09 * paulwal off to find a git tutorial 2011-10-26T04:55:10 WadimX: shouldn't be 2011-10-26T04:55:19 mine should be coming from a google server 2011-10-26T04:55:44 so how slow/fast are the test servers? no way my bot is going to timeout on my own machine, but it does on the servers 2011-10-26T04:56:03 paulwal: you'll need to fork the repo then add and commit the starter bot under a dist/starter_bots/tcl directory 2011-10-26T04:56:26 ok 2011-10-26T04:56:54 *** SharkMonkey has joined #aichallenge 2011-10-26T04:57:08 then you push that commit back to your own github repo and send a pull request back to the main repo 2011-10-26T04:57:18 *** racko has joined #aichallenge 2011-10-26T04:57:26 http://help.github.com/send-pull-requests/ 2011-10-26T04:57:42 has details on the github steps for sending the pull request 2011-10-26T04:58:01 Are the compile flags used for a given language documented anywhere? How are haskell bots compiled? C++? 2011-10-26T04:58:14 SharkMonkey: see worker/compiler.py 2011-10-26T04:58:16 thanks, janzert 2011-10-26T04:58:17 @repo 2011-10-26T04:58:18 janzert: repo could be https://github.com/aichallenge/aichallenge. 2011-10-26T04:58:21 janzert, thanks 2011-10-26T04:58:37 * janzert is off to bed 2011-10-26T04:58:39 *** jezuz` has joined #aichallenge 2011-10-26T04:59:46 sleep is for the weak! 2011-10-26T04:59:48 ta 2011-10-26T05:00:52 well it's already 5am here so... :P 2011-10-26T05:01:43 Human inadequacy. MyBot doesn't need sleep. 2011-10-26T05:01:52 *** WadimX has quit IRC (Ping timeout: 265 seconds) 2011-10-26T05:02:38 *** jezuz has quit IRC (Ping timeout: 260 seconds) 2011-10-26T05:04:28 *** olexs has joined #aichallenge 2011-10-26T05:05:43 *** Puj has joined #aichallenge 2011-10-26T05:09:28 says bobby at 8pm aha 2011-10-26T05:10:23 lol 2011-10-26T05:11:01 answer to self: in about.config, set security.fileuri.strict_origin_policy -> false 2011-10-26T05:11:28 *** bearoff_w has joined #aichallenge 2011-10-26T05:12:01 *** bearoff_w has left #aichallenge 2011-10-26T05:12:21 *** Egg5 has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20110928134238]) 2011-10-26T05:13:28 *** ashoka has quit IRC (Ping timeout: 265 seconds) 2011-10-26T05:13:57 *** rufus_ has quit IRC (Ping timeout: 265 seconds) 2011-10-26T05:14:05 janzert: What's the problem with the visualizer? 2011-10-26T05:14:19 mleise he's gone to sleep 2011-10-26T05:14:33 *** SharkMonkey has quit IRC (Ping timeout: 245 seconds) 2011-10-26T05:18:18 *** kaemo has quit IRC (Ping timeout: 245 seconds) 2011-10-26T05:19:34 *** barsandcat has quit IRC (Quit: Page closed) 2011-10-26T05:25:55 does data about water come in on every turn even if i've seen it already? or do they only send the ones i haven't seen? 2011-10-26T05:27:10 only unseen water 2011-10-26T05:27:16 any prolog bot develoipers in here? 2011-10-26T05:28:23 well i coded i think the first alicebot in prolog 2011-10-26T05:28:52 passes the programQ/D tests 2011-10-26T05:29:35 oh but it runs out of memory on the chomskyAIML set 2011-10-26T05:30:25 but anyhow alicebots are not the di8rrection i intend to leave off on this bot.. but being able to handle sucha aiml set is part of the baseline 2011-10-26T05:31:42 the initial in->out idioms (that what aiml is) .. get converted to a special deep structure intermediate langauge 2011-10-26T05:31:58 the aiml is a templete of " 2011-10-26T05:32:21 [possible viabble and maybe even legal discourse examples" 2011-10-26T05:32:45 the format actualyl store is the deep 2011-10-26T05:33:00 deepIn->deepOut "examples" 2011-10-26T05:33:50 users anotate such transitions into a special ontology that has a set planner mode3l set of flows 2011-10-26T05:34:29 then durrent runtime discourse such flows are attempted to be followed 2011-10-26T05:35:12 i am mentionihng all this to recruite some people who are prolog hyackers 2011-10-26T05:36:51 good morning everyone 2011-10-26T05:37:15 oh hey its the Blkt of another channel i've seen before 2011-10-26T05:37:16 morning! 2011-10-26T05:37:44 dmiles_afk: which channel? 2011-10-26T05:37:53 I didn't quite understand all of that, dmiles_afk. But I'm curios, are you writing a prolog starter package for the AI Challenge? 2011-10-26T05:37:56 the lisp subproject 2011-10-26T05:37:57 curious* 2011-10-26T05:38:11 (the lisp subproject to Blkt) 2011-10-26T05:38:40 paulwal: i should be! 2011-10-26T05:39:10 that'd be cool! i just finished one in Tcl. 2011-10-26T05:40:16 paulwal: the ants problem.. is the viusualizer a tcl app? 2011-10-26T05:41:10 i asking becasue i am trying to decide if the prolog code needs glued to somethjing 2011-10-26T05:41:59 dmiles_afk, I believe the visualizer is in java. I wrote a new starter package for Tcl. It's not uploaded to the website yet. 2011-10-26T05:42:01 http://aichallenge.org/starter_packages.php 2011-10-26T05:42:08 if the developer starting in prollog would be focusing not on the UI .. but more their brain code 2011-10-26T05:42:11 https://github.com/aichallenge/aichallenge/wiki/Ants-Starter-Pack-Guide 2011-10-26T05:42:33 you don't need to code the visualizer 2011-10-26T05:42:48 the game engine takes your inputs and translate it into a map replay that the java visualizer plays 2011-10-26T05:43:06 You don't need to write a visualizer. Your prolog bot will be executed by a python script, which can optionally send your bot's output to the java visualizer 2011-10-26T05:43:30 what bobby said 2011-10-26T05:44:22 *** barsandcat has joined #aichallenge 2011-10-26T05:44:58 excellent .. so instailly the prolog starter pack would make sure the doeveloper was writing out toi the right format 2011-10-26T05:45:02 initially 2011-10-26T05:45:24 *** v_paul_v_ has joined #aichallenge 2011-10-26T05:45:34 To give you an idea, my Tcl starter bot is about ~375 lines of code, comments, and whitespace 2011-10-26T05:45:35 hi 2011-10-26T05:46:04 python tools/playgame.py "python MyBot.py" "python tools/sample_bots/python/HunterBot.py" --map_file tools/maps/multi_hill_maze/maze_02p_02.map --log_dir game_logs --turns 60 --scenario --food randome --player_seed 7 --verbose -e 2011-10-26T05:46:14 give me error 2011-10-26T05:46:20 randome 2011-10-26T05:47:18 *** xathis has quit IRC (Ping timeout: 265 seconds) 2011-10-26T05:47:38 that sucks, v_paul_v_ 2011-10-26T05:48:03 there is an API that ants of two language implenmtations can battle it out? 2011-10-26T05:48:11 yep 2011-10-26T05:48:20 what to write on --food option& 2011-10-26T05:48:22 ? 2011-10-26T05:48:30 random not randome 2011-10-26T05:48:31 symmetric 2011-10-26T05:48:33 none 2011-10-26T05:48:34 random 2011-10-26T05:48:35 sections 2011-10-26T05:48:39 I guess i am wonmdering if java shoulsd present almost like a telnet server that each langauge clients into for controll of their army 2011-10-26T05:48:42 thanks 2011-10-26T05:49:17 if there is that type of world serfver presently 2011-10-26T05:49:34 dmiles_afk these are all the currently supported languages: http://aichallenge.org/starter_packages.php Tcl will soon be added to that and I saw in the forums someone just finished an Ada starter bot. 2011-10-26T05:49:55 dmiles_afk: there are TCP servers available for testing that work over the same text protocol 2011-10-26T05:50:21 how is the landing page game chosen? 2011-10-26T05:50:31 The official AI Challenge server runs your bot locally on their server, and communication is via stdout and stdin 2011-10-26T05:50:33 ok then in that case it might make sense for the prolog code to concentrate on connecting to that realtime server? 2011-10-26T05:50:57 no 2011-10-26T05:51:03 no need to connect to server 2011-10-26T05:51:03 ok good.. that case is stdin/out but it would be channeled to a tcp se4rver 2011-10-26T05:51:24 pretty much. 2011-10-26T05:51:58 again error :( 2011-10-26T05:52:23 python tools/playgame.py "python MyBot.py" "python tools/sample_bots/python/HunterBot.py" --map_file tools/maps/multi_hill_maze/maze_02p_02.map --log_dir game_logs --turns 60 --scenario --food random --player_seed 7 --verbose -e 2011-10-26T05:52:24 there are 3rd party tcp servers for testing for when the official server is backed up. There is a (python?) script available that acts as a layer between the tcp server and your bot's stdin/stdout 2011-10-26T05:52:33 heelp :) 2011-10-26T05:52:50 v_paul_v, what error are you getting? 2011-10-26T05:53:09 running for 60 turns 2011-10-26T05:53:11 Traceback (most recent call last): 2011-10-26T05:53:11 File "/home/paul/Документы/Ants/tools/engine.py", line 114, in run_game 2011-10-26T05:53:11 game.start_game() 2011-10-26T05:53:11 File "/home/paul/Документы/Ants/tools/ants.py", line 1399, in start_game 2011-10-26T05:53:11 self.do_food_visible(self.food_visible * self.num_players) 2011-10-26T05:53:12 File "/home/paul/Документы/Ants/tools/ants.py", line 1066, in do_food_visible 2011-10-26T05:53:14 self.food_sets_visible = deque(self.get_symmetric_food_sets(True)) 2011-10-26T05:53:16 File "/home/paul/Документы/Ants/tools/ants.py", line 1270, in get_symmetric_food_sets 2011-10-26T05:53:18 self.map_symmetry = choice(self.get_map_symmetry()) 2011-10-26T05:53:20 *** Saulzar has quit IRC (Ping timeout: 260 seconds) 2011-10-26T05:53:21 File "/home/paul/Документы/Ants/tools/ants.py", line 1226, in get_map_symmetry 2011-10-26T05:53:23 if self.map_similar(hill_aims[0][0], loc, aim, enemy_map[0]) is None: 2011-10-26T05:53:25 KeyError: 0 2011-10-26T05:54:04 It seems your map is not loading correctly. 2011-10-26T05:54:17 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-10-26T05:54:19 if i set food to "none" - i havn't error 2011-10-26T05:54:37 ok i wont make any promises.. but this project fits into something i am already doing slightly.. I have a prolog bot library that connects to secondlife .. sort of robocup like bots.. I see that happening with ants 2011-10-26T05:55:40 ants will help sure up the skeleton of what these prolog bots represent 2011-10-26T05:56:48 sounds cool 2011-10-26T05:57:09 how they represent data from the server into a known state to the action .. triggers update to next state 2011-10-26T05:57:14 v_paul_v_, remove the scenario option 2011-10-26T05:58:08 Puj, ok, it's work, thank you 2011-10-26T05:58:14 Sure Vivitron 2011-10-26T05:58:18 v_paul_v_ 2011-10-26T06:00:01 so the endgoal is to get the developer writing moderately highlevel rules.. like: self(X),nearBy(X,Y),enimy(Y),bigger(X,Y)->turnTo(Xm,Y),distance(X,Y,D),walkFw(D) ? 2011-10-26T06:00:09 *** replore has quit IRC (Remote host closed the connection) 2011-10-26T06:03:25 umm: self(X),enimyOf(X,Y),nearBy(X,Y),biggerThan(X,Y),distance(X,Y,D)->turnTo(X,Y),walkFwd(D) 2011-10-26T06:03:40 the prolog person concots the best set of rules and things 2011-10-26T06:04:14 *** kaemo has joined #aichallenge 2011-10-26T06:09:13 *** FusionX has quit IRC (Excess Flood) 2011-10-26T06:09:35 *** FusionX has joined #aichallenge 2011-10-26T06:12:08 *** Areks|2 has joined #aichallenge 2011-10-26T06:13:58 someone broke the httpd? 2011-10-26T06:15:01 *** Areks has quit IRC (Ping timeout: 240 seconds) 2011-10-26T06:15:01 of http://aichallenge.org 2011-10-26T06:15:40 dmiles_afk: working fine for me 2011-10-26T06:23:01 fine for me now 2011-10-26T06:23:08 *** UncleVasya has joined #aichallenge 2011-10-26T06:23:19 hahaha your game on the frontpage, olexs 2011-10-26T06:23:31 you beat GarySWest 2011-10-26T06:23:55 *** Kingpin13 has joined #aichallenge 2011-10-26T06:24:11 he seems to have weird bugs in handling large ant groups, he could've steamrolled my ants but got stuck :) 2011-10-26T06:24:54 the line of ants at turn 240 2011-10-26T06:25:05 row 12 col 109 2011-10-26T06:25:08 was that deliberate? 2011-10-26T06:25:15 or was it just quirk of ai? 2011-10-26T06:25:42 more or less deliberate 2011-10-26T06:25:47 nice 2011-10-26T06:26:07 i can't even move my ants properly yet 2011-10-26T06:26:08 lol 2011-10-26T06:26:11 one of my possible moves is "hold", where all ants in a group will stay at minimum "safe" distance... that seems to be it :) 2011-10-26T06:27:11 cool 2011-10-26T06:27:39 @seen rio{ 2011-10-26T06:27:39 Puj: I have not seen rio{. 2011-10-26T06:27:42 Hmf 2011-10-26T06:32:17 any tips for the amateurs, olexs? :D 2011-10-26T06:33:25 bobbydroptable: don't time out, don't lose ants when it's unnecessary, collect all food, and keep maximum map area visible :) 2011-10-26T06:33:30 lol 2011-10-26T06:33:45 anything more technical? data structures, algorithms etc. 2011-10-26T06:35:17 bobbydroptable: A* 2011-10-26T06:35:21 path finding 2011-10-26T06:35:33 delt0r_: you'd be surprised, but I don't use A* 2011-10-26T06:35:45 olexs: plenty of reasons not to 2011-10-26T06:35:51 *** alc has joined #aichallenge 2011-10-26T06:35:53 simple breadth-first search is all I need 2011-10-26T06:35:55 but its a good place to start 2011-10-26T06:36:37 olexs: this is a A* with a particular heuristic. Or the old fashion Dis*** never can remember the name 2011-10-26T06:36:48 Djikstra 2011-10-26T06:36:51 Thanks 2011-10-26T06:36:55 lol 2011-10-26T06:36:58 :D 2011-10-26T06:37:34 delt0r_: A* is BFS + heuristic. 2011-10-26T06:37:42 That is why A* is nice, by changing the Heuristic you can get quite different behaviors including plain old Djikstra 2011-10-26T06:37:44 olexs: you going backwards from 'objectives' breadth first? 2011-10-26T06:37:58 Six: depends on the application. in many cases, yes 2011-10-26T06:39:19 bobbydroptable: if you want more technical tips about battles, what I use is a alpha-beta search over a game tree for every local conflict 2011-10-26T06:39:36 olexs: A* can be a DFS too with a non admissible H 2011-10-26T06:39:48 or whatever the proper terminology is 2011-10-26T06:39:55 delt0r_: true 2011-10-26T06:40:39 here of course we have these "maze" maps 2011-10-26T06:40:41 I tried recently implementing an A* to replace the BFS in some applications in my bot, but I can't get in quite as optimal as my BFS is :) little implementation details make a lot of difference in sheer speed 2011-10-26T06:40:43 haha i'm afriad i haven't progressed that far yet 2011-10-26T06:41:03 i have just learnt of a particular data structure which would hopefully solve the problem I was trying to solve 2011-10-26T06:41:09 if i can get that working, the rest is pretty much set :D 2011-10-26T06:41:11 so the open lists are going to be small anyway 2011-10-26T06:41:29 bobbydroptable: what problem was that? 2011-10-26T06:41:56 eliminating "random walks" 2011-10-26T06:42:12 i.e. "oh crap nothing to do, walk in a random direction" 2011-10-26T06:42:22 and also global path finding 2011-10-26T06:42:58 right 2011-10-26T06:43:21 yeh 2011-10-26T06:43:40 *** halligalli has joined #aichallenge 2011-10-26T06:43:53 my first bot worked fine, if there was a hill near an ant, it would attack. if there was food near an ant it would pick it up 2011-10-26T06:44:11 but if there weren't anything of the sort, then it would walk in a fashion i could not control 2011-10-26T06:44:33 and i could not direct ants to a distant goal 2011-10-26T06:44:44 because it was so focused on its immediate vision 2011-10-26T06:45:21 bobbydroptable: what I do in such a situation is direct the ant to the closest square not yet explored 2011-10-26T06:45:39 *** Blkt has quit IRC (Quit: brb) 2011-10-26T06:45:45 not just not visible, but not explored at all 2011-10-26T06:46:43 do you just point it at a particular direction or do you plan a path there 2011-10-26T06:47:12 olexs but if that location is far away and there is an invisible area at the time - how do you handle that? Do you save map information over time? 2011-10-26T06:47:21 does anyone know if it's legal/possible to include a shared library with your bot and dynamically load it? 2011-10-26T06:47:43 *** justasu has joined #aichallenge 2011-10-26T06:48:00 bobbydroptable: I do my breadth-first search from the ant to the next unexplored square in a limited radius, and direct it there. my bfs will always save the direction it started the search with for every found target, that is all I need 2011-10-26T06:48:32 halligalli: I limit the radius of this search. If there are no unexplored squares near an ant but invisible squares, it will go to the invisible ones 2011-10-26T06:48:36 *** justasu has quit IRC (Client Quit) 2011-10-26T06:49:11 and yes, I save map information over time, I always know what my ants already have seen and what they can see right now 2011-10-26T06:49:24 okay, scenario: 2 ants are at a T intersection 2011-10-26T06:49:38 both does the BFS 2011-10-26T06:49:47 olexs i am wondering if it makes sense to save paths over time to be able to have them instantly instead of computing them via A* or alike 2011-10-26T06:50:13 does your ants both go in the same direction, or do they go in different directions, or do you have no control over which direction they decide on 2011-10-26T06:50:16 halligalli: I don't do that, but it certainly can make sense in some situations 2011-10-26T06:50:46 halligalli: it is horribly difficult ==" i tried doing it for the past 3 days, and I've switched strategies now. 2011-10-26T06:51:04 bobbydroptable you mean it is difficult to save paths? 2011-10-26T06:51:11 *** Blkt has joined #aichallenge 2011-10-26T06:51:21 its difficult to save paths to be used in future searches. 2011-10-26T06:51:40 if you just want to safe a path from a single search ,that's not too bad 2011-10-26T06:51:42 bobbydroptable it is memory intensive but why should it be difficult? 2011-10-26T06:52:06 bobbydroptable: in a rare case the ants are equally distant from both sides, they might go in the same direction. but since one ant will almost always be behind the other or near it, they will ultimately choose different paths, since the first ant will pick a direction and go, when the next ant arrives at the intersection the other unexplored side will be 1 pixel closer to him 2011-10-26T06:52:07 well i'm not particularly skilled in this area, so if there was an easy way, i'm all ears 2011-10-26T06:52:23 olexs, do you wipe clean the location of hills, food, and ants before the start of a turn, or do you remember their last known locations? 2011-10-26T06:53:17 bobbydroptable neither am i ;) maybe i'll try it and run into the same difficulties ;) 2011-10-26T06:53:29 paulwal: I wipe everything, it makes it easier right now. I will probably modify the bot later to remember food and hills locations outside of visible area, since there are some problems with that now 2011-10-26T06:54:11 can anyone point me in a direction of what kind of algorithms might be helpful for fighting? 2011-10-26T06:54:22 some clustering kinds? 2011-10-26T06:54:48 what problems, olexs? is it too cpu intensive to figure out which tiles are no longer visible? 2011-10-26T06:54:52 halligalli: clustering is a good start. again, for me it's BFS with a limited radius, repeated multiple times 2011-10-26T06:55:31 olexs you mean bfs from ants to find other ants and move in that direction? 2011-10-26T06:55:46 halligalli: maybe we are talking abotu separate problems or maybe i am just thinking too much over a non-problem 2011-10-26T06:56:17 paulwal: no, nothing like that. but there is a rare condition where my ant will see a food in another corridor in a maze that he cannot reach yet; he will start going there, but as soon as he loses that food item from sight, he might turn around and go see it again, and again, and again :) 2011-10-26T06:56:44 halligalli: bfs from ants to find other ants, to build groups of them, no moving involved on this step yet 2011-10-26T06:57:19 after I build clusters, the complex part comes into play, with the state-space search in a game tree :) 2011-10-26T06:57:23 olexs ah, okay, so finding possible fighting gangs and then afterwards applying algorithms to move them in a fight-appropriate manner 2011-10-26T06:57:28 exactly 2011-10-26T06:57:44 olexs phew, game tree, fancy 2011-10-26T06:57:57 olexs probably time-limits is the biggest problem 2011-10-26T06:58:08 olexs: might be an idea to give them a 'goto' goal 2011-10-26T06:58:08 yep... this is the most time-consuming part 2011-10-26T06:59:03 olexs so something like min-max adversary i guess? 2011-10-26T06:59:24 halligalli: yes. alpha-beta is an optimization of a min-max search, that's what I do 2011-10-26T06:59:33 cool 2011-10-26T06:59:40 i still have a lot of stuff to do ;) 2011-10-26T07:00:03 currently i have a global heat map and each ant is doing bfs to find the hotspot around 2011-10-26T07:00:14 it seems I'm giving away all my secrets here :) 2011-10-26T07:00:37 heatmap is a nice idea 2011-10-26T07:00:50 hehe, don't worry, i think the implementation is the hardest part in terms of efficiency in time 2011-10-26T07:00:52 could have pretty simple ant logic with a good heatmap 2011-10-26T07:00:56 still enough technicalities behind everything that actually make it work... 2011-10-26T07:01:11 yeah, my heatmap is pretty simple, yet it performs decent 2011-10-26T07:01:53 but i guess at least the fights have to be treated separately 2011-10-26T07:03:26 yeah 2011-10-26T07:03:47 has anyone tried to try to optimize the whole game state, i.e. have a heuristic judging the whole game state and then minimizing it by moving the ants? 2011-10-26T07:03:58 heatmap sounds good for overall goal planning, but local vision is better for particular goals 2011-10-26T07:04:56 this challenge is awesome, one can try so many concepts 2011-10-26T07:05:06 halligalli: that sounds tough to do with unknown similtanious moves 2011-10-26T07:05:58 like, enemy ants can move in a number of different ways 2011-10-26T07:06:01 Six well it would be a state search over all your ants making all possible moves, try to minimize that - i assume it will time out if you do not have good heuristics 2011-10-26T07:06:26 for starters you could ignore the enemy moves 2011-10-26T07:06:54 halligalli: I think I can tell you right now that will time out instantly :) I only consider moves in local clusters, and only a limited number of group moves, not all possible moves by all ants, and it is still very time-consuming 2011-10-26T07:07:16 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-10-26T07:07:26 olexs yeah, i guess so ;) 2011-10-26T07:08:01 *** xathis has joined #aichallenge 2011-10-26T07:08:20 if you try to search all ant moves, it would be a 4^ants branching factor, much higher then even chess, and in chess they use move generators with optimizations :) 2011-10-26T07:08:25 olexs but again maybe not finding the optimal, but using a heuristic guiding the choices, i.e. decide on a random move per ant and then randomly change moves of ants, kind of a random walk like manner 2011-10-26T07:09:24 man amazing how your algorithm craps out when your parameters are in the wrong order. 2011-10-26T07:09:35 5^ants (no move) 2011-10-26T07:09:40 true 2011-10-26T07:10:02 they use game trees in chess too, don't they 2011-10-26T07:10:08 yes 2011-10-26T07:11:02 in fact I got the idea from chess AIs, and still think of some optimizations they use, like quiescence search 2011-10-26T07:11:36 olexs: how many ants are you allowing in each partition? 2011-10-26T07:12:31 antimatroid: there are no limits set per se. it wil cluster all ants roughly "in one fight together", the bfs search radius is double attack radius iirc, with iterative extension for every found ant 2011-10-26T07:12:33 I think if you want to use strategies from abstract logic game AIs (like chess etc) go would be the game to look at... it also has the problem of a very very high branching factor 2011-10-26T07:13:10 game trees are always what one wants, it's just a matter of where you're going to prune things 2011-10-26T07:13:27 or if that's just never going to happen heuristics <3 2011-10-26T07:13:56 simple pruning a-la alpha-beta works rather well 2011-10-26T07:14:34 i don't trust alpha beta pruning when generating a tree 2011-10-26T07:14:41 why not? 2011-10-26T07:14:42 but i guess one has to throw stuff out somewhere 2011-10-26T07:14:48 because it isn't optimal 2011-10-26T07:14:54 alpha beta pruning is 2011-10-26T07:15:12 it's alright when evaluating a tree, but if you use alpha beta to prune stuff while generating the tree you may miss optimal strategies 2011-10-26T07:15:29 no, thats simply not true 2011-10-26T07:15:30 olexs: if they altered the attackradius/viewradius would you need to redo your gametrees? 2011-10-26T07:15:44 yes it is 2011-10-26T07:15:50 alpha-beta pruning is safe, equivalent to minimax 2011-10-26T07:16:01 ergh, you're not reading me 2011-10-26T07:16:03 bobbydroptable: no, all the parameters are dynamic 2011-10-26T07:16:23 it's alright when evaluating a tree you have, but not when you're still going to generate more tree 2011-10-26T07:16:24 oh so its not loading from a file, its generating it on the fly? 2011-10-26T07:16:53 bobbydroptable: nothing's loading from a file :) the trees are generated for every battle on the fly, yes 2011-10-26T07:17:20 sounds computationally expensive 2011-10-26T07:17:35 it is, but it fits in the time frame 2011-10-26T07:17:38 i'd rather a "numerical superiority" heuristic lol 2011-10-26T07:17:55 and in java too :| 2011-10-26T07:18:18 yeh the number of java in the top 10 is really really surprising 2011-10-26T07:18:18 lol 2011-10-26T07:18:23 i tried doing java for planetwars 2011-10-26T07:18:27 timedout all the time 2011-10-26T07:18:32 then again, i suck so it's probably just me 2011-10-26T07:18:47 but hey, python is on top now, so that gives me hoep 2011-10-26T07:19:28 I don't know what everybody has got against java :) it's as fast as you program it to be... as long as you know to use the correct data structures and not overload the GC, you're just fine with java 2011-10-26T07:19:43 well yeh, thus i put that disclaimer there :D 2011-10-26T07:20:04 i don't even really know it 2011-10-26T07:20:25 used it in first year programming classes as a first language, used c in second year then didn't look back once i met c++ 2011-10-26T07:20:30 If I do scriptable fighting and after gathering food, exploring, haunting hills, processing battles I still have ~350ms more, for what I must use them? :) 2011-10-26T07:20:48 UncleVasya: game tree 2011-10-26T07:20:50 tell your ants to form letters 2011-10-26T07:20:52 lol 2011-10-26T07:21:05 *** randll has joined #aichallenge 2011-10-26T07:21:22 The map is too small to write interesting message :) 2011-10-26T07:21:33 draw a penis 2011-10-26T07:21:45 form your ants into squadrons of penii 2011-10-26T07:21:57 Maybe to draw a heart... but still it is not time-consuming. 2011-10-26T07:22:01 i want to see block formations of dicks go into battle 2011-10-26T07:22:03 make it happen 2011-10-26T07:22:11 lol 2011-10-26T07:22:16 someone on the TCP server had ants form "GG" when winning a game 2011-10-26T07:22:23 that's nice 2011-10-26T07:22:23 http://ants.fluxid.pl/replay.5621 2011-10-26T07:22:35 Garf: yep, exactly that :) 2011-10-26T07:23:00 *** barsandcat has quit IRC (Ping timeout: 265 seconds) 2011-10-26T07:23:25 now i gotta see htis 2011-10-26T07:23:35 GG... Now I need my dota-dose :( 2011-10-26T07:24:13 oh man that is epic 2011-10-26T07:24:28 impressive how his ants just shift 2011-10-26T07:26:14 haha GG 2011-10-26T07:28:28 well my data structure is done ... 2011-10-26T07:28:33 now its just what to do with it ... 2011-10-26T07:29:04 *** svujic has joined #aichallenge 2011-10-26T07:34:46 Can a player have more than one hives? 2011-10-26T07:38:05 hive* 2011-10-26T07:38:08 GeorgeJ yes 2011-10-26T07:38:15 GeorgeJ: hives no, hills yes 2011-10-26T07:38:47 hive is generally one per player, it is not a map object but rather a display for how many ants you can spawn 2011-10-26T07:38:59 hills are the "spawners", on some maps players have many 2011-10-26T07:39:07 I see, I ment hills, yes 2011-10-26T07:41:41 Have you guys got some more funny replays? 2011-10-26T07:42:34 http://aichallenge.org/visualizer.php?game=13905&user=53 2011-10-26T07:44:55 *** savaron has joined #aichallenge 2011-10-26T07:46:27 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 258 seconds) 2011-10-26T07:49:31 i saw a game with 4 starter packs 2011-10-26T07:49:35 that was pretty funny 2011-10-26T07:50:32 Could you recommend an active tcp server? 2011-10-26T07:53:38 UncleVasya: interesting there when GreenTea's pairing up code caused 4 of his ants to cycle 2011-10-26T07:56:35 *** foRei has joined #aichallenge 2011-10-26T07:56:53 GeorgeJ: ants.fluxid.pl 2011-10-26T07:57:17 tyvm 2011-10-26T07:57:27 *** yoden has joined #aichallenge 2011-10-26T08:00:50 attackradius2 is 5, sqrt(5) is 2.23, does that mean there ants have to be in a 2pixel range to attack eachother? 2011-10-26T08:01:26 Ah, nvm 2011-10-26T08:02:15 int(sqrt(5)) == 2 2011-10-26T08:03:05 *** replore_ has joined #aichallenge 2011-10-26T08:03:39 hmm, maybe it should be rounded 2011-10-26T08:03:44 *** kofman has quit IRC (Ping timeout: 276 seconds) 2011-10-26T08:09:02 nah, don't think so 2011-10-26T08:11:20 *** savaron has quit IRC (Ping timeout: 265 seconds) 2011-10-26T08:13:32 Does the food disappear if its not eaten? 2011-10-26T08:14:01 don't think so 2011-10-26T08:14:39 nope 2011-10-26T08:15:32 it only appears 2011-10-26T08:17:33 *** alc has quit IRC (Ping timeout: 252 seconds) 2011-10-26T08:17:44 i wish my bot was smarter 2011-10-26T08:18:45 Why? 2011-10-26T08:19:10 *** alc has joined #aichallenge 2011-10-26T08:19:18 me too 2011-10-26T08:19:30 I don't want my bot to be smarter than I wrote it to be :) 2011-10-26T08:20:04 my bot only knows how to get food and sometimes even tries to raze hills 2011-10-26T08:20:41 what pathfinding algorithms do you folks use? 2011-10-26T08:20:54 There is near 2 mons more, you have a time to teach it do other things. 2011-10-26T08:21:06 tiered a* is my plan 2011-10-26T08:21:13 Generally it's BFS or A*. 2011-10-26T08:21:39 my but just tries to follow walls, and when it sees it food or hill it uses BFS 2011-10-26T08:22:21 at least it doesn't time out anymore 2011-10-26T08:26:41 A* always better than BFS right? 2011-10-26T08:27:01 someone was here earlier saying he didn't need to do A* at all 2011-10-26T08:27:08 and he's 4th on the table 2011-10-26T08:27:10 so i can't say 2011-10-26T08:27:23 hmmm 2011-10-26T08:27:41 olexs said that 2011-10-26T08:28:05 A* is better if you need long paths around obstacles 2011-10-26T08:28:41 since most searches I do are rather small (within view radius or less), BFS suffices, and it is less complex implementation-wise 2011-10-26T08:29:30 A* is faster and as optimal as BFS in the case you've done the hieristic function right. If There is a mistake than A*'s path will be longer than the one found by BFS. 2011-10-26T08:29:55 I think in Ants you can have 2 reasons to choose A*: 2011-10-26T08:29:58 I will use A* in some cases later, for now I can't get it to be as fast in reality as my BFS is because of data structure overhead needed 2011-10-26T08:30:12 1. You want to try it. 2011-10-26T08:30:17 Do you guys cache ? 2011-10-26T08:30:51 why shouldn't you if you can? 2011-10-26T08:31:50 2. You know well what you do and you really need these few miliseconds (wich is the difference between A* and BFS). 2011-10-26T08:32:05 *** pvarga has quit IRC (Quit: pvarga) 2011-10-26T08:32:16 You could compute a*, cache the result, and invalidate it if water or an enemy appears 2011-10-26T08:33:19 2nd means you can optimize well on your language of choise and will use smart enought even few released miliseconds. 2011-10-26T08:33:56 If you care only about implementing your ideas to make it just work, than BFS is enough I believe. 2011-10-26T08:34:10 good then I'll go for BFS for now 2011-10-26T08:35:37 just make a get_path_toN function, and implement BFS at first if you find it easier, then a* as get_path_to2 or something, you should be able to use them interchangeably 2011-10-26T08:37:43 *** delt0r_ has quit IRC (Ping timeout: 252 seconds) 2011-10-26T08:39:49 *** xathis has quit IRC (Quit: Page closed) 2011-10-26T08:39:58 I have a Search class in my bot, that encapsulates all search details. in the experimental A* branch, it will switch to A* if the radius is > viewradius and there is an admissible heuristic available, otherwise it'll use BFS. the code that calls the search does not need to worry about which algorithm is used, unless is wants to force a certain one 2011-10-26T08:41:49 olexs: You coding in java? 2011-10-26T08:41:56 GeorgeJ: yes. 2011-10-26T08:42:22 Meh, people could cheat on tcp servers 2011-10-26T08:43:00 *** dwins has joined #aichallenge 2011-10-26T08:43:01 sure, that's why those are for testing only, and the proper contest compiles and runs your code locally 2011-10-26T08:44:00 And broke they own results on TCP then => the get not a good feedback about they'll bot perfomance. 2011-10-26T08:44:50 Teoreticaly people can cheat for fun, but it is aichallenge, not the CS1.6 server for 13-years okld kids :) 2011-10-26T08:45:42 Is there a memory limit? 2011-10-26T08:45:52 So I think the assumption that we all are good guys here is valid ;) 2011-10-26T08:46:03 1.5 gb approx 2011-10-26T08:46:11 Whoa, nice 2011-10-26T08:48:20 1.5 gb :O 2011-10-26T08:50:04 *** master_ninja has quit IRC (Quit: ...) 2011-10-26T08:51:14 *** UncleVasya has quit IRC (Ping timeout: 244 seconds) 2011-10-26T08:51:30 *** delt0r_ has joined #aichallenge 2011-10-26T08:51:38 *** Harpyon has joined #aichallenge 2011-10-26T08:51:47 oh well, I can't use more memory anyway :-) 2011-10-26T09:01:38 *** alc has quit IRC (Ping timeout: 245 seconds) 2011-10-26T09:06:34 *** xathis has joined #aichallenge 2011-10-26T09:07:04 *** cyphase has joined #aichallenge 2011-10-26T09:10:20 *** olexs1 has joined #aichallenge 2011-10-26T09:10:46 http://aichallenge.org/visualizer.php?game=15893&user=4513 <- dangerous bot, I expect this one will be the number 1 very soon 2011-10-26T09:12:33 *** olexs has quit IRC (Ping timeout: 260 seconds) 2011-10-26T09:15:57 I wonder if any bots go for map coverage 2011-10-26T09:16:02 oh thanks avdg2 :) 2011-10-26T09:16:36 nice bot btw ;-) 2011-10-26T09:18:06 nice 2011-10-26T09:18:06 swarming 2011-10-26T09:18:09 battle strategy is good 2011-10-26T09:18:24 i'm trying to do that too :( 2011-10-26T09:19:07 took out greentea and momobot 2011-10-26T09:19:20 definitely going to rank high 2011-10-26T09:19:27 but i have no defence strategy at all atm, so i'm loosing low food spawn games and the multi-hill-maps 2011-10-26T09:19:33 meh, that bot collected so much food, the hill became a bottleneck 2011-10-26T09:19:33 nice i'm jealous 2011-10-26T09:20:20 early offence is often good 2011-10-26T09:20:20 *** olexs1 has quit IRC (Quit: Leaving.) 2011-10-26T09:21:00 *** morphism has joined #aichallenge 2011-10-26T09:21:09 *** morphism has left #aichallenge 2011-10-26T09:21:49 15:10:46 < avdg2> http://aichallenge.org/visualizer.php?game=15893&user=4513 ← awesome battle! 2011-10-26T09:22:16 xathis: how do you keep map coverage? 2011-10-26T09:22:27 do you specialize ants? 2011-10-26T09:23:53 at the beginning of each turn no ant is specialized 2011-10-26T09:24:17 woah 2011-10-26T09:24:30 ruthless 2011-10-26T09:24:34 yeh 2011-10-26T09:24:34 xathis: turn ot game? 2011-10-26T09:24:40 where do people learn this shiz lol 2011-10-26T09:24:48 hehe 2011-10-26T09:24:50 then i look for food and enemy and stuff like that and decide for each ant individually 2011-10-26T09:25:12 sorry, gtg 2011-10-26T09:25:19 cya 2011-10-26T09:25:23 cya 2011-10-26T09:25:25 I'm thinking of making ant armies, each army can either control each ant individualy or as a group 2011-10-26T09:25:26 *** xathis has quit IRC (Quit: Page closed) 2011-10-26T09:25:52 he has each ant cover one part of map 2011-10-26T09:26:56 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2011-10-26T09:26:59 Make a "scout" army, divide map space to squares small enough for vision of an ant, and send each ant in a square 2011-10-26T09:27:35 yeah really nice strategy :D 2011-10-26T09:27:47 i wish the squares things worked 2011-10-26T09:27:49 *** hehe has joined #aichallenge 2011-10-26T09:28:05 bobbydroptable: Why wouldn't it? 2011-10-26T09:28:05 i kept having problem separating my map 2011-10-26T09:28:11 because of the water tiles causing problems 2011-10-26T09:28:15 feels like playing an rts 2011-10-26T09:28:18 meh, the squares fits in my strategy thoug 2011-10-26T09:28:20 :D 2011-10-26T09:28:20 h 2011-10-26T09:28:35 so i am in a square, then food appears across a river, then my ant can't get there 2011-10-26T09:28:50 because the square is split by the river 2011-10-26T09:28:56 bobbydroptable: ant_closest_to(row, col) 2011-10-26T09:29:07 bobbydroptable: thats the trick ;-) 2011-10-26T09:29:07 you need to do it more like space filling as you go 2011-10-26T09:29:23 GeorgeJ we are talking about separating the grid and having an ant in each square right? 2011-10-26T09:29:39 bobbydroptable: For vision, yes, not neccesarely for gathering 2011-10-26T09:29:49 ah 2011-10-26T09:29:52 fair enough then 2011-10-26T09:30:07 *** paulwal has quit IRC (Ping timeout: 265 seconds) 2011-10-26T09:30:53 *** Garf has joined #aichallenge 2011-10-26T09:32:03 *** hehe has quit IRC (Client Quit) 2011-10-26T09:35:42 *** olexs has joined #aichallenge 2011-10-26T09:36:55 *** v_paul_v_ has quit IRC (Remote host closed the connection) 2011-10-26T09:37:09 *** geekchimp has joined #aichallenge 2011-10-26T09:37:28 *** olexs has quit IRC (Client Quit) 2011-10-26T09:37:45 jow do i debug my code? i use python 2011-10-26T09:39:10 *** olexs has joined #aichallenge 2011-10-26T09:40:05 *** Darkstorm369 has joined #aichallenge 2011-10-26T09:41:27 geekchimp: you could output debug data in stderr 2011-10-26T09:42:55 i know nothing about python, but in java i user remote debugging 2011-10-26T09:43:19 i think this is also possible for python 2011-10-26T09:43:51 Or have a seppare file named yourbot_pid.log 2011-10-26T09:44:02 And dump log data there, if you don't want to use stderr 2011-10-26T09:44:50 meh, I believe stderr is the best way to do it 2011-10-26T09:45:08 you can specify the log options by hand (see "./playgame.py" help screen) 2011-10-26T09:45:49 if you have to write files, you have to modify it as well before submitting it to the server 2011-10-26T09:45:56 (which doesn't allow writes) 2011-10-26T09:46:05 http://pydev.org/manual_adv_remote_debugger.html 2011-10-26T09:48:26 avdg2: Or you could juse use a debug switch 2011-10-26T09:48:37 yeah 2011-10-26T09:48:45 Either at compile time or runtime 2011-10-26T09:48:50 but I'm using tests as well 2011-10-26T09:49:03 *** ikaros has joined #aichallenge 2011-10-26T09:49:27 *** geekchimp has quit IRC (Ping timeout: 265 seconds) 2011-10-26T09:51:16 *** TerryVB has joined #aichallenge 2011-10-26T09:54:34 *** DeeDee has joined #aichallenge 2011-10-26T09:55:27 hi, please is here any org? 2011-10-26T09:55:46 organisation? 2011-10-26T09:56:25 admin.. or some judge... 2011-10-26T09:56:52 ah, I believe that they are mostly around here a bit later 2011-10-26T09:57:14 but I don't know if there is any of them online atm 2011-10-26T09:58:26 thank you 2011-10-26T09:59:37 *** ajhager has joined #aichallenge 2011-10-26T10:00:19 *** Darkstorm369 has quit IRC () 2011-10-26T10:15:20 *** bearoff_w has joined #aichallenge 2011-10-26T10:15:59 *** bearoff_w has left #aichallenge 2011-10-26T10:16:32 *** sigh has quit IRC (Remote host closed the connection) 2011-10-26T10:16:36 *** he_the_great has joined #aichallenge 2011-10-26T10:17:04 *** xathis has joined #aichallenge 2011-10-26T10:17:04 *** sigh has joined #aichallenge 2011-10-26T10:21:24 *** sigh has quit IRC (Ping timeout: 255 seconds) 2011-10-26T10:22:18 *** Harpyon has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-10-26T10:22:31 *** Harpyon has joined #aichallenge 2011-10-26T10:32:47 janzert shoudl be awake soon. 2011-10-26T10:33:13 *** svujic has left #aichallenge 2011-10-26T10:36:42 lol 2011-10-26T10:36:48 he went to bed at 5am 2011-10-26T10:37:27 bobbydroptable: now you can call people week for being in bed :) 2011-10-26T10:37:43 haha 2011-10-26T10:37:44 yeh 2011-10-26T10:37:46 sigh 2011-10-26T10:38:07 no, it appears he is actually in bed :P 2011-10-26T10:38:22 sigh has been ruined for me 2011-10-26T10:38:29 lol 2011-10-26T10:38:33 i've never met him 2011-10-26T10:38:36 what are you doing up late 2011-10-26T10:38:42 i'm coding and watching gruen transfer 2011-10-26T10:38:46 http://ants.fluxid.pl/replay.7092 i love his GG at the end 2011-10-26T10:39:00 i mean flashm's 2011-10-26T10:39:03 who's fluxid? 2011-10-26T10:39:13 him 2011-10-26T10:39:18 is that accoun? 2011-10-26T10:39:26 people said accoun was hosting that tcp server? 2011-10-26T10:39:39 ants.fluxid.pl is mine instance 2011-10-26T10:39:47 Fluxid: are you accoun? 2011-10-26T10:39:51 no 2011-10-26T10:39:56 hmm okay 2011-10-26T10:40:07 sorry :P 2011-10-26T10:40:14 instance, but i used someone's else code 2011-10-26T10:40:15 lol 2011-10-26T10:40:58 i wonder if rufes band knows we've launched 2011-10-26T10:41:04 a few of them were around during beta 2011-10-26T10:42:57 mmm 2011-10-26T10:43:01 i wouldn't know them :) 2011-10-26T10:43:59 *** he_the_great has quit IRC (Ping timeout: 248 seconds) 2011-10-26T10:44:28 anybody knows if there's way how to download replays? 2011-10-26T10:45:42 there has usually been a way 2011-10-26T10:45:46 but i don't know sorry 2011-10-26T10:46:40 *** aerique has quit IRC (Quit: ...) 2011-10-26T10:47:56 it's realy fascinating that atns :D 2011-10-26T10:48:05 better than most of tv programs :D 2011-10-26T10:48:57 DeeDee: aichallenge.org/game/15876 is an example, the number is the game id 2011-10-26T10:49:25 thanks :) 2011-10-26T10:50:05 you can view it by copying the data into game_logs/replay.0.html right after visualizer.loadReplayData( 2011-10-26T10:50:20 yeye.. i already found it :) 2011-10-26T10:50:25 i'm sure there's a more elegant way 2011-10-26T10:55:29 *** Egg5 has joined #aichallenge 2011-10-26T10:55:52 *** fercho has joined #aichallenge 2011-10-26T10:55:56 meh, I wish I was much more stupid, so I just cared more about coding... 2011-10-26T10:56:09 * avdg2 brain is exploding 2011-10-26T10:56:10 *** Egg5 has quit IRC (Client Quit) 2011-10-26T10:56:16 hi there! 2011-10-26T10:56:22 got a question about the ai challenge 2011-10-26T10:56:28 wondering if any one could help? 2011-10-26T10:56:32 here ;-) 2011-10-26T10:56:38 thanks 2011-10-26T10:56:50 i just downloaded the starterpack for javascript (node) 2011-10-26T10:57:01 yeah 2011-10-26T10:57:05 * avdg2 is doing js 2011-10-26T10:57:06 in the list it says that the hills are not implemented 2011-10-26T10:57:24 *** Egg5 has joined #aichallenge 2011-10-26T10:57:32 does that means that when fetching the enemyHills() object, i don't get any results? 2011-10-26T10:57:52 fercho: if you want to do some js programming with that package, its better to begin with the game specifications ;-) 2011-10-26T10:57:58 http://aichallenge.org/specification.php 2011-10-26T10:58:10 so you know the protocol and also what the code is trying to do 2011-10-26T10:58:24 because that js package isn't the best one to start with 2011-10-26T10:58:48 but that doesn't mean that js is a bad language ;-) 2011-10-26T10:58:51 I understand 2011-10-26T10:58:57 no, i get that 2011-10-26T10:59:12 *** savaron has joined #aichallenge 2011-10-26T10:59:16 anyone know anything about quadtrees? :( 2011-10-26T10:59:50 I wondered if everything was fine, because i tried to fetch the enemyhills and access the row properties but i get errors all the time 2011-10-26T11:00:00 so i thought they may not be implemented yet 2011-10-26T11:00:16 I'll take a look at the game specifications first 2011-10-26T11:00:28 fercho: the best way to learn the spec and write your own engine, thats how I've done it 2011-10-26T11:00:39 *** onensora has quit IRC () 2011-10-26T11:00:41 and copy code :-) 2011-10-26T11:01:41 hehehe 2011-10-26T11:01:43 thanks a lot 2011-10-26T11:01:48 amazing! 2011-10-26T11:01:55 have a nice day 2011-10-26T11:02:03 good luck :-) 2011-10-26T11:02:49 *** dvladim has joined #aichallenge 2011-10-26T11:02:51 *** barsandcat has joined #aichallenge 2011-10-26T11:04:00 are maps always square? 2011-10-26T11:04:08 *** Egg5 has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20110928134238]) 2011-10-26T11:04:09 no 2011-10-26T11:04:13 sigh 2011-10-26T11:04:17 ok 2011-10-26T11:04:33 *** fercho has quit IRC (Quit: Page closed) 2011-10-26T11:05:34 *** olexs has quit IRC (Ping timeout: 244 seconds) 2011-10-26T11:05:49 *** randll has quit IRC (Ping timeout: 265 seconds) 2011-10-26T11:10:47 *** tncardoso has joined #aichallenge 2011-10-26T11:11:51 cya 2011-10-26T11:11:57 \logout 2011-10-26T11:12:10 *** DeeDee has quit IRC (Quit: Page closed) 2011-10-26T11:12:33 *** fercho has joined #aichallenge 2011-10-26T11:12:42 hi there! 2011-10-26T11:12:46 hi 2011-10-26T11:12:49 sorry to disturb, got another question 2011-10-26T11:13:14 can i drop it out? 2011-10-26T11:13:14 i'll try to help 2011-10-26T11:13:23 amazing, thanks 2011-10-26T11:13:23 drop what out? 2011-10-26T11:13:28 the quesiton =) 2011-10-26T11:13:34 go 2011-10-26T11:13:35 *** dvladim has quit IRC (Ping timeout: 260 seconds) 2011-10-26T11:13:35 *** theonewolf has quit IRC (Ping timeout: 260 seconds) 2011-10-26T11:13:39 oh 2011-10-26T11:13:56 I can't find the hills array in the javascript starterpack 2011-10-26T11:14:12 *** theonewolf has joined #aichallenge 2011-10-26T11:14:41 its in the github sources 2011-10-26T11:14:42 hmmm 2011-10-26T11:14:55 website prolly needs updating with new package 2011-10-26T11:15:05 i see 2011-10-26T11:15:08 there you go 2011-10-26T11:15:09 lol 2011-10-26T11:15:15 do u have the link to the repo? 2011-10-26T11:16:33 https://github.com/aichallenge/aichallenge 2011-10-26T11:17:05 fercho, Ants.js l 63 2011-10-26T11:17:29 *** aavikkojootti has joined #aichallenge 2011-10-26T11:19:11 hi!. i have problem with visualizer. It doesen't pop up after play_one_game nor after executing ./tutorial.sh 2011-10-26T11:20:59 http://pastie.org/2762459 2011-10-26T11:21:04 weird 2011-10-26T11:21:12 I can't find the starterpacks in the github repo 2011-10-26T11:21:15 * avdg2 remembers the hill problem 2011-10-26T11:21:28 just define this.hills on top of the file 2011-10-26T11:21:31 its defined 2011-10-26T11:21:35 I mean used 2011-10-26T11:21:38 fercho: https://github.com/aichallenge/aichallenge/tree/epsilon/ants/dist/starter_bots/javascript 2011-10-26T11:21:39 but not defined 2011-10-26T11:22:50 *** delt0r__ has joined #aichallenge 2011-10-26T11:25:16 *** delt0r_ has quit IRC (Ping timeout: 260 seconds) 2011-10-26T11:26:00 *** analyst74 has quit IRC (Ping timeout: 240 seconds) 2011-10-26T11:26:58 *** Areks has joined #aichallenge 2011-10-26T11:27:05 Garf, thank you for the link 2011-10-26T11:27:11 avdg2 thanks for the info 2011-10-26T11:27:27 In the github repo the ants.js file is complete 2011-10-26T11:27:28 *** Six has quit IRC (Ping timeout: 265 seconds) 2011-10-26T11:27:29 can a dead ant be at the same place as a live ant in the input? 2011-10-26T11:27:30 thanks! 2011-10-26T11:27:50 i.e. an ant dies the turn before someone moves into its square 2011-10-26T11:28:04 garf: in theory yes, but I have to check the order of the engine first 2011-10-26T11:28:08 let me check 2011-10-26T11:28:26 doesn't seem possible though 2011-10-26T11:28:31 with attack range 2 2011-10-26T11:28:40 no, it seems its not possible 2011-10-26T11:28:41 there's no way it would get in range before dying 2011-10-26T11:28:43 first step is move 2011-10-26T11:28:52 I think 2011-10-26T11:29:01 *** Areks|2 has quit IRC (Ping timeout: 240 seconds) 2011-10-26T11:29:01 so hills is the only thing that can overlap with anything else 2011-10-26T11:29:09 at least if I can believe http://aichallenge.org/specification.php#Turns-and-Phases 2011-10-26T11:29:25 well at the current attack radius of 2.?? 2011-10-26T11:29:39 that means the safest distance would be A..B 2011-10-26T11:29:40 two dead ants of the same team can overlap 2011-10-26T11:29:54 oh yeah of course :-) 2011-10-26T11:30:12 hello? am i in ignore? 2011-10-26T11:30:13 overlooked that 2011-10-26T11:30:24 how long do dead ants last? 2011-10-26T11:30:25 aavikkojootti: sorry? 2011-10-26T11:30:33 yes but if an ant is dead on a tile there can't be a live one there? 2011-10-26T11:30:34 *** fercho has quit IRC (Quit: Page closed) 2011-10-26T11:30:36 did you guys saw my last message? 2011-10-26T11:30:41 i just registered 2011-10-26T11:30:42 aavikkojootti yes 2011-10-26T11:30:45 ok 2011-10-26T11:30:59 * avdg2 didn't read it 2011-10-26T11:31:09 Garf you only see them just as they die. 2011-10-26T11:31:37 then how can 2 dead ants be on the same square? 2011-10-26T11:31:46 suicide 2011-10-26T11:31:57 ah, a collision 2011-10-26T11:32:01 ok, I wont worry about that 2011-10-26T11:32:13 yeh cuz it should never happen :) 2011-10-26T11:32:37 unless there is a noob bot 2011-10-26T11:33:02 (and a good bot, but that doesn't account 2 dead bots on the same location, so it crashes :o- 2011-10-26T11:33:06 ) 2011-10-26T11:33:43 I wonder if this really happens though 2011-10-26T11:33:53 i don't think you see enemy dead ants. 2011-10-26T11:33:54 ? 2011-10-26T11:34:13 if its in your radius, then yes 2011-10-26T11:34:30 *** Larose has joined #aichallenge 2011-10-26T11:34:35 ah ok 2011-10-26T11:34:47 that's probably why my ant tracker screws up! 2011-10-26T11:34:48 lol 2011-10-26T11:36:31 *** cobracom has joined #aichallenge 2011-10-26T11:36:32 *** canadiancow|work has joined #aichallenge 2011-10-26T11:36:45 its scary how good my ants are doing (relative to other people) considering they dont know what a wall is 2011-10-26T11:37:01 meh, my fails to track ants 2011-10-26T11:37:07 it stores undefined objects 2011-10-26T11:38:31 *** Knekkebjoern has quit IRC (Quit: Leaving.) 2011-10-26T11:39:09 good plan, bad execution I'm afraid 2011-10-26T11:40:24 canadiancow|work: try a maze map ;) 2011-10-26T11:41:40 oh im not saying mine are good at all 2011-10-26T11:44:12 *** theonewolf has quit IRC (Ping timeout: 255 seconds) 2011-10-26T11:44:50 *** theonewolf has joined #aichallenge 2011-10-26T11:44:54 there is a whole lot of time outs :D, time for a compiled language. http://aichallenge.org/visualizer.php?game=16020&user=4677 2011-10-26T11:45:23 why on earth does python store dead ants in a map ? 2011-10-26T11:45:23 :-) 2011-10-26T11:45:34 starter kit i mean 2011-10-26T11:45:43 again a case of "who do you expect who will be the winner?" 2011-10-26T11:46:16 oh cuz its using it to zero the map 2011-10-26T11:49:06 *** dmiles_afk has quit IRC (Ping timeout: 258 seconds) 2011-10-26T11:52:03 *** smiley1983 has joined #aichallenge 2011-10-26T11:52:26 guys, is there a way to tell how many ants are waiting in my hills? (or in my enemy's hills?) 2011-10-26T11:53:00 you could keep track of how much food you collect 2011-10-26T11:54:19 *** JamesMG_ has joined #aichallenge 2011-10-26T11:54:46 well, you can estimate it 2011-10-26T11:55:03 there is always the risk of getting spawned food 2011-10-26T11:55:07 anyone knows if the server sends my own hills every round or only if they are in sight? 2011-10-26T11:55:33 and there's a problem that a food can be destroyed ... 2011-10-26T11:55:53 well, thats if the food is outside sight or you don't care about enemies 2011-10-26T11:56:16 no, I mean that both me and my enemy are both eligible to eat the food 2011-10-26T11:56:25 and thus it is destroyed 2011-10-26T11:56:33 http://aichallenge.org/forums/viewtopic.php?f=25&t=1559&p=9631&hilit=visualizer#p9627 this seems like my problem. where firefox used? 2011-10-26T11:56:50 (and hills can also eat food :-)) 2011-10-26T11:56:57 really? 2011-10-26T11:57:04 *** Knekkebjoern has joined #aichallenge 2011-10-26T11:57:07 ant spawned, eats food 2011-10-26T11:57:27 I think the game should have given us that information (at least for our own hills) 2011-10-26T11:57:37 read the specification ;-) 2011-10-26T11:57:40 or the source 2011-10-26T11:57:48 *** JamesMG has quit IRC (Read error: Connection reset by peer) 2011-10-26T11:57:51 http://aichallenge.org/specification.php 2011-10-26T11:58:05 it's not there, as far as I can tell 2011-10-26T11:58:49 *** olexs has joined #aichallenge 2011-10-26T11:59:08 I think there are also a lot of thing caused by the specification, but not necessarily written 2011-10-26T11:59:15 cobra, in the significant majority of cases, all except for 0-1 players, should have zero food in their hive 2011-10-26T11:59:21 *** aavikkojootti has left #aichallenge 2011-10-26T11:59:43 *** Vivitron has quit IRC (Ping timeout: 248 seconds) 2011-10-26T11:59:55 if you have more it shouldn't matter what you do in the most cases 2011-10-26T12:00:00 *** dmiles_afk has joined #aichallenge 2011-10-26T12:00:41 *** bobbby_ has joined #aichallenge 2011-10-26T12:00:47 *** Saulzar has joined #aichallenge 2011-10-26T12:01:22 well, maybe I just like to complain :) 2011-10-26T12:01:35 *** smiley1983 has quit IRC (Ping timeout: 252 seconds) 2011-10-26T12:01:42 I really enjoyed Planet Wars and have hard time adjusting 2011-10-26T12:02:01 who not? :-) (without shouting complains included) 2011-10-26T12:02:33 *** bobbby_ has quit IRC (Client Quit) 2011-10-26T12:02:55 cobra, I'd say for enemy hills, instead of trying to predict how many food reserves it has.. predict the proobability that an ant will spawn there in the next round(s) 2011-10-26T12:04:44 well i solved my problem. 2011-10-26T12:04:47 *** barsandcat has quit IRC (Ping timeout: 265 seconds) 2011-10-26T12:04:52 hopefully 2011-10-26T12:05:05 i may finally rest without feeling useless 2011-10-26T12:05:06 lol 2011-10-26T12:05:10 see you all tomorrow 2011-10-26T12:05:24 *** bobbydroptable has left #aichallenge 2011-10-26T12:05:49 burny - good idea, I don't have a clue how to implement it :). 2011-10-26T12:06:47 well, the easiest methods I can think of are.. 2011-10-26T12:07:03 measure your own ant per turn ratio.. and assume your enemy is doing similar 2011-10-26T12:07:14 or, have ants in range for awhile, and count 2011-10-26T12:07:38 *** sunshowers has joined #aichallenge 2011-10-26T12:08:43 yep, those were my immediate ideas as well 2011-10-26T12:10:01 *** analyst74 has joined #aichallenge 2011-10-26T12:13:56 *** SharkMonkey has joined #aichallenge 2011-10-26T12:16:43 *** muha has joined #aichallenge 2011-10-26T12:16:47 *** JamesMG has joined #aichallenge 2011-10-26T12:17:06 *** Blkt has quit IRC (Quit: cya) 2011-10-26T12:17:08 *** ztfw has joined #aichallenge 2011-10-26T12:17:28 . 2011-10-26T12:19:49 *** JamesMG_ has quit IRC (Ping timeout: 240 seconds) 2011-10-26T12:20:19 I think that this might be all I do on my ant for a while 2011-10-26T12:20:28 I have gotten really behind in life, watching antss :-( 2011-10-26T12:20:37 [btw, is there a way to still access the old Planet Wars games (at least some of them)? Wanted to show it to a friend ...] 2011-10-26T12:23:12 My ant still only take 3 ms per move max 2011-10-26T12:23:24 do the ants move? 2011-10-26T12:23:31 McLeopold: ? 2011-10-26T12:24:40 *** Areks|2 has joined #aichallenge 2011-10-26T12:27:48 *** sunshowers has quit IRC (Quit: Page closed) 2011-10-26T12:28:27 *** smiley1983 has joined #aichallenge 2011-10-26T12:31:19 *** mviel has quit IRC (Remote host closed the connection) 2011-10-26T12:35:51 *** smiley1983 has quit IRC (Quit: leaving) 2011-10-26T12:35:54 *** Garf has quit IRC (Read error: Connection reset by peer) 2011-10-26T12:36:49 *** Garf has joined #aichallenge 2011-10-26T12:40:48 *** Puj has quit IRC (Ping timeout: 240 seconds) 2011-10-26T12:48:40 *** posthuman0 has joined #aichallenge 2011-10-26T12:50:19 *** moi has joined #aichallenge 2011-10-26T12:51:40 *** cobracom has quit IRC (Ping timeout: 265 seconds) 2011-10-26T12:56:48 *** moi has quit IRC (Quit: Page closed) 2011-10-26T13:00:59 *** posthuman0 has quit IRC (Quit: Page closed) 2011-10-26T13:01:58 *** xar0l has joined #aichallenge 2011-10-26T13:03:10 are there any flyers for this that i could put up in the local college computer department? 2011-10-26T13:03:39 *** Migi32 has joined #aichallenge 2011-10-26T13:04:30 *** Macuyiko has joined #aichallenge 2011-10-26T13:05:24 hi. Is c++0x supported? Are there any plans for supporting it? 2011-10-26T13:08:35 *** analyst74 has quit IRC (Ping timeout: 276 seconds) 2011-10-26T13:15:48 Migi32: you only have to send binaries right? 2011-10-26T13:16:06 oh has that changed since the last contest? 2011-10-26T13:16:22 avdg2: no, you send your code 2011-10-26T13:16:26 oh 2011-10-26T13:16:29 the server compiles and runs it 2011-10-26T13:16:45 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-10-26T13:16:56 ah, didn't know that 2011-10-26T13:17:02 *** dvladim has joined #aichallenge 2011-10-26T13:17:05 *** g0llum has joined #aichallenge 2011-10-26T13:17:43 olexs: if you aren't sure, just try it 2011-10-26T13:17:59 it shouldn't hurt 2011-10-26T13:20:05 *** analyst74 has joined #aichallenge 2011-10-26T13:21:09 *** TerryVB has quit IRC (Ping timeout: 265 seconds) 2011-10-26T13:22:22 for python the code is the binaries ;-) 2011-10-26T13:23:19 I think I am additcted to ants. 2011-10-26T13:23:36 I am behind in all of my work, because I spend all of my time thinking about my ants 2011-10-26T13:24:08 *** pipeep has quit IRC (Ping timeout: 245 seconds) 2011-10-26T13:24:09 I don't even have one of the top ants! 2011-10-26T13:25:17 XD 2011-10-26T13:26:23 f4hy i know what you mean.... ;) 2011-10-26T13:26:48 For the C++ one you submit a Makefile though right? 2011-10-26T13:27:05 yup 2011-10-26T13:28:43 So if gcc 4.5 supports 0x .. yes 2011-10-26T13:30:37 oh, yeah if you can submit a makefile then it's no problem 2011-10-26T13:30:48 *** g0llum has quit IRC (Ping timeout: 240 seconds) 2011-10-26T13:30:50 it's just a -std=c++0x option in gcc 2011-10-26T13:32:14 *** levis501 has joined #aichallenge 2011-10-26T13:33:42 *** g0llum has joined #aichallenge 2011-10-26T13:39:09 *** sunshowers has joined #aichallenge 2011-10-26T13:44:27 *** brainfarm has quit IRC (Quit: Page closed) 2011-10-26T13:44:43 *** seth1010 has quit IRC (Quit: Page closed) 2011-10-26T13:44:48 *** dvladim has quit IRC (Ping timeout: 255 seconds) 2011-10-26T13:47:03 *** sunshowers has quit IRC (Quit: Page closed) 2011-10-26T13:47:46 *** bysin has quit IRC (Read error: Connection reset by peer) 2011-10-26T13:49:58 aichallenge: McLeopold epsilon * rbce8b0a / website/sql.php : limit sql fields for memcache storage - http://git.io/3qiyqA 2011-10-26T13:53:28 *** amstan has joined #aichallenge 2011-10-26T13:53:28 *** ChanServ sets mode: +o amstan 2011-10-26T13:53:31 *** Redgis has joined #aichallenge 2011-10-26T13:56:10 *** levis501 has quit IRC (Quit: levis501) 2011-10-26T13:58:41 *** GeorgeSebastian has joined #aichallenge 2011-10-26T14:00:02 i'm starting to get into the fun stuff after writing my own gamerunner/parsers 2011-10-26T14:00:52 *** GeorgeSebastian has quit IRC (Max SendQ exceeded) 2011-10-26T14:02:25 *** oscar has joined #aichallenge 2011-10-26T14:02:27 amstan: there are a lot of incomplete redirection target errors on the website related to the forums 2011-10-26T14:02:44 McLeopold: which ones? 2011-10-26T14:02:50 there's only one redirect 2011-10-26T14:04:32 just look at the logs/error.log 2011-10-26T14:04:54 McLeopold: btw, are the game servers 32-bit or 64-bit? 2011-10-26T14:05:11 64 2011-10-26T14:05:59 ouch :( 2011-10-26T14:06:55 *** oscar has quit IRC (Client Quit) 2011-10-26T14:06:57 McLeopold: they're not really errors, just warnings: http://enc.com.au/2011/03/apache-and-incomplete-redirection-messages/ 2011-10-26T14:07:10 i added the absulute stuff so it stops 2011-10-26T14:08:05 *** dwins has quit IRC (Ping timeout: 252 seconds) 2011-10-26T14:08:34 *** tncardoso has quit IRC (Quit: Leaving) 2011-10-26T14:08:53 *** dwins has joined #aichallenge 2011-10-26T14:13:53 it feels good when you write a long algorithm/function, and the only errors when you finally test are typos and other trivial things 2011-10-26T14:14:49 how well have genetic algorithms done in past challenges? Anyone know? 2011-10-26T14:16:09 speaking of past challenges, how many people have published their source code at the end? particularly among top bots 2011-10-26T14:16:54 http://aichallenge.org/visualizer.php?game=17360 deadlock! 2011-10-26T14:18:36 *** GeorgeSebastian has joined #aichallenge 2011-10-26T14:18:51 aichallenge: McLeopold epsilon * rfebfdc7 / worker/worker.py : delete __MACOSX folders from archives - http://git.io/yDG42w 2011-10-26T14:21:14 are there any maps with un-explorable blocks? i.e. surrounded by so much water they're too far to see from anywhere 2011-10-26T14:21:23 cyphase: no 2011-10-26T14:21:28 maybe 2011-10-26T14:21:32 where cannot be any island 2011-10-26T14:21:38 according to game spec 2011-10-26T14:21:40 water can be in there 2011-10-26T14:21:42 doesnt have to be an island 2011-10-26T14:21:54 ah 2011-10-26T14:22:01 If there is a path then it is explorable 2011-10-26T14:22:27 if you have a large square of water 15x15, the center can never be explored 2011-10-26T14:22:34 SharkMonkey, i mean if there is no path? like imagine a 20*20 square of water 2011-10-26T14:22:42 or yea, 15*15 would do it 2011-10-26T14:22:44 Oh, I see 2011-10-26T14:22:51 *** dvladim has joined #aichallenge 2011-10-26T14:23:14 oops, added an unnecessary question mark there. i do that on occasion :P 2011-10-26T14:23:16 er 17*17 with radius2=77 2011-10-26T14:24:35 @later tell janzert some mac users were having compile issues, I believe because of the __MACOSX folder. Can you review the change? 2011-10-26T14:24:35 McLeopold: Job's done. 2011-10-26T14:24:41 *** cybsy has joined #aichallenge 2011-10-26T14:25:07 How is the job done, contestbot? 2011-10-26T14:28:48 *** guest___ has joined #aichallenge 2011-10-26T14:30:15 SharkMonkey: as in, he remembered what he needs to do 2011-10-26T14:31:56 *** thestinger has joined #aichallenge 2011-10-26T14:32:00 *** kiv has quit IRC (Ping timeout: 240 seconds) 2011-10-26T14:32:46 *** ZiyaoWei has joined #aichallenge 2011-10-26T14:33:55 *** kiv has joined #aichallenge 2011-10-26T14:34:51 Yeah. But I wouldn't accept that response from fedex when I drop of a package. :) 2011-10-26T14:35:08 Should I be reading that with the voice of a warcraft peon in my head? 2011-10-26T14:35:53 wow, my crappy bot is somehow doing well 2011-10-26T14:35:55 :\ 2011-10-26T14:36:38 By that you mean beating all the starter package bots and the ones timing out? 2011-10-26T14:36:57 I guess :) 2011-10-26T14:36:58 *** irchs has joined #aichallenge 2011-10-26T14:37:15 I think it just got lucky 2011-10-26T14:38:05 *** Kingpin13 has quit IRC (Quit: quit) 2011-10-26T14:38:13 SharkMonkey: yes, you should be 2011-10-26T14:41:01 One of my bots that just collected food and went north(Always north). Did fairly well. Made it up to rank 77 at least, before trending back downard. 2011-10-26T14:41:33 On some maps it would create this hard to penetrate vertical wall of ants 2011-10-26T14:42:19 Anyone else having problems with the starter Java Package? 2011-10-26T14:42:34 guest___: problems how? 2011-10-26T14:42:38 duh, after reupload i'm on 125 place... i wonder if it can climb back to 30 2011-10-26T14:42:56 Fluxid: theoretically 2011-10-26T14:43:07 It seems to be crashing on the first turn. 2011-10-26T14:43:16 *** craklyn has joined #aichallenge 2011-10-26T14:43:19 amstan: i wonder mostly because i;m not sure if it is really better 2011-10-26T14:43:27 oh, ok 2011-10-26T14:44:00 I've got the latest tools and package as well. 2011-10-26T14:45:07 ok, any hints on how to predict focus attack on neighbouring cells and predicting if enemy has larger weakness than my ant(s) around them? 2011-10-26T14:45:28 *** racko has quit IRC (Quit: Page closed) 2011-10-26T14:45:36 *** chris_p has joined #aichallenge 2011-10-26T14:45:56 Fluxid, can you rephrase the question? 2011-10-26T14:47:03 my english is poor... i want to know if my ant goes, for example, west, will it be weaker than enemy that is in the same direction 2011-10-26T14:47:25 if there are two my ants and one enemy, i want to attack, if other way around, to flee 2011-10-26T14:47:37 right 2011-10-26T14:48:17 so, any hints? :) 2011-10-26T14:48:23 hmm. i haven't gotten there yet, but.. 2011-10-26T14:48:37 depends how smart you want it to be 2011-10-26T14:48:38 Flxuid: Look into the idea of "Min-Max" 2011-10-26T14:49:02 Fluxid: You are probably concerned with something like this - http://www.ocf.berkeley.edu/~yosenl/extras/alphabeta/alphabeta.html 2011-10-26T14:49:57 cool, thanks. i'll look into it 2011-10-26T14:50:35 i guess the top bots implemented it, they behave in similar way when around an enemy 2011-10-26T14:50:44 You predict the way the board might look in the future and you calculate the "value" of those future boards based on some rule you make, or heuristic. You then choose the board which is optimal for you. 2011-10-26T14:51:26 Hopefully what I said makes sense. I can try to rephrase it if you need. 2011-10-26T14:51:55 *** ybit has joined #aichallenge 2011-10-26T14:52:28 i think i get it 2011-10-26T14:53:01 *** mleise has quit IRC (Ping timeout: 240 seconds) 2011-10-26T14:53:34 but now, how to predict enemy's moves... 2011-10-26T14:54:54 Fluxid: with minmax/alphabeta you already do that. you take into account all possible moves you and your enemy can do and choose an optimal (for you, that is) strategy, even if the enemy does "the perfect move"(tm) 2011-10-26T14:55:06 *** frost_ has joined #aichallenge 2011-10-26T14:55:09 McLeopold: I don't think that will hurt, but I'm also having a hard time imagining what it will help 2011-10-26T14:55:20 what sort of errors were being seen? 2011-10-26T14:55:47 *** isbric_ has joined #aichallenge 2011-10-26T14:56:39 hi, does anyone know if im allowed to save data to a textfile in my directory of each match? 2011-10-26T14:56:43 bqf: thanks 2011-10-26T14:56:50 no, you are not 2011-10-26T14:56:51 isbric_: you are not allowed to write data 2011-10-26T14:57:16 janzert: no errors 2011-10-26T14:57:26 so there is basicly no way of keeping information from game to game? 2011-10-26T14:57:31 or, no error messages, just failed compiles 2011-10-26T14:57:43 correct 2011-10-26T14:58:19 *** jcdjcd has quit IRC (Ping timeout: 252 seconds) 2011-10-26T14:58:45 thus, removing the "I" from "AI", but meh, it's still basically the same principal 2011-10-26T14:59:06 burny: is it posible to connect to some external web or such? 2011-10-26T14:59:16 no 2011-10-26T14:59:18 no 2011-10-26T14:59:23 that would make cheating be pretty trivial 2011-10-26T14:59:27 learning must be done offline 2011-10-26T14:59:37 ah :/ 2011-10-26T14:59:46 *** McLeopold has quit IRC (Read error: Connection reset by peer) 2011-10-26T14:59:52 no, you can make an artificial intelligence without needing to record data 2011-10-26T14:59:53 or id have to implement a realy fast lerning alg :P 2011-10-26T14:59:59 it is just forgetting each time 2011-10-26T15:00:08 Zannick: ^^ 2011-10-26T15:00:10 alternately, run the learning algorithm yourself and save it 2011-10-26T15:00:22 you can supply data files, you just can't write to them 2011-10-26T15:00:30 *** Saulzar has quit IRC (Ping timeout: 260 seconds) 2011-10-26T15:00:31 zannick: doing that, makes your bot extremly vulnerable to last minute changes 2011-10-26T15:00:34 ofc, the hard part is finding good competition for the training process 2011-10-26T15:00:47 that's what the tcp servers are for 2011-10-26T15:00:50 burny: not necessarly 2011-10-26T15:01:02 burny: i don't know what you mean 2011-10-26T15:01:16 all bots are vulnerable to "last minute changes" 2011-10-26T15:01:27 burny: you could calculate all possible arrangements of ants in a small area with info on what your optimal move is 2011-10-26T15:01:32 burny: store that all in files 2011-10-26T15:01:41 *** ZiyaoWei has quit IRC (Ping timeout: 265 seconds) 2011-10-26T15:01:47 bots that aren't tuned to their currently existing enemys, aren't susceptible to any changes, let alone last minute changes 2011-10-26T15:01:54 anyone that reads russian around? 2011-10-26T15:02:20 janzert: google translate? 2011-10-26T15:02:39 well, then don't have your learning algorithm tune to specific opponents 2011-10-26T15:02:46 amstan: yeah, the result is ambiguous 2011-10-26T15:03:00 Zannick: tcp servers? 2011-10-26T15:03:05 the small lookup is not a bad idea, similar to chess endgame tables 2011-10-26T15:03:07 *** mcstar has joined #aichallenge 2011-10-26T15:03:08 *** savaron has quit IRC (Ping timeout: 265 seconds) 2011-10-26T15:03:11 @tcp 2011-10-26T15:03:12 Zannick: tcp could be http://ants.fluxid.pl/howto. 2011-10-26T15:03:29 *** Bluedgis has quit IRC (Quit: leaving) 2011-10-26T15:03:47 Are the radii rounded down or up? 2011-10-26T15:04:10 they aren't rounded 2011-10-26T15:04:18 don't use sqrt 2011-10-26T15:04:22 I think our registration letter got hung up in an anti-spam system and this is instructions for it to be delivered :/ http://pastebin.com/kd1sueGm 2011-10-26T15:04:46 So effectively up then then? 2011-10-26T15:05:04 janzert: well, i think we can ignore it 2011-10-26T15:05:20 Zannick: What then? 2011-10-26T15:05:44 if x and y are the axis distances in question 2011-10-26T15:05:45 unless we feel really nice and find someone that speaks russian to confirm on what exactly we have to do 2011-10-26T15:06:14 instead of comparing sqrt(x^2 + y^2) to sqrt(r^2), just compare (x^2+y^2) to r^2 2011-10-26T15:06:17 well I think mail.ru might be fairly large russian email host so if we just need to reply with a letter and then we'll be good to get past their spam system it would be nice 2011-10-26T15:06:29 remember, the radius values you are given are r^2, not r. 2011-10-26T15:06:54 I'll try and catch UncleVasya later 2011-10-26T15:07:03 janzert: such antispam features can't expect all automated emails to conform to their instructions 2011-10-26T15:07:22 i think he's ukrainian 2011-10-26T15:07:43 nvm 2011-10-26T15:07:54 yes, I do to :) 2011-10-26T15:07:54 Zannick: You raise a valid point 2011-10-26T15:10:51 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-10-26T15:10:56 @later tell antimatroid Here's a question for you: http://aichallenge.org/forums/viewtopic.php?f=24&t=1599 2011-10-26T15:10:56 amstan: I think that worked... 2011-10-26T15:12:18 @later tell McLeopold the thing I saw in macosx zips that would likely still be causing compilation failures is the inclusion of __SourceFile.java files (source files with __ prefixed) that looked to be coming from an IDE or editor as backup files. I don't think we should do anything to try and automatically handle these though since it seems too likely to also catch a legitimate file (e.g. __init__.py) 2011-10-26T15:12:18 janzert: Ready to serve, my lord. 2011-10-26T15:13:15 who is behind that tcp server? 2011-10-26T15:13:27 isbric_: code or instance? 2011-10-26T15:13:39 instance 2011-10-26T15:13:53 ants.fluxid.pl is mine, code from https://github.com/Ashod/ants-tcp 2011-10-26T15:17:37 *** yohann has joined #aichallenge 2011-10-26T15:17:41 hello 2011-10-26T15:17:43 how meny uses it? and how hard culd i be on it? 2011-10-26T15:19:05 Fluxid: ^ 2011-10-26T15:19:23 isbric_: what do you mean? 2011-10-26T15:19:32 *** jcdny has joined #aichallenge 2011-10-26T15:19:38 for sure it's better to test bot there than with randombots and hunterbots 2011-10-26T15:19:43 i use the python starter pack, and followed all the strategy tutorial, and uploaded my code, my bot is doing a lot of timeout, is is normal ? 2011-10-26T15:20:05 yohann: no, it is bad 2011-10-26T15:20:09 there is no A* nor anything time consuming yet 2011-10-26T15:20:16 the way the python tutorial gets you to deal with unseen locations is pretty bad 2011-10-26T15:20:22 yohann: maybe you have a bug 2011-10-26T15:21:19 fluxid, i juste fallowed strategy guide step by step and get the same result as in the example.map 2011-10-26T15:21:24 it calculates the distance from each ant to each unseen location (most of the map at the start), keeps it all in a list, then sorts it 2011-10-26T15:21:33 *** bhasker has joined #aichallenge 2011-10-26T15:21:47 thestinger: so that is the time consuming thing 2011-10-26T15:21:54 *** Accoun has quit IRC () 2011-10-26T15:22:13 yeah, I had the same problem when I was still using my original python bot 2011-10-26T15:22:47 I thought it was my graph searches doing it, but it was the code for sending ants to unseen squares 2011-10-26T15:23:22 thank you, i think i need to fix that before going into more interesting things then. 2011-10-26T15:24:10 *** janzert has quit IRC (Read error: Connection reset by peer) 2011-10-26T15:24:24 I have similar code in my current bot, but it's C++ so I can get away with it - I also remove unseen locations from the list if pathfinding determines they are unreachable (water squares in the middle of a pond) 2011-10-26T15:24:38 i think i'll start by calculate the nearest unseen square from each unordered ants 2011-10-26T15:24:55 it should save a lot of time 2011-10-26T15:24:56 *** michaelpjohnson has joined #aichallenge 2011-10-26T15:25:30 *** danielharan has joined #aichallenge 2011-10-26T15:25:31 Fluxid: do you have a rate limit of the games or is it queued like the real thing? 2011-10-26T15:26:03 isbric_: no, no limits 2011-10-26T15:26:13 you get into game immediately after you connect 2011-10-26T15:26:19 fifo 2011-10-26T15:26:35 is there any risk of overloading your server? 2011-10-26T15:26:38 *** epicmonkey has joined #aichallenge 2011-10-26T15:26:57 eventually you need to wait until game finishes when game queue is full (5 games at once) 2011-10-26T15:27:01 isbric_: none 2011-10-26T15:27:12 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-26T15:27:17 your bot runs locally on your computer when you use the tcp server 2011-10-26T15:27:19 how good is the tcp? 2011-10-26T15:27:25 does it use the official engine? 2011-10-26T15:27:31 isbric_: most load is on your side, server eats as much as nothing 2011-10-26T15:28:00 Fluxid: ok, but i can do a maximum of 5 simutanius games? 2011-10-26T15:28:02 yes, but is not a replacement 2011-10-26T15:28:29 of game server where the official ranking is 2011-10-26T15:28:49 i didnt imply that it is 2011-10-26T15:29:07 im talking about the quality of the code 2011-10-26T15:30:43 mcstar: quality of code should not matter as long as the server is acting acording to specs 2011-10-26T15:30:47 the same engine, quality - sometimes it prints that Nonetype has no attrbute readline, but it works somehow 2011-10-26T15:31:17 "as long as the server is acting acording to specs" 2011-10-26T15:31:25 yes exactly 2011-10-26T15:31:29 Fluxid: thx 2011-10-26T15:31:46 *** TTE has joined #aichallenge 2011-10-26T15:32:19 ugh... there's some subtle bug in my A* implementation I think 2011-10-26T15:32:36 sometimes it can't find a path to reachable places 2011-10-26T15:32:43 very, very rarely though 2011-10-26T15:32:59 *** janzert has joined #aichallenge 2011-10-26T15:34:33 *** guest___ has quit IRC (Ping timeout: 265 seconds) 2011-10-26T15:35:26 does your own hill block your movements? 2011-10-26T15:35:33 no 2011-10-26T15:37:07 hmm I must be missing something that can result in moves being blocked... because the tools say a move is blocked where I'm certain it isn't 2011-10-26T15:37:11 Fluxid, i'm just playing with that tcp-code, it seems, the error, you're refering to comes from tcpserver.py, line 325, where there's no stderr output(no error) from jskills 2011-10-26T15:37:57 ARGH 2011-10-26T15:37:58 g0llum: no idea, didn't dive into code 2011-10-26T15:38:04 *** Migi32 has quit IRC (Quit: bye) 2011-10-26T15:38:10 food can block your movement 2011-10-26T15:38:14 I had problems with that 2011-10-26T15:38:20 while I did fix the row/column vs x/y conversion in my code... I swapped east and west 2011-10-26T15:38:29 Fluxid: how often do you update the rankings? 2011-10-26T15:38:32 thestinger: yes 2011-10-26T15:38:38 food blocks movement 2011-10-26T15:39:02 isbric_: i don't update anything :) 2011-10-26T15:39:21 it's all done by the ghost in the machine 2011-10-26T15:40:05 ghost in the shell! 2011-10-26T15:40:33 Fluxid: then let me just ask how the score output is calculated? 2011-10-26T15:40:34 it'd done by lonely unfreed memory from killed zombie processes 2011-10-26T15:40:38 *** Accoun has joined #aichallenge 2011-10-26T15:41:09 isbric_: lol, no idea :D the name of algorithm is trueskill, it is the same which used in official ranking 2011-10-26T15:41:16 well, I figured out where the bug is in my pathfinding :) 2011-10-26T15:41:44 lol, the cursor is blinking halfway through an unfinished line of code because i got distracted by researching algorithms 2011-10-26T15:43:45 it's like giving up on dinner halfway because you need to read more recipes 2011-10-26T15:44:26 reading recipes definitely beats eating 2011-10-26T15:45:39 food == '*' == ascii(42) 2011-10-26T15:47:30 *** aslvrstn has joined #aichallenge 2011-10-26T15:47:42 *** dvladim has quit IRC (Read error: Operation timed out) 2011-10-26T15:50:14 *** shlhu has joined #aichallenge 2011-10-26T15:50:36 Quick question, are there any limits on group entries? 2011-10-26T15:50:54 *** sunshowers has joined #aichallenge 2011-10-26T15:51:56 identical to non-group entries I think 2011-10-26T15:51:56 Fluxid: how long does a run of 1000 turn usualy take? 2011-10-26T15:52:01 ie. only 1 account 2011-10-26T15:53:24 *** TTE has quit IRC (Ping timeout: 265 seconds) 2011-10-26T15:54:05 isbric_: it depends on connection, but my guess is as long as locally +10% 2011-10-26T15:54:25 i didn't go as far as 1000 turns, and didn't time it, just left it working 2011-10-26T15:54:27 *** shlhu has quit IRC (Client Quit) 2011-10-26T15:56:25 so how wuld i get my score back from the run? the last output is score 2 while the scoreboard says my skill is > 20.0 2011-10-26T15:57:41 *** onensora has joined #aichallenge 2011-10-26T15:57:47 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-10-26T15:58:09 *** g0llum has joined #aichallenge 2011-10-26T16:00:30 *** sunshowers has quit IRC (Quit: Page closed) 2011-10-26T16:01:14 *** janzert has quit IRC (Read error: Connection reset by peer) 2011-10-26T16:01:52 score != skill 2011-10-26T16:02:44 ok, then i think i might need some better code sins i only get score 1 :P 2011-10-26T16:03:18 *** janzert has joined #aichallenge 2011-10-26T16:14:53 so there's no constraints atall on what the varius radius might be? 2011-10-26T16:16:31 burny: its unlikely to change much 2011-10-26T16:16:50 like its not going to be made 1000 or something 2011-10-26T16:17:02 *** ztfw has quit IRC (Remote host closed the connection) 2011-10-26T16:17:03 *** chris_p has quit IRC (Quit: Page closed) 2011-10-26T16:17:19 i think,y ou can look into, what's used in the contest at http://aichallenge.org/game_settings.php 2011-10-26T16:17:29 *** Gazillo has joined #aichallenge 2011-10-26T16:17:37 I would go so far as to say that even if it does change it will be comparable to the setting today 2011-10-26T16:18:12 *** Gazillo has quit IRC (Client Quit) 2011-10-26T16:18:33 if it does change, it can potentially have huge impact on bot design though 2011-10-26T16:21:10 *** doricatron has joined #aichallenge 2011-10-26T16:21:36 *** levis501 has joined #aichallenge 2011-10-26T16:21:41 hi admins ... I have a question: We are multiple people who want to make 1 team. How do I add multiple people on 1 account? 2011-10-26T16:25:06 doricatron: everyone just shares the one account and you can list the team members in the bio (About Me) area 2011-10-26T16:25:29 *** xathis_ has joined #aichallenge 2011-10-26T16:25:56 How well have genetic algorithm approaches done in past challenges? Anyone know? 2011-10-26T16:26:14 *** Savaron has joined #aichallenge 2011-10-26T16:26:15 a good bot shouldn't be that dependent on those particular numbers :P 2011-10-26T16:26:25 isn't there a mistake in the pyton tutoral bot? 2011-10-26T16:26:35 http://aichallenge.org/ants_tutorial_step_5.php 2011-10-26T16:26:48 what mistake? 2011-10-26T16:26:58 for dist, ant_loc in ant_dist: 2011-10-26T16:26:58 do_move_location(ant_loc, hill_loc) 2011-10-26T16:27:10 hill_loc should be stored in the sorted array and unpacked here again 2011-10-26T16:27:29 I guess its ok because the tutorial only has 1 hill 2011-10-26T16:28:24 *** xathis has quit IRC (Ping timeout: 240 seconds) 2011-10-26T16:28:24 well, for instance.. if attackradius remains as is.. it's stronger, and faster, to manually solve the combat patterns, but if it has a range, or will change, then manually doing it would be a waste of time 2011-10-26T16:29:04 *** irchs has quit IRC (Quit: irchs) 2011-10-26T16:29:25 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-10-26T16:30:28 *** doricatron has quit IRC (Quit: Page closed) 2011-10-26T16:31:08 sure, but that shouldn't require a rewrite of the bot 2011-10-26T16:33:20 hypothetically.. people don't have time to make changes during the last days/weeks 2011-10-26T16:33:32 or, don't want to write stuff twice 2011-10-26T16:33:43 *** levis501 has left #aichallenge 2011-10-26T16:34:08 or, even though the code doesn't need to be redone, it completely changes the balance numbers, that the rest of the logic is using 2011-10-26T16:34:09 etc. 2011-10-26T16:34:34 *** _ulises <_ulises!~ulises@cmi/researcher/ulises> has joined #aichallenge 2011-10-26T16:34:40 *** dinosoep has joined #aichallenge 2011-10-26T16:34:42 <_ulises> evening 2011-10-26T16:35:34 <_ulises> are there any admins of the website about? I'm having issues loging in (even though I have my password saved in my browser, etc.) 2011-10-26T16:35:51 I'm having problems with my bot 2011-10-26T16:36:32 I think the mistake is in the starter package and not my bot cause I've been looking a damn long time and can't spot the mistake 2011-10-26T16:37:00 have you tried submittig the starter as-is? 2011-10-26T16:37:16 no 2011-10-26T16:37:24 but everything was working great 2011-10-26T16:37:25 *** Macuyiko has quit IRC (Ping timeout: 240 seconds) 2011-10-26T16:37:27 but now I get more ants 2011-10-26T16:37:29 *** irchs has joined #aichallenge 2011-10-26T16:37:38 and it seems after a lot of commands it suddenly crashes 2011-10-26T16:38:35 I'll show the part containing the problem, if that doesn't contain the problem then it must be in the starter 2011-10-26T16:40:53 have you tried to duplicate the problem and manually walk-through the failure? 2011-10-26T16:42:22 I'm not able to duplicate the problem, I'll try to make a very short example 2011-10-26T16:42:37 here is the code that when commented out doesn't make my bot crash:http://dl.dropbox.com/u/15024434/errors.d 2011-10-26T16:42:55 but there is no obvious thing wrong with it 2011-10-26T16:43:05 and it only crashes around 30 turns 2011-10-26T16:43:06 pastebin is probably easier for posting code 2011-10-26T16:43:27 you're timing out perhaps? 2011-10-26T16:43:38 thats the code in the examples that gets very slow on large maps/ants 2011-10-26T16:43:38 *** higgs has joined #aichallenge 2011-10-26T16:44:16 Alright, how the hell are the good ants so good at attacking? What are people using? min max? 2011-10-26T16:44:23 *** higgs is now known as thimbleBrain 2011-10-26T16:44:31 but when you time out it should say "timed out" and not "crashed"? 2011-10-26T16:44:40 or am I wrong? 2011-10-26T16:44:46 and here is pastebin:http://pastebin.com/UcK1e090 2011-10-26T16:45:07 *** Savaron has quit IRC (Ping timeout: 265 seconds) 2011-10-26T16:45:18 and it is indeed very suboptimal, I'm still learning :) 2011-10-26T16:46:18 aichallenge: McLeopold epsilon * rfc916ba / (ants/submission_test/TestBot.py ants/submission_test/ants.py): update submission test bot - http://git.io/L4bAVw 2011-10-26T16:47:58 I doubt such popular starter package would have serious bug (C++, is it?) 2011-10-26T16:48:00 *** ltriant has joined #aichallenge 2011-10-26T16:48:29 *** Puj has joined #aichallenge 2011-10-26T16:48:41 if you can't replicate the problem locally, try tcp server, it'll give you different opponent, and hopefully trigger the problem 2011-10-26T16:49:00 it's the d programming language 2011-10-26T16:49:10 it's not very popular and it's new :p 2011-10-26T16:49:23 ah 2011-10-26T16:49:24 had to steal the isvissible and unocupied method 2011-10-26T16:50:37 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2011-10-26T16:51:31 you can always check out the complete source : http://pastebin.com/KwJgf65n but I doubt it's very usefull. 2011-10-26T16:52:36 lol, my bot is crashing too much atm 2011-10-26T16:53:11 and the annoying thing is that it gives you no information whatsoever as to how it crashed :s 2011-10-26T16:53:22 You probably know way more than I do about D (which is nothing) :p 2011-10-26T16:53:33 its a language 2011-10-26T16:53:47 arh, now I know something about D! 2011-10-26T16:54:01 but I assume it would have some sort of debugging/logging technique 2011-10-26T16:54:09 yes, it has 2011-10-26T16:54:12 that you can use to do tracing 2011-10-26T16:54:22 but I never get an exception 2011-10-26T16:54:28 or something like that 2011-10-26T16:54:36 it just "crashed" 2011-10-26T16:55:32 when running locally? 2011-10-26T16:55:35 *** JamesMG_ has joined #aichallenge 2011-10-26T16:55:49 yes 2011-10-26T16:55:59 I've not uploaded a single thing yet 2011-10-26T16:57:16 *** irchs has quit IRC (Quit: irchs) 2011-10-26T16:57:45 *** JamesMG has quit IRC (Read error: Connection reset by peer) 2011-10-26T16:58:02 tried to wrap whole program in a try catch? 2011-10-26T16:58:12 no 2011-10-26T16:58:28 *** Palmik has quit IRC (Remote host closed the connection) 2011-10-26T16:58:33 I'll do it now 2011-10-26T16:58:45 *** sunshowers has joined #aichallenge 2011-10-26T16:58:46 oh, one thing, when running with the playgame.py 2011-10-26T16:58:53 dinosoep: when running locally i have a little script like: tee bot.in.log | dwinsbot | tee bot.out.log instead of just my bot 2011-10-26T16:58:59 Where is stderr redirected to when running test_bot.bat? 2011-10-26T16:59:06 so i can review stdin and stdout afterward 2011-10-26T16:59:16 yeah and stderr 2011-10-26T16:59:30 (which is useful for debug messages) 2011-10-26T16:59:53 there is an option in playgame.py that allows output error message, otherwise all errors will be swollen 2011-10-26T17:00:09 yeah? how do you make it spit out errors? 2011-10-26T17:00:15 use -E 2011-10-26T17:00:27 *** oddmunds has joined #aichallenge 2011-10-26T17:00:39 (you can check that by running playgame.py without parameters) 2011-10-26T17:00:55 thank you :) I'll also store my stdout/in but I don't understand a single thing of it's output :p 2011-10-26T17:01:11 -I will give you all the stdin 2011-10-26T17:01:13 dinosoep: learn the game specification 2011-10-26T17:01:31 *** Knekkebjoern has quit IRC (Quit: Leaving.) 2011-10-26T17:01:40 if you want a good bot, you better know all the layers involved ;-) 2011-10-26T17:01:44 avdg2, to much to learn in to less time :p I'll have a look at it 2011-10-26T17:01:54 *** Knekkebjoern has joined #aichallenge 2011-10-26T17:02:09 and I don't think I'll get a very good bot :) I'm still 16 xd 2011-10-26T17:02:17 16 is old enough 2011-10-26T17:02:33 yeah, if you're smart I guess :p 2011-10-26T17:02:51 Disadvantage: Incomplete education. Advantage: Less responsibilities. More time to devote to winning 2011-10-26T17:02:56 dinosoep: just build it and we'll see (we don't care about age so far right?) 2011-10-26T17:03:32 meh, I also have incomplete education, but there are personal reasons for that ;-) 2011-10-26T17:03:40 *** ajhager has quit IRC (Quit: ajhager) 2011-10-26T17:04:44 tried wrapping in try catch and running with -E but it still crashed 2011-10-26T17:04:58 -E doesn't prevent crashes ;-) 2011-10-26T17:05:06 no but should give useful output 2011-10-26T17:05:13 at least I hoped 2011-10-26T17:05:13 (oh, I wish I had a flag that build the bot for me) 2011-10-26T17:05:14 What language? 2011-10-26T17:05:17 D 2011-10-26T17:05:28 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-10-26T17:05:57 dinosoep: try to output some data in stderr 2011-10-26T17:05:57 Add having few people to help you when you have troubles to your disadvantage list :P 2011-10-26T17:06:41 x) it's indeed a big disadvantage but beeing more productive is a way bigger advantage :) 2011-10-26T17:06:41 why are the people in the chat for? 2011-10-26T17:06:48 isn't it for ask/give help? 2011-10-26T17:06:59 =) 2011-10-26T17:07:11 I just mean, I dont' think a lot of people know D.. 2011-10-26T17:07:20 So smaller pool of people who are qualified to help 2011-10-26T17:07:35 wow, i thought the scala list was short 2011-10-26T17:07:42 actually this channel shouldn't care about languages right? there are other channels for this ;-) 2011-10-26T17:07:42 http://aichallenge.org/language_profile.php?language=D 2011-10-26T17:07:54 I've yet to meet the first people programming in D for this challenge 2011-10-26T17:08:08 13 sounds good enough 2011-10-26T17:08:09 haha 2011-10-26T17:08:10 wait, they exist? 2011-10-26T17:08:51 at least, if they really want to build a bot with it 2011-10-26T17:08:52 :p 2011-10-26T17:08:55 we'll see 2011-10-26T17:09:24 it's a great language but it's simply not finished 2011-10-26T17:09:44 *** dr_eyebrows has joined #aichallenge 2011-10-26T17:09:46 I hope for you its stable enough ;-) 2011-10-26T17:10:01 or that you know the internals somewhat 2011-10-26T17:10:05 normally, it should be 2011-10-26T17:10:24 *** djstrong has quit IRC (Ping timeout: 240 seconds) 2011-10-26T17:10:30 and I don't know the internals thouroughly but I think good enough for what I'm still doing with it 2011-10-26T17:10:55 http://aichallenge.org/language_profile.php?language=CoffeeScript 2011-10-26T17:11:07 meh :-) 2011-10-26T17:11:13 i think that's the smallest language at the moment 2011-10-26T17:11:19 *** ikaros has joined #aichallenge 2011-10-26T17:11:40 nope. groovy! 2011-10-26T17:11:51 haha 2011-10-26T17:11:51 so: making descisions based on identiy of opponent is bad.. what about making a descision based on the general percieved performance of the enemy? 2011-10-26T17:12:01 http://aichallenge.org/language_profile.php?language=Groovy 2011-10-26T17:12:11 so far no one 2011-10-26T17:12:18 Woah, so many 2011-10-26T17:12:29 *** mcstar has quit IRC (Quit: WeeChat 0.3.5) 2011-10-26T17:12:33 but a lot of those starter kits are new. I've only heard about d's one 1.5 weeks ago 2011-10-26T17:12:39 Burny: I think there is no problem with that, though I doubt it's easily done 2011-10-26T17:12:55 and I'll submit the groovy starter kit so I'll be the first :) 2011-10-26T17:13:46 craklyn, doing it well is probably hard, but it would be semi easy, in some cases 2011-10-26T17:14:55 If you see the enemy willingly trade an ant for no gain, you could track that. But there may be a bot which is super smart and sacrifices for the long-term advantage (grabbing a defensible nook in the terrain, getting two ants the next turn, getting a food so no actual net loss, etc) 2011-10-26T17:15:04 *** tdubellz has quit IRC (Quit: reconnecting) 2011-10-26T17:15:05 *** SnapJudgment has joined #aichallenge 2011-10-26T17:15:09 Lots of situations like that make it hard to judge, in my opinion 2011-10-26T17:15:41 *** irchs has joined #aichallenge 2011-10-26T17:15:59 meh, the current js leader is very lucky: http://aichallenge.org/visualizer.php?game=17597&user=577 2011-10-26T17:16:42 *** foRei has quit IRC (Quit: Bye) 2011-10-26T17:17:04 python newbie question: In the starter kit, ants.enemy_ants() returns a tuple. How do I get the location of an enemy ant to give to ants.distance()? 2011-10-26T17:17:24 *** tdubellz has joined #aichallenge 2011-10-26T17:17:31 test_bot.bat can't seem to be able to handle relative paths 2011-10-26T17:17:43 *** tdubellz is now known as Guest82457 2011-10-26T17:17:53 calling test_bot.bat ../some/other/dir/bot.exe tries to call ./bot.exe 2011-10-26T17:17:59 *** sunshowers has quit IRC (Quit: Page closed) 2011-10-26T17:17:59 *** halligalli has quit IRC (Ping timeout: 265 seconds) 2011-10-26T17:18:53 *** cyphase has quit IRC (Read error: Connection reset by peer) 2011-10-26T17:18:57 *** aslvrstn has quit IRC (Ping timeout: 265 seconds) 2011-10-26T17:20:05 advg, xeorbot in that game, has some pretty obvious weaknesses.. is that the lucky one you were refering to? 2011-10-26T17:20:15 *** Guest82457 is now known as tdubellz 2011-10-26T17:20:31 yeah 2011-10-26T17:20:37 check its current rank 2011-10-26T17:25:22 *** yohann has quit IRC (Quit: Lost terminal) 2011-10-26T17:25:51 hmm I started from scratch instead of using a starter pack and locally everything works fine... but on uploading I fail the tests with this: turn 0 bot 0 timed out 2011-10-26T17:26:22 any idea what could cause that? I just have dummy logic that takes almost no time at all 2011-10-26T17:26:53 what for language are you using jix_ 2011-10-26T17:26:57 *** rufus_ has joined #aichallenge 2011-10-26T17:26:58 scala 2011-10-26T17:27:14 did you compared version numbers? 2011-10-26T17:28:31 *** cyphase has joined #aichallenge 2011-10-26T17:28:32 I have 2.9.1 the server has 2.9.0.1 .. I don't think that small difference should matter.. also it compiles fine 2011-10-26T17:28:51 *** Lithosphere has quit IRC (Remote host closed the connection) 2011-10-26T17:28:59 *** ltriant has quit IRC (Read error: Connection reset by peer) 2011-10-26T17:29:05 *** Lithosphere has joined #aichallenge 2011-10-26T17:29:05 *** Lithosphere has joined #aichallenge 2011-10-26T17:30:19 Is there a way to run the upload tests locally? 2011-10-26T17:31:16 jix_: the test_bot.sh script is the same one used on the server, according to http://aichallenge.org/using_the_tools.php 2011-10-26T17:31:31 *** t_ has joined #aichallenge 2011-10-26T17:31:37 hi there 2011-10-26T17:32:46 oh how could I not see that one... 2011-10-26T17:33:54 I'm on the first page of the Ants Tutorial, I copied the Tools folder, the starter files, and using Python 3.2, I've got the error : 2011-10-26T17:34:27 Traceback (most recent call last): 2011-10-26T17:34:29 File "C:\Python32\Lib\threading.py", line 736, in _bootstrap_inner 2011-10-26T17:34:31 *** ltriant has joined #aichallenge 2011-10-26T17:34:31 self.run() 2011-10-26T17:34:32 File "C:\Python32\Lib\threading.py", line 689, in run 2011-10-26T17:34:35 self._target(*self._args, **self._kwargs) 2011-10-26T17:34:37 File "C:\aichallenge\tools\sandbox.py", line 330, in _monitor_file 2011-10-26T17:34:39 line = fd.readline() 2011-10-26T17:34:40 AttributeError: '_io.FileIO' object has no attribute 'read1' 2011-10-26T17:34:50 dwins: hmm I do pass that one locally though.. 2011-10-26T17:35:04 if you're going to paste more than 5 lines at a time could you use pastebin? 2011-10-26T17:35:05 and I can't find the solution... :S 2011-10-26T17:35:13 jix_: so does my built-from-scratch scala bot 2011-10-26T17:35:18 woops sorry :s 2011-10-26T17:35:34 also have you tried running it with python 2.7? 2011-10-26T17:35:37 I will think to pastebin for the next time 2011-10-26T17:35:45 i think the tools are python2, not 3 :/ 2011-10-26T17:35:56 dwins: and yours does work on the server too? 2011-10-26T17:36:01 yep 2011-10-26T17:36:17 actually, i'm not sure whether that explains it 2011-10-26T17:36:29 so why is there a python 3.2 section for the tools ? 2011-10-26T17:36:37 Zannick: the server expects a MyBot.scala file and a MyBot object with a main method, that tripped me up a bit 2011-10-26T17:36:51 but once i got things compiling and named right it ran ok 2011-10-26T17:37:08 hm. maybe this? http://bugs.python.org/issue4996 2011-10-26T17:37:10 oops, jix_ ^^ 2011-10-26T17:37:38 t_, there is python 3.2 section for the tools? 2011-10-26T17:38:02 well, that's an old bug 2011-10-26T17:38:17 but ought to be fixed in 3.2? 2011-10-26T17:38:22 dwins: hmm there seem to be more online tests... for my first upload I got a 64bit seed an it failed as I parsed it as an Int.. so my code does run... and the local script doesn't pass such a seed 2011-10-26T17:38:40 interesting 2011-10-26T17:38:51 jix_: oops! 2011-10-26T17:39:07 jix_: well the specification does say that it's a 64-bit int for the seed. other numeric values are all 32-bit ints though 2011-10-26T17:39:20 * dwins suddenly worries about signs 2011-10-26T17:39:32 yeah I know so I fixed it and now I get a timeout in turn 0 that I can't reproduce 2011-10-26T17:39:54 dwins: the seed that I got was negative, so just using toLong should work 2011-10-26T17:41:15 Zannick, more, if y avec this error, the script terminate with no more problem, and launch a web application that is frozen (i can click "play", nothing move) ... 2011-10-26T17:41:21 I'm cursed. 2011-10-26T17:41:50 ho ! I forgot : I'm use windows XP ... 2011-10-26T17:41:59 (is that a curse ?) 2011-10-26T17:42:11 *** mbaricevic has joined #aichallenge 2011-10-26T17:42:28 no, the problem is within your python3.2 lib 2011-10-26T17:42:34 analyst74: there is a python 3.2 version for the starter pack, at least 2011-10-26T17:43:17 that starter bot is for 3.2, but the bot engine is supposed to run with 2.x I suppose 2011-10-26T17:43:28 yup 2011-10-26T17:43:38 t_: the python3 bot does not include python3 tools. the tools are python2 only 2011-10-26T17:44:26 earlier my default python was python3 2011-10-26T17:44:33 had point it to python2 2011-10-26T17:44:34 I don't get the point : must I have the two versions of python to launch the bot ?? 2011-10-26T17:44:40 for the tools to run 2011-10-26T17:44:46 I wasn't able to find this anywhere on the main site (e.g. in the problem description, etc) but after a search on the Forums it looks like the time limit for a turn is 1 second? I'm just poking my head in to verify this. 2011-10-26T17:45:08 mbaricevic: it's specified in the initial bot input 2011-10-26T17:45:38 http://aichallenge.org/specification.php#Bot-Input 2011-10-26T17:45:44 "turntime" 2011-10-26T17:45:55 Oh!!! Okay there we go. Thanks! :) 2011-10-26T17:46:38 mbaricevic: it's 500ms http://aichallenge.org/game_settings.php 2011-10-26T17:47:20 *** GMatBM has joined #aichallenge 2011-10-26T17:47:44 See, that's why I popped in here. You guys are so smart. Thanks again. 2011-10-26T17:49:36 *** djstrong has joined #aichallenge 2011-10-26T17:50:36 *** mbaricevic has quit IRC (Quit: Page closed) 2011-10-26T17:52:31 *** t_ has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20110928134238]) 2011-10-26T17:57:29 *** SharkMonkey has quit IRC (Quit: /QUIT) 2011-10-26T18:00:11 *** tdubellz has quit IRC (Remote host closed the connection) 2011-10-26T18:01:00 *** dinosoep has quit IRC (Ping timeout: 265 seconds) 2011-10-26T18:01:10 *** GMatBM has quit IRC (Quit: .) 2011-10-26T18:01:24 *** Harpyon has quit IRC (Quit: Computer has gone to sleep.) 2011-10-26T18:01:26 *** Glert has joined #aichallenge 2011-10-26T18:02:02 *** Cyndre has quit IRC (Read error: Connection reset by peer) 2011-10-26T18:02:16 *** Cyndre has joined #aichallenge 2011-10-26T18:02:23 *** irchs has quit IRC (Quit: irchs) 2011-10-26T18:04:18 the other day, a couple people said it's possible to block a path of width 3, I don't see how that's possible though 2011-10-26T18:05:01 *** SnapJudgment has quit IRC (Quit: This computer has gone to sleep) 2011-10-26T18:07:21 *** olexs has quit IRC (Quit: Leaving.) 2011-10-26T18:08:40 burny: I don't think 3 is possible < 3 is 2011-10-26T18:08:56 hmm can anyone point me at what exactly is tested on submission? 2011-10-26T18:09:19 jix_: look at the test_bot script 2011-10-26T18:09:32 janz, that's what I think also 2011-10-26T18:10:21 janzert: but on submission I got passed a large negative seed which made my bot fail... the test_bot script doesn't do that... now I fixed that and it still fails on the server... while it does pass the test_bot script locally 2011-10-26T18:11:07 sounds like you're running locally on a 32bit system while the server is 64bit so I'd look for other differences related to that 2011-10-26T18:12:28 janzert: no I am running 64bit... and I'm using scala which is the same on both... but as I said I fixed that part... it fails on a later test now and I can't reproduce that 2011-10-26T18:13:00 *** Apophis has quit IRC (Ping timeout: 255 seconds) 2011-10-26T18:13:05 *** craklyn has quit IRC (Ping timeout: 265 seconds) 2011-10-26T18:14:44 maybe try adding some log messages to stderr and resubmitting to help you narrow down what your code is doing 2011-10-26T18:16:15 *** epicmonkey has quit IRC (Ping timeout: 248 seconds) 2011-10-26T18:18:19 janzert, adding logging and resubmitting? is there a way to get logs from the server? 2011-10-26T18:18:39 the last bit of stderr will be given to you 2011-10-26T18:18:51 actually I think the first bit and last bit 2011-10-26T18:19:08 what does first and last bit mean? 2011-10-26T18:19:33 first x bytes and last x bytes where x is I believe around 500 2011-10-26T18:19:54 so several lines 2011-10-26T18:20:27 *** akosch has quit IRC (Ping timeout: 260 seconds) 2011-10-26T18:20:27 and how do you get this? 2011-10-26T18:20:55 it is displayed on your profile along with the errors 2011-10-26T18:21:09 ah 2011-10-26T18:21:32 *** kaemo has quit IRC (Read error: Connection reset by peer) 2011-10-26T18:21:50 thanks :) 2011-10-26T18:23:32 *** Minthos has joined #aichallenge 2011-10-26T18:24:09 *** rufus_ has quit IRC (Quit: Page closed) 2011-10-26T18:24:13 *** kaemo has joined #aichallenge 2011-10-26T18:24:53 *** akosch has joined #aichallenge 2011-10-26T18:35:40 *** retybok has joined #aichallenge 2011-10-26T18:38:13 *** thimbleBrain has quit IRC (Ping timeout: 265 seconds) 2011-10-26T18:39:04 *** replore_ has quit IRC (Remote host closed the connection) 2011-10-26T18:43:36 what information is correct the scala starter package's readme says the server has scala 2.8.1 the website says it has scala 2.9.0.1 2011-10-26T18:44:28 *** xar0l has left #aichallenge 2011-10-26T18:44:55 the website 2011-10-26T18:46:00 *** jtza8 has joined #aichallenge 2011-10-26T18:47:10 *** thestinger has quit IRC (Quit: WeeChat 0.3.6) 2011-10-26T18:49:11 Interesting, on OS X, when running tools/play_one_game.sh from my source dirrectory, "." refers to the source dir in the bash file. 2011-10-26T18:49:27 Obviously it doesn't in the tutorial. 2011-10-26T18:49:48 . is always the working directory i thought? e.g. when executing a bash script, the directory from where the script was called 2011-10-26T18:50:03 not the directory of the script? 2011-10-26T18:50:22 in windows batch scripting i think it is the same, you use %~dp0 for the file location directory 2011-10-26T18:51:22 Strange that it works in the tutorial. 2011-10-26T18:51:37 *** ajhager has joined #aichallenge 2011-10-26T18:52:28 meh, my code is completely not behaving like in the test :-( 2011-10-26T18:55:36 *** canadiancow|work has quit IRC (Remote host closed the connection) 2011-10-26T18:55:36 my code crashes during the server's unit tests but not my own, because I have a newer compiler 2011-10-26T18:56:14 *** djr_ has joined #aichallenge 2011-10-26T18:56:31 *** mcstar has joined #aichallenge 2011-10-26T18:56:52 antimatroid: 2011-10-26T18:56:57 *** cleverneuron has joined #aichallenge 2011-10-26T18:56:59 mcstar: 2011-10-26T18:57:01 hello 2011-10-26T18:57:04 hey 2011-10-26T18:57:07 hey 2011-10-26T18:57:11 stop this 2011-10-26T18:57:17 stop this 2011-10-26T18:57:19 Anyone have problems with the visualizer in Ubuntu 64bit with Sun java vm? 2011-10-26T18:57:19 Minthos: it's unlikely the D compiler will be updated 2011-10-26T18:57:22 :D 2011-10-26T18:57:31 :( 2011-10-26T18:57:31 i wrote a dijkstra in c++ 2011-10-26T18:57:35 k 2011-10-26T18:57:47 we tried shortly before the contest started but the newer versions available at that point at least had other problems 2011-10-26T18:57:53 i can do 12000 , 60 step long path searches on a maze map in 1 sec 2011-10-26T18:58:21 oh I get it :/ I created objects, but the objects are sharing the same data somehow 2011-10-26T18:58:34 meh, prototyping 2011-10-26T18:58:39 just started today, but it's crashing whenever I run ./play_one_game.sh -So | java -jar visualizer.jar 2011-10-26T18:58:50 org.mozilla.javascript.EcmaError: TypeError: Cannot read property "cols" from null (Application.js#1325) 2011-10-26T18:59:01 antimatroid: do you remember what were your timings back than? 2011-10-26T18:59:19 mcstar: no idea sorry 2011-10-26T18:59:31 *** retybok has quit IRC (Quit: leaving) 2011-10-26T18:59:37 anyway, it seems pretty fast 2011-10-26T18:59:49 any ideas or tips? 2011-10-26T18:59:52 it beats my common lisp a* as far as i can remember 2011-10-26T19:00:09 im not using queues, or vectors 2011-10-26T19:00:38 Is there a list of common problems/errors with the test code? 2011-10-26T19:00:40 and tools 2011-10-26T19:00:56 everything works in the right hands 2011-10-26T19:01:02 didn't find anything in the forum 2011-10-26T19:01:53 antimatroid: i started solving project euler, do you? 2011-10-26T19:01:54 cleverneuron: just adapt to it 2011-10-26T19:02:02 ? 2011-10-26T19:02:09 mcstar: i've done like 50 something problems ages ago 2011-10-26T19:02:15 sigh has done hundreds of them i think 2011-10-26T19:02:21 heh 2011-10-26T19:02:22 have no intensions of writing my own visualizer app since the default one is broken 2011-10-26T19:02:23 nice 2011-10-26T19:02:30 ive done 12 in 2 days 2011-10-26T19:02:49 i will grow old till i finish 2011-10-26T19:03:10 antimatroid: what is your nick there? 2011-10-26T19:03:17 can you give me a friend key? 2011-10-26T19:03:34 antimatroid 2011-10-26T19:03:43 Im guessing I'm not meeting a dependency, but all I saw was Python 2.6 and Java 6 which I'm using the most up to date versions under Ubuntu including using the Sun JVM instaed of IcedTea 2011-10-26T19:03:59 *** cybsy has quit IRC (Ping timeout: 256 seconds) 2011-10-26T19:04:00 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-10-26T19:04:12 ah, i didnt know they added a friends feature 2011-10-26T19:04:18 was that with the redesign, or later? 2011-10-26T19:04:20 antimatroid: so you know what a friend key is? 2011-10-26T19:04:26 no 2011-10-26T19:04:32 i just logged in and had to verify my email :P 2011-10-26T19:04:36 ah 2011-10-26T19:04:37 http://projecteuler.net/friends 2011-10-26T19:04:56 do i give you my friend key? 2011-10-26T19:05:02 *** loktar has joined #aichallenge 2011-10-26T19:05:04 if you want 2011-10-26T19:05:09 83302450121588_188b46d295e58d78153bbd8065225d6b 2011-10-26T19:05:12 we arent exactly "fiends" 2011-10-26T19:05:20 you made it public 2011-10-26T19:05:24 now change it fast 2011-10-26T19:05:32 ruuun 2011-10-26T19:05:34 fiends 2011-10-26T19:05:39 :D 2011-10-26T19:05:52 there are good typos and bad typos 2011-10-26T19:05:58 lol 2011-10-26T19:06:09 why does it matter if others have it? 2011-10-26T19:06:20 dunno 2011-10-26T19:06:37 you cant identify people than 2011-10-26T19:06:39 i guess 2011-10-26T19:06:49 done 2011-10-26T19:06:57 http://pastebin.com/1yUFrUHa 2011-10-26T19:06:57 so how do i add you? 2011-10-26T19:07:41 i've only solved 39 problems 2011-10-26T19:08:02 there you are 2011-10-26T19:08:03 i usually do 1-3 at a time 2011-10-26T19:08:25 anyone here actually to discuss the aichallenge? 2011-10-26T19:08:39 <_ulises> cleverneuron: I wouldn't mind ;) 2011-10-26T19:08:40 cleverneuron: what are the -So options on play_one_game 2011-10-26T19:09:09 none that I'm aware that is the exact line the website said to run as a test 2011-10-26T19:09:27 39problems? 2011-10-26T19:09:29 Zannick: saves output as a stream 2011-10-26T19:09:41 ./playgame gives you the full explanation 2011-10-26T19:09:47 right 2011-10-26T19:09:50 burny, at projecteuler.net 2011-10-26T19:10:05 i presume that is not a complete commandline 2011-10-26T19:10:16 i do see it on the site now 2011-10-26T19:10:30 Will try and figure it out. was just following this http://aichallenge.org/using_the_tools.php 2011-10-26T19:10:37 specifically the Unix use of the tools 2011-10-26T19:11:00 *** nand has joined #aichallenge 2011-10-26T19:11:45 Zannick: thank you will dig through this for a bit. Have a good night everyone 2011-10-26T19:11:49 *** cleverneuron has quit IRC (Quit: Page closed) 2011-10-26T19:11:51 7 submissions and still no idea why I fail the tests on the server :( 2011-10-26T19:12:16 jix_: did you test locally? 2011-10-26T19:12:17 cyphase, hm.. some of those would be hard without a computer, are they supposed to be done with or without? 2011-10-26T19:12:25 *** dwins has quit IRC (Ping timeout: 252 seconds) 2011-10-26T19:12:26 it seems my bot gets no input past the player_seed line 2011-10-26T19:12:30 amstan: yeah it runs games fine here 2011-10-26T19:12:30 the interesting thin is, there arent many people with problems 2011-10-26T19:12:31 <_ulises> jix_: the server usually gives you logs as to why tests failed I believe 2011-10-26T19:12:44 _ulises: yeah I get some debug output and then a timeout 2011-10-26T19:12:48 jix_: same os, same cpu? 2011-10-26T19:12:49 there are over 4000 users, and the stream of problems is like the very end of beta 2011-10-26T19:13:04 <_ulises> jix_: care to paste them on pastebin? 2011-10-26T19:13:23 amstan: I'm using scala so it shouldn't be too os/cpu sensitive but I'm running arch linux 64bit with scala 2.9.1 here 2011-10-26T19:13:34 _ulises: the output? it's only my own debug output 2011-10-26T19:13:34 arch ftw 2011-10-26T19:14:00 <_ulises> jix_: oh, you mean your local tests give you timeouts? or is it when you submit your code to the server? 2011-10-26T19:14:17 _ulises: on the server 2011-10-26T19:14:17 jix_: ok, seems fine then 2011-10-26T19:14:29 jix_: define test though 2011-10-26T19:14:44 jix_: there's test_bot.sh or something that runs the exact test that the server runs 2011-10-26T19:14:51 amstan: locally I can run games with the live visualizer and I can run test_bot.sh 2011-10-26T19:14:56 its late though, bb all 2011-10-26T19:14:59 *** mcstar has left #aichallenge ("WeeChat 0.3.5") 2011-10-26T19:15:12 <_ulises> jix_: so on your profile you should see the logs for why your code failed on the server, please paste that on pastebin? 2011-10-26T19:15:19 jix_: so your profile page doesn't include logs? 2011-10-26T19:15:35 Do positions start from 0? 2011-10-26T19:15:40 burny, some need a computer to be practical, some can be done by hand but are easier with a computer, and some are relatively easy by hand (if you know the math) 2011-10-26T19:15:45 *** oops has joined #aichallenge 2011-10-26T19:16:00 amstan: it does... but from that it seems I get no input past the player_seed line 2011-10-26T19:16:11 burny, of course some problems can only be done naively with a computer but are very quick if you know the proper formula 2011-10-26T19:16:35 I output line for each settings line... followed by a line for when I got ready. line player_seed is the last output I get 2011-10-26T19:16:37 k, wasn't sure if maybe they were intended as pure maths 2011-10-26T19:16:45 jix_: i remember there being a problem with the player_seed, the specs say that all numbers are 32 bit, but the seed was 64bit for some reason 2011-10-26T19:17:15 amstan: the specs do say it's 64bit... and that was why my first upload failed... I fixed that though... it even fails when I don't parse the line at all 2011-10-26T19:17:25 oh, ok 2011-10-26T19:17:49 burny, no, there's definitely a programming bias, but its not 100%. also, here's an example of a problem that naively is hard by hand and easy with a computer, but that can be done much faster and for much larger numbers if you know the proper formula: http://projecteuler.net/problem=1 2011-10-26T19:17:51 amstan: also the server tests with a number that doesn't fit into 32bit... the script bundled with the tools doesn't... so it's not doing the exact same thing 2011-10-26T19:18:24 *** nand has quit IRC (Quit: Page closed) 2011-10-26T19:18:59 burny, all the problems are intended to be solvable on an average computer in under one minute of run time 2011-10-26T19:19:45 wouldn't that be 233159? 2011-10-26T19:19:54 ahh 2011-10-26T19:19:58 close 2011-10-26T19:20:00 *** Larose has quit IRC (Ping timeout: 240 seconds) 2011-10-26T19:20:38 oops, do I have half the answer? 2011-10-26T19:21:05 *** ademar has quit IRC (Read error: Operation timed out) 2011-10-26T19:21:26 nope, just did the maths wrong 2011-10-26T19:22:28 234168 2011-10-26T19:22:30 ? 2011-10-26T19:22:31 meh, I just saw the problem, but I can't do it on paper right now 2011-10-26T19:22:45 *** xar0l has joined #aichallenge 2011-10-26T19:22:45 just do it in your head then :P 2011-10-26T19:22:49 paper is for suckers 2011-10-26T19:23:06 no, I want to solve it vectorized 2011-10-26T19:23:14 lol 2011-10-26T19:23:22 bring out the big guns 2011-10-26T19:23:56 I'm quite good in finding structure, but sad I haven't got many mathematical exercises the last years 2011-10-26T19:23:58 burny, your answers are interesting considering the correct answer :P 2011-10-26T19:24:39 i need sigh aha 2011-10-26T19:24:51 (and I have a bad memory for numbers) 2011-10-26T19:24:55 can anyone remember why exactly you can't just search out from all ants simultaneously to calculate vision? 2011-10-26T19:25:01 *** delt0r__ has quit IRC (Ping timeout: 240 seconds) 2011-10-26T19:25:12 hrm, so 234168 was wrong? lames 2011-10-26T19:25:31 i know it has something to do with the euclidean and manhatten metrics not being equivalent, but i don't see it from the example: 2011-10-26T19:25:32 a.* 2011-10-26T19:25:32 .a. 2011-10-26T19:25:32 *** onensora has quit IRC () 2011-10-26T19:26:08 hmmm I'm not supposed to output "go" before I get the "ready" am I? 2011-10-26T19:26:32 jix_, no 2011-10-26T19:26:43 jix_, though, i'm not sure it would matter 2011-10-26T19:26:57 actually, yea it would 2011-10-26T19:27:02 oh, oops.. I misread the question.. so I'm off by 1000 :P 2011-10-26T19:27:16 *** Sakura_ has joined #aichallenge 2011-10-26T19:27:26 'go' means you're ready for turn1 input 2011-10-26T19:27:36 wait, gah 2011-10-26T19:27:39 *thinks* 2011-10-26T19:27:40 hey people, how's your coding going? 2011-10-26T19:28:25 Having some questioning here. 2011-10-26T19:28:32 poorly - my bot is awesome but the server fails to run it 2011-10-26T19:28:40 jix_, 'go' tells the engine you're ready for your turn to begin. if you send it early, you could be doing setup during your valuable turn time 2011-10-26T19:28:43 I'm really out of ideas... it works locally with the exact same input.... 2011-10-26T19:29:09 cyphase: I'm not sending it early... but I get a timeout before I get a ready (at least it seems so to my bot) on the server 2011-10-26T19:29:15 *** rul has joined #aichallenge 2011-10-26T19:29:17 What would give 'deadtiles'? I fail to see what could possible be /That/ interesting into them. Any ideas? 2011-10-26T19:29:38 Minthos- that sucks. can't you optimize further? It's about timeout/crashes? 2011-10-26T19:30:13 it's about crashing, I'm trying to debug it 2011-10-26T19:30:51 I'm pretty sure it must be something in the scala lib that changed between 2.9.0.1 and 2.9.1... so I'm probably using something in a way it wasn't intended to be used... as they are supposed to be compatible 2011-10-26T19:31:14 Minthos: Pretty hard to debug crashes in that kind of remote setting. 2011-10-26T19:31:15 jix_: so what might be happening is that you stop reading from input 2011-10-26T19:31:22 and it must be something with the way I'm reading stuff... 2011-10-26T19:31:36 maybe I should use the plain java stuff instead of the nicer iterator based stuff 2011-10-26T19:32:01 and so you don't get the 'ready' and the engine thinks you've timed out 2011-10-26T19:32:32 So... nobody can tell me why || how the DeadTiles are relevant to anything? 2011-10-26T19:32:32 <_ulises> jix_: perhaps the input is being buffered behind the scenes? 2011-10-26T19:32:55 Sakura_: I can reproduce it locally when I use the server's compiler version.. but I still don't know why it crashes 2011-10-26T19:33:11 _ulises: yeah but that alone shouldn't be a problem as long as it allows emptying the buffer without blocking 2011-10-26T19:33:20 but yeah it probably doesn't 2011-10-26T19:33:33 or in 2.9.1 it does and in 2.9.0.1 it doesn't 2011-10-26T19:33:40 Minthos: Ha! You have much more chances to fix your problem then. 2011-10-26T19:34:03 <_ulises> jix_: are you providing the scala jars? if so, try with a different version of scala, you never know 2011-10-26T19:34:14 *** danielharan has joined #aichallenge 2011-10-26T19:34:47 25*25+8*8 2011-10-26T19:34:50 bah 2011-10-26T19:34:53 Okay, here's another question then: Can we include in our code some 3rdParty code if we credit them? Or are every single pieces of code required to be original (exclusively written by you) ? 2011-10-26T19:36:19 *** antimatroid1 has joined #aichallenge 2011-10-26T19:36:25 *** antimatroid has quit IRC (Read error: No route to host) 2011-10-26T19:36:46 Sakura_, you can use 3rd party code 2011-10-26T19:37:06 cyphase: Okay, thanks. 2011-10-26T19:37:50 *** delt0r__ has joined #aichallenge 2011-10-26T19:38:10 Any memory restriction? I can't remember having read anything about that. 2011-10-26T19:38:10 would be interesting if there was an open source bot team 2011-10-26T19:38:15 1GB 2011-10-26T19:38:19 Indeed 2011-10-26T19:38:58 1GB? Damn. I'll have to optimize... 2011-10-26T19:39:23 lol 2011-10-26T19:39:26 really? 2011-10-26T19:39:31 ;) naw 2011-10-26T19:39:48 but I see what I could do with that much memory 2011-10-26T19:39:56 i remember last time open sourcing a bot during the competition was frowned upon 2011-10-26T19:40:17 I read that you shouldn't do that on the website today somewhere 2011-10-26T19:40:29 I'd expect my AI to become sentient around the 700 mb mark... 2011-10-26T19:40:36 *** Sunshowers has joined #aichallenge 2011-10-26T19:40:49 Yes, in the forum: It's prohibited. 2011-10-26T19:41:06 it was along the lines of "sharing non strategy code is ok, discussing strategy too, but no sharing strategy code" 2011-10-26T19:41:44 Even sharing strategies implementations is discouraged. Weird since the web site seem to have been designed with that in mind... (forums..) 2011-10-26T19:41:54 exact 2011-10-26T19:42:01 *** dr_eyebrows has quit IRC (Ping timeout: 265 seconds) 2011-10-26T19:42:35 Jix_, do you know if there's a clarification about the use of map data (map data embedding) inside the code? 2011-10-26T19:43:04 *** ccc has joined #aichallenge 2011-10-26T19:43:16 *** McLeopold has joined #aichallenge 2011-10-26T19:43:41 I wish the rules would be clear about this. If they don't plan on releasing the Final's map before we can update our code, then they should be clear about it. 2011-10-26T19:44:02 Sakura_: I think they stated that the final map's will be new somewhere 2011-10-26T19:44:07 Final maps will be generated after submission are closed 2011-10-26T19:44:35 yes, but we don't know if we'll get access to them before we can do a last update. 2011-10-26T19:44:51 Sakura_: where on the web site would you have expected to read this? And, no, you don't get access. 2011-10-26T19:44:57 McLeopold: Alright. Now I know. Thanks. 2011-10-26T19:45:05 Sakura_: the point is you should not code your bot to specific maps 2011-10-26T19:45:12 hmm is there a fixed number of maps for the final competition? 2011-10-26T19:45:17 no 2011-10-26T19:45:28 or, we don't know what the number may be 2011-10-26T19:45:49 :-) 2011-10-26T19:45:50 sorry for the stupid question but when will the submissions be closed? 2011-10-26T19:45:55 dec 18th 2011-10-26T19:45:55 In the rules, it would be good to indicate that as many people could bet on that strategy, thinking they're doing good before the finals, just to do bad in the finals. 2011-10-26T19:46:43 my only big worry is implementation :-) 2011-10-26T19:46:59 its always my worst part 2011-10-26T19:47:07 Zannick: then THAT should be told in the rules. Simple as you formulated it: Coding specifically for maps is prohibited. Know that you won't have access to Final's map before the your last chance submission. 2011-10-26T19:47:42 it's not prohibited, just useless for the finals 2011-10-26T19:47:45 meh, prohibited is a strong word 2011-10-26T19:47:56 you just get stupid bots anyway :-) 2011-10-26T19:48:36 avdg2: but it would distort the live rankings which is bad IMHO 2011-10-26T19:48:44 who's afraid of strong words? :-) 2011-10-26T19:48:55 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-26T19:49:03 jix_: the current live ranking is just an indication for the finals ;-) 2011-10-26T19:49:07 *** jtza8 has quit IRC (Remote host closed the connection) 2011-10-26T19:49:20 knowing right from the start the enemies hills and how to get there optimally makes bots stupid? hmmm. 2011-10-26T19:49:39 Sakura_: well you don't 2011-10-26T19:49:55 in the last contest all of the starting maps had exactly 19 planets 2011-10-26T19:49:56 Sakura_: you'd still have to discover the map, but you would probably be able to guess in the first few turns 2011-10-26T19:50:00 but 19 was not in the spec 2011-10-26T19:50:02 jix_: I sure won't 2011-10-26T19:50:13 so when new maps were generated with a different number, people's bots broke 2011-10-26T19:50:18 because they hardcoded the 19 2011-10-26T19:50:19 planets? wut 2011-10-26T19:50:50 McLeopold: It's pretty quick to know in which map you are; takes only a few turns. 2011-10-26T19:50:59 Success: ready to play FINALLY 2011-10-26T19:51:12 jix_: go4it :-) 2011-10-26T19:51:21 well the only problem is I have no AI yet ;) 2011-10-26T19:51:23 Sakura_: so then, we agree? :) "few turns" 2011-10-26T19:52:20 McLeopold: which I wouldn't consider "exploring".. really few turns. But yes, we agree. :) 2011-10-26T19:53:12 Sakura_: the last contest involved planets, yes. 2011-10-26T19:53:21 Unless we release a bunch of maps with no water nearby 2011-10-26T19:53:23 what are planets? 2011-10-26T19:53:35 the previous contest, Sakura_ 2011-10-26T19:53:40 as in not relevant for ants 2011-10-26T19:53:44 mecury, venus, earth, mars...etc 2011-10-26T19:53:53 lol McLeo 2011-10-26T19:53:55 what he said :) 2011-10-26T19:54:12 :) I never know where the language barriers are here 2011-10-26T19:54:42 *** thisisit has joined #aichallenge 2011-10-26T19:54:43 I see. I'm new to this contest. I thought you said the Finals maps would have 'planets'. *PANIC* 2011-10-26T19:54:54 haha my random walk won it's first game 2011-10-26T19:55:00 next contest: Ants Planet Wars. 2011-10-26T19:55:00 French is my personal barrier. 2011-10-26T19:55:04 jix_: link? 2011-10-26T19:55:10 You are playing both ants and planet wars 2011-10-26T19:55:11 http://aichallenge.org/visualizer.php?game=18551&user=4501 2011-10-26T19:55:13 Planet of the Ants? 2011-10-26T19:55:23 gah, so annoying when i look at old, uncommented code and i dont know what it's doing 2011-10-26T19:55:25 *** danielharan has joined #aichallenge 2011-10-26T19:55:34 your attacking ships travel between planets, and on the planets are games of ants 2011-10-26T19:55:36 lol nice win 2011-10-26T19:56:00 *** danielharan has joined #aichallenge 2011-10-26T19:56:03 congrat jix 2011-10-26T19:56:04 my random walk bot even destroyed 3 ant hills, what are the odds 2011-10-26T19:56:11 *** iAnti has joined #aichallenge 2011-10-26T19:56:15 any moderators or anyone associated with the chalenge in the house? 2011-10-26T19:57:07 meh, why are the bots acting like mine? 2011-10-26T19:57:08 *** Sakura_ is now known as Sakura_Away 2011-10-26T19:57:18 MirrorBot 2011-10-26T19:57:23 * avdg2 gives up the theory they listened to me 2011-10-26T19:57:39 might release my current code as an alternative scala starter package... it's a very thin wrapper... it doesn't do any bookkeeping, just parses stuff and puts it into lists 2011-10-26T19:58:36 meh, I'm happy that I found the real bug in my code, which was nothing more than an understanding in js :p 2011-10-26T19:58:45 *misunderstanding 2011-10-26T19:58:46 *** Sunshowers has left #aichallenge 2011-10-26T19:58:56 I have a question... does anyone know if we are allowed to save info from match to match? on the competition servers? 2011-10-26T19:59:04 thisisit: what are you looking for help with? 2011-10-26T19:59:07 thisisit: you aren't allowed to write anything 2011-10-26T19:59:12 thisisit: no, you can't write anything to disk 2011-10-26T19:59:24 aww 2011-10-26T19:59:54 *** craklyn has joined #aichallenge 2011-10-26T19:59:59 *** Exception has quit IRC (Quit: Page closed) 2011-10-26T20:00:20 It's just that I was designing a bot that learns with every match... but it needs to store what it's leaned every end of a match 2011-10-26T20:00:56 "You may not write to files. You may however read from files within your submission directory, which will be the current directory." 2011-10-26T20:01:33 alright... 2011-10-26T20:01:35 thisisit: what you'd want to do, say, is run the bot on your own machine, against the unofficial tcp servers, and then for the official one, upload a fixed data file (and disable writing) 2011-10-26T20:01:46 this is completely legit 2011-10-26T20:02:08 *** iAnti has quit IRC (Quit: Page closed) 2011-10-26T20:02:25 Zannick: Unofficial tcp servers? where can I find information about those? Are those running the submitted bots? 2011-10-26T20:02:30 @tcp 2011-10-26T20:02:32 Zannick: tcp could be http://ants.fluxid.pl/howto. 2011-10-26T20:02:36 oh ok. so it would be playing against real player made bots? 2011-10-26T20:02:52 basically your bot receives and sends it output across tcp 2011-10-26T20:02:58 yes, with tcp you run your bot locally against others 2011-10-26T20:03:05 you don't upload it somewhere like the official servers 2011-10-26T20:03:35 @tcp ? Not sure I understand what means "@tcp". 2011-10-26T20:03:36 Sakura_Away: I have no idea what you mean. 2011-10-26T20:03:37 pefect that's what I need. 2011-10-26T20:03:43 hmm, is the ranking page broken? 2011-10-26T20:03:49 so no one gets your code, but you cant confirm someone isn't controlling it behind the scenes 2011-10-26T20:03:51 you may wish to check out the forums, there may be more info there 2011-10-26T20:03:56 or running it on a cluster 2011-10-26T20:03:57 :P 2011-10-26T20:04:21 Sakura_Away: '@' is the prefix for a command for our irc bot contestbot. i asked it to give the link to one of the tcp servers 2011-10-26T20:04:30 @supg 2011-10-26T20:04:31 Puj: You have no gotten any error messages recently, so here's a random one just to let you know that we care. 2011-10-26T20:05:26 Zannick: Haaa.. I see. Thanks. Btw, can we pass some parameters to it before the simulation? Like adjusting timeout to allow for local debugging? 2011-10-26T20:05:34 hmm.. i just thought that someone could precompute a lot about the various known maps and include it with their bot 2011-10-26T20:05:39 Sakura_Away: play_game.py --help 2011-10-26T20:05:41 :-) nice quote 2011-10-26T20:06:14 thx 2011-10-26T20:06:20 cyphase: right, but running it on a cluster only helps you in the unofficial server 2011-10-26T20:06:26 Zannick, i know 2011-10-26T20:06:35 humm. I was talking about the contestbot amstan 2011-10-26T20:07:03 and there's nothing preventing running stuff on a cluster to precompute some things ;) 2011-10-26T20:07:05 is there some safeguard against bots communicating with the outside world, or will you just be disqualified if it's found out 2011-10-26T20:07:12 i only ask out of curiosity :) 2011-10-26T20:07:14 safeguards, yes. 2011-10-26T20:07:24 safeguards and disqualification 2011-10-26T20:07:39 safeguards, disqualification, and looks of disapproval. 2011-10-26T20:07:49 our chief weapons 2011-10-26T20:07:56 Speaking of which, I wonder if there is an explicit code size limit? 2011-10-26T20:08:02 cyphase: we will post you on the frontpage that you tried to cheat 2011-10-26T20:08:10 there's a limit on the number of files i think 2011-10-26T20:08:15 Zannick: Does the remote tcp bot accepts some parameters regarding timeout etc? 2011-10-26T20:08:15 Puj, the archive has to be <= 2MB i believe 2011-10-26T20:08:18 cyphase: maybe even on reddit, so you loose all the precious karma 2011-10-26T20:08:22 and did we ever make it so that people could break their sources up into multiple folders? 2011-10-26T20:08:26 okay, okay, i'm not planning anything 2011-10-26T20:08:27 :D 2011-10-26T20:08:34 Hmm, okay 2011-10-26T20:08:35 Hmmmmmmm, Kaaarma 2011-10-26T20:08:38 Sakura_Away: the tcp server decides all the game parameters, your bot has to go with them 2011-10-26T20:08:38 lol 2011-10-26T20:08:38 *** Cyndre has quit IRC (Quit: Leaving) 2011-10-26T20:08:55 antimatroid1: my source is distributed among multiple folders 2011-10-26T20:08:59 so yeah that works 2011-10-26T20:09:00 My prefered fix for having no true self-esteem :D 2011-10-26T20:09:01 i guess my point with the precomputing is, will the final competition be using fresh maps? 2011-10-26T20:09:11 jix_: awesome 2011-10-26T20:09:17 i couldn't do that with planet wars, it was a pita 2011-10-26T20:09:18 you can make local settings on your own bot but they are independent of the game settings like turn time limits 2011-10-26T20:09:26 Zannick: Alright. 2011-10-26T20:09:27 final comp will be new maps, yes 2011-10-26T20:09:31 cyphase: yes. 2011-10-26T20:09:35 otherwise there would be no need for map discovery 2011-10-26T20:09:51 burny: new maps will be generated 2011-10-26T20:09:54 yea. i guess i didnt think that question through 2011-10-26T20:09:58 and only new maps will be used 2011-10-26T20:10:10 Zannick: Too bad cause it would have allowed for sweet local debugging. 2011-10-26T20:10:25 you could pre-compute moves for specific situations 2011-10-26T20:10:38 but then you need to find the moves too :P 2011-10-26T20:10:39 it's still good debugging 2011-10-26T20:10:44 but you can't control everything 2011-10-26T20:10:57 I'm going to try to make my AI take the form of a PRGN.. using the visibility as a seed, and the generated numbers, as actions 2011-10-26T20:11:25 hmm 2011-10-26T20:11:30 i wonder what the probability of the attack_radius and the visibility_radius changing for the final competition 2011-10-26T20:11:37 Zannick: Sure, but I can't step-in in real time and see exactly what's going on, which I do locally now that I have changed the Turn Time / Timeout values in the engine code. 2011-10-26T20:11:54 that would be interesting, write a bot to play thousands of games locally against other bots with a huge time limit, and precompute moves for any battle sitations that it has not come across otherwise use pre computed moved 2011-10-26T20:11:56 moves* 2011-10-26T20:12:00 sure 2011-10-26T20:12:05 machine learning 2011-10-26T20:12:17 Hence my wonderance. 2011-10-26T20:12:19 logging is your friend 2011-10-26T20:12:22 puj, I think they are planing to not change(unless there's some 'broken' thing found 2011-10-26T20:12:50 yeah, we consider it unideal to change teh parameters but we reserve the right to if we think the game is broken for some reason 2011-10-26T20:12:54 *** dd has joined #aichallenge 2011-10-26T20:13:02 Zannick: Of course, but tracing and live debugging is my BEST friend :-) 2011-10-26T20:13:03 i would be interested in seeing machine learning used against player within a game 2011-10-26T20:13:09 in other words, you shouldn't hard code for specific parameters 2011-10-26T20:13:11 *** thisisit has quit IRC (Quit: Page closed) 2011-10-26T20:13:20 e.g. player 2 seems to be very aggressive, so lets draw him out.. 2011-10-26T20:13:22
Is the game server down? 2011-10-26T20:13:36 against a player* 2011-10-26T20:13:47 I suggest we have an unofficial final with larger attack radii in order to place the best adaptive bots as well. But it must be the same submission as the official finals 2011-10-26T20:13:52 *** survivor has joined #aichallenge 2011-10-26T20:14:15 puj: just have a crazy tcp server 2011-10-26T20:14:25 *** grom358 has joined #aichallenge 2011-10-26T20:14:29 where you give ridiculous maps/parameters etc. 2011-10-26T20:14:39 And open mah internetz to tha w0rld?! 2011-10-26T20:15:03
Guys, any idea why there are zero ranking servers playing the game at the moment? 2011-10-26T20:15:07 Good call, I imagine people would use it regardless. 2011-10-26T20:15:09 i would love to see a tcp game with like 30+ bots on a huge map 2011-10-26T20:15:10 that's dangerous, you could catch something 2011-10-26T20:15:15 and a huge timeout 2011-10-26T20:15:17 *** Apophis has joined #aichallenge 2011-10-26T20:15:24 *** user1 has joined #aichallenge 2011-10-26T20:15:32 although probably hard to visualise aha 2011-10-26T20:15:35 zoom zoom zoom 2011-10-26T20:15:36 hoping someone can point me in the right direction here.. I got a raytrace function.. Now I want to scan in clockwise or anticlockwise direction looking for non-wall tile at edge of the ant's vision 2011-10-26T20:15:52 trig 2011-10-26T20:15:55 slow 2011-10-26T20:15:56 *cylon scanning noise* 2011-10-26T20:15:59 is there more efficient way then doing a ray trace to edge tile on edge of the ant's vision? 2011-10-26T20:16:33 Precompute the subset of tiles you want to look at and store them as like vision_radius_precomputed[][] 2011-10-26T20:16:36 *** Areks|2 has quit IRC (Ping timeout: 252 seconds) 2011-10-26T20:16:43 Any ideas on java visualizer crashing on windows? "Cannot find default value for object" CanvasElement.js line 1038 2011-10-26T20:16:46 Then the first index is your radius 2011-10-26T20:17:19 I use the .jar file user1 2011-10-26T20:17:25 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-10-26T20:17:31 Puj: yeah I already got function to know what tiles are in the ant's vision 2011-10-26T20:18:11 i can ray trace to each tile on the edge of the vision looking for path with no wall tiles 2011-10-26T20:18:11 user1: probably bad input 2011-10-26T20:18:14 *** djr_ has joined #aichallenge 2011-10-26T20:18:25 I run "play_one_game_live.cmd", the visualizer appears, and when I move the mouse over it it crashes with that error. 2011-10-26T20:18:26 but wondering if better way then ray tracing to each tile like that 2011-10-26T20:18:30 user1: redirect the output to a file, then pastebin 2011-10-26T20:18:33 I haven't changed any files, or put my own AI in yet 2011-10-26T20:18:35 oh, wait 2011-10-26T20:18:51 what browser and version and os? 2011-10-26T20:18:55 *** Redgis has quit IRC (Ping timeout: 244 seconds) 2011-10-26T20:19:09 Win 7, Firefox 7.0.1 2011-10-26T20:19:09 *** dd has quit IRC (Quit: Page closed) 2011-10-26T20:19:15 but it doesn't bring the browser up at all 2011-10-26T20:19:19 grom358, what are you trying to achieve? 2011-10-26T20:19:25 McLeopold: when you have some time, can we ponder upon archiving contests? 2011-10-26T20:19:35 ok 2011-10-26T20:19:43 *** rul has quit IRC (Quit: Saliendo) 2011-10-26T20:19:44 start pondering... 2011-10-26T20:19:48 e1.. 2011-10-26T20:20:06 user1: you're right, no browser... 2011-10-26T20:20:09 Damn, I was about to type d5, but realized you weren't playing chess. 2011-10-26T20:20:17 trying to find a tile on edge of vision that isn't blocked by wall 2011-10-26T20:20:31 Puj: I was :) 2011-10-26T20:20:32 *** xathis_ has quit IRC () 2011-10-26T20:20:35 Via a direct path grom358? 2011-10-26T20:20:43 I meant e4 2011-10-26T20:20:45 yeah via a direct path 2011-10-26T20:20:49 Oh okay, McLeopold d5 2011-10-26T20:20:59 I already using raytrace to get to food 2011-10-26T20:21:21 Is there anyone to donate to, or do you guys accept donations? 2011-10-26T20:21:30 loktar, i do 2011-10-26T20:21:39 heh I mean Im sure everyone does :P 2011-10-26T20:21:41 loktar: yes, j3camero 2011-10-26T20:21:44 but the aichallenge runners 2011-10-26T20:21:46 :) 2011-10-26T20:21:48 ok cool 2011-10-26T20:21:49 *** replore_ has joined #aichallenge 2011-10-26T20:21:50 That should be the most efficient, you could make it more efficient with dp, storing the already computed rays for later use when computing subrays and superays of those stored rays. 2011-10-26T20:22:15 loktar: you can send an email to the staff email list, I'm sure he can get back to you 2011-10-26T20:22:23 Yeah Ill do that, thanks 2011-10-26T20:22:38 McLeopold: well, the idea was to get a separate server where we'll have a don't touch enviroment 2011-10-26T20:22:39 Puj: yeah I did have idea to only calculate the ray trace once I reach the target 2011-10-26T20:22:45 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-26T20:22:49 a new ray I mean 2011-10-26T20:22:51 If someoen throws a tcp game with 1000x1000 map, my bot can handle it 2011-10-26T20:22:56 McLeopold: but jeff doesn't like that, he seems to think it'll be costly in the long run 2011-10-26T20:23:13 I would do the separate machine 2011-10-26T20:23:24 but I was hoping todo 360 degree ray trace every turn so I can see which tiles lead to walls 2011-10-26T20:23:41 maybe not an ec2 instance though 2011-10-26T20:23:42 McLeopold: janzert wants that too, can you guys reply to him supporting what i said? 2011-10-26T20:23:43 *** Knekkebjoern has quit IRC (Ping timeout: 258 seconds) 2011-10-26T20:23:52 McLeopold: i don't need ec2 2011-10-26T20:23:55 what about his current vps? 2011-10-26T20:24:17 it went down a few times on its own, i don't trust it being stable 2011-10-26T20:24:24 and i don't want to tell him that 2011-10-26T20:24:29 does that matter for the archive? 2011-10-26T20:24:31 Well, grom358, you would be scanning at least every tile in vision in that case, which is at best worst than bfs and you would need some trig to get the rotational angles correct, so you may be limiting yourself. 2011-10-26T20:24:53 bfs? 2011-10-26T20:24:53 McLeopold: it does, because someone will have to actively work on keeping it up 2011-10-26T20:25:01 Breadth-first search 2011-10-26T20:25:07 *** Apophis has quit IRC (Ping timeout: 244 seconds) 2011-10-26T20:25:22 *** bobbydroptable has joined #aichallenge 2011-10-26T20:25:30 Puj: actually I don't need trig to rotate.. I just get the set of tiles at the edge of vision and ray trace to each and see if the path was clear 2011-10-26T20:26:25 Looks like it works if I open the replay html file in firefox instead of using the jar. Thanks 2011-10-26T20:26:26 *** Knekkebjoern has joined #aichallenge 2011-10-26T20:26:27 Cool. I don't know if it is going to be faster though, grom358 2011-10-26T20:27:16 mmm.. where on the website did it have diagram of vision? 2011-10-26T20:27:32 xathis 3rd now. well done! 2011-10-26T20:27:50 *** Six has joined #aichallenge 2011-10-26T20:27:57 As an aside, google image search is not entirely valuable. 2011-10-26T20:29:39 amstan: so, what does jeff want to do? 2011-10-26T20:29:41 http://aichallenge.org/visualizer.php?game=18428&user=4513 wow that last 10 moves was BRUTAL 2011-10-26T20:30:00 McLeopold: aren't you on the mailing list? 2011-10-26T20:30:13 yes, hold on... 2011-10-26T20:30:43 uh, maybe not 2011-10-26T20:31:27 McLeopold: /notice me your email 2011-10-26T20:31:42 its my nick at gmail 2011-10-26T20:32:27 1 person is already a member and will not receive an invitation 2011-10-26T20:32:45 is it a google group? 2011-10-26T20:32:49 yes 2011-10-26T20:33:12 i forwarded the thread 2011-10-26T20:33:15 to you 2011-10-26T20:33:54 ok 2011-10-26T20:34:11 so, a subdomin hosted on the current server would work for now 2011-10-26T20:34:37 McLeopold: another issue is that the planetwars stuff is big(read 50GB) 2011-10-26T20:34:47 McLeopold: ideally we should pick a server and stick to it 2011-10-26T20:35:17 yeah, idealy we have 1 site that hosts all contests and keeps them all going 2011-10-26T20:37:15 gtg 2011-10-26T20:37:19 *** McLeopold has quit IRC (Quit: Leaving.) 2011-10-26T20:37:55 *** Apophis has joined #aichallenge 2011-10-26T20:40:14 *** pgoldbr has joined #aichallenge 2011-10-26T20:40:28 hm, i am not on this mailing list 2011-10-26T20:40:30 *** isbric_ has quit IRC (Ping timeout: 265 seconds) 2011-10-26T20:40:58 Hi there! Did anyone else noticed they aren't running matches right now? 2011-10-26T20:42:02 good. i need time to catch up 2011-10-26T20:42:02 :D 2011-10-26T20:43:19 *** andresilva91 has joined #aichallenge 2011-10-26T20:45:21 *** andresilva91 has quit IRC (Client Quit) 2011-10-26T20:46:12 gah.. how to calculate which tiles are on the border of the vision radius 2011-10-26T20:48:32 there should be a function called vision() or get_vision() or the like 2011-10-26T20:48:42 there's some code there that computes the view circle 2011-10-26T20:49:02 yeah.. it returns all visible vision tiles.. I just want ones on the edge 2011-10-26T20:49:24 i'm saying , disect that code and figure out how to get only the edges ? 2011-10-26T20:49:43 yeah I am.. and I'm stuck 2011-10-26T20:49:47 oh ok 2011-10-26T20:49:48 lol 2011-10-26T20:49:51 why do you need it for? 2011-10-26T20:50:19 cause I want to ray trace to each edge to see if can reach it 2011-10-26T20:50:42 trying to find edge node with no walls between antLoc and it 2011-10-26T20:50:54 *** danielharan has joined #aichallenge 2011-10-26T20:51:07 oh... 2011-10-26T20:51:13 haha you're at my stage! :D 2011-10-26T20:51:21 i used A* to solve that 2011-10-26T20:51:27 I got A* too 2011-10-26T20:51:35 but its too slow for general movement 2011-10-26T20:51:50 yeh i abandoned it eventually gone on to other things 2011-10-26T20:51:54 only gonna use A* for certain things 2011-10-26T20:52:02 so it hasn't gone to waste 2011-10-26T20:52:22 funnily enough the A* isn't the most complicated thing to code lol 2011-10-26T20:52:29 but general exploring I got working.. but want to improve it by looking out a bit more then one tile 2011-10-26T20:52:31 after you'd written it over and over 4-5 times 2011-10-26T20:52:55 and I already got raytracing working to food 2011-10-26T20:53:06 oh nice 2011-10-26T20:53:17 *** joakimar has joined #aichallenge 2011-10-26T20:53:45 but to integrate it into the explore code.. I need to figure out which tiles make the edge of the ant's vision 2011-10-26T20:54:24 and ray trace to each one (going in clockwise/anticlockwise direction) until I find a passable path 2011-10-26T20:55:50 Site seems to not play games right now? 2011-10-26T20:56:34 *** herpaderp has joined #aichallenge 2011-10-26T20:56:53 hey, anyone here? 2011-10-26T20:58:37 hello 2011-10-26T20:58:41 not sure about admins 2011-10-26T21:00:43 so in my profile page, it says that there are 0.0 games per minute, and my next game 'could take a while...' 2011-10-26T21:01:06 i'm guessing that's not normal? 2011-10-26T21:02:15 *** user1 has quit IRC (Ping timeout: 265 seconds) 2011-10-26T21:02:40 That's not normal. Usually there are ongoing games and an estimate for your next game is shown 2011-10-26T21:03:20 visualize.jar started failing for me just today 2011-10-26T21:03:46 even with the supplied play_one_game_live.sh 2011-10-26T21:03:49 anyone else seeing 0.0 games per minute in their profiles? 2011-10-26T21:04:00 http://aichallenge.org/server_stats.php 2011-10-26T21:04:04 herpaderp: yes me too, so I think it's out of order 2011-10-26T21:04:08 workers have gone down 2011-10-26T21:04:18 \help 2011-10-26T21:04:29 contestbot: seen janzert 2011-10-26T21:04:29 amstan: janzert was last seen in #aichallenge 2 hours, 6 minutes, and 42 seconds ago: we tried shortly before the contest started but the newer versions available at that point at least had other problems 2011-10-26T21:04:43 @later tell janzert workers seem to be down 2011-10-26T21:04:43 amstan: I come to serve. 2011-10-26T21:04:59 contestbot never sleeps. 2011-10-26T21:05:00 bobbydroptable: User error, it's not my fault. 2011-10-26T21:05:16 contestbot: uptime 2011-10-26T21:05:16 amstan: I have been running for 5 days, 5 hours, 48 minutes, and 16 seconds. 2011-10-26T21:07:13 anybody seen problems with visualizer.jar? It might be due most recent Mac OS X update:( 2011-10-26T21:08:19 *** he_the_great has joined #aichallenge 2011-10-26T21:08:44 *** herpaderp has quit IRC (Quit: Page closed) 2011-10-26T21:10:32 joakimar local or on the server? 2011-10-26T21:10:40 janzert: looks like our workers are down 2011-10-26T21:11:31 joakimar: I didn't tried visualize.jar yet, but I have an osx update on idle 2011-10-26T21:12:51 mmm.. I guess what I need is to draw a circle onto tiles.. lol can't even figure out how todo that 2011-10-26T21:14:03 pythagoras? meh, I lack basic knowledge about the circle 2011-10-26T21:14:20 I don't know whats better 2011-10-26T21:14:36 mmm.. http://en.wikipedia.org/wiki/Midpoint_circle_algorithm 2011-10-26T21:14:55 joakimar visualize.jar crashing on macos 10.7.2 :( 2011-10-26T21:15:31 joakimar CanvasElement.js:1038 Cannot find default value for object 2011-10-26T21:15:50 *** letmetellyou has joined #aichallenge 2011-10-26T21:16:13 when you have something like 2011-10-26T21:16:15 a row col owner # live ant 2011-10-26T21:16:24 then what value can owner have, and what does each value mean? 2011-10-26T21:17:03 survivor: yes that's what I'm seeing in visualize.jar too, for 10.6.8. Probably due to latest webkit I guess. 2011-10-26T21:17:41 survivor: I found a workaround – play the game non-live, and the replay opens automatically in my browser (Chrome) which works 2011-10-26T21:18:24 0 to n-1 2011-10-26T21:18:43 letmetellyou 0 to n-1, where n is the number of players on the map 2011-10-26T21:18:55 *** aarossig has quit IRC (Ping timeout: 252 seconds) 2011-10-26T21:19:00 0 means your ant. everything else is other players 2011-10-26T21:19:04 *** Six has quit IRC (Ping timeout: 265 seconds) 2011-10-26T21:19:07 *** Chirmaya has joined #aichallenge 2011-10-26T21:19:08 ok thx 2011-10-26T21:19:55 I have a general programming question if someone could help me (or try to ^^) 2011-10-26T21:20:21 Chirmaya go 2011-10-26T21:20:44 *** aarossig has joined #aichallenge 2011-10-26T21:21:18 sweet.. !! 2011-10-26T21:21:45 *** Knekkebjoern has quit IRC (Quit: Leaving.) 2011-10-26T21:21:46 I just implemented an A* pathfinding algorithm, and it seems to work fine until my ants get vision on the solid dividing wall on that first map 2011-10-26T21:22:07 Right then I get a timeout, but I am not sure how I should go about debugging it 2011-10-26T21:22:33 I wonder how many of the top100 bots, make/made their own path finder 2011-10-26T21:22:57 chirmaya: chances are, it is timing out because it is taking too long to work its way around the wall 2011-10-26T21:23:11 output some messages to stderr 2011-10-26T21:23:17 So it could be timing out even though it isn't in an infinite loop? 2011-10-26T21:23:23 yup 2011-10-26T21:23:27 *** danielharan has quit IRC (Ping timeout: 248 seconds) 2011-10-26T21:23:34 find out the relevant method for outputting to stderr 2011-10-26T21:23:45 and put debugging messages inside to see what's taking so long 2011-10-26T21:24:03 Is there a processing timelimit per turn then? 2011-10-26T21:24:06 burny: seeing as there is no pathfinding code in the starterpack, probably a huge proportion? 2011-10-26T21:24:10 yes 2011-10-26T21:24:11 Chimaya: 1000ms i believe 2011-10-26T21:24:17 i.e. 1sec 2011-10-26T21:24:43 Oh, nice. Thanks a lot! Now I totally know where to start =) 2011-10-26T21:24:55 its worst 2011-10-26T21:24:59 500 2011-10-26T21:25:01 500ms official on the server 2011-10-26T21:25:03 i thought 2011-10-26T21:25:15 http://aichallenge.org/game_settings.php 2011-10-26T21:25:26 (hmm, I wonder...) 2011-10-26T21:25:29 @settings 2011-10-26T21:25:30 avdg2: settings could be http://aichallengebeta.hypertriangle.com/game_settings.php. 2011-10-26T21:25:34 :-) 2011-10-26T21:26:05 that's still plenty of time XD 2011-10-26T21:26:05 haha 2011-10-26T21:26:30 viewradius has gone up 2011-10-26T21:26:35 used to be 55 2011-10-26T21:26:36 i think 2011-10-26T21:26:50 lots of time for loading though, should precompute stuff there... 2011-10-26T21:26:53 ya, half the docs sugest it's 55 2011-10-26T21:26:54 bobbydroptable: check again the settings ;-) 2011-10-26T21:27:03 its 77 2011-10-26T21:27:16 i thought it was 55 2011-10-26T21:27:23 *** joakimar has quit IRC (Ping timeout: 265 seconds) 2011-10-26T21:27:38 just check the link above, it contains the configuration settings ;-) 2011-10-26T21:27:48 i said it USED to be 55 :p 2011-10-26T21:27:51 *** dominater has joined #aichallenge 2011-10-26T21:28:13 oh well 2011-10-26T21:28:35 lol 2011-10-26T21:30:43 *** danielharan has joined #aichallenge 2011-10-26T21:32:10 *** alc has joined #aichallenge 2011-10-26T21:32:13 *** loktar has quit IRC (Ping timeout: 265 seconds) 2011-10-26T21:34:26 *** survivor has left #aichallenge 2011-10-26T21:36:36 *** [1]analyst74 <[1]analyst74!~analyst74@74-115-199-33.eng.wind.ca> has joined #aichallenge 2011-10-26T21:38:24 *** analyst74 has quit IRC (Ping timeout: 240 seconds) 2011-10-26T21:38:24 *** [1]analyst74 is now known as analyst74 2011-10-26T21:42:29 *** jasper has joined #aichallenge 2011-10-26T21:43:42 how long does it usually take for a zip to be opened and built? 2011-10-26T21:43:56 depends on the size, mode, etc. 2011-10-26T21:44:29 its the sample pack for c++, with one extra file that is around 100 lines or so 2011-10-26T21:44:38 and disk speed in case the cpu is fast enough ;-) 2011-10-26T21:44:40 and its been about an hour? 2011-10-26T21:44:54 compiling? 2011-10-26T21:45:06 approximatly zero seconds then 2011-10-26T21:45:13 +/- 1 second 2011-10-26T21:45:25 or servers 2011-10-26T21:45:33 that are sleeping 2011-10-26T21:45:46 ahh sleeping servers 2011-10-26T21:45:48 hmm 2011-10-26T21:45:50 @stats 2011-10-26T21:45:50 avdg2: I have 1 registered users with 0 registered hostmasks; 1 owner and 0 admins. 2011-10-26T21:45:55 bleh 2011-10-26T21:46:00 http://aichallenge.org/server_stats.php 2011-10-26T21:46:04 @serverstats 2011-10-26T21:46:05 avdg2: I'm sorry Dave, err avdg2; I cannot 'serverstats'. 2011-10-26T21:46:35 *** canadiancow|work has joined #aichallenge 2011-10-26T21:46:53 games in the last hour is 0.0 2011-10-26T21:47:05 0.0 2011-10-26T21:47:28 (its an emotic) 2011-10-26T21:47:40 o.0 ok, i assume something is happening on server side, i'll just wait it out till morning 2011-10-26T21:49:46 give me time to refactor my path finding :) 2011-10-26T21:50:00 jasper : as i was saying, gives me time to catch up 2011-10-26T21:50:02 :D 2011-10-26T21:52:07 *** Phaedrus_ has joined #aichallenge 2011-10-26T21:53:20 *** ChrisH_ has joined #aichallenge 2011-10-26T21:54:10 *** tdubellz has joined #aichallenge 2011-10-26T21:54:23 *** Agenttokyo has joined #aichallenge 2011-10-26T21:54:46 *** Six has joined #aichallenge 2011-10-26T21:56:04 *** alc has quit IRC (Remote host closed the connection) 2011-10-26T21:56:19 *** clckwrk has joined #aichallenge 2011-10-26T21:57:00 *** clckwrk has quit IRC (Client Quit) 2011-10-26T21:57:15 *** danielharan has quit IRC (Ping timeout: 258 seconds) 2011-10-26T21:59:02 *** sunshowers_ has joined #aichallenge 2011-10-26T21:59:07 *** alc has joined #aichallenge 2011-10-26T21:59:19 *** sunshowers_ has quit IRC (Client Quit) 2011-10-26T21:59:29 *** Sunshowers has joined #aichallenge 2011-10-26T22:02:46 yay, I finally see the word I put in :-) 2011-10-26T22:03:01 lol 2011-10-26T22:03:15 http://aichallenge.org/visualizer.php?game=18554 2011-10-26T22:03:26 that's the fad now, huh 2011-10-26T22:03:26 okay I am using http://en.wikipedia.org/wiki/Midpoint_circle_algorithm .. but I need the tiles in clockwise, anticlockwise order 2011-10-26T22:03:35 any ideas how todo that? 2011-10-26T22:04:00 avdg2 lol awesome game 2011-10-26T22:04:08 grom358: I was using midpoint circle, but gave up and did BFS to do the same thing 2011-10-26T22:04:10 grom358 no idea sorry 2011-10-26T22:04:14 probably some sort of trig 2011-10-26T22:04:22 grom358: since you dont really want midpoint circle since there are wallss 2011-10-26T22:04:28 f4hy: BFS? 2011-10-26T22:04:34 breadth firsst search 2011-10-26T22:04:37 100% what you want 2011-10-26T22:05:04 f4hy: what are u doing exactly though.. I want to search in clockwise/anticlockwise order 2011-10-26T22:05:29 grom358: oh... why? 2011-10-26T22:05:57 grom358: why not search radially outward, then tag which is the mosst clockwise afterwards 2011-10-26T22:06:13 probably more effiecent anyway 2011-10-26T22:06:14 f4hy: most clockwise?? 2011-10-26T22:06:24 from wherever you call calling zero 2011-10-26T22:06:33 say north, or wahtever you are searching clockwise 2011-10-26T22:07:12 well I have working explore code now.. but it works on single tiles.. trying to extend it to use its field of vision 2011-10-26T22:07:18 do BFS tag everything in the radius you want, then order them based on angle 2011-10-26T22:07:32 what is the 'problem' with circles? 2011-10-26T22:07:35 f4hy: order based on angle? 2011-10-26T22:07:47 grom358: well, how else are your searching clockwise? 2011-10-26T22:07:50 or, is it not related to tants? 2011-10-26T22:08:18 kk.. what I do is ray trace in a direction.. if there is wall.. rotate clockwise 2011-10-26T22:08:28 *** Agenttokyo has quit IRC (Ping timeout: 265 seconds) 2011-10-26T22:09:12 I want to generate the order at loadtime.. eg. Have tile offsets that draw a circle in clockwise order 2011-10-26T22:09:26 if I have that.. I just raytrace to each of the tiles 2011-10-26T22:09:29 what are you drawing for? 2011-10-26T22:09:37 why do you want to raytrace? 2011-10-26T22:09:46 well.. not draw as such. 2011-10-26T22:10:00 but have list of tiles in order 2011-10-26T22:10:15 f4hy: I raytracing to see if I can travel in that direction 2011-10-26T22:10:28 probably has the same problem i did, which is figuring out where to go when an ant has nothing to do, i guess 2011-10-26T22:10:51 you can't move on angles anyway 2011-10-26T22:10:57 only horintal/verticle 2011-10-26T22:11:01 bobbydroptable: yes.. but i have working code for that.. its just I only look at adjacent tiles 2011-10-26T22:11:12 burny: yes the raytrace takes care of that 2011-10-26T22:11:23 does your ray need to be exactly the right length? 2011-10-26T22:11:38 and, do you need to hit exactly every tile? 2011-10-26T22:11:42 burny: my raytrace works fine already.. I got it going to food with it 2011-10-26T22:11:50 if not, there are some real easy approximations 2011-10-26T22:11:57 *** kornork has joined #aichallenge 2011-10-26T22:12:11 grom358: I have no idea how raytracting could possibly be helpful. If you want to generate circles, BFS is the easiest way to load a 'circles' worth of tiles around a point. 2011-10-26T22:12:16 got no problem with my raytrace.. but I'm trying to use it for exploring now 2011-10-26T22:12:41 its even more "extermination" if 8 bots are doing it the same time :-) http://aichallenge.org/visualizer.php?game=18431 2011-10-26T22:12:45 f4hy: how with the BFS are u working out what direction to travel in?? 2011-10-26T22:13:28 grom358: search for target in BFS, find target, move towards it 2011-10-26T22:14:06 atm I keep track of visited nodes.. then I try to travel in a certain direction (say NORTH). If can't go that way go to next direction in clockwise/anticlockwise direction 2011-10-26T22:14:20 that is not actually what i do, because my ants don't scan for directions where to go as such, but I do something using 'circles' 2011-10-26T22:14:25 and don't travel back to visited nodes unless at deadend 2011-10-26T22:14:31 worker pricing shot up to over 10 times the normal price, looks like it's coming back down now though 2011-10-26T22:14:38 so they should be back soon 2011-10-26T22:14:42 amstan: ^^ 2011-10-26T22:15:06 janzert: wow, I wonder what happened. 2011-10-26T22:15:18 which works fine already.. but what happens is it will go down one row.. then loop back on a row down.. cause it has to physically go to the tile to add it to visited nodes 2011-10-26T22:15:23 yeah, I don't remember seeing it go that high before 2011-10-26T22:15:43 I personally never do any searches from an ant. Instead I search outwards from goals 2011-10-26T22:15:51 *** ademar has joined #aichallenge 2011-10-26T22:16:00 but most people do BFS from each ant (or something like a* which is just weighted BFS) 2011-10-26T22:16:05 also, do you think ec2 pricing will go up as the holidays approach and Amazon starts getting busier. I thought I heard that they consumed ec2 resources to deal with peak holiday traffic. 2011-10-26T22:16:17 f4hy: yeah I got a working a* too 2011-10-26T22:16:28 I dont have an a* 2011-10-26T22:16:35 I dont need it I dont think 2011-10-26T22:16:46 *** Larose has joined #aichallenge 2011-10-26T22:17:06 eg, given x2,y2: with center x,y. the next point could be x+(x2-x)*0.99-(y2-y)*0.14, y+(y2-y)*0.99+(x2-x)*0.14 2011-10-26T22:17:17 My ant does alright without it, I have a pretty decent ant, which gathers food, attacks, defeneds, with no explcit pathing 2011-10-26T22:17:44 this is for my explore code.. I want to look around from start direction.. say NORTH (12 oclock) and raytrace in clockwise direction until I find path with no walls 2011-10-26T22:17:44 pathin is more important to centralized planning 2011-10-26T22:18:25 *** avdg2 is now known as avdg 2011-10-26T22:18:47 I don't have path finding either, but I assume pure agent-based AI would have limitation on how well it performs 2011-10-26T22:18:58 like u pointing a laser and turning 2011-10-26T22:19:24 analyst74: what is agent-based mean? 2011-10-26T22:19:36 meaning treat every ant as its own individual 2011-10-26T22:20:01 hardest part about per-agent, would be makign them not step on each other 2011-10-26T22:20:06 or, block each other 2011-10-26T22:20:21 well some sort of centralized thing would be necessary 2011-10-26T22:20:35 analyst74: http://ants.fluxid.pl/replay.6652 My bot vs yours ;-) 2011-10-26T22:21:31 but ultimately, in agent -oriented 2011-10-26T22:21:57 you assign individual goal/task/objective to each ant and each attempts to achieve it on their own accord 2011-10-26T22:22:10 I dont do anything like that 2011-10-26T22:22:35 most people won't, they just loop through, say, all the food, then move the nearest ant to the food 2011-10-26T22:22:53 I am a physicist, so I don't really know much about AI, so I treated it as a physics problem and had all things that I desire emit a potential field of various strength and radius, and then ants just follow the gradiant to higher potential. 2011-10-26T22:22:58 agent oriented would say "okay ant, here's a list of food, decide which food you want to get to" 2011-10-26T22:23:14 bobbydroptable: ahh ok makes sense 2011-10-26T22:23:19 that was how i did my first bot :D almost 2011-10-26T22:23:23 didn't work very well for me 2011-10-26T22:23:37 bobbydroptable: I just needed stuff to get out of local extrema 2011-10-26T22:23:40 ChrisH_: historically it hasn't 2011-10-26T22:23:42 came up with something clever, works fairly well 2011-10-26T22:23:49 f4hy: My first bot worked like that too. In AI that's called influence mapping. 2011-10-26T22:24:04 ChrisH_: sweet, I learned what my bot does! 2011-10-26T22:24:23 I redsecovered influence mapping 2011-10-26T22:24:34 *** Phaedrus_ has quit IRC (Quit: Page closed) 2011-10-26T22:24:53 f4hy there is more you can do with that technique, so now you know what to plug into goole to get ideas. :) 2011-10-26T22:25:31 *** djr__ has joined #aichallenge 2011-10-26T22:25:37 ChrisH_: Thanks! all I have been doing is tuing parameters and setting up specific situations where I change what influences what. 2011-10-26T22:26:12 such as if you are near an enemy hive, increase the influence to group up, otherwise spread out. So the various potential fields interact in fun ways 2011-10-26T22:26:20 here is another great site for ideas that I've seen others post here: http://www-cs-students.stanford.edu/~amitp/gameprog.html 2011-10-26T22:27:17 f4hy: yeah, it can be fun, but I found it is also difficult to get it tuned in a way that avoids certain bad behaviors 2011-10-26T22:27:39 f4hy: but I'm not an expert on the technique either 2011-10-26T22:28:26 ChrisH_: Ya I had some fun isssues to work out. Such has getting stuck at a midpoint between two goals and such 2011-10-26T22:28:33 I had a problem early on where four ants would all fight over the same food because they were simultaneously attracted to the food, but repelled from each other 2011-10-26T22:28:35 so.. with midpoint circle I get the tiles that make up the edge of the vision and can raytrace to each one.. but its not in clockwise order 2011-10-26T22:28:38 after solving the local extrma isssues htough it works great 2011-10-26T22:29:11 *** djr_ has quit IRC (Ping timeout: 252 seconds) 2011-10-26T22:29:22 ChrisH_: which bot is yours? is it on the tcp server? 2011-10-26T22:29:42 does your ray need to be exactly the right length? and, do you need to hit exactly every tile? 2011-10-26T22:30:01 *** mleise has joined #aichallenge 2011-10-26T22:30:08 My bot is "ChrisH" on the main server. I played a few on the TCP server last week under a different name, but I don't have it up there now. 2011-10-26T22:30:24 I'll show u some sample code in a minute.. just packaging it together 2011-10-26T22:30:34 ChrisH_: aww, I leave mine running on the TCP server all day. I just like watching it fight. 2011-10-26T22:30:47 eg, instead of hitting 10,10 10,11 9,11... could it do 10,10 9,11? 2011-10-26T22:30:58 It is not the best bot, but I think fairly decent considering I only spend 3ms/turn tops. 2011-10-26T22:31:27 burny: no.. the path it returns does not have diagonal moves 2011-10-26T22:31:29 f4hy: it can be addictive, but I'm doing my development in a single CPU VM, so I don't have the resources to keep a bot running while I code. 2011-10-26T22:31:44 nevermind 2011-10-26T22:32:35 ChrisH_: I just have the tcp running on a stupid lab computer. copied the binary over and started the script. uses bassically zero cpu so I am not worried. 2011-10-26T22:33:14 *** McLeopold has joined #aichallenge 2011-10-26T22:33:41 f4hy: I definitely use more than 3ms per turn, and my code is pretty optimized where it needs to be, but I'm doing a LOT of pathfinding, and it's gotten to the point where I need to look for alternatives to avoid timing out on bit maps with long twisty hallways and trying to path 120 ants halfway across the map. 2011-10-26T22:33:44 ya, running the bot on TCP uses less than 1% cpu on a pretty old machine 2011-10-26T22:34:12 I mean, big maps 2011-10-26T22:34:19 *** DecodedCortex has joined #aichallenge 2011-10-26T22:34:22 Zero pathfinding, which is why mine is fast. 2011-10-26T22:34:28 ChrisH i have that same problem, and I hope i've solved it ^^ 2011-10-26T22:34:30 I will probably have to add pathfinding eventually 2011-10-26T22:34:36 but not sure why I need it 2011-10-26T22:35:41 bleh I hate numeric strings :-) *shames red-dead for not converting strings to integers* 2011-10-26T22:35:49 bobbydroptable: I missed your solution, do you mind summarizing? 2011-10-26T22:36:31 http://pastebin.com/MttHnFeU 2011-10-26T22:37:07 You will see I get possible paths to travel with it.. 2011-10-26T22:37:12 grom358: hmmmm, would that be a hint for others? :-) 2011-10-26T22:37:15 ChrisH_ quadtree. 2011-10-26T22:37:59 bobbydroptable: you mean as a way to do hierarchical search? 2011-10-26T22:38:03 fixed my problem of building a global graph over a partially observable environment. 2011-10-26T22:38:07 oh man, defense for the win http://ants.fluxid.pl/replay.7712 2011-10-26T22:38:11 yup! 2011-10-26T22:38:14 *** letmetellyou has quit IRC (Quit: Page closed) 2011-10-26T22:38:45 which means long distances shouldn't be a problem anymore, because I won't need to A* groups of unpassable obstacles 2011-10-26T22:38:58 that being said, I'm still working out the pathfinding 2011-10-26T22:39:10 finding neighbours in a tree isn't very intuitive 2011-10-26T22:39:20 :-) 2011-10-26T22:39:32 bobbydroptable: I implemented A*, but it turns out that many-to-many Djikstra's (my old way) is slower than many to many A* because of the added overhead of the heuristic function. 2011-10-26T22:39:39 *** mellodroit has joined #aichallenge 2011-10-26T22:39:58 bobby, it's intuitive if you're a monkey 2011-10-26T22:40:08 I think Dijkstra's only when you have different node costs 2011-10-26T22:40:08 ouch. burn 2011-10-26T22:40:10 :( 2011-10-26T22:40:17 lol 2011-10-26T22:40:19 bobbydroptable: But I've got some other ideas that I think will move my code forward. 2011-10-26T22:40:28 sweet. let me know how they go then 2011-10-26T22:41:12 lol avdg poor yello 2011-10-26T22:41:15 McLeopold: It's really a BFS with a priority queue so I can have multiple start points, so it's not full Djikstra's with weights (yet) 2011-10-26T22:42:05 bobbydroptable, in case you didn't get the joke.. monkey.. tree.. 2011-10-26T22:42:14 cyphase i did :P 2011-10-26T22:42:17 lol 2011-10-26T22:44:47 mmm.. i just came up with really hacky way of making this circle in clockwise order 2011-10-26T22:45:28 ohh, woops, in my earlier comment, I meant that my new A* code turned out slower than my old code, even after tuning. It surprised me how much expensive computing lots of manhattan distances with wrapping could be. 2011-10-26T22:46:26 "how expensive", I should go back to typing code, my english typing is bad tonight 2011-10-26T22:46:44 hehe 2011-10-26T22:46:50 brainfart 2011-10-26T22:47:01 more than one 2011-10-26T22:50:11 *** maccheese has joined #aichallenge 2011-10-26T22:53:11 *** maccheese has quit IRC (Client Quit) 2011-10-26T22:56:11 one server is back online 2011-10-26T22:56:55 and there was much rejoicing 2011-10-26T22:57:12 :-) 2011-10-26T22:58:01 ChrisH_ is it true that there is no way to do packages in go in the contest given how things are compiled? 2011-10-26T22:58:25 jcdny: yes, you get your one main package and that's all 2011-10-26T22:58:27 :( 2011-10-26T22:58:50 pesky 2011-10-26T22:59:01 Maybe they will let us have packages for the next contest. 2011-10-26T22:59:18 or rather, maybe the code that compiles Go can be improved to allow it. 2011-10-26T22:59:39 this is my motivation to learn Go although I think the version I am learning not doing packages is less joyful than it could be 2011-10-26T23:00:07 I helped them fix some bugs with it this time around, so maybe I'll contribute some stuff to get packages working for next time. 2011-10-26T23:00:18 jcdny: same here 2011-10-26T23:00:57 I absolutely love working in go though. 2011-10-26T23:00:59 jcdny: have you tried doing anything with go routines? 2011-10-26T23:01:10 I did some toy stuff. 2011-10-26T23:01:14 amstan: ping 2011-10-26T23:01:20 McLeopold: pong 2011-10-26T23:01:31 jcdny: but nothing in the contest? 2011-10-26T23:01:36 can you get the visualizer working in the forums? 2011-10-26T23:01:37 planning on having the turn read nonblocking with channels 2011-10-26T23:02:14 and possibly attempting some sort of preemption for idle time calcs. 2011-10-26T23:02:22 jcdny: My turn read is all done with channels underneath, but not go routines to drive it 2011-10-26T23:02:57 jcdny: yeah, i've considered it too, but so far I'm focusing on the bot logic more 2011-10-26T23:03:39 McLeopold: yes, can you get me the javascript that invokes it? 2011-10-26T23:03:49 jcdny: what's your bot's name? 2011-10-26T23:03:51 me too. I think the non blocking reads is really easy though so I might do that soon 2011-10-26T23:03:55 bugnuts 2011-10-26T23:04:16 amstan: you need the visualizer-min.js and something from the widget.php 2011-10-26T23:04:49 cool, hope to see you continue up the charts. 2011-10-26T23:05:04 right now it has no pathfinding so is necessarily pretty limited 2011-10-26T23:05:15 McLeopold: i can make the game replay to be inline, or you could give it a game number 2011-10-26T23:05:52 yeah, I'm trying to improve the performance of my path finding right now in fact, it is already working and pretty fast, but it needs to be faster so I have more time for combat tactics and other cool stuff on the bigger maps 2011-10-26T23:05:57 I have a good start on prioritizing goals but w/o pathfinding it's not doing much. 2011-10-26T23:06:14 hmmm, phpbb doesn't do pre tags? 2011-10-26T23:06:34 jcdny: yeah, I've been there. 2011-10-26T23:06:57 *** jasper has quit IRC (Ping timeout: 265 seconds) 2011-10-26T23:07:18 you converted your planet wars code to go didn't you? 2011-10-26T23:07:22 jcdny: gotta love slices though, my code is mostly just slice manipulation 2011-10-26T23:07:51 amstan: I'll have to look at that later, the js needs a pre tag to inspect. it looks like phpbb does code tags with
tags in the middle 2011-10-26T23:07:56 jcdny: I converted about half of it, I had the attack logic working, but not the defense when I moved on to other projects 2011-10-26T23:08:29 yeah it makes things nice. leary about whats under the hood though I am mostly trying to avoid premature optimization. 2011-10-26T23:08:40 did you put it up anywhere? 2011-10-26T23:08:45 jcdny: no 2011-10-26T23:09:26 I put all my code in a private repo on github & plan to make it public when things are done. 2011-10-26T23:10:09 McLeopold: i'm pretty close 2011-10-26T23:10:15 McLeopold: i just need the syntax for the inline version 2011-10-26T23:10:21 McLeopold: i'm almost done with the provide id part 2011-10-26T23:10:25 jcdny: I think slices are pretty good under the covers. I've only been optimizing when forced to, but that has happened multiple times now. I worked through the gopprof examples a few weeks ago, though, and that tool has made figuring out what needs speed improvements so much easier. 2011-10-26T23:11:29 and now that you can download the bot input for any game, it is easy to debug timeouts locally by running my code with the input stream and profiling turned on. 2011-10-26T23:11:45 visualizer-min.js:1Uncaught TypeError: Cannot call method 'hasChildNodes' of null 2011-10-26T23:11:50 I wrote my first hello world program on wednesday so I am still learning all the idioms and tools 2011-10-26T23:12:24 *** Vivitron has joined #aichallenge 2011-10-26T23:13:11 jcdny: oh, then you're doing great already up in the 500s in a week on a new language. 2011-10-26T23:13:24 jcdny: did you compete in PW? 2011-10-26T23:13:53 no I only heard about the contest on hacker news last week... 2011-10-26T23:14:07 *** adrian has joined #aichallenge 2011-10-26T23:14:15 contestbot: seen mleise 2011-10-26T23:14:15 amstan: mleise was last seen in #aichallenge 11 hours, 29 minutes, and 20 seconds ago: there is a whole lot of time outs :D, time for a compiled language. http://aichallenge.org/visualizer.php?game=16020&user=4677 2011-10-26T23:14:17 hello all 2011-10-26T23:14:30 McLeopold: doesn't seem to like me: http://pastehtml.com/view/bby4xhqkj.html 2011-10-26T23:14:33 *** adrian is now known as Guest50273 2011-10-26T23:14:45 jcdny: so you're also learning the ropes of the contest as well, wow, keep at it and I bet you'll do well 2011-10-26T23:14:51 I have a question 2011-10-26T23:15:04 does any one know how to do looging in python 2011-10-26T23:15:07 oh.. that's why! 2011-10-26T23:15:08 yeah gl to you too 2011-10-26T23:15:10 how do we set the level 2011-10-26T23:15:15 logging level 2011-10-26T23:15:17 ?? 2011-10-26T23:15:41 amstan: did you place the visualizer in an existing html tag? 2011-10-26T23:16:21 mleise: i forgot
I let the parsing of the engine feed to the starter pack I've downloaded, but it suddently crashed because it didn't know how to process those lines.... 2011-10-26T23:25:09 I've been coding my bot for a few days now and it's the first time it happens. 2011-10-26T23:25:16 bah, the nerves in my tooth are dead, root canal surgery :( 2011-10-26T23:25:26 Any thought would be nice. 2011-10-26T23:25:34 that should be normal happens to me too 2011-10-26T23:25:40 antimatroid1 ouch 2011-10-26T23:25:51 mleise: any ideas? 2011-10-26T23:26:57 I modified my starter kit to ignore commands with first token longer than 1 character 2011-10-26T23:27:37 *** chris__0076 has joined #aichallenge 2011-10-26T23:27:44 mleise: apparently the visualizer div has to be before the javascript stuff 2011-10-26T23:28:51 *** avdg has quit IRC (Quit: Leaving.) 2011-10-26T23:30:34 amstan: it runs in *fullscreen* for me. check out the options, and set the embedded flag: http://aichallenge.org/visualizer/jsdoc/symbols/Options.html 2011-10-26T23:30:49 *** avdg has joined #aichallenge 2011-10-26T23:31:10 *** Chris_0076 has quit IRC (Ping timeout: 258 seconds) 2011-10-26T23:31:12 mleise: i copied the code from the widget in the main page 2011-10-26T23:31:19 yeah, i should probably force fullscreen off 2011-10-26T23:32:11 or at least override the user setting in the other parameter called 'config', then the user can still activate it. 2011-10-26T23:32:35 That would be 'fullscreen' here: http://aichallenge.org/visualizer/jsdoc/symbols/Config.html 2011-10-26T23:32:36 but var config = {}; 2011-10-26T23:32:51 *** chris__0076 is now known as chris_0076 2011-10-26T23:32:56 var config = { fullscreen: false }; 2011-10-26T23:32:56 bobby & analyst: thanks. 2011-10-26T23:32:57 i set options.embedded = true now 2011-10-26T23:34:22 mleise: any other suggestions for the settings? 2011-10-26T23:34:44 mleise: i should probably disable the buttons and the keyboard control 2011-10-26T23:34:48 *** avdg has quit IRC (Ping timeout: 240 seconds) 2011-10-26T23:35:18 mleise: yes, i have to, it makes it impossible to type when replying otherwise. 2011-10-26T23:36:42 are you going to support inline games or just maps? 2011-10-26T23:37:10 McLeopold: i can do replacement of phpbb tags to arbitrary html, so anything is possible 2011-10-26T23:39:49 then you could do what the wiki does, but probably inforce embedde 2011-10-26T23:39:50 d 2011-10-26T23:40:24 *** canadiancow|work has quit IRC (Remote host closed the connection) 2011-10-26T23:42:39 *** FlyinFish has joined #aichallenge 2011-10-26T23:47:55 McLeopold: where can i find the code for that? 2011-10-26T23:47:59 *** jterracina has joined #aichallenge 2011-10-26T23:48:00 man, the map wrapping adds such an extra complexity ==" 2011-10-26T23:48:04 for what? 2011-10-26T23:48:09 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-10-26T23:48:13 McLeopold: for your wiki visualizer embeding 2011-10-26T23:48:57 *** FlyinFish has quit IRC (Quit: HydraIRC -> http://www.hydrairc.com <- The alternative IRC client) 2011-10-26T23:50:29 *** McLeopold has quit IRC (Read error: Connection reset by peer) 2011-10-26T23:50:42 *** jterracina has quit IRC (Client Quit) 2011-10-26T23:51:18 *** McLeopold has joined #aichallenge 2011-10-26T23:51:21 amstan: https://github.com/aichallenge/aichallenge/blob/epsilon/website/visualizer_widget.php#L19 2011-10-26T23:51:41 McLeopold: that's big 2011-10-26T23:51:44 McLeopold: what does it do? 2011-10-26T23:51:55 does it handle urls, gameids and data? 2011-10-26T23:52:01 looks for pre tags, then replaces them with the visualizer 2011-10-26T23:52:05 no 2011-10-26T23:52:13 I wouldn't allow game id's in the forums 2011-10-26T23:52:20 why not? 2011-10-26T23:52:37 because links are good enough, just the ability to describe maps and situations 2011-10-26T23:54:13 *** Larose has quit IRC (Quit: Leaving.) 2011-10-26T23:56:17 *** dvladim has joined #aichallenge 2011-10-26T23:56:41 *** jimi_hendrix has joined #aichallenge 2011-10-26T23:58:24 *** ccc_ has joined #aichallenge 2011-10-26T23:59:28 hi, how to turnoff the turntime value in the tools for local testing, I tried giving large value but it seems to not work (meaning my bot is always crashing)