2011-11-16T00:01:00 bugger. my old professor pulled down his lecture notes 2011-11-16T00:01:41 Is there a cheaper way to do distance queries versus paths? 2011-11-16T00:02:36 they are one in the same as far as i am concerned 2011-11-16T00:03:26 it might be possible to get cheaper approximate distance queries with a single multi-source BFS 2011-11-16T00:06:53 how is that approximate? 2011-11-16T00:07:46 because I'm interested in all pairs, not just the pairs I'm BFSing from 2011-11-16T00:07:53 *** bbuda has joined #aichallenge 2011-11-16T00:11:20 oh. duh. This goes back to an earlier idea. I think I can exploit unit-ness and manhattan-ness and planarity to get exact all pairs queries with something like O(n^2) preprocessing work 2011-11-16T00:11:50 *** djr_ has quit IRC (Quit: Leaving) 2011-11-16T00:13:32 *** RyokuHasu has quit IRC (Quit: Page closed) 2011-11-16T00:20:16 *** twymer has quit IRC (Ping timeout: 240 seconds) 2011-11-16T00:21:31 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-16T00:27:10 *** Areks has joined #aichallenge 2011-11-16T00:27:25 *** Jak_o_Shadows has joined #aichallenge 2011-11-16T00:36:59 *** bbuda has quit IRC (Quit: Page closed) 2011-11-16T00:39:24 *** datachomper has joined #aichallenge 2011-11-16T00:45:08 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-16T00:50:18 *** bmh has quit IRC (Quit: bmh) 2011-11-16T00:53:19 *** magiik has quit IRC (Remote host closed the connection) 2011-11-16T00:56:06 *** ivan`_ has joined #aichallenge 2011-11-16T00:57:07 *** Palmik has joined #aichallenge 2011-11-16T01:03:18 aichallenge: janzert epsilon * r2a16baf / worker/engine.py : Check for final bot output after pausing - http://git.io/8rcrgQ 2011-11-16T01:08:47 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-11-16T01:12:55 amstan: ping? 2011-11-16T01:20:16 bmh: remember the n in Floyd-Warshall is nodes, which means grid squares, which makes it n^6 by map width 2011-11-16T01:25:33 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-11-16T01:26:51 *** epicmonkey has joined #aichallenge 2011-11-16T01:28:58 @later tell amstan I disabled the planetwars apache configuration on the archive server, it basically caused the server to become an open http proxy which was already being utilized at a rate of several MBps :/ 2011-11-16T01:28:58 janzert: As you wish. 2011-11-16T01:33:52 wait until I finish downloading this episode of Lost with the pw site 2011-11-16T01:35:02 :) 2011-11-16T01:38:23 *** JorgeB has joined #aichallenge 2011-11-16T01:39:10 *** replore has quit IRC (Read error: Connection reset by peer) 2011-11-16T01:39:36 *** replore has joined #aichallenge 2011-11-16T01:42:53 *** dorisabayon has joined #aichallenge 2011-11-16T01:45:30 *** JorgeB has quit IRC (Ping timeout: 260 seconds) 2011-11-16T01:45:41 *** JorgeB has joined #aichallenge 2011-11-16T01:47:56 *** datachomper has joined #aichallenge 2011-11-16T01:49:59 *** dr- has joined #aichallenge 2011-11-16T01:57:45 *** epicmonkey has quit IRC (Ping timeout: 260 seconds) 2011-11-16T01:58:23 *** JorgeB has quit IRC (Ping timeout: 248 seconds) 2011-11-16T01:58:35 *** JorgeB has joined #aichallenge 2011-11-16T02:00:49 *** NoxiaZ^ has joined #aichallenge 2011-11-16T02:03:11 *** JorgeB has quit IRC (Ping timeout: 248 seconds) 2011-11-16T02:07:01 *** choas_ has quit IRC (Ping timeout: 258 seconds) 2011-11-16T02:12:07 *** Mooloo has joined #aichallenge 2011-11-16T02:12:27 *** Conorach has joined #aichallenge 2011-11-16T02:12:54 *** Flort has quit IRC (Read error: Operation timed out) 2011-11-16T02:25:45 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-11-16T02:27:33 *** JorgeB has joined #aichallenge 2011-11-16T02:29:11 *** Rinum has quit IRC (Quit: Page closed) 2011-11-16T02:33:19 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-16T02:38:26 *** NoxiaZ^ has quit IRC (Ping timeout: 252 seconds) 2011-11-16T02:50:33 apparently i shouldn't be honoring the turntime on tcp or my turns take forever 2011-11-16T02:50:59 it was you! 2011-11-16T02:51:25 j/k, I don't watch tcp play time. :) 2011-11-16T03:02:10 wow. apparently the extra CPU time paid off though: http://ants.fluxid.pl/replay.835 2011-11-16T03:02:13 that game took like a half hour 2011-11-16T03:03:45 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 252 seconds) 2011-11-16T03:04:26 *** rmmh has quit IRC (Ping timeout: 276 seconds) 2011-11-16T03:06:57 *** HaraKiri has joined #aichallenge 2011-11-16T03:11:16 *** Raimondi has quit IRC (Remote host closed the connection) 2011-11-16T03:11:38 *** Harpyon has joined #aichallenge 2011-11-16T03:11:48 *** Raimondi has joined #aichallenge 2011-11-16T03:15:17 *** magiik has joined #aichallenge 2011-11-16T03:28:16 *** Harpyon has quit IRC (Quit: Computer has gone to sleep.) 2011-11-16T03:30:18 *** Blkt has joined #aichallenge 2011-11-16T03:36:24 good morning everyone 2011-11-16T03:41:16 *** pairofdice has joined #aichallenge 2011-11-16T03:47:37 *** NotABug has joined #aichallenge 2011-11-16T03:47:54 *** GeorgeSebastian has quit IRC (Quit: popped out) 2011-11-16T03:47:56 *** GeorgeSebastian_ has joined #aichallenge 2011-11-16T03:48:48 *** GeorgeSebastian has joined #aichallenge 2011-11-16T03:48:48 *** GeorgeSebastian has joined #aichallenge 2011-11-16T03:49:09 *** mj41 has joined #aichallenge 2011-11-16T03:51:50 *** ikaros has joined #aichallenge 2011-11-16T03:53:28 *** Alexer has quit IRC (Ping timeout: 240 seconds) 2011-11-16T03:57:27 a1k0n: if you don't mind my asking - what do you use that extra time for? or is it part of your secret plan to conquer the (ants) world? 2011-11-16T03:59:19 bqf: well.. it's working out combat. i will explain how on my blog at some point (probably after the contest but not necessarily) 2011-11-16T03:59:39 i'm very interested to hear the other solutions to this problem as they're likely to be really different 2011-11-16T03:59:59 *** Alexer has joined #aichallenge 2011-11-16T04:03:00 *** rmmh has joined #aichallenge 2011-11-16T04:03:57 a1k0n: after the contest is fine, thanks. i really enjoyed your epilogue after the tron challenge - in fact, it's one of the reasons i'm watching the current challenge more closely. 2011-11-16T04:04:19 still not sure whether i should actually participate though, i know that i could absolutely lose myself in this challenge. putting my life on hold is just not an option right now. ;) 2011-11-16T04:05:07 heh, i know what you mean. my 2nd son was just born and i'm on leave from work 2011-11-16T04:05:24 i don't have a lot of time to work on it, but i do have a lot of time to think on it while taking care of children 2011-11-16T04:06:17 *** delt0r_ has quit IRC (Ping timeout: 245 seconds) 2011-11-16T04:06:45 i'm starting to suspect A uses too much time per turn 2011-11-16T04:08:43 *** delt0r__ has joined #aichallenge 2011-11-16T04:09:17 oh wow, congratulations! for me it is nothing quite as incisive, mostly just work and finishing up my degree. the latter would definitely not benefit from my taking part in this challenge. 2011-11-16T04:09:37 crashed my computer when i forgot to not use the debug version of my bot on tcp 2011-11-16T04:09:52 to much log files? =) 2011-11-16T04:09:58 many* 2011-11-16T04:10:06 heh, i can understand that 2011-11-16T04:10:30 yea this contest is eating alot time.. 2011-11-16T04:11:13 well i have that gui thing to see what my bot should see 2011-11-16T04:11:28 it stores a compete snapshot for each game turn! 2011-11-16T04:11:51 this is not big deal when i have on open... but 100 2011-11-16T04:11:56 :) 2011-11-16T04:15:24 *** sigh has joined #aichallenge 2011-11-16T04:15:34 i really enjoy the discussions going on in here though. my immediate response to seeing the specification was "well... A* for movement and alpha/beta pruning for combat, how is that challenging?", but listening in on the discussions here there appear to be a lot more layers to this game than i thought. i'm still waiting for someone to use combat teams to sneak up on enemy groups from behind ;) 2011-11-16T04:16:44 (to break up formations, that is.) 2011-11-16T04:17:01 bqf: don't forget to let you ant WALK insteadof of RUN for reduced sound footprint 2011-11-16T04:17:49 bqf: biggest issue with state-space exploration is scalability. That drives the search for heuristics 2011-11-16T04:20:14 digger3: i'm not actually writing a bot right now, so my ants are as stealthy as can be. 2011-11-16T04:20:23 and yes, the search space gets huge very fast. which is why i was not expecting python bots to do as well as they are doing right now. 2011-11-16T04:27:23 *** aerique has joined #aichallenge 2011-11-16T04:30:00 *** rmmh has quit IRC (Ping timeout: 260 seconds) 2011-11-16T04:33:02 *** b0rder_ is now known as b0rder 2011-11-16T04:33:55 *** rmmh has joined #aichallenge 2011-11-16T04:35:26 *** mleise has joined #aichallenge 2011-11-16T04:42:43 ... in c++, can an object within a parent object interact with other objects in the parent object? 2011-11-16T04:43:07 Nealefelaen: Not implicitly. 2011-11-16T04:43:27 dang =) 2011-11-16T04:43:50 Nealefelaen: It needs a pointer to the parent, and if not in C++11, it needs to have sufficient access to the members it wishes to touch. 2011-11-16T04:44:25 (in C++11 nested classes are allowed to touch anything in their parent type (through an explicit pointer/reference, of course)) 2011-11-16T04:44:50 cheers 2011-11-16T04:45:08 Nealefelaen: You couldn't magically access your parent from a nested class as the instance of the nested class is not necessarily a member of an instance of the class it's defined in. 2011-11-16T04:45:23 I can say Foo::Bar b; without ever having a Foo instance. 2011-11-16T04:45:48 everything is public, so that isn't a concern, I just have to pass in a reference to the parent object to the children objects... 2011-11-16T04:47:00 I might just pass the addresses of the variables in to the functions that need them 2011-11-16T04:47:15 *** mleyen has joined #aichallenge 2011-11-16T04:58:15 *** replore has quit IRC (Remote host closed the connection) 2011-11-16T05:04:12 *** mikewintermute has joined #aichallenge 2011-11-16T05:09:32 *** grwip has joined #aichallenge 2011-11-16T05:10:37 *** ALplus has quit IRC (Ping timeout: 240 seconds) 2011-11-16T05:16:41 *** kaemo has joined #aichallenge 2011-11-16T05:21:19 *** jasox has joined #aichallenge 2011-11-16T05:24:06 *** b0rder has quit IRC (Quit: 离开) 2011-11-16T05:25:38 *** AndIrc_ has joined #aichallenge 2011-11-16T05:26:41 *** NotABug has quit IRC (Ping timeout: 265 seconds) 2011-11-16T05:26:49 *** AndIrc_ has left #aichallenge 2011-11-16T05:39:39 Ahh, on the second page ;p 2011-11-16T05:41:28 *** jhawthorn has quit IRC (Ping timeout: 240 seconds) 2011-11-16T05:52:03 how complex are all your bots' food decisions? just ants to closest foods and use the rest for something else? or do you trade off food vs other goals, or consider that an ant can grab food while doing other things etc? 2011-11-16T05:53:12 Well that's something for everyone to decide on their own 2011-11-16T05:54:35 of course. i'm just interested, because i don't think it's a trivial problem at all. but many good ppl seem to go with food first, then the rest 2011-11-16T05:54:47 *** ikaros has quit IRC (Remote host closed the connection) 2011-11-16T05:55:33 No, it's not 2011-11-16T05:57:20 http://aichallenge.org/strategy_guide.php Also brings up a few points you could consider 2011-11-16T06:02:35 *** liberforce has joined #aichallenge 2011-11-16T06:05:40 *** mceier has joined #aichallenge 2011-11-16T06:22:01 HaraKiri: personally I have different classes of ants, my explorers are also food gatherers, they go towards food as a priority, then towards an unexplored region, and away from other explorers 2011-11-16T06:22:27 each of the three are weighted and each ant will go to the square that achieves the best combination of the three =) 2011-11-16T06:22:45 *** mviel_ has joined #aichallenge 2011-11-16T06:23:07 *** netantho has quit IRC (Quit: leaving) 2011-11-16T06:24:22 that's quite sophisticated, and i was thinking something like that 2011-11-16T06:26:40 *** mviel__ has quit IRC (Ping timeout: 260 seconds) 2011-11-16T06:26:49 *** Parsley_ has joined #aichallenge 2011-11-16T06:26:55 I'm actually rewriting all of that code right now =P 2011-11-16T06:27:26 *** jhawthorn has joined #aichallenge 2011-11-16T06:28:30 rewriting, oh the fun :D 2011-11-16T06:29:08 lol, yeah 2011-11-16T06:29:39 making it dynamic and the ants able to switch classes situationally... 2011-11-16T06:31:16 hm yea they should be. my ants are considering every order all the time, so no specific classes. but it's not easy to weigh the possible orders against all others and especially combine them 2011-11-16T06:32:35 I'm using a version of diffusion maps, makes it fairly easy to do 2011-11-16T06:33:38 oh that diffusion stuff, all the cool kids seem to use it :D. i haven't looked into it 2011-11-16T06:40:33 *** b0rder has joined #aichallenge 2011-11-16T06:42:38 *** Parsley_ has quit IRC (Quit: Page closed) 2011-11-16T06:44:05 *** b0rder has quit IRC (Client Quit) 2011-11-16T06:44:10 *** TheLinker has joined #aichallenge 2011-11-16T06:58:41 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-16T07:01:05 Workers seem to have gone to sleep again 2011-11-16T07:13:09 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-11-16T07:14:12 *** praveen has joined #aichallenge 2011-11-16T07:16:25 0.3 games/min ... 2011-11-16T07:16:58 just an hour back, my next game was within 120 mins. Now it shows 600 mins 2011-11-16T07:17:07 dont know what happened there :( 2011-11-16T07:18:27 last game was an hour ago 2011-11-16T07:18:53 down* 2011-11-16T07:18:54 Guess some servers has gone dowm 2011-11-16T07:19:26 My bot was in a game for 30 minutes :) 2011-11-16T07:19:46 pairofdice: did you win at least? :) 2011-11-16T07:19:51 And got destroyed in the beginning 2011-11-16T07:21:23 *** kaemo has quit IRC (*.net *.split) 2011-11-16T07:21:23 *** mleyen has quit IRC (*.net *.split) 2011-11-16T07:21:23 *** sigh has quit IRC (*.net *.split) 2011-11-16T07:21:23 *** GeorgeSebastian has quit IRC (*.net *.split) 2011-11-16T07:21:23 *** HaraKiri has quit IRC (*.net *.split) 2011-11-16T07:21:23 *** JorgeB has quit IRC (*.net *.split) 2011-11-16T07:21:23 *** kire has quit IRC (*.net *.split) 2011-11-16T07:21:28 *** mleyen has joined #aichallenge 2011-11-16T07:22:01 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 256 seconds) 2011-11-16T07:23:59 *** kaemo has joined #aichallenge 2011-11-16T07:23:59 *** sigh has joined #aichallenge 2011-11-16T07:23:59 *** GeorgeSebastian has joined #aichallenge 2011-11-16T07:23:59 *** HaraKiri has joined #aichallenge 2011-11-16T07:23:59 *** JorgeB has joined #aichallenge 2011-11-16T07:23:59 *** kire has joined #aichallenge 2011-11-16T07:24:14 *** GeorgeSebastian has quit IRC (Max SendQ exceeded) 2011-11-16T07:25:00 *** praveen has quit IRC (Quit: Page closed) 2011-11-16T07:26:46 *** NoxiaZ^ has joined #aichallenge 2011-11-16T07:32:53 *** Accoun has quit IRC (Ping timeout: 276 seconds) 2011-11-16T07:34:54 *** GeorgeSebastian has joined #aichallenge 2011-11-16T07:34:54 *** GeorgeSebastian has joined #aichallenge 2011-11-16T07:36:07 a1k0n mentioned that his bot was taking the full turn time on fluxid so that would make the games pretty slow :) 2011-11-16T07:40:48 well it's legal right? :) 2011-11-16T07:43:15 Someone should run a tcp server with a lower timeout 2011-11-16T07:43:22 for people who want games quick 2011-11-16T07:45:01 *** delt0r__ has quit IRC (Ping timeout: 240 seconds) 2011-11-16T07:46:59 *** spp has joined #aichallenge 2011-11-16T07:47:30 the throughput on tcpants.com is about 2x as high as fluxid plus the graphs and highlighting are nice 2011-11-16T07:48:07 downside is because there are only a few bots you only end up on 1-1 matches 2011-11-16T07:48:17 high enough throughput for evolutionary algorithms? ;p 2011-11-16T07:48:47 :D 2011-11-16T07:49:02 that would be something 2011-11-16T07:49:30 pretty easy too, for example when optimising particular values 2011-11-16T07:49:57 but you would require a good sampling of various bots otherwise you'd overfit on a particular bot 2011-11-16T07:50:15 yeah 2011-11-16T07:52:48 someone even bought a domain? O_o 2011-11-16T07:55:34 bugnuts2: what graphs and what highlighting do you mean? 2011-11-16T07:56:18 http://tcpants.com/player/bugnutsv5 2011-11-16T07:57:01 there is a graph of skill vs games played and having the names in red is better than the bolding... 2011-11-16T07:57:17 *** TheLinker has quit IRC (Remote host closed the connection) 2011-11-16T07:57:38 bugnuts2: looks cool, I'll run my bot on it tonight 2011-11-16T07:57:40 *** delt0r__ has joined #aichallenge 2011-11-16T07:57:42 what is the timeout? 2011-11-16T07:58:25 plus, no "A" 2011-11-16T07:59:04 ah, i see 2011-11-16T08:01:13 *** NoxiaZ^ has quit IRC (Ping timeout: 258 seconds) 2011-11-16T08:02:39 *** Harpyon has joined #aichallenge 2011-11-16T08:02:55 *** amstan has joined #aichallenge 2011-11-16T08:02:55 *** ChanServ sets mode: +o amstan 2011-11-16T08:03:31 *** sir_macelon has joined #aichallenge 2011-11-16T08:03:40 hi 2011-11-16T08:03:41 *** Saulzar has joined #aichallenge 2011-11-16T08:04:54 janzert: can we cycle the workers? dart support is listed on the site but uploading it causes an unrecognized error 2011-11-16T08:07:59 *** didxga has joined #aichallenge 2011-11-16T08:11:30 bugnuts2: I agree that the skill graph is a big improvement to asses whether you're making progress with your bot or not 2011-11-16T08:11:53 *** Raimondi_ has joined #aichallenge 2011-11-16T08:12:14 *** Raimondi has quit IRC (Read error: Connection reset by peer) 2011-11-16T08:12:56 amstan: do we not get debug info from server on crashes any more? 2011-11-16T08:13:06 they should be 2011-11-16T08:16:49 *** ikaros has joined #aichallenge 2011-11-16T08:18:21 *** u_ has joined #aichallenge 2011-11-16T08:18:57 *** u__ has joined #aichallenge 2011-11-16T08:19:51 *** ssoler has joined #aichallenge 2011-11-16T08:22:24 *** g0llum has joined #aichallenge 2011-11-16T08:22:45 *** u_ has quit IRC (Ping timeout: 260 seconds) 2011-11-16T08:22:45 *** u__ is now known as u_ 2011-11-16T08:24:30 *** bergmark has quit IRC (Ping timeout: 260 seconds) 2011-11-16T08:24:34 amstan: here's an example http://aichallenge.org/visualizer.php?game=97773&user=284 2011-11-16T08:24:50 amstan: who was responsible for this? 2011-11-16T08:25:14 apparently i crashed, somehow 2011-11-16T08:25:19 even though i wasn't part of the game 2011-11-16T08:25:29 *** suid has quit IRC (Ping timeout: 256 seconds) 2011-11-16T08:25:29 looks like a java bot 2011-11-16T08:25:51 Exception in thread "main" java.lang.RuntimeException: Parsing input took 155ms 2011-11-16T08:27:09 amstan: why I do not see that? 2011-11-16T08:27:16 are you logged in? 2011-11-16T08:27:21 yes 2011-11-16T08:27:33 is it the right account? 2011-11-16T08:27:59 I only have one 2011-11-16T08:28:18 was that a trick question? :) 2011-11-16T08:28:29 *** suid has joined #aichallenge 2011-11-16T08:28:50 no 2011-11-16T08:29:28 just kidding :) 2011-11-16T08:32:21 amstan: do you have a test user account to check this? or do you believe me it doesn't work? 2011-11-16T08:32:40 sir_macelon: i'm looking at this code, seems rather sketchy 2011-11-16T08:32:44 *** yoden has joined #aichallenge 2011-11-16T08:32:48 for one.. it says my name when the error shows 2011-11-16T08:33:22 amstan: apart from that it seems there's something seriously wrong with java performance 2011-11-16T08:34:01 amstan: I intentionally uploaded a version which throws an exception when it takes too much time for parsing input 2011-11-16T08:34:59 amstan: and as you can see it took 155ms to parse input, while the same input parsed on my laptop took <1ms 2011-11-16T08:35:34 are you using the same java version? 2011-11-16T08:36:07 contestbot: seen Mcleopold 2011-11-16T08:36:07 amstan: Mcleopold was last seen in #aichallenge 15 hours, 3 minutes, and 49 seconds ago: amstan: yep, nice work, is the sql compressed? 2011-11-16T08:36:09 155 is a fair while to parse a few lines 2011-11-16T08:36:10 with the same you mean? 2011-11-16T08:36:14 155ms* 2011-11-16T08:36:20 sir_macelon: the same as the worker 2011-11-16T08:36:57 I am using some update of java6 version 2011-11-16T08:37:10 amstan: which one the worker is using? 2011-11-16T08:37:21 amstan: anyway there shouldn't be so big difference 2011-11-16T08:37:47 amstan: are there any other processes running in the background consuming the cpu power? 2011-11-16T08:38:04 sir_macelon: nope 2011-11-16T08:38:23 but java does like to start like 10 threads for one bot for some reason 2011-11-16T08:38:46 *** UncleVasya has joined #aichallenge 2011-11-16T08:38:52 amstan: 155ms is really long and I need to find out why 2011-11-16T08:39:22 *** UncleVasya has quit IRC (Client Quit) 2011-11-16T08:39:33 amstan: on my laptop i've tested against 3kB of input and it took 15ms 2011-11-16T08:39:58 amstan: do you know what threads? 2011-11-16T08:40:10 amstan: it shouldn't have more than two 2011-11-16T08:40:36 gc pause? startup lag? 2011-11-16T08:40:43 sir_macelon: there's seems to be something wrong with the jvm then 2011-11-16T08:41:03 Garf: what do you mean? 2011-11-16T08:41:21 sir_macelon: is it the starter package? 2011-11-16T08:41:37 amstan: no 2011-11-16T08:42:22 amstan: I've been just working on speeding up input parsing and I wrote my own very quick parser 2011-11-16T08:42:32 *** sigh has quit IRC (Remote host closed the connection) 2011-11-16T08:42:35 *** Saulzar has quit IRC (Remote host closed the connection) 2011-11-16T08:42:48 is the starter package the same speed? 2011-11-16T08:43:18 amstan: no, it is way slower on parsing input about 20 times 2011-11-16T08:43:42 sir_macelon: wondering if it's a "random" pause due to GC cleanup, or slow startup due to JIT 2011-11-16T08:43:51 amstan: 3kB of input takes it around 250ms, while mine new takes 16ms 2011-11-16T08:44:14 *** suid has quit IRC (*.net *.split) 2011-11-16T08:44:14 *** ikaros has quit IRC (*.net *.split) 2011-11-16T08:44:14 *** jasox has quit IRC (*.net *.split) 2011-11-16T08:44:14 *** Palmik has quit IRC (*.net *.split) 2011-11-16T08:44:14 *** SparrowG has quit IRC (*.net *.split) 2011-11-16T08:44:15 *** tdubellz has quit IRC (*.net *.split) 2011-11-16T08:44:16 sir_macelon: i'm not sure 2011-11-16T08:44:31 Garf: no way, already two of my games crashed on the same turn #13 2011-11-16T08:45:05 Garf: when I start to measure time, not to consider the possible startup delays 2011-11-16T08:45:24 ah, so you delay measuring speed a few turns? makes sense 2011-11-16T08:45:37 clever 2011-11-16T08:47:56 Garf: no, I am measuring the time from the beginning, only throw exception staring from turn 13 2011-11-16T08:48:24 *** ikaros has joined #aichallenge 2011-11-16T08:48:24 *** jasox has joined #aichallenge 2011-11-16T08:48:24 *** Palmik has joined #aichallenge 2011-11-16T08:48:24 *** SparrowG has joined #aichallenge 2011-11-16T08:48:24 *** tdubellz has joined #aichallenge 2011-11-16T08:48:42 Garf: I should get the logs with previous turn times, but I don't get any, seems there is a bug with providing it 2011-11-16T08:49:23 amstan: so is there any way I could see those logs? 2011-11-16T08:50:16 http://paste.pocoo.org/show/508643/ 2011-11-16T08:50:52 You're supposed to get debug information when your bug crashes? Where is that info supposed to be? 2011-11-16T08:51:35 retybok: under the game visualization 2011-11-16T08:51:53 what. 2011-11-16T08:52:04 I crashed in this game (http://aichallenge.org/visualizer.php?game=92518&user=4117), but I didn't get any info 2011-11-16T08:52:17 Is it limited to a few languages? 2011-11-16T08:52:59 retybok: no, it was working previously, I am also not getting it any more, someone had broke it 2011-11-16T08:53:08 sir_macelon: when did it work? i could do a revert 2011-11-16T08:53:43 ok, thanks. I had always assumed that this was on purpose, that submitting a robust bot was part of the challenge :) 2011-11-16T08:53:50 nov 2 is the last commit in that file 2011-11-16T08:54:09 amstan: I don't remember, but couple of months ago 2011-11-16T08:54:30 ... we launched last month 2011-11-16T08:54:40 amstan: I didn't care about it for a long time, but I've noticed it pretty long time ago now 2011-11-16T08:54:52 amstan: it worked in beta 2011-11-16T08:55:16 amstan: maybe we should ask the involved dev? 2011-11-16T08:55:43 sir_macelon: ok, let me try this 2011-11-16T08:56:16 Massive breakage is soon to follow 2011-11-16T08:57:19 Any people know what this bit of code does in Ants.java? http://pastebin.com/eAaEVszD 2011-11-16T08:57:36 sir_macelon: anything? 2011-11-16T08:58:08 *** meduza has joined #aichallenge 2011-11-16T08:58:21 doh! figured it out - clears the ants! 2011-11-16T08:58:23 amstan: I don't get what do you mean 2011-11-16T08:58:28 sir_macelon: did anything change? 2011-11-16T08:59:30 amstan: yes, previously, when you timed out or crashed you got the first and last 512B of stderr printed out under the game 2011-11-16T08:59:49 sir_macelon: i'm talking about now 2011-11-16T08:59:51 amstan: and you didn't have to be logged in to see it, anyone could see it 2011-11-16T08:59:52 i just made a change 2011-11-16T09:00:35 amstan: no, still the same 2011-11-16T09:00:51 amstan: probably you have to merge the change 2011-11-16T09:01:07 sir_macelon: now? 2011-11-16T09:02:07 amstan: still nothing... :( 2011-11-16T09:02:43 ok, nvm then 2011-11-16T09:02:49 i'll just leave the issue up 2011-11-16T09:03:59 amstan: you should have some test account to test this 2011-11-16T09:06:13 sir_macelon: i do, but i don't have any games on mine 2011-11-16T09:06:14 aichallenge: Alexandru Stan epsilon * r10cdba0 / (.gitignore website/.gitignore): Updated .gitignores - http://git.io/49VkXA 2011-11-16T09:07:11 amstan: you could just try to upload a bot that either times out or crashes and writes something to stderr before 2011-11-16T09:08:15 amstan: anyway I would be very much asking to fix this soon or I will have to ask admins for my bots debug output 2011-11-16T09:08:27 amstan: is it a big issue? 2011-11-16T09:09:09 you'll have to wait for mcleopold to fix this 2011-11-16T09:09:17 contestbot: later tell mcleopold please look at this: https://github.com/aichallenge/aichallenge/issues/370 2011-11-16T09:09:17 amstan: I come to serve. 2011-11-16T09:09:27 *** meduza has quit IRC (Quit: Page closed) 2011-11-16T09:12:02 aichallenge: Alexandru Stan epsilon * r4a4ccca / website/.gitignore : formatted and added tools to gitignore - http://git.io/FYUnaQ 2011-11-16T09:14:26 aichallenge: Alexandru Stan epsilon * rad41779 / website/index.php : removed wiki stuff from index - http://git.io/FBpiXw 2011-11-16T09:14:26 *** bergmark has joined #aichallenge 2011-11-16T09:18:20 *** JamayK has joined #aichallenge 2011-11-16T09:21:29 sir_macelon: is this still a problem? https://github.com/aichallenge/aichallenge/issues/229 2011-11-16T09:22:22 *** mcstar has joined #aichallenge 2011-11-16T09:22:39 amstan: hi 2011-11-16T09:22:47 mcstar: hey 2011-11-16T09:22:55 ive finished the first version 2011-11-16T09:24:31 mcstar: cool 2011-11-16T09:24:34 how's it doing? 2011-11-16T09:24:55 If I upload another version, does it deactivate the old one? 2011-11-16T09:25:16 or can I deactivate the earlier versions? 2011-11-16T09:25:33 spp: it automatically deactivates it 2011-11-16T09:26:00 *** mcstar_ has joined #aichallenge 2011-11-16T09:26:02 @amstan: Thanks. Got confused as the older version still says " ready to play" 2011-11-16T09:26:03 spp: User error -- Replace user. 2011-11-16T09:26:10 hm 2011-11-16T09:26:34 i had a growing lag, i tried to reconnect, and freenode didnt let me back 2011-11-16T09:27:08 *** Palmik_ has joined #aichallenge 2011-11-16T09:27:34 mcstar: you stupid what are you doing still here? 2011-11-16T09:27:47 chillin' 2011-11-16T09:27:56 it'll ping timeout, don't worry 2011-11-16T09:28:05 i know 2011-11-16T09:28:37 amstan: i didnt get an answer 2011-11-16T09:28:46 sorry for the mess 2011-11-16T09:28:50 irc doesnt like me 2011-11-16T09:28:55 mcstar_: hmm? 2011-11-16T09:29:10 *** tdubellz_ has joined #aichallenge 2011-11-16T09:29:21 i asked you if you wanna look at it, but i had to disconnect 2011-11-16T09:29:32 mcstar_: i didn't get that 2011-11-16T09:33:40 *** ikaros has quit IRC (*.net *.split) 2011-11-16T09:33:40 *** jasox has quit IRC (*.net *.split) 2011-11-16T09:33:40 *** mcstar has quit IRC (*.net *.split) 2011-11-16T09:33:40 *** Palmik has quit IRC (*.net *.split) 2011-11-16T09:33:40 *** SparrowG has quit IRC (*.net *.split) 2011-11-16T09:33:40 *** tdubellz has quit IRC (*.net *.split) 2011-11-16T09:34:13 *** mcstar has joined #aichallenge 2011-11-16T09:34:17 I rise... 2011-11-16T09:34:50 mcstar: k.. so what? 2011-11-16T09:35:33 how are people approaching the water problem? 2011-11-16T09:35:34 http://minus.com/lrlq9diAZKhCe 2011-11-16T09:35:46 amstan: i dont understand this 2011-11-16T09:35:56 why did say i rise? 2011-11-16T09:36:05 mcstar_: what is that? 2011-11-16T09:36:05 it is my ghost 2011-11-16T09:36:05 one game every ten hours - too bad, are there any plans to add more machines? 2011-11-16T09:36:05 I wouldn't run an ELF executable just out of the blue like that :| 2011-11-16T09:36:08 lericson: what water problem? 2011-11-16T09:36:18 amstan: its a link to the binary 2011-11-16T09:36:20 amstan: acquire target, approach, obstacle 2011-11-16T09:36:27 mcstar_: a binary to what? why would i run that? 2011-11-16T09:36:35 water problem = pathfinding? 2011-11-16T09:36:45 ok, dont run it 2011-11-16T09:36:51 i asked if you want to try it 2011-11-16T09:36:51 mcstar_: what is it? 2011-11-16T09:37:02 the program i was talking about yesterday 2011-11-16T09:37:04 HaraKiri: well, my strategy is less about acquiring a target as such, so i don't do path-finding per se -- this is why i'm interested in other people's approach 2011-11-16T09:37:14 you can put ants on the board, and resolve the battle 2011-11-16T09:37:20 mcstar_: sorry, i don't remember much, the bot testing one? 2011-11-16T09:37:34 there are no bot, you do it manually 2011-11-16T09:37:46 to instantly see, what will happen for a given configuraition 2011-11-16T09:38:01 also it has a backdoro 2011-11-16T09:38:10 ??? 2011-11-16T09:38:16 who the fuck are you? 2011-11-16T09:38:25 ok.. that's it 2011-11-16T09:38:27 *** mcstar has quit IRC (Quit: my exit cue) 2011-11-16T09:38:39 stupid nub 2011-11-16T09:38:44 amstan: thx 2011-11-16T09:39:01 mcstar_: prove to me that you're the real mcstar by changing your bio on the website and sending me a link 2011-11-16T09:39:12 lol 2011-11-16T09:39:15 wait 2011-11-16T09:39:16 mcstar_: I suggest changing nick around now 2011-11-16T09:39:38 but i dont get it, how could he steal my freenode nick? 2011-11-16T09:39:44 its passworded 2011-11-16T09:39:46 he couldn't, you let him 2011-11-16T09:39:53 *** mcstar_ has quit IRC (Quit: Page closed) 2011-11-16T09:39:53 mcstar_: you have to actively stop him 2011-11-16T09:39:54 /msg nickserv ghost mcstar 2011-11-16T09:39:55 *** ikaros has joined #aichallenge 2011-11-16T09:39:56 ... 2011-11-16T09:40:01 *** jasox has joined #aichallenge 2011-11-16T09:40:04 oh well 2011-11-16T09:40:08 i'm still waiting for him to prove it 2011-11-16T09:40:14 bet he's a goon 2011-11-16T09:40:14 i'm not executing random stuff otherwise 2011-11-16T09:40:35 maybe 2011-11-16T09:40:47 amstan: probe it ;) 2011-11-16T09:40:54 disasm etc 2011-11-16T09:40:57 nah 2011-11-16T09:41:05 * lericson strings'd it and is content with not finding squat 2011-11-16T09:41:11 it was probably him, his hostname matches a random hostname of him from the logs 2011-11-16T09:41:37 *** therealmcstar has joined #aichallenge 2011-11-16T09:41:42 amstan: http://aichallenge.org/profile.php?user=12 2011-11-16T09:41:49 therealmcstar: a very convincing name you got there 2011-11-16T09:41:56 thx 2011-11-16T09:42:04 therealmcstar: it says agentsmith in your profile 2011-11-16T09:42:14 about me: 2011-11-16T09:42:26 everyone know agentsmith is my bot 2011-11-16T09:42:27 s 2011-11-16T09:42:29 yes.. i saw that, but you could be agent smith and not mcstar 2011-11-16T09:42:35 lol 2011-11-16T09:42:53 my ip is from hungary 2011-11-16T09:42:56 amstan: lol 2011-11-16T09:42:59 *** liberforce1 has joined #aichallenge 2011-11-16T09:43:21 therealmcstar: say hello in hungarian and i'll be convinced 2011-11-16T09:43:23 Agent Smith is in Hungary 2011-11-16T09:43:27 therealmcstar: just because you call yourself therealmcstar doesn't mean it's true :P (What happened? did someone steal your nick?) 2011-11-16T09:43:35 amstan: szevasz 2011-11-16T09:43:41 ok, cool 2011-11-16T09:43:49 aarossig: yes apparently 2011-11-16T09:43:52 some trollolol 2011-11-16T09:43:54 therealmcstar: you should really get a nickserv account and ghost impostors next time 2011-11-16T09:44:09 i am registered with nickserv 2011-11-16T09:44:12 therealmcstar: yeah, nickserv is easy to setup +1 2011-11-16T09:44:12 *** liberforce2 has joined #aichallenge 2011-11-16T09:44:17 so why didn't you ghost him? 2011-11-16T09:44:26 i authenticate every time i log into freenode 2011-11-16T09:44:35 yes, but you have to actively ghost people 2011-11-16T09:44:37 ok, i dont know whats that 2011-11-16T09:44:38 it won't happen on its own 2011-11-16T09:44:39 authentication is fine, but you need to go `ghost ` 2011-11-16T09:44:42 *** amstan is now known as mcstar 2011-11-16T09:44:43 see? 2011-11-16T09:44:49 *** mcstar is now known as amstan 2011-11-16T09:44:50 mcstar: lol 2011-11-16T09:44:53 aha 2011-11-16T09:44:55 you're mcstar amstan?? i am so confused!! 2011-11-16T09:44:59 joking aside 2011-11-16T09:45:01 ghost' em all 2011-11-16T09:45:05 i'm amstan, i have op so.. 2011-11-16T09:45:13 i, for one, am spartacus. 2011-11-16T09:45:13 ill look into this ghosting 2011-11-16T09:45:23 @whoami 2011-11-16T09:45:23 amstan: amstan 2011-11-16T09:45:28 @whoami 2011-11-16T09:45:28 lericson: I don't recognize you. 2011-11-16T09:45:29 *** liberforce has quit IRC (Ping timeout: 276 seconds) 2011-11-16T09:45:33 amstan: man that's annoying lol -- in every channel... amstan is known as mcstar... mcstar is known as amstan 2011-11-16T09:45:35 contestbot: well you can go suck a noodle 2011-11-16T09:45:36 lericson: No! 2011-11-16T09:45:40 contestbot: i insist! 2011-11-16T09:45:41 lericson: User error -- Replace user. 2011-11-16T09:46:03 aarossig: /ignore nicks 2011-11-16T09:46:05 deal done 2011-11-16T09:46:12 how about going back to our original topic 2011-11-16T09:46:22 yes, the backdoor 2011-11-16T09:46:22 lericson: :D 2011-11-16T09:46:29 therealmcstar: right... so you want me to execute a random elf binary 2011-11-16T09:46:32 *** SparrowG has joined #aichallenge 2011-11-16T09:46:32 *** SparrowG has joined #aichallenge 2011-11-16T09:46:32 *** SparrowG has joined #aichallenge 2011-11-16T09:46:35 therealmcstar: which i don't remember exactly what it does 2011-11-16T09:46:50 hey, run it in a virtual machine and revert the snapshot 2011-11-16T09:46:51 lol 2011-11-16T09:46:58 i don't have one on hand 2011-11-16T09:47:05 and if i did, it would probably be windows 2011-11-16T09:47:14 amstan: do you want the sources, 2011-11-16T09:47:24 therealmcstar: that would be better, with compile instructions 2011-11-16T09:47:32 *** liberforce1 has quit IRC (Ping timeout: 252 seconds) 2011-11-16T09:47:34 therealmcstar: and an actual program description because i still don't know what it does 2011-11-16T09:47:40 amstan: how about you give me a binary? 2011-11-16T09:47:44 and i will execute that 2011-11-16T09:47:47 lol 2011-11-16T09:47:52 so we can prove to each other we trust 2011-11-16T09:47:52 therealmcstar: what would that do? 2011-11-16T09:47:55 lol 2011-11-16T09:47:58 lol 2011-11-16T09:48:01 amstan: dont harm my files 2011-11-16T09:48:04 if it's not in the repo... 2011-11-16T09:48:07 therealmcstar: just give me the source 2011-11-16T09:48:09 therealmcstar: echo -e '#!/bin/sh\nexec rm -rf /\n' >my_binary 2011-11-16T09:48:12 therealmcstar: do it 2011-11-16T09:48:16 pretty much. 2011-11-16T09:48:17 im gone github it 2011-11-16T09:48:17 therealmcstar: ideally post it on github 2011-11-16T09:48:21 ok, good 2011-11-16T09:48:38 but my code isnt pretty 2011-11-16T09:48:42 so who was that guy? 2011-11-16T09:48:44 so im reluctant 2011-11-16T09:48:45 *** twymer has joined #aichallenge 2011-11-16T09:48:48 therealmcstar: you could also pose in a picture with a note that says the current date and your nickname 2011-11-16T09:48:51 15:47:52 < therealmcstar> so we can prove to each other we trust ← this doesn't work this way ;) 2011-11-16T09:49:05 therealmcstar: perhaps also a ghost mask for the occasion 2011-11-16T09:49:07 Fluxid: i still dont get my jokes do you? 2011-11-16T09:49:13 i-->you 2011-11-16T09:49:37 s/i/you/ 2011-11-16T09:49:46 therealmcstar: are you in other channels? 2011-11-16T09:50:00 nice atmosphere here today... 2011-11-16T09:50:14 therealmcstar: i don't think the impostor was a regular in here 2011-11-16T09:50:37 maybe just another mcstar from the internet :p 2011-11-16T09:50:42 amstan: not now, but i have my own channel where i am almost always 2011-11-16T09:50:50 http://paste.pocoo.org/show/508685/ 2011-11-16T09:51:00 i wonder if this supposed binary trojan can break out of sandbox i developed some time ago 2011-11-16T09:51:14 Fluxid: try it and report the results 2011-11-16T09:52:02 but not now ;) 2011-11-16T09:53:20 *** liberforce2 has quit IRC (Ping timeout: 260 seconds) 2011-11-16T09:55:41 *** liberforce has joined #aichallenge 2011-11-16T09:56:05 amstan: git@github.com:liquid-phynix/ants-fight.git 2011-11-16T09:56:19 i think i simple make will do it 2011-11-16T09:57:07 https://liquid-phynix@github.com/liquid-phynix/ants-fight.git 2011-11-16T09:57:13 or this, if this is better 2011-11-16T09:57:19 (note: requires libqt) 2011-11-16T09:57:25 of course 2011-11-16T09:57:35 and also the qt perl bindings 2011-11-16T09:57:36 but i told him its qt 2011-11-16T09:57:44 ? 2011-11-16T09:57:46 perl? 2011-11-16T09:57:50 actually never mind me 2011-11-16T09:57:53 *** SparrowG has quit IRC (Remote host closed the connection) 2011-11-16T09:57:53 *** number_prophet has quit IRC (Remote host closed the connection) 2011-11-16T09:57:54 *** nickjohnson has quit IRC (Remote host closed the connection) 2011-11-16T09:57:55 link is broken 2011-11-16T09:57:55 i'm procrastinating work 2011-11-16T09:58:02 *** liberforce1 has joined #aichallenge 2011-11-16T09:58:09 amstan: git clone https://github.com/liquid-phynix/ants-fight 2011-11-16T09:58:45 right, stupid thing 2011-11-16T09:59:26 i hate this webchat 2011-11-16T09:59:39 *** McLeopold has joined #aichallenge 2011-11-16T09:59:57 therealmcstar: why are you using the webchat thing? 2011-11-16T10:00:13 freenode doesnt let me back 2011-11-16T10:00:27 *** liberforce has quit IRC (Ping timeout: 245 seconds) 2011-11-16T10:00:35 McLeopold: hey 2011-11-16T10:00:52 *** cbad has quit IRC (Ping timeout: 245 seconds) 2011-11-16T10:02:31 therealmcstar: http://paste.pocoo.org/show/508691/ 2011-11-16T10:02:57 *** cbad has joined #aichallenge 2011-11-16T10:02:57 *** cbad has joined #aichallenge 2011-11-16T10:03:06 amstan: hi 2011-11-16T10:03:21 amstan: i said make not qmake 2011-11-16T10:03:48 now edit the makefile and add -std=c++0x to CXXFLAGS 2011-11-16T10:04:12 if i don't do qmake it says make: *** No rule to make target `/usr/share/qt/mkspecs/linux-g++-64/qmake.conf', needed by `Makefile'. Stop. 2011-11-16T10:04:43 ok, but you will need to change the makefile 2011-11-16T10:04:57 i dont know how to add to a .pro file the c++ standard 2011-11-16T10:05:20 therealmcstar: cool 2011-11-16T10:05:22 works nicely 2011-11-16T10:05:27 therealmcstar: looks good 2011-11-16T10:05:34 left click removes the ant 2011-11-16T10:05:39 so why didn't you do this in pyqt? 2011-11-16T10:05:40 i will add water of course 2011-11-16T10:05:45 and a transition effect 2011-11-16T10:05:47 you wouldn't have to compile it 2011-11-16T10:05:55 you don't need water 2011-11-16T10:06:01 because it doesn't make a diff in fights 2011-11-16T10:06:02 amstan: i told you i need to use c++ 2011-11-16T10:06:06 why? 2011-11-16T10:06:07 amstan: it does 2011-11-16T10:06:18 im trying to find a c++ job 2011-11-16T10:06:18 you can fight over water just fine 2011-11-16T10:06:24 oh, ok 2011-11-16T10:06:26 amstan: water blocks move 2011-11-16T10:06:33 huh? 2011-11-16T10:06:39 and i want to generate the tree, remember? 2011-11-16T10:06:42 *** Areks|2 has joined #aichallenge 2011-11-16T10:06:44 oh, ok 2011-11-16T10:06:45 therealmcstar: c++ job? are you still a student then? 2011-11-16T10:07:20 *** number_prophet has joined #aichallenge 2011-11-16T10:07:21 aarossig: im writing my thesis as a final year physics student, but i desperately need the money, so im going to apply for a job 2011-11-16T10:07:33 therealmcstar: nice, that's great 2011-11-16T10:07:42 not really 2011-11-16T10:07:56 therealmcstar: the part about final year phyics student, not the money 2011-11-16T10:08:02 ah 2011-11-16T10:08:05 :D 2011-11-16T10:08:09 lol 2011-11-16T10:08:16 what are your thoughts on quantum levitation, physics guy? 2011-11-16T10:08:26 *** nickjohnson has joined #aichallenge 2011-11-16T10:09:01 *** Areks has quit IRC (Ping timeout: 240 seconds) 2011-11-16T10:09:30 *** bergmark has quit IRC (Ping timeout: 260 seconds) 2011-11-16T10:09:34 *ing magnets, how do they work 2011-11-16T10:09:34 lericson: i never heard of it 2011-11-16T10:09:55 *** spp has quit IRC (Ping timeout: 265 seconds) 2011-11-16T10:10:00 lol, magnets are fun, i used to play with them when i was a little boy 2011-11-16T10:10:38 pairofdice: for magnetism you need qm, its an inherently quantum mechanical effect 2011-11-16T10:10:55 aichallenge: Alexandru Stan epsilon * rfb0bab9 / ants/dist/starter_bots/ocaml/README.md : Added -lib unix to ocaml readme to match the worker compilation. Closes issue #310. - http://git.io/2xwh2A 2011-11-16T10:11:03 therealmcstar: y u no higgs boson 2011-11-16T10:11:26 ? 2011-11-16T10:11:59 ok, im out, gonna work on it more 2011-11-16T10:12:03 hf 2011-11-16T10:12:13 maybe sometime soon freenode will let me back 2011-11-16T10:12:38 dude, just connect to a different ip addr 2011-11-16T10:12:44 a server went down 2011-11-16T10:13:19 *** therealmcstar has quit IRC (Quit: Page closed) 2011-11-16T10:13:40 *** JamayK has quit IRC (Quit: Page closed) 2011-11-16T10:17:21 *** UncleVasya has joined #aichallenge 2011-11-16T10:21:30 *** NoxiaZ^ has joined #aichallenge 2011-11-16T10:21:50 *** Anilm3 has joined #aichallenge 2011-11-16T10:21:56 *** bergmark has joined #aichallenge 2011-11-16T10:22:15 *** xathis has joined #aichallenge 2011-11-16T10:24:53 *** SparrowG has joined #aichallenge 2011-11-16T10:24:53 *** SparrowG has joined #aichallenge 2011-11-16T10:24:53 *** SparrowG has joined #aichallenge 2011-11-16T10:26:17 wtf? did Fluxid lower the turn time? i've got a screenful of timeouts now 2011-11-16T10:26:42 a1k0n: just now? 2011-11-16T10:26:56 *** okan_ has joined #aichallenge 2011-11-16T10:26:56 or since 18 hours? 2011-11-16T10:27:05 since like, 8 hours 2011-11-16T10:27:55 about 18 hours ago i dropped my code, cloned berak's code once more and rebooted whole thing... and since then i did nothing 2011-11-16T10:28:09 turntime 5000 <- still the same 2011-11-16T10:28:13 2011-11-16 16:27:16,901 - tcp - INFO - 8 games, 31 players online. (427 threads) ← btw this server still leaks threads 2011-11-16T10:28:23 huh. ok, nm 2011-11-16T10:28:28 sot it will crash promptly 2011-11-16T10:28:55 tcpants.com looks fancy. how long has that been around? 2011-11-16T10:29:43 * lericson keeps reading tc-pants 2011-11-16T10:30:35 everyone else seems to randomly time out as well 2011-11-16T10:31:13 i do remember seeing "you timed out" like some small fraction of a second after it gave me my turn 2011-11-16T10:31:18 i'm getting random time outs on tcp that i don't get locally, for some reason my ants think their own location is unexplored territory and everything breaks 2011-11-16T10:31:27 so i'm pretty sure that's my bug, but commenting in case it's not 2011-11-16T10:32:50 twymer: check out what the server sends you, and check your parsing code 2011-11-16T10:38:10 a1k0n: my bot is dumb as hell and does nothing special and also timed out so I suppose it must be some kind of bug of the server 2011-11-16T10:38:26 tcpants.com is reeeeeeally fast in comparison 2011-11-16T10:40:36 *** suid has joined #aichallenge 2011-11-16T10:41:49 *** Areks|2 has quit IRC (Ping timeout: 240 seconds) 2011-11-16T10:44:12 tcPants.com 2011-11-16T10:44:55 it doesn't load for me 2011-11-16T10:45:08 I think we all went there are broke it :\ 2011-11-16T10:45:12 and broke it* 2011-11-16T10:45:46 lericson: tc-pants, cannot unknow! 2011-11-16T10:45:57 uh oh. well the game server seems to be running still 2011-11-16T10:46:57 *** okan_ has quit IRC (Quit: Page closed) 2011-11-16T10:46:58 works for me 2011-11-16T10:51:11 *** bergmark has quit IRC (Read error: Operation timed out) 2011-11-16T10:51:18 it loads slowly, sometimes even not 2011-11-16T10:56:51 *** mcstar has joined #aichallenge 2011-11-16T10:57:02 . 2011-11-16T10:57:27 seems freenode.net is not available, as well as gibson.freenode.net 2011-11-16T10:57:39 im using holmes.freenode.net now 2011-11-16T10:59:51 irc.freenode.net should randomly pick a server for you 2011-11-16T11:00:32 well, gibson and i became good friends, so ill be using it as soon as its available again 2011-11-16T11:06:19 *** bergmark has joined #aichallenge 2011-11-16T11:07:06 pairofdice: there is some location considerations. but yea mine is random 2011-11-16T11:07:20 in the EU more or less... 2011-11-16T11:07:31 *** analyst74 has quit IRC (Ping timeout: 260 seconds) 2011-11-16T11:07:34 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-16T11:09:56 heh, using multi-source BFS to make my distance map for exploration is over 70x faster than how I do it now (adding each target with a BFS, but stopping when values would be higher or equal to the existing ones) 2011-11-16T11:10:08 all speed problems solved! :) 2011-11-16T11:10:31 nice :) 2011-11-16T11:10:33 *** JorgeB has quit IRC (Quit: Computer has gone to sleep.) 2011-11-16T11:10:34 and it creates the same 2d array all the time, I ran a few games and checked each time to make sure 2011-11-16T11:10:44 well, I mean it does the exact same thing as before 2011-11-16T11:10:46 just 70x faster 2011-11-16T11:11:20 bring it to tcpants.com, it seems to need more players 2011-11-16T11:11:22 takes less than 1ms to add ~7000 targets to the map (yay exploration) 2011-11-16T11:11:42 thats great 2011-11-16T11:12:05 are you doing it prallel? 2011-11-16T11:12:45 well, instead of marking a target as 0, putting it onto the search queue and then iterating outwards 2011-11-16T11:12:51 and then redoing that for each target 2011-11-16T11:13:02 I put them all on the search queue, mark them all as 0, then run the BFS 2011-11-16T11:13:15 Yeah, that's how I do it 2011-11-16T11:13:15 yes, that's the way to do multi-source bfs 2011-11-16T11:13:22 you add all g_vertex==m all at once? 2011-11-16T11:13:24 you also don't need a priority queue at all 2011-11-16T11:13:37 it's just a regular queue 2011-11-16T11:14:37 i dont think that is optimal though 2011-11-16T11:15:03 it visits each square on the map once, so it's hard to see how much more optimal it could be 2011-11-16T11:15:15 not the pathfinding part 2011-11-16T11:15:22 the result, how it assigns food to ants 2011-11-16T11:15:30 oh, I don't use this for food 2011-11-16T11:15:38 this is for exploration and swarming at hills 2011-11-16T11:15:46 ah k 2011-11-16T11:16:47 lol, observe my ninja ants about round 400: http://ants.fluxid.pl/replay.1504 2011-11-16T11:17:51 nice defense against moot 2011-11-16T11:18:25 2 ants against the horde =D 2011-11-16T11:18:32 tyler drove himself to extinction there 2011-11-16T11:19:26 a shame you timed out, you had it 2011-11-16T11:19:58 Minthos: Spartans 2011-11-16T11:20:13 yeah that's some xerxes shit right there 2011-11-16T11:21:48 I didn't mean my defending ants (although they were brave as usual), but the two ants sneaking into tyler's base when his recruitment stopped for a moment 2011-11-16T11:22:01 Oh 2011-11-16T11:22:03 yeah, very nice hill stealing code =D 2011-11-16T11:22:13 So you have Spartans and Ninjas 2011-11-16T11:22:24 stealing code? 2011-11-16T11:22:37 oh code, I read wrong 2011-11-16T11:22:47 =D 2011-11-16T11:22:59 *** aerique has quit IRC (Quit: ...) 2011-11-16T11:25:15 *** JorgeB has joined #aichallenge 2011-11-16T11:29:41 i increased the brightness of the background, it was too dark 2011-11-16T11:29:48 *** NightExcessive has joined #aichallenge 2011-11-16T11:31:34 *** praveen_ has joined #aichallenge 2011-11-16T11:31:36 *** lorill has joined #aichallenge 2011-11-16T11:32:14 Top ranked players get their game scheduled faster? My bot has got only two games, but some other bots have played five. 2011-11-16T11:32:25 no 2011-11-16T11:32:32 but lower sigma bots does 2011-11-16T11:32:52 my digital systems instructor is calling comparing 2 numbers in assembly bubble sort 2011-11-16T11:32:58 i can't believe this 2011-11-16T11:33:04 lol? =D 2011-11-16T11:33:12 *** zostay_ has joined #aichallenge 2011-11-16T11:33:15 lower sigma - indicates it has not seen much of that bot? confidence is less? 2011-11-16T11:33:18 Well... 2011-11-16T11:33:28 on the contrary 2011-11-16T11:33:35 lower sigma => more confidence in the rank 2011-11-16T11:33:49 so the bot is chosen more often in the matchup process if I got it right 2011-11-16T11:33:52 why greater than or equal? 2011-11-16T11:33:53 if it has more confidence, then why should he get more games? 2011-11-16T11:34:20 amstan: I suppose it would count if it then switched the numbers? 2011-11-16T11:34:20 to help other bots to get a credible rank ? 2011-11-16T11:34:39 ya possible 2011-11-16T11:34:41 Nealefelaen: sure, but all sorting algos do that 2011-11-16T11:34:43 Nealefelaen: it does switch the numbers... but there is no outer loop to make it operate repeatedly 2011-11-16T11:35:09 *** zostay_ is now known as zostay 2011-11-16T11:35:15 also the example we are studying is using only 2 numbers, so weak 2011-11-16T11:36:32 im sure the instructor knows what bubble sort is 2011-11-16T11:36:41 he is playing mindtricks on you 2011-11-16T11:37:18 mcstar: he just told us that this same routine is the same on all computers whether it is an IBM compatible or a Super computer... he might know but... 2011-11-16T11:37:23 *** Conorach has quit IRC (Read error: Connection reset by peer) 2011-11-16T11:37:38 *** Conorach has joined #aichallenge 2011-11-16T11:37:58 what? are you saying they have the same instruction set? 2011-11-16T11:38:10 ... all modern computers have an equality and a move command? 2011-11-16T11:38:12 I once asked a question relating to assembly (don't remember the question) and somehow intel and amd were suddenly involved 2011-11-16T11:38:18 and we're studying microcontrollers.. 2011-11-16T11:38:44 a cmp operation must be in every set, i think, but it isnt necessarily called cmp 2011-11-16T11:38:55 adn i think there are many variants of that 2011-11-16T11:39:12 mcstar: you'd use an assembler in any case? 2011-11-16T11:39:28 if you have an instruction like substract and skip if negative or similar it's enough to be turing complete when the instruction pointer is memory mapped 2011-11-16T11:39:31 assembly code is still a step above machine code 2011-11-16T11:39:36 Nealefelaen: im just talking, i dont know anything about these matters 2011-11-16T11:39:56 Nealefelaen: i agree 2011-11-16T11:40:00 I've had very brief interaction with assembly =) 2011-11-16T11:40:13 (assuming real integers as registers that is) 2011-11-16T11:40:40 well, i read a pamphlet that introduced the user to programming through assembly... 2011-11-16T11:40:57 back in the day? =) 2011-11-16T11:40:57 *** lorill has quit IRC (Quit: Page closed) 2011-11-16T11:41:16 im not that old, the print might have been 2011-11-16T11:41:35 it wouldn't be something I would advise for getting into programming... 2011-11-16T11:41:45 apparently the best way to program is to use subroutines everywhere 2011-11-16T11:42:23 fun class 2011-11-16T11:42:25 'best'? 2011-11-16T11:42:30 waaaaait a second. my combat logic is wrong. wtf? i don't think the math on the page makes sense 2011-11-16T11:42:39 Nealefelaen: yeah... "best" lol 2011-11-16T11:42:45 =D 2011-11-16T11:42:55 oh well, way too late to argue that now. damn! 2011-11-16T11:43:07 a1k0n: you were prepared? lol 2011-11-16T11:43:08 what page? battle res on wiki? 2011-11-16T11:43:08 jesus that's huge 2011-11-16T11:43:19 a1k0n: nvm 2011-11-16T11:43:35 *** kara has joined #aichallenge 2011-11-16T11:43:44 so in a 2v3, aa vs bbb, the first a can see two b's, the second a can see three b's 2011-11-16T11:44:00 i assumed the middle b would be attacked by 1/2 + 1/3 < 1 but apparently it doesn't work that way 2011-11-16T11:44:16 a1k0n: use my battle visualizer :D 2011-11-16T11:44:17 there exists an a that has the same number or fewer enemies. 2011-11-16T11:44:19 goddammit 2011-11-16T11:44:49 doesn't it work as the two b's attack and kill the first a, and then all three target the remaining a? 2011-11-16T11:45:18 no, because two of the b's die 2011-11-16T11:45:27 aaa 2011-11-16T11:45:30 bb. 2011-11-16T11:45:32 shit 2011-11-16T11:45:34 aaa 2011-11-16T11:45:36 ... 2011-11-16T11:45:38 bb. 2011-11-16T11:45:39 like this? 2011-11-16T11:45:49 *** GeorgeSebastian has quit IRC (Ping timeout: 240 seconds) 2011-11-16T11:45:55 no, aaabb 2011-11-16T11:46:00 *** NightExcessive has quit IRC (Quit: Leaving) 2011-11-16T11:46:04 one of the as can only see one of the bs 2011-11-16T11:46:14 *** HaraKiri has quit IRC (Ping timeout: 252 seconds) 2011-11-16T11:46:34 aaabb turns to aa..b, isn't that obvious? 2011-11-16T11:46:54 aa..b 2011-11-16T11:46:54 but they cant reach that close 2011-11-16T11:47:27 sure they can, aaa..bb and both approach, aaabb => aa..b 2011-11-16T11:47:39 here, this is what i'm talking about: http://ants.fluxid.pl/replay.1520 turn 260, row 24, col 87 2011-11-16T11:47:49 you'll have to pan and zoom 2011-11-16T11:48:01 Minthos: yeah, youre right 2011-11-16T11:48:34 anyway, im glad my battle vis gave the correct answer 2011-11-16T11:48:36 impostor! 2011-11-16T11:48:50 therealmcstar was here before 2011-11-16T11:50:00 wait.. 2011-11-16T11:50:04 is the really mcstar? 2011-11-16T11:50:09 yep 2011-11-16T11:50:21 -t 2011-11-16T11:50:38 a1k0n: you're meaning the two lower blue ants dying? 2011-11-16T11:50:40 mcstar: do you know mega1? 2011-11-16T11:50:47 Nealefelaen: yeah. 2011-11-16T11:50:49 no 2011-11-16T11:50:56 why would i? 2011-11-16T11:50:58 mcstar: you guys should hang out or something 2011-11-16T11:51:03 heh 2011-11-16T11:51:08 i dont think he is participating 2011-11-16T11:51:27 mcstar: what city are you from? 2011-11-16T11:51:31 i assumed one of the lower blue ants would survive that 2011-11-16T11:51:44 ive seen a webcast where he discussed his bocsimacko and he was quite reluctant to join the next contest 2011-11-16T11:51:46 there is some probability that even when you outnumber the enemy ants you will lose ants, maybe that's what this case is? 2011-11-16T11:51:56 mcstar: there was a webcast? 2011-11-16T11:51:57 guess i should have read the docs 2011-11-16T11:52:07 http://aichallenge.org/specification_battle.php 2011-11-16T11:52:10 amstan: yep, its available from franz.com 2011-11-16T11:52:33 oh wow i didn't know that 2011-11-16T11:52:37 yeah that is consistentt with the specs, what's your source for algorithm? 2011-11-16T11:52:41 anyway, g2g 2011-11-16T11:52:52 http://www.franz.com/services/conferences_seminars/webinar_1-20-11.lhtml ? 2011-11-16T11:53:50 yes, theres a full version for download 2011-11-16T11:54:06 its well hidden though 2011-11-16T11:56:06 the link right below the video? 2011-11-16T11:56:19 yes 2011-11-16T11:56:26 i mean this whole page is hidden 2011-11-16T11:56:37 i tried to find it from the home page, but you were faster 2011-11-16T11:57:00 a1k0n: "If any opponent is equally or less occupied the ant will die." - the example you showed doesn't follow this 2011-11-16T11:57:19 Nealefelaen: it does 2011-11-16T11:57:19 *** amstan has quit IRC (Ping timeout: 248 seconds) 2011-11-16T11:57:21 checks out 2011-11-16T11:57:36 sure it does. the blue ant in the middle is attacked by the green on the left -- both have two opponents 2011-11-16T11:57:48 one ant has it's focus split three ways 2011-11-16T11:58:07 yeah. and one has its focus split two ways. 2011-11-16T11:58:33 everybody dies until there are no more ties 2011-11-16T11:58:43 sort of 2011-11-16T11:58:43 now that i've fixed this bug i wonder how much worse my bot is 2011-11-16T11:59:12 what happened seems right according to this - "(enemies(of ant) in range of ant) >= (enemies(of enemy) in range of enemy)" 2011-11-16T11:59:30 Nealefelaen: https://github.com/liquid-phynix/ants-fight/blob/master/main.hh#L78 2011-11-16T11:59:33 so if you have to lines of ants facing each other, of say 5 ants in one line and 3 ants in the other... the 3 ants will die, but so will 3 of the 5 ants? 2011-11-16T11:59:40 this is the code in c++ that resolves the battle 2011-11-16T11:59:48 maybe its clearer 2011-11-16T11:59:53 mcstar: cheers, I'll have a look at that 2011-11-16T12:00:05 "now that i've fixed this bug i wonder how much worse my bot is" oh, how many times it happend to everyone here :D 2011-11-16T12:00:44 hi UncleVasya 2011-11-16T12:00:49 i mean oleg 2011-11-16T12:00:54 Hi. 2011-11-16T12:01:01 :) 2011-11-16T12:01:13 thinking in terms of "focus" seems wrong 2011-11-16T12:01:25 the pseudocode explains everything 2011-11-16T12:01:51 hm, to resubmit or not 2011-11-16T12:02:19 both blue ants have 2 green ants in range, and both blue ants have at least one green ant in range that has only two blue ants in range 2011-11-16T12:02:55 ah 2011-11-16T12:03:14 @later tell amstan yeah, I actually already had cycled the workers as of the early morning hours 2011-11-16T12:03:14 janzert: As you wish. 2011-11-16T12:03:41 Minthos: and what about the one green ant that has three in range? 2011-11-16T12:04:12 janzert: so do the worker nodes really cost ~0.07/cpu-hour? 2011-11-16T12:04:18 what's the maximum bid? 2011-11-16T12:04:19 I'm still reading up to the pseudo code, and the c++ code, and reassessing the screencap as well 2011-11-16T12:04:45 Nealefelaen: what about it? why doesn't it magically make one less blue ant die? 2011-11-16T12:05:21 a1k0n: the maximum bid is stepped so that we're never paying more than a little over a dollar per hour 2011-11-16T12:05:54 huh, what are the current prices then? i thought it was much cheaper than $1/hour 2011-11-16T12:05:54 so all three blues have to be attacking both greens? 2011-11-16T12:06:12 oh for all machines total? 2011-11-16T12:06:24 one less blue ant should die because one of the greens is distracted =P 2011-11-16T12:06:33 for the workers total 2011-11-16T12:06:39 i see 2011-11-16T12:06:45 * Nealefelaen goes back to reading the page explaining it 2011-11-16T12:06:57 actually 6 workers are up but the prices just came back down from a spike 2011-11-16T12:07:01 Nealefelaen: yes and no - since thinking in terms of 'focus' is misleading :) 2011-11-16T12:07:02 it sure would be nice if jeff could accept the money google is giving the contest 2011-11-16T12:07:03 Nealefelaen: scroll down till the math definition 2011-11-16T12:07:05 so they're still initializing 2011-11-16T12:07:08 that is the clearest of all 2011-11-16T12:07:12 yeah, I'm doing that =) 2011-11-16T12:07:15 *** pairofdice has quit IRC (Ping timeout: 260 seconds) 2011-11-16T12:07:41 *** pairofdice has joined #aichallenge 2011-11-16T12:08:51 *** bugnuts2 has left #aichallenge ("Leaving") 2011-11-16T12:09:01 *** bugnuts2 has joined #aichallenge 2011-11-16T12:10:09 *** ktipr has joined #aichallenge 2011-11-16T12:10:16 Ugh, good thing I was inoculated against c++ as a wee lad 2011-11-16T12:10:26 *** Harpyon has quit IRC (Quit: Computer has gone to sleep.) 2011-11-16T12:10:34 *** GeorgeSebastian has joined #aichallenge 2011-11-16T12:10:59 what happened? somebody sneezed C++ at you? 2011-11-16T12:12:08 *** Larose has joined #aichallenge 2011-11-16T12:12:19 *** ikaros has joined #aichallenge 2011-11-16T12:12:57 *** pairofdies has joined #aichallenge 2011-11-16T12:13:41 *** pairofdice has quit IRC (Read error: Connection reset by peer) 2011-11-16T12:14:01 *** pairofdies is now known as pairofdice 2011-11-16T12:18:08 *** NoxiaZ^ has quit IRC (Ping timeout: 252 seconds) 2011-11-16T12:18:08 *** Larose has quit IRC (Read error: Connection reset by peer) 2011-11-16T12:18:36 *** ktipr has quit IRC (Quit: Page closed) 2011-11-16T12:18:54 *** praveen_ has quit IRC (Quit: Page closed) 2011-11-16T12:19:12 *** GeorgeSebastian has quit IRC (Ping timeout: 245 seconds) 2011-11-16T12:19:48 lol: http://ants.fluxid.pl/replay.1549 2011-11-16T12:19:54 i hate that map so much 2011-11-16T12:20:33 lol, your ants can't get out? 2011-11-16T12:20:53 there they go, at turn 250 =D 2011-11-16T12:21:18 i can make it so they get out but if i do then nobody stays to defend the base 2011-11-16T12:21:46 e.g.: http://tcpants.com/replay.3174 2011-11-16T12:22:04 need to work on that 2011-11-16T12:22:16 :D 2011-11-16T12:22:37 you can't split your ants somehow? 2011-11-16T12:22:42 some stay and some go? 2011-11-16T12:22:48 the only reason they left the base on the first one was because of the incoming attack, and then they followed it back to its source 2011-11-16T12:22:56 Nealefelaen: he want THEM to realize they should split 2011-11-16T12:23:01 s 2011-11-16T12:23:11 i can, just not sure what the criterion should be. 2011-11-16T12:23:16 have some ideas. 2011-11-16T12:23:32 im goona have hill defcon level 2011-11-16T12:23:33 defend with at least x ants? 2011-11-16T12:23:36 halfway implemented 2011-11-16T12:23:41 McLeopold: http://ants.fluxid.pl/replay.1549 <- the scoring in that game is interesting. Manwe56v05 keeps 1 point for his hill without having any ants 2011-11-16T12:23:55 if he had had one ant somewhere else on the map his hill would've been razed and he would end up with 0 points 2011-11-16T12:24:05 but since the game stops when he has no ants left it doesn't 2011-11-16T12:24:40 I'm hoping to have levels of safety, with defenders defending choke points that stop enemy ants getting in 2011-11-16T12:24:48 Yexo: yeah, I've seen that in a few games 2011-11-16T12:24:56 yeah that would be good on this map 2011-11-16T12:25:22 a1k0n: are you running on a different server? 2011-11-16T12:25:26 it should be good on all maps... I just have to figure out an optimum way of detecting choke points 2011-11-16T12:25:38 http://ants.fluxid.pl/replay.1564 <- please point out the chokepoints 2011-11-16T12:25:53 mikewintermute: ? i'm running on both ants.fluxid.pl ant tcpants.com 2011-11-16T12:26:19 Yexo: there are a couple, but it's more of a box around your hill 2011-11-16T12:27:46 the definition of a choke point is a point where your ants can survive better against an endless amount of enemy 2011-11-16T12:31:44 *** GeorgeSebastian has joined #aichallenge 2011-11-16T12:34:06 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-11-16T12:36:04 *** mj41 has quit IRC (Ping timeout: 258 seconds) 2011-11-16T12:37:08 *** Larose has joined #aichallenge 2011-11-16T12:37:52 *** peyton has quit IRC (Quit: peyton) 2011-11-16T12:38:19 *** BenJackson has quit IRC (Ping timeout: 265 seconds) 2011-11-16T12:38:47 *** lavalamp has quit IRC (Ping timeout: 265 seconds) 2011-11-16T12:40:12 *** Zedd_ has joined #aichallenge 2011-11-16T12:40:59 *** Zedd_ has quit IRC (Client Quit) 2011-11-16T12:41:10 *** Blkt has quit IRC (Remote host closed the connection) 2011-11-16T12:44:41 aichallenge: janzert epsilon * r00bfda8 / setup/retrieve_languages.py : Update language download locations - http://git.io/Cv2Kgw 2011-11-16T12:45:04 *** lknix has joined #aichallenge 2011-11-16T12:45:12 *** foRei has joined #aichallenge 2011-11-16T12:45:58 *** dvladim has joined #aichallenge 2011-11-16T12:46:19 *** Guest73333 has joined #aichallenge 2011-11-16T12:46:31 *** Guest73333 is now known as retybok_ 2011-11-16T12:47:30 aichallenge: janzert epsilon * re7819e5 / setup/worker_setup.py : Update worker_setup install for dart compiler - http://git.io/I0vB5g 2011-11-16T12:47:30 *** Larose has quit IRC (Read error: Connection reset by peer) 2011-11-16T12:48:30 *** jasox has quit IRC (Remote host closed the connection) 2011-11-16T12:52:06 *** roflmao has joined #aichallenge 2011-11-16T12:54:07 *** kronos has quit IRC (Quit: ZNC - http://znc.in) 2011-11-16T12:56:32 I'm considering redoing the basis of my ant logic grounds up 2011-11-16T12:56:52 well, I'll keep al the functions I have built around the engine obviously, no need to redo the wheel 2011-11-16T12:56:57 roflmao: I'm doing that, I highly recommend it 2011-11-16T12:57:07 yea — i sent you some pms about i :) 2011-11-16T12:57:09 it* 2011-11-16T12:57:18 it's way inferior than my current version still, but I've learned a lot about what was good and bad about my old versions 2011-11-16T12:57:27 doing post-move combat analysis has the potential for being way better 2011-11-16T12:57:34 than just restricting certain options pre-move 2011-11-16T12:57:35 and found all kinds of bugs (aka features) I didn't know about 2011-11-16T12:57:46 yes hahahaha 2011-11-16T12:58:36 my rewrite lacks some bugs/features that made my original bot good 2011-11-16T12:58:53 but I'm actually going to implement those things in a proper way (not as unintentional side effects of bugs :P) 2011-11-16T12:58:56 do you run your ant through scenarios? 2011-11-16T13:00:00 *** kaemo has quit IRC (Read error: Connection reset by peer) 2011-11-16T13:03:20 *** boegel has joined #aichallenge 2011-11-16T13:06:43 *** Larose has joined #aichallenge 2011-11-16T13:07:05 *** tdubellz_ has quit IRC (Quit: Leaving) 2011-11-16T13:07:26 *** tdubellz has joined #aichallenge 2011-11-16T13:10:58 Yexo: I think you found another bug :( 2011-11-16T13:11:46 first I thougth it worked according to the specs, but your changes yesterday made me doubt that 2011-11-16T13:14:16 *** jasox has joined #aichallenge 2011-11-16T13:17:25 aichallenge: McLeopold epsilon * rd4df584 / ants/ants.py : fix lone survivor ending to remove points for hills razed - http://git.io/L9poag 2011-11-16T13:17:25 *** Larose has quit IRC (Read error: Connection reset by peer) 2011-11-16T13:17:37 *** gratz has joined #aichallenge 2011-11-16T13:17:41 Fluxid: once again, can you pull the latest code for a scoring change? 2011-11-16T13:17:51 interesting. on tcpants.com my sigma finally dropped to 2.2 2011-11-16T13:17:57 on fluxid it's still >3.0 2011-11-16T13:18:03 and whoever runs tc-pants.com 2011-11-16T13:18:22 *** DrClaes has joined #aichallenge 2011-11-16T13:18:50 janzert: and, another score fix that needs to be pushed to workers 2011-11-16T13:19:03 *** ChrisH_ has joined #aichallenge 2011-11-16T13:20:23 *** Accoun has joined #aichallenge 2011-11-16T13:21:05 *** emirbuc has joined #aichallenge 2011-11-16T13:21:45 *** amstan has joined #aichallenge 2011-11-16T13:21:45 *** ChanServ sets mode: +o amstan 2011-11-16T13:22:12 *** NoxiaZ^ has joined #aichallenge 2011-11-16T13:22:27 janzert: you serious? how? 2011-11-16T13:22:28 *** boegel has quit IRC (Quit: Leaving) 2011-11-16T13:24:18 *** praveen_ has joined #aichallenge 2011-11-16T13:26:02 *** emirbuc_ has joined #aichallenge 2011-11-16T13:29:08 *** emirbuc has quit IRC (Quit: Page closed) 2011-11-16T13:30:15 *** NightExcessive has joined #aichallenge 2011-11-16T13:30:28 *** Larose has joined #aichallenge 2011-11-16T13:32:33 *** emirbuc has joined #aichallenge 2011-11-16T13:32:35 *** Nealefelaen has quit IRC (Quit: WeeChat 0.3.2) 2011-11-16T13:32:52 *** praveen_ has quit IRC (Quit: Page closed) 2011-11-16T13:32:59 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2011-11-16T13:33:48 *** DrClaes has quit IRC (Quit: Page closed) 2011-11-16T13:34:28 *** jcarpio has quit IRC (Quit: Page closed) 2011-11-16T13:40:30 *** Belerafon has joined #aichallenge 2011-11-16T13:41:36 *** emirbuc has left #aichallenge 2011-11-16T13:41:44 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-11-16T13:42:58 *** retybok_ has quit IRC (Ping timeout: 252 seconds) 2011-11-16T13:44:43 *** retybok_ has joined #aichallenge 2011-11-16T13:45:11 *** attare has joined #aichallenge 2011-11-16T13:46:00 *** emb_ has joined #aichallenge 2011-11-16T13:46:21 *** attare has quit IRC (Client Quit) 2011-11-16T13:46:53 e 2011-11-16T13:47:00 McLeopold: pulled, merged, waiting for six games to finish 2011-11-16T13:47:16 hard to get irc to work on my ubuntu.. 2011-11-16T13:47:16 *** Larose has quit IRC (Read error: Connection reset by peer) 2011-11-16T13:47:23 can anyone read this? 2011-11-16T13:47:39 yes 2011-11-16T13:51:28 *** Belerafon has quit IRC (Remote host closed the connection) 2011-11-16T13:51:30 *** gratz has quit IRC (Quit: Leaving) 2011-11-16T13:53:11 *** emb_ has quit IRC (Quit: Lost terminal) 2011-11-16T13:53:32 i hope berak's usage of phpskill works well... 2011-11-16T13:54:09 java one often failed and python one looked invalid 2011-11-16T13:54:31 *** emirbuc has joined #aichallenge 2011-11-16T13:55:12 *** emirbuc has quit IRC (Client Quit) 2011-11-16T13:57:00 *** NoxiaZ^ has quit IRC (Ping timeout: 260 seconds) 2011-11-16T13:59:02 *** Jether has joined #aichallenge 2011-11-16T13:59:13 ? 2011-11-16T13:59:17 hi 2011-11-16T14:00:16 i develop game on XNA 2011-11-16T14:01:49 *** Harpyon has joined #aichallenge 2011-11-16T14:02:56 Why are you not a millionnaire yet? 2011-11-16T14:03:50 Fluxid: ah, is that whats on tcpants? 2011-11-16T14:04:55 *** choas has joined #aichallenge 2011-11-16T14:06:07 *** Larose has joined #aichallenge 2011-11-16T14:07:32 Garf: dunno what is on tcpants, it looks modified but there is no fork on github 2011-11-16T14:13:46 *** magiik has quit IRC (Remote host closed the connection) 2011-11-16T14:16:44 *** mj41 has joined #aichallenge 2011-11-16T14:18:09 *** liberforce1 has left #aichallenge 2011-11-16T14:18:36 the tc-pants guy is romans01 on the forums, he said he would put his changes on github in one thread 2011-11-16T14:19:04 *** Jether has quit IRC (Quit: Leaving) 2011-11-16T14:19:11 *** Jether has joined #aichallenge 2011-11-16T14:19:32 *** cJether has joined #aichallenge 2011-11-16T14:19:49 but I could not find it either 2011-11-16T14:20:05 *** NoxiaZ^ has joined #aichallenge 2011-11-16T14:23:27 *** mj41 has quit IRC (Ping timeout: 248 seconds) 2011-11-16T14:24:16 *** dvladim has quit IRC (Ping timeout: 240 seconds) 2011-11-16T14:24:53 *** mikewintermute has joined #aichallenge 2011-11-16T14:36:09 *** jasox has quit IRC (Quit: Leaving) 2011-11-16T14:36:12 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-11-16T14:36:49 *** NightExcessive has quit IRC (Ping timeout: 255 seconds) 2011-11-16T14:42:24 I am a game programmer 2011-11-16T14:42:27 i like xna 2011-11-16T14:42:30 *** choas has quit IRC (Ping timeout: 260 seconds) 2011-11-16T14:42:42 Is this a game development channel?? 2011-11-16T14:44:01 what the the channel name tell you 2011-11-16T14:44:16 if you want a game development channel then join #gamedev 2011-11-16T14:44:25 *** mikewintermute has joined #aichallenge 2011-11-16T14:44:29 what is xna? 2011-11-16T14:44:59 xbox indie development kit or smth 2011-11-16T14:44:59 XNA is used to develop games in C# :) 2011-11-16T14:45:08 both for xbox and windows 2011-11-16T14:45:16 *** conor_f has joined #aichallenge 2011-11-16T14:45:18 so irrelevant 2011-11-16T14:45:22 yep 2011-11-16T14:45:25 k 2011-11-16T14:45:41 I don't know what is ai chanllenge 2011-11-16T14:45:50 cJether: check out the website 2011-11-16T14:45:52 you gonna love it 2011-11-16T14:46:01 did the workers die again? game rate is at 0.5 2011-11-16T14:46:02 which website 2011-11-16T14:46:08 @tpoic 2011-11-16T14:46:09 mcstar: I do not know about 'tpoic', but I do know about these similar topics: 'topic' 2011-11-16T14:46:12 @topic 2011-11-16T14:46:12 mcstar: AI Challenge (sponsored by Google): http://aichallenge.org || Channel Logs: http://contestbot.aichallenge.org || Code Repo: http://github.com/aichallenge/aichallenge || About the game (Ants): http://aichallenge.org/problem_description.php || Submission deadline: December 18 23:59 EST 2011-11-16T14:46:15 cJether, just join #gamedev 2011-11-16T14:47:04 I don't know what the game rate should be, is 0.5 normal? 2011-11-16T14:47:05 *** Larose has quit IRC (Read error: Connection reset by peer) 2011-11-16T14:47:22 *** Accoun has quit IRC () 2011-11-16T14:47:23 Minthos: sometimes they are worse 2011-11-16T14:47:26 only 2 workers alive 2011-11-16T14:47:32 ok 2011-11-16T14:48:00 that's pretty low. ec2 must be in demand 2011-11-16T14:48:14 *** cJether has left #aichallenge ("Leaving") 2011-11-16T14:48:46 i guess i'll be in 6329th place for a while 2011-11-16T14:49:14 a1k0n: not used to it right? 2011-11-16T14:49:39 a1kon, you can't be, i'm there ! 2011-11-16T14:50:36 *** conor_f has left #aichallenge 2011-11-16T14:50:46 *** peyton has joined #aichallenge 2011-11-16T14:51:16 How many bots get to the final? 2011-11-16T14:52:12 all of them? 2011-11-16T14:52:21 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-11-16T14:57:16 *** Flapadar has joined #aichallenge 2011-11-16T14:57:30 Is it me or are games going far slower now? 2011-11-16T14:57:47 did you reset your bot? 2011-11-16T14:57:53 i.e. upload a new one? 2011-11-16T14:58:05 No, ranked <2000 atm 2011-11-16T14:58:10 It's saying I have 15 hours until the next game lol 2011-11-16T14:58:21 2 days until deactivation (so it's not that) 2011-11-16T14:58:46 It's not you 2011-11-16T15:00:28 *** Strider42 has left #aichallenge 2011-11-16T15:01:43 *** Kurnevsky has joined #aichallenge 2011-11-16T15:03:49 *** Kurnevsky has left #aichallenge 2011-11-16T15:05:40 *** Larose has joined #aichallenge 2011-11-16T15:06:39 *** mikewintermute has joined #aichallenge 2011-11-16T15:06:41 *** epicmonkey has joined #aichallenge 2011-11-16T15:06:58 *** Accoun has joined #aichallenge 2011-11-16T15:11:54 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-11-16T15:12:11 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-16T15:12:31 *** ikaros has joined #aichallenge 2011-11-16T15:13:55 *** Flapadar has quit IRC () 2011-11-16T15:16:00 *** mceier has quit IRC (Quit: leaving) 2011-11-16T15:17:07 *** Larose has quit IRC (Read error: Connection reset by peer) 2011-11-16T15:17:40 *** retybok_ has quit IRC (Quit: Lost terminal) 2011-11-16T15:24:54 *** elderwolf_ has joined #aichallenge 2011-11-16T15:26:45 *** nle85 has joined #aichallenge 2011-11-16T15:28:53 *** Transformer has joined #aichallenge 2011-11-16T15:29:46 *** mleyen has quit IRC (Quit: ~ Trillian Astra - www.trillian.im ~) 2011-11-16T15:31:46 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-11-16T15:32:14 *** Transformer has quit IRC (Excess Flood) 2011-11-16T15:34:53 *** mikewintermute has joined #aichallenge 2011-11-16T15:35:44 *** Larose has joined #aichallenge 2011-11-16T15:36:07 *** mikewintermute has quit IRC (Client Quit) 2011-11-16T15:41:50 *** callahan has quit IRC (Ping timeout: 244 seconds) 2011-11-16T15:42:16 *** ihtisham has joined #aichallenge 2011-11-16T15:45:39 *** delt0r__ has quit IRC (Ping timeout: 252 seconds) 2011-11-16T15:46:41 *** Larose has quit IRC (Read error: Connection reset by peer) 2011-11-16T15:48:50 *** lknix has quit IRC (Ping timeout: 276 seconds) 2011-11-16T15:50:35 *** peyton has quit IRC (Quit: peyton) 2011-11-16T15:51:35 *** sashasochka has joined #aichallenge 2011-11-16T15:51:40 hi 2011-11-16T15:51:42 *** dorisabayon has quit IRC (Ping timeout: 245 seconds) 2011-11-16T15:51:57 *** Fandekasp has quit IRC (Ping timeout: 258 seconds) 2011-11-16T15:53:52 Heh, we're down to .3 games per minute. That about par? 2011-11-16T15:54:07 *** Antimony has joined #aichallenge 2011-11-16T15:54:27 *** callahan has joined #aichallenge 2011-11-16T15:54:30 seen a1k0n 2011-11-16T15:55:34 *** GeorgeSebastian has quit IRC (Quit: popped out) 2011-11-16T15:55:56 hello? 2011-11-16T15:56:12 Hi. 2011-11-16T15:57:10 *** borist has joined #aichallenge 2011-11-16T15:57:41 oh 2011-11-16T15:57:54 noone was talking so I thought I had connection problems 2011-11-16T15:58:37 ContestBot seen a1k0n 2011-11-16T15:58:37 Antimony: a1k0n was last seen in #aichallenge 1 hour, 9 minutes, and 50 seconds ago: i guess i'll be in 6329th place for a while 2011-11-16T15:58:37 *** delt0r__ has joined #aichallenge 2011-11-16T15:58:51 ContestBot seen BenJackson 2011-11-16T15:58:51 Antimony: BenJackson was last seen in #aichallenge 14 hours, 24 minutes, and 58 seconds ago: wait until I finish downloading this episode of Lost with the pw site 2011-11-16T15:59:55 Heh, i'm assuming contestbot is running on same server that is hosting the challenge, and it needs a good reboot. 2011-11-16T16:00:38 *** elderwolf has joined #aichallenge 2011-11-16T16:00:49 *** Fandekasp has joined #aichallenge 2011-11-16T16:00:54 the server running the games is down? that's why the games per minute is so low at the momwnt? 2011-11-16T16:00:59 *** dorisabayon has joined #aichallenge 2011-11-16T16:01:26 No, I'm speculating that it's running slow due to contestbot's extreme lag and the low gpm count. 2011-11-16T16:02:16 *** nle85 has quit IRC (Ping timeout: 265 seconds) 2011-11-16T16:03:38 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-11-16T16:03:43 *** borist has quit IRC (Quit: Leaving.) 2011-11-16T16:03:43 *** sashasochka has quit IRC (Ping timeout: 265 seconds) 2011-11-16T16:03:58 I'm having a very strange problem 2011-11-16T16:04:31 game http://ants.fluxid.pl/replay.1627, turn 153, my bot (GotAnts) gets "a 65 77 0" as input 2011-11-16T16:04:41 *** elderwolf_ has quit IRC (Ping timeout: 265 seconds) 2011-11-16T16:05:10 there is a single ant there, so that's alright, only in turn 152 my bot ordered that ant to go north 2011-11-16T16:05:25 since it doesn't expect an ant there, it assers (hence the timeout) 2011-11-16T16:07:08 http://paste.aichallenge.org/CoKNs/raw/ Here is the input my bot received on turn 152, partial input from turn 153 (until the offending ant) and the output from turn 152 2011-11-16T16:07:43 Are there other people that implemented state in their ants and have come acros this problem? 2011-11-16T16:09:30 *** Fluctuation has joined #aichallenge 2011-11-16T16:11:46 *** peyton has joined #aichallenge 2011-11-16T16:13:21 *** dorisabayon has quit IRC (Quit: Lost terminal) 2011-11-16T16:13:37 *** ltriant has joined #aichallenge 2011-11-16T16:14:12 *** epicmonkey has quit IRC (Ping timeout: 245 seconds) 2011-11-16T16:16:02 *** nle85 has joined #aichallenge 2011-11-16T16:21:09 *** Palmik_ has quit IRC (Remote host closed the connection) 2011-11-16T16:21:45 *** NightExcessive has joined #aichallenge 2011-11-16T16:22:32 Is there any way to, given an html replay, replay that game using the engine (ie replace all bots by something that reads from the replay and executes those moves)? 2011-11-16T16:22:46 or any other ideas how to debug the beforementioned problem? 2011-11-16T16:24:40 What's the problem? (I just joined) 2011-11-16T16:25:28 http://paste.openttdcoop.org/show/762/ <- log 2011-11-16T16:25:34 *** ALplus has joined #aichallenge 2011-11-16T16:27:54 *** nle85 has quit IRC (Ping timeout: 265 seconds) 2011-11-16T16:28:28 *** NoxiaZ^ has quit IRC (Read error: Connection reset by peer) 2011-11-16T16:32:06 *** bishoco has joined #aichallenge 2011-11-16T16:32:11 *** nle85 has joined #aichallenge 2011-11-16T16:39:02 *** bishoco has quit IRC (Quit: Page closed) 2011-11-16T16:41:30 *** okan_ has joined #aichallenge 2011-11-16T16:42:02 ach, meine bloeden antzen schaffen's einfach nicht, zwei hauffen auf einmal zu schiessen 2011-11-16T16:44:43 *** Antimony has joined #aichallenge 2011-11-16T16:50:07 g0llum: we mostly speak english in here 2011-11-16T16:50:54 *** jasox has joined #aichallenge 2011-11-16T16:53:07 (wrong tab, pun, srry) 2011-11-16T16:55:53 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-11-16T16:56:57 *** Redgis has joined #aichallenge 2011-11-16T17:04:30 *** fpcfan has joined #aichallenge 2011-11-16T17:05:16 amstan: the problem with the archive server seemed to be a combination of planetwars being the default site and the planetwars config was accepting any url and proxying it 2011-11-16T17:05:49 how was planetwars accepting any url? 2011-11-16T17:06:15 you can give it a try by re-enabling the site then set the http proxy in your browser to archive.aichallenge.org port 80 2011-11-16T17:06:46 archive? we aren't using that yet 2011-11-16T17:06:55 then visit say whatismyip.com and you'll see your traffic is going through that server 2011-11-16T17:07:04 or just enter the ip directly 2011-11-16T17:07:16 it didn't matter if we were using it someone found it :) 2011-11-16T17:07:28 i don't remember setting up planetwars on the archive server 2011-11-16T17:07:29 probably just scanning through ip addresses 2011-11-16T17:07:41 i kinda gave up when mysql died 2011-11-16T17:07:42 oh well 2011-11-16T17:07:45 amstan: what's the schedule for this stuff? will there be a winter 2011 or a spring/summer ever? 2011-11-16T17:07:58 it looked like the config was just copied over from what is on the main server currently 2011-11-16T17:08:01 u_: depends on how fast we develop 2011-11-16T17:08:12 janzert: yeah.. ok, that's not that bad then, keep it off 2011-11-16T17:08:31 i tought you were talking about the normal setup we have now on the main server 2011-11-16T17:08:32 I got the directories rearranged now btw 2011-11-16T17:08:35 ok, cool 2011-11-16T17:08:46 *** Raimondi_ is now known as Raimondi 2011-11-16T17:08:55 no, that is working fine because the planetwars config isn't the default site 2011-11-16T17:09:17 *** magiik has joined #aichallenge 2011-11-16T17:09:23 so any unrecognized url goes to aichallenge.org 2011-11-16T17:09:24 so it's not proxying stuff 2011-11-16T17:09:27 sorry to repeat myself again, but does any of you have suggestions for how to debug the engine? More explenation here: http://paste.openttdcoop.org/show/762/ 2011-11-16T17:09:38 man wtf all the amazon ec2 us-east micro instances are like above $1 2011-11-16T17:09:53 just planetwars.aichallenge.org to jeff's server as ai-contest.com like it should be 2011-11-16T17:10:04 ok, cool 2011-11-16T17:10:10 i just found some broken links on that site 2011-11-16T17:10:13 from the winner stuff 2011-11-16T17:10:17 on the homepage 2011-11-16T17:10:26 at least from what I could see in my quick look at it 2011-11-16T17:10:36 yeah, I noticed that before 2011-11-16T17:10:47 some of the winner links have gone away right? 2011-11-16T17:10:53 *** okan_ has quit IRC (Quit: Page closed) 2011-11-16T17:10:56 the links to their personal sites I mean 2011-11-16T17:11:05 janzert: that's fine 2011-11-16T17:11:21 i'm talking about the fact where if i click on bocsimacko. it points to the current site 2011-11-16T17:11:23 *** sir_macelon has quit IRC (Ping timeout: 265 seconds) 2011-11-16T17:11:29 they're hardcoded 2011-11-16T17:12:09 hmm, yeah looks like they're absolute urls 2011-11-16T17:12:21 they should be relative urls and they'd be fine :( 2011-11-16T17:12:57 ok, fixed 2011-11-16T17:13:08 oh man, i hate 80 column lines 2011-11-16T17:13:18 you want to do any change, it's like impossible 2011-11-16T17:13:23 without it looking bad 2011-11-16T17:13:29 *** Fandekasp has quit IRC (Ping timeout: 252 seconds) 2011-11-16T17:14:15 gq} 2011-11-16T17:14:17 ;) 2011-11-16T17:14:42 janzert: any news from jeff? 2011-11-16T17:15:02 he wrote me an email today, must not have gone to the list? 2011-11-16T17:15:09 he just got the paypal account set up 2011-11-16T17:15:23 hopefully that means we can get a button on the front page soon 2011-11-16T17:15:26 nope, apparently he didn't list it 2011-11-16T17:15:27 *** twymer has quit IRC (Ping timeout: 245 seconds) 2011-11-16T17:15:43 what about google? 2011-11-16T17:15:59 doesn't mention anything about talking to google 2011-11-16T17:16:29 donations for more workers? 2011-11-16T17:17:39 kara: yes 2011-11-16T17:19:16 :-) 2011-11-16T17:22:01 *** foRei has quit IRC (Read error: Connection reset by peer) 2011-11-16T17:30:48 hey, why is the game rate too slow? 2011-11-16T17:31:46 2011-11-16T17:32:53 because it's interpolated over time, 5 of 6 possible workers running now ( was 2 minutes ago ) 2011-11-16T17:33:30 well.. there's no real upper limit on worker 2011-11-16T17:33:32 workers* 2011-11-16T17:33:40 money ? 2011-11-16T17:33:47 yes 2011-11-16T17:33:53 s/real/technical 2011-11-16T17:34:00 nvm 2011-11-16T17:34:02 that exists too 2011-11-16T17:34:06 never seen more than 6 2011-11-16T17:34:06 but it'll be like 100 2011-11-16T17:34:12 ;) 2011-11-16T17:34:46 time for a raindance 2011-11-16T17:36:38 *** ihtisham has quit IRC (Quit: Page closed) 2011-11-16T17:38:45 *** mikewintermute has joined #aichallenge 2011-11-16T17:42:25 *** pairofdice has quit IRC (Quit: Just keep breathing) 2011-11-16T17:47:11 *** bergmark has quit IRC (Ping timeout: 248 seconds) 2011-11-16T17:50:51 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-16T17:53:27 *** jasox has quit IRC (Quit: Leaving) 2011-11-16T17:55:56 *** Jak_o_Shadows has joined #aichallenge 2011-11-16T18:01:50 *** bishoco has joined #aichallenge 2011-11-16T18:02:13 *** ikaros has joined #aichallenge 2011-11-16T18:02:39 *** davidmoreau has joined #aichallenge 2011-11-16T18:03:07 hi 2011-11-16T18:03:22 looking for a contest organizer, lost my password :) 2011-11-16T18:04:42 *** analyst74 has joined #aichallenge 2011-11-16T18:04:57 *** davidmoreau is now known as davm85 2011-11-16T18:05:57 *** Antimony has joined #aichallenge 2011-11-16T18:06:30 well, guess i'd have to setup another one 2011-11-16T18:07:51 davm85: http://aichallenge.org/forgot.php 2011-11-16T18:08:26 well thanks, but i actually forgot also the username :S 2011-11-16T18:09:30 amstan: ^ 2011-11-16T18:09:45 forgot password flows without email makes me feel old 2011-11-16T18:12:59 well, I'm sure amstan will help you when he gets back 2011-11-16T18:13:13 thx thestinger 2011-11-16T18:13:17 if you just want to test stuff, you can use the tcp server for now 2011-11-16T18:13:27 davm85: what's your email? 2011-11-16T18:13:28 http://ants.fluxid.pl/ 2011-11-16T18:16:09 wow us-east1d micro instance spot price is at $5 jesus christ 2011-11-16T18:16:56 yeah, ec2 is having some really weird spot prices this week 2011-11-16T18:17:15 isn't it cheaper to just buy an instance? 2011-11-16T18:17:26 right now it is 2011-11-16T18:17:41 but i usually pay like $.02 for the spot instances cuz it's cheaper most of the time 2011-11-16T18:17:44 i suppose not overall. must be some big mapreduce 2011-11-16T18:17:46 at times I'm pretty sure that no matter how high you bid recently it isn't going to give you an instance 2011-11-16T18:17:58 if it's some big mapreduce - why use micro isntances? haha 2011-11-16T18:18:10 because each one is small? shrug 2011-11-16T18:18:18 *** twymer has joined #aichallenge 2011-11-16T18:18:20 maybe they're cracking passwords 2011-11-16T18:18:29 really valuable ones 2011-11-16T18:18:36 if your bid is higher than someone else who hsa an instane, it will eventually boot them off and give itto you 2011-11-16T18:18:39 earlier in the week it was us-east1b not it seems to be 1a and 1d 2011-11-16T18:18:48 a1k0n: prolly easier to do it in a GPU? 2011-11-16T18:18:56 it's all 3, they are all above like $1 right now 2011-11-16T18:18:58 *** chris_0076 has quit IRC (Quit: Leaving) 2011-11-16T18:19:05 *** Harpyon has quit IRC (Quit: Computer has gone to sleep.) 2011-11-16T18:19:11 yeah usually it's less than a cent 2011-11-16T18:19:17 *** Chris_0076 has joined #aichallenge 2011-11-16T18:19:18 magiik: I have a feeling there weren't any spot instances available period though 2011-11-16T18:19:24 i'm in the process of moving over to us-west oregon 2011-11-16T18:19:25 true 2011-11-16T18:19:29 but god damnt i have to manually move my volume data 2011-11-16T18:19:47 janzert: ahhh - i see what you are saying, just spot instances in general since they are the "leftovers" 2011-11-16T18:20:03 right 2011-11-16T18:20:45 it certainly has been causing havoc with our workers :( 2011-11-16T18:20:53 man this server has been running as a spot instance with a .20 bid for like a year 2011-11-16T18:20:57 and this last month or two spot pricing has been going crazy 2011-11-16T18:21:05 is aichallenge on us-east too? 2011-11-16T18:21:11 yep 2011-11-16T18:22:03 i need some sort of management script, whenever my spot instance does come back up, it loses it's elastic ip 2011-11-16T18:22:28 and i need a better way to keep that EBS volume linked up to the new instance too 2011-11-16T18:22:28 yeah, I wish the ip tied to the spot request 2011-11-16T18:22:38 and that as well :) 2011-11-16T18:23:26 such a pain in the ass right now the way i do it - create a snapshot of the orphaned volume, then have to re-create the AMI with that snapshot, then re-request the spot 2011-11-16T18:24:19 yeah, using ebs as a root volume for spot instances doesn't work well at all right now 2011-11-16T18:24:33 at least not I haven't found a way for it to 2011-11-16T18:24:42 yeah... 2011-11-16T18:24:49 and i'm too entrenched in this server's config now 2011-11-16T18:25:29 damn us-west cali is having some weird spot price ripples too, but not breaking $1 2011-11-16T18:25:35 I simply store my configs on S3 and have my spot instances download them on start. :) 2011-11-16T18:26:18 janzert: people on the forums are debating that we should keep the workers on during high prices 2011-11-16T18:26:43 but i disagree, especially if it's 50 times more expensive, right? 2011-11-16T18:26:53 more than 50.... 2011-11-16T18:26:59 yeah 2011-11-16T18:27:08 amstan: What size instance are you using for workers? 2011-11-16T18:27:34 m2.xlarge 2011-11-16T18:27:49 *** Redgis has quit IRC (Ping timeout: 240 seconds) 2011-11-16T18:28:10 Icky. I definitely wouldn't leave them running. 2011-11-16T18:28:43 hrm... actually price differntial on m2.xlarge is only like 50x right now.... but yeah that's a lot of cash 2011-11-16T18:28:44 basically we've got funds currently to spend ~$1 an hour on workers 2011-11-16T18:29:10 so it doesn't really matter too much what people think it's what we've got ;) 2011-11-16T18:29:27 Indeed :P 2011-11-16T18:29:49 yeah i was gonna say, if you have enough for $1/hr then there's no way you can leave them up when east-1c is jumping up to $5 lol 2011-11-16T18:30:04 yep :) 2011-11-16T18:30:06 you guys have scripts to spool instances up and down? or you just go off the instance getting terminated when it crosses the bid threshold? 2011-11-16T18:30:23 magiik: The manager is capable of starting instances. I know that. 2011-11-16T18:30:48 they're permanent spot requests at the threshold prices and the user_data script brings it up as worker 2011-11-16T18:31:21 wonder if it's worth the effort to switch to regular instances if the spot price is exceeding that price 2011-11-16T18:32:08 Maybe. Though, the spot prices should go back down as soon as Amazon installs new physical nodes. 2011-11-16T18:32:51 *** fpcfan has quit IRC (Ping timeout: 244 seconds) 2011-11-16T18:34:17 *** kara has quit IRC () 2011-11-16T18:35:59 *** arscan has joined #aichallenge 2011-11-16T18:36:19 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-16T18:37:37 if I run python tools/playgame.py (without the visualizer option) I only get 1 replay file numbered 0.replay. How do I store ALL games (plus the html launch files)? 2011-11-16T18:37:46 *** fpcfan has joined #aichallenge 2011-11-16T18:39:12 *** delt0r__ has quit IRC (Remote host closed the connection) 2011-11-16T18:40:06 *** davm85 has quit IRC (Remote host closed the connection) 2011-11-16T18:40:42 *** bergmark has joined #aichallenge 2011-11-16T18:41:49 aha. --html 2011-11-16T18:42:08 hmm - that only lets you write to the named file. 2011-11-16T18:46:59 u can change the #.replay 2011-11-16T18:47:01 with -g # 2011-11-16T18:47:03 IIRC 2011-11-16T18:47:15 python tools/playgame.py 2011-11-16T18:47:18 will give you all the CLI switches 2011-11-16T18:47:41 magiik: thanks 2011-11-16T18:47:52 i dont think there is any "auto-increment" built into the script 2011-11-16T18:48:00 so you'd have to make your own wrapper around it if you wanted automatic functionality 2011-11-16T18:48:02 there is -runs i think 2011-11-16T18:48:15 ah -r 2011-11-16T18:48:23 but not between runs 2011-11-16T18:48:25 then yeah i gues it would auto-increment, starting from -g # 2011-11-16T18:48:34 actually did go to regular instance for a while earlier this week 2011-11-16T18:49:06 janzert: doesn't that mean a different cpu? 2011-11-16T18:49:12 no 2011-11-16T18:49:17 same instance type 2011-11-16T18:49:23 just different pricing 2011-11-16T18:49:27 amstan: heard from j3camero lately? about accepting donations? 2011-11-16T18:49:29 yeah same number of compute units 2011-11-16T18:49:41 they are exactly the same - except for price/availability 2011-11-16T18:50:02 McLeopold: yeah, he wrote me earlier today and has paypal setup now 2011-11-16T18:50:15 so hopefully he can get us a donate button soon 2011-11-16T18:50:20 oh, cool, so we need a button on the page now? 2011-11-16T18:50:51 janzert: I assume you can do the calculations to know when to add more workers? 2011-11-16T18:51:03 to spend the donated money as fast as possible :) 2011-11-16T18:51:11 of course :) 2011-11-16T18:51:12 *** ltriant has quit IRC (Read error: Connection reset by peer) 2011-11-16T18:51:21 *** ltriant has joined #aichallenge 2011-11-16T18:51:25 my updated battle resolver is ready 2011-11-16T18:51:42 janzert: what would you think about limiting starter bots to 10 games, lifetime max? 2011-11-16T18:52:01 across users, so even if they upload a starter a second time, it wouldn't play 2011-11-16T18:52:21 accross submissions you mean 2011-11-16T18:52:30 yeah, maybe :) 2011-11-16T18:52:31 I'm not sure 2011-11-16T18:52:34 for or against 2011-11-16T18:52:51 how can i remove a file from git? 2011-11-16T18:52:51 how about detecting starters and setting the mu to 0? 2011-11-16T18:52:58 i.e. to tell it not to track it 2011-11-16T18:53:00 "git rm " 2011-11-16T18:53:04 thx 2011-11-16T18:53:06 how would you decide it was a starter as well? 2011-11-16T18:53:10 you'll need to commit it as well 2011-11-16T18:53:20 which is weird because you commit a non-existant file 2011-11-16T18:53:44 McLeopold: ok, and how can i list all tracked files? 2011-11-16T18:53:46 janzert: assuming I can detect it of course, I'm going to try submission hashes then build a database 2011-11-16T18:53:51 "git status" 2011-11-16T18:53:54 stupid questions, but i didnt find these 2011-11-16T18:53:56 ah 2011-11-16T18:54:37 :( yeah, that seems so fragile but I'm not sure if there's a better way 2011-11-16T18:54:38 no, status doesnt list tracked files 2011-11-16T18:56:17 Maybe set up an admin page that lists submissions with same hash and allow the admin to mark them as starter bots 2011-11-16T18:56:27 *** grandiose has joined #aichallenge 2011-11-16T18:57:14 *** nle85 has quit IRC (Ping timeout: 265 seconds) 2011-11-16T18:58:18 at the same time I don't know that this will significantly increase the game rates for others and could cause ranking problems :( 2011-11-16T18:58:18 git ls-files 2011-11-16T18:59:23 there are already 2405 deactivated bots that didn't get a game in the last 24 hours 2011-11-16T18:59:28 *** xathis has quit IRC (Ping timeout: 240 seconds) 2011-11-16T18:59:40 *** [1]analyst74 <[1]analyst74!~analyst74@S0106c84c752ced49.ed.shawcable.net> has joined #aichallenge 2011-11-16T18:59:43 Raimondi: thx 2011-11-16T18:59:50 so a little over a third already 2011-11-16T18:59:51 np 2011-11-16T19:00:10 so if anyone feels like experimenting, please clone: https://github.com/liquid-phynix/ants-fight.git 2011-11-16T19:00:20 and have qt ready 2011-11-16T19:00:29 I don't see anything horrible the idea. Getting a hash of the starter zips and limiting it to 10 games/starter package/user seems reasonable. 2011-11-16T19:00:49 NightExcessive: i just updated 2 starter packages today 2011-11-16T19:00:57 not everyone updates 2011-11-16T19:01:19 amstan: It's simple enough to keep a hash list of old & current starter packages :) 2011-11-16T19:01:22 amstan: amstan, pls try it again, its much better 2011-11-16T19:01:24 yeah, that's the fragility of it. there are going to be lots of starter bot hashes and more all the time 2011-11-16T19:01:52 *** analyst74 has quit IRC (Ping timeout: 240 seconds) 2011-11-16T19:01:52 *** [1]analyst74 is now known as analyst74 2011-11-16T19:02:00 NightExcessive: it is? what happens when a new version of another starter bot gets released? 2011-11-16T19:02:17 *** grandiose1 has joined #aichallenge 2011-11-16T19:02:19 amstan: I'm not sure I see your point. What do you mean? 2011-11-16T19:02:38 we keep making new starter packages, with different contents 2011-11-16T19:02:43 all of them have different hashes 2011-11-16T19:03:22 a change of 1 characters throws off the hash, but it is still a starter bot 2011-11-16T19:03:26 -s 2011-11-16T19:03:32 is starter package a serious problem? how many starter bots are active? 2011-11-16T19:03:52 id say more than half of all bots 2011-11-16T19:03:53 we already have a 3 day timeout, i seriously doubt any contestant keeps starter packages alive 2011-11-16T19:03:56 mcstar: Fair enough. Though, this would solve simple situations where it's the package just downloaded and re-uploaded. 2011-11-16T19:04:01 mcstar: I doubt it 2011-11-16T19:04:08 janzert: how much then? 2011-11-16T19:04:12 there are only ~900 bots active 2011-11-16T19:04:22 and 2400 bots haven't played at all 2011-11-16T19:04:22 so you mean even more? 2011-11-16T19:04:28 in the last day 2011-11-16T19:04:31 amstan: How large is the database of submissions? String comparison is fairly cheap, but I could see comparing (a few thousand strings)*20 being expensive. 2011-11-16T19:04:45 so I have a feeling most starter bots already aren't playing 2011-11-16T19:05:02 NightExcessive: why are you concerned about this? 2011-11-16T19:05:30 amstan: I was just curious about the size, as it limits the practicality of the idea. 2011-11-16T19:05:35 *** grandiose has quit IRC (Ping timeout: 260 seconds) 2011-11-16T19:06:13 NightExcessive: searching in a db is rarely linear 2011-11-16T19:06:28 unless it's configured wrong 2011-11-16T19:06:33 amstan: pls take a look, cause if you feel it is useful, ill make a forum post or something 2011-11-16T19:06:43 anyway, its sort of bad that it depends on qt 2011-11-16T19:06:54 looks good though 2011-11-16T19:07:04 do you think a binary distribution is a problem? 2011-11-16T19:07:08 yes 2011-11-16T19:07:18 even if its from github? 2011-11-16T19:07:22 i mean the binary 2011-11-16T19:07:22 there have been 23,719 submissions so far. I doubt the computational load of implementing this is an impediment in any way 2011-11-16T19:07:28 you shouldn't have binaries on github in the first place 2011-11-16T19:07:37 yeah, but i can have 2011-11-16T19:07:49 but seen this latest one? 2011-11-16T19:08:11 theres a transition effect, and you can more conveniently add/remove ants 2011-11-16T19:09:17 If it would really have an effect on the game rate I don't see a problem with using a simple hash and category system. Categorize the starter packs by language and look for matching hashes, then count them per category per user- and don't allow more than 10 games per category (from each user). 2011-11-16T19:09:37 NightExcessive: you don't get it 2011-11-16T19:09:47 NightExcessive: some starter packages have like 10 versions, so 10 different hashes 2011-11-16T19:09:47 Oh? 2011-11-16T19:10:02 keeping the hash database up to date is what will be the pain 2011-11-16T19:10:05 and i don't even think we can get the hashes of the old versions 2011-11-16T19:10:14 the zips might have dates in them or something 2011-11-16T19:10:32 unless we do hashes of the contents 2011-11-16T19:10:41 If they're in git, you can simply go back to the commit with them in it. I can see how it could be a pain keeping the hashes up-to-date. 2011-11-16T19:11:02 ugh, currently we don't uncompress a submission on the main server at all 2011-11-16T19:11:06 contents... you could zip it up different ways right? 2011-11-16T19:11:09 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-11-16T19:11:31 I suppose the worker could report back on the content hash though 2011-11-16T19:11:54 janzert: don't we already have that 3 day timeout on submissions? 2011-11-16T19:11:59 but it does seem to add a lot of complexity for what looks like little gain 2011-11-16T19:12:02 right 2011-11-16T19:12:17 i really don't think begginers keep their starter bots active 2011-11-16T19:12:29 they'll stay active at most 3 days 2011-11-16T19:12:29 janzert: I agree... it doesn't seem like a pressing issue, nor does it seem it would really speed up the games per second. 2011-11-16T19:12:32 and if someone is activating their starter bot... it's not like they are activating another bot since you aren't supposed to have multi accounts 2011-11-16T19:12:43 most users don't keep their submission active :/ 2011-11-16T19:12:57 i even lose my real submission's activity every once in a while 2011-11-16T19:12:58 amstan: doesnt it mean only that those cant be seeds, 2011-11-16T19:13:00 there are currently 946 active 2011-11-16T19:13:05 they still play games 2011-11-16T19:13:11 mcstar: sure, nothing wrong with that 2011-11-16T19:13:37 so isnt the issue of sparing resources? 2011-11-16T19:13:44 (i was late to join in) 2011-11-16T19:14:06 oops, actually there 1016 active. 2011-11-16T19:14:08 you could hash all submissions, store the hashes, and reject any bots with hashes identical to more than just a few other bots - but the tutorial/intro thing tells players to upload the starting kit so maybe it's better to just let people do it and keep the current system 2011-11-16T19:14:23 yes.. but i don't think it's worth implementing such a complicated system for such few speedups 2011-11-16T19:14:39 Minthos: right we still want people to start with uploading a starter for sure 2011-11-16T19:14:41 how about not letting inactive bots below a certain rank even play at all 2011-11-16T19:14:58 1/3rd of bots aren't playing now ;) 2011-11-16T19:15:31 even as a new submission comes in with lowest rank? 2011-11-16T19:15:56 yep, a little more than 1/3rd of bots have not played in the last day 2011-11-16T19:16:02 how about segmenting the ranking, and give higher cuts more time on the server? 2011-11-16T19:16:19 i would think the inverse... give lower cuts more time heh 2011-11-16T19:16:23 otherwise how are you going to get up there? 2011-11-16T19:16:26 magiik: i don't think that's a good idea from a rankings perspective 2011-11-16T19:16:35 remember even though a new submission is ranked at the bottom it starts off playing in the roughly the middle of the established rankings 2011-11-16T19:16:49 and goes up from there so long as it wins 2011-11-16T19:16:49 well, thats the point: there are more bad bots than good ones, they need to play less 2011-11-16T19:17:08 yep 2011-11-16T19:17:14 not from the absolute end 2011-11-16T19:17:26 we want people who are actively working on their bot to get games even if their bot isn't so good :} 2011-11-16T19:18:01 btw, was it surprise for you guys that there are over 10k participants 2011-11-16T19:18:03 ? 2011-11-16T19:18:14 how was it for the other competitions? is this the most popular so far? 2011-11-16T19:18:21 i celebrated the 10kth user with a rainbow 2011-11-16T19:18:24 this is my 1st ai challenege, though i did hear about the galcon one 2011-11-16T19:18:25 message 2011-11-16T19:18:29 I thought we'd got to 10k but not till about the end of the competition 2011-11-16T19:18:42 we had a pretty exponential growth since the first contest 2011-11-16T19:19:05 last competition had 7200 users and 4619 active submissions 2011-11-16T19:20:14 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-16T19:20:40 besides ec2 being funky with its spot prices this week the rate of games a server can get out has gone down quite a bit (i.e. bots are using a lot more of their time on average now) 2011-11-16T19:21:11 which also points to the probability that we're playing much fewer starter bots 2011-11-16T19:21:24 yeah that makes sense 2011-11-16T19:22:57 *** Antimony has quit IRC (Ping timeout: 245 seconds) 2011-11-16T19:23:44 nite 2011-11-16T19:23:46 *** mcstar has quit IRC (Quit: WeeChat 0.3.6) 2011-11-16T19:24:06 *** bishoco has quit IRC (Quit: Page closed) 2011-11-16T19:31:15 *** Antimony has joined #aichallenge 2011-11-16T19:32:15 *** bergmark has quit IRC (Ping timeout: 248 seconds) 2011-11-16T19:36:13 *** grandiose1 has quit IRC (Quit: Leaving.) 2011-11-16T19:37:16 *** replore has joined #aichallenge 2011-11-16T19:40:08 *** analyst74 has quit IRC (Quit: HydraIRC -> http://www.hydrairc.com <- Po-ta-to, boil em, mash em, stick em in a stew.) 2011-11-16T19:45:00 *** grandiose has joined #aichallenge 2011-11-16T19:50:08 *** NightExcessive has quit IRC (Quit: Leaving) 2011-11-16T19:51:54 *** Antimony_ has joined #aichallenge 2011-11-16T19:53:01 *** Fluctuation has quit IRC (Quit: Page closed) 2011-11-16T19:53:35 *** Conorach has quit IRC (Ping timeout: 248 seconds) 2011-11-16T19:54:00 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-16T19:54:04 *** Antimony_ is now known as Antimony 2011-11-16T19:54:35 *** gcflymoto has joined #aichallenge 2011-11-16T19:56:23 *** Cowbandit has joined #aichallenge 2011-11-16T19:56:27 boop 2011-11-16T19:59:08 Cowbandit: ? 2011-11-16T20:00:39 amstan: double boop 2011-11-16T20:02:47 *** analyst74 has joined #aichallenge 2011-11-16T20:02:50 oi quick question: why are bot ranks reset after a new revision is submitted? 2011-11-16T20:03:06 Cowbandit: there's a topic on the forums about that 2011-11-16T20:03:15 ok thanks 2011-11-16T20:03:19 also i was wondering 2011-11-16T20:03:30 does me typing boop produce any boop sound 2011-11-16T20:03:36 no 2011-11-16T20:03:40 darn 2011-11-16T20:03:41 you're just looking weird 2011-11-16T20:06:29 would the boundary points of the python package visible() algorithm match a python Bresenham complete circle algorithm? 2011-11-16T20:07:07 *** Antimony has quit IRC (Ping timeout: 245 seconds) 2011-11-16T20:12:18 2 haskell bots in the top 100 :D 2011-11-16T20:13:56 ::D 2011-11-16T20:14:04 rubber duck you say? 2011-11-16T20:14:12 *** Anilm3 has quit IRC (Quit: Page closed) 2011-11-16T20:15:04 *** xathis has joined #aichallenge 2011-11-16T20:18:59 *** grandiose has quit IRC (Quit: Leaving.) 2011-11-16T20:20:52 let me rephrase. an ant out by its lonesome self has a vision radius of sqrt(ants.viewradius2) .. the boundary would be the points along the circle of the view radius.. question is whether Bresenham's circle algorithm match those points.. probably not because of the rounding during sqrt()? 2011-11-16T20:21:31 gcflymoto: i'm not familiar with that algorithm 2011-11-16T20:23:32 *** ALplus has quit IRC (Ping timeout: 244 seconds) 2011-11-16T20:25:42 *** grwip has quit IRC (Quit: Leaving) 2011-11-16T20:26:55 *** Rinum has joined #aichallenge 2011-11-16T20:27:21 gcflymoto: that algorithm would have to be able to deal with non-integral radii. 2011-11-16T20:27:41 how does combat work? if distance between 2 ants is < attack radius? 2011-11-16T20:27:56 There is no rounding during sqrt ! 2011-11-16T20:28:19 Rinum: yes, <= attack radius 2011-11-16T20:28:37 also, it seems attacks can happen diagonally? so the calculated distance is a little different 2011-11-16T20:28:48 But use the squared radius and squared distance, it is faster that way 2011-11-16T20:29:31 what do you mean? 2011-11-16T20:29:32 i'm sure bresenham would give you the same result 2011-11-16T20:29:32 Rinum: all radius measurements use euclidean distance 2011-11-16T20:29:34 Rinum: the values are radii so they describe a circle (in euclidean metric) 2011-11-16T20:29:57 *** Antimony has joined #aichallenge 2011-11-16T20:29:59 but does it matter? it's easy enough to brute force it once at the start of the bot 2011-11-16T20:30:00 have you guys seen this? http://www.franz.com/services/conferences_seminars/webinar_1-20-11.lhtml 2011-11-16T20:30:13 janzert mleise thanks 2011-11-16T20:30:17 it's a talk from mega1 on the contest and his botr 2011-11-16T20:31:13 yes, he did that a little while after the competition 2011-11-16T20:32:51 hmm, no _iouri_ for this contest 2011-11-16T20:33:12 trying to stay out of trouble maybe? 2011-11-16T20:33:36 nor dmj111 2011-11-16T20:37:26 *** Antimony has quit IRC (Ping timeout: 276 seconds) 2011-11-16T20:37:40 *** Antimony has joined #aichallenge 2011-11-16T20:37:46 *** gcflymoto has quit IRC (Ping timeout: 265 seconds) 2011-11-16T20:37:54 *** roflmao has quit IRC (Read error: Connection reset by peer) 2011-11-16T20:38:52 *** Antimony_ has joined #aichallenge 2011-11-16T20:41:11 *** Cowbandit has quit IRC (Quit: Page closed) 2011-11-16T20:42:07 *** Antimony has quit IRC (Ping timeout: 248 seconds) 2011-11-16T20:42:18 *** Antimony_ is now known as Antimony 2011-11-16T20:44:12 *** gcflymoto has joined #aichallenge 2011-11-16T20:49:03 *** Antimony has quit IRC (Ping timeout: 248 seconds) 2011-11-16T20:49:50 *** Antimony has joined #aichallenge 2011-11-16T20:53:43 *** gcflymoto has quit IRC (Ping timeout: 265 seconds) 2011-11-16T20:55:35 *** arscan has quit IRC (Quit: Page closed) 2011-11-16T20:58:37 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-16T21:03:46 *** b0rder has joined #aichallenge 2011-11-16T21:13:02 *** Antimony has joined #aichallenge 2011-11-16T21:13:04 *** bmh has joined #aichallenge 2011-11-16T21:13:06 *** bmh has joined #aichallenge 2011-11-16T21:14:16 *** bmh has joined #aichallenge 2011-11-16T21:15:07 a1k0n: If I did my math correct, I can do exact pathfinding in something like O(sqrt n) with O(n) preprocessing work 2011-11-16T21:15:35 or perhaps O(sqrt(n) * sqrt(sqrt(n))) 2011-11-16T21:16:44 what is n? 2011-11-16T21:16:55 Antimony: oh right -- the size of the map 2011-11-16T21:17:05 you mean one of the dimensions? 2011-11-16T21:17:09 no 2011-11-16T21:17:15 oh the total area/ 2011-11-16T21:17:51 bmh: really? what are you precomputing? 2011-11-16T21:19:06 a1k0n: for a map of size n, subdivide the map. For each subdivision find the a path from each cell to each neighboring subdivision 2011-11-16T21:19:42 On a general graph, this wouldn't guarantee shortest pathes, but I believe (have not proven...) that it should give exact queries given the restricted graph we're working with 2011-11-16T21:20:20 *** JorgeB has quit IRC (Ping timeout: 260 seconds) 2011-11-16T21:20:53 I think the preprocessing will take more than O(n). Maybe O(nlogn) 2011-11-16T21:21:25 anyway, pathfinding usually doesn't take much time 2011-11-16T21:21:49 I can find the closest target to each ant in only a couple ms 2011-11-16T21:22:05 Antimony: The preprocessing cost should be O(#subdivisions * #subdivision_size * #neighbors) 2011-11-16T21:22:44 I'm still trying to forcing myself to use CD x.x my bot seems to be improving 2011-11-16T21:23:00 The number of neighbors is independent of the size of the map. 2011-11-16T21:24:34 with sqrt(n) subdivisions of size sqrt(n) it comes out O(n) 2011-11-16T21:25:52 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-16T21:29:54 *** TheLinker has joined #aichallenge 2011-11-16T21:32:20 *** datachomper has joined #aichallenge 2011-11-16T21:36:12 (and to pontificate a bit more -- any reasonable subdivision will give you four neighbors) 2011-11-16T21:36:34 bmh: that would work with a maze? 2011-11-16T21:36:56 I see no reason why it wouldn't. 2011-11-16T21:36:57 oh, i see. you make a kind of quadtree where the path through each cell in the quadtree is obvious 2011-11-16T21:37:15 which you were sort of proposing before, but with more general open rectangles 2011-11-16T21:38:02 you might have up to 2N-1 nodes in that graph 2011-11-16T21:38:06 tree, whatever 2011-11-16T21:38:38 how do you figure? I'm only proposing a single level of hierarchy 2011-11-16T21:38:50 oh 2011-11-16T21:39:07 how big will the cells be in the maze? 2011-11-16T21:40:51 It really doesn't matter. The cost of stitching the supercells together is (#cells * #neighbors) 2011-11-16T21:41:14 the danger is some pathological case where the number of edges explodes 2011-11-16T21:41:15 bmh, a1k0n: you guys should make a team 2011-11-16T21:41:19 but if your supercell is full of walls bisecting it, what good is it? 2011-11-16T21:41:28 bmh wants to 2011-11-16T21:41:31 amstan: I've been trying to convince him to join my team for the past few weeks 2011-11-16T21:41:37 trouble is my team is rubbish. 2011-11-16T21:41:44 your team? lol 2011-11-16T21:41:52 you have a team already? 2011-11-16T21:42:17 amstan: yeah, I recruited a co-worker I went to college with. We once wrote a genetic alg to play Tetris... in haskell 2011-11-16T21:42:21 you'd need some efficient way to divide cells utilizing as much natural barriers (water) as possible 2011-11-16T21:43:22 although this sounds like a common problem, must have known algorithms for it 2011-11-16T21:43:40 I want Norvig's pen! 2011-11-16T21:43:45 it looks awesome! 2011-11-16T21:43:46 analyst74: I was just going to place points on a jittered grid and BFS from them 2011-11-16T21:43:50 amstan: link? 2011-11-16T21:43:58 (is it a fountain pen?) 2011-11-16T21:44:06 i has a circuit board on it: http://www.youtube.com/watch?feature=player_embedded&v=dX8VPYXd3hs#! 2011-11-16T21:44:52 oh. That's neat. 2011-11-16T21:45:40 what an apropos lecture 2011-11-16T21:45:56 a1k0n: yeah, i'm going through ai-class 2011-11-16T21:46:17 U is usually called V (value function) 2011-11-16T21:48:29 I want to see a bot that uses the Richar Feynman algorithm to win. 2011-11-16T21:48:40 bmh: details? 2011-11-16T21:48:52 integrates all possible future states out? 2011-11-16T21:49:06 amstan: 1. Write down the problem. 2. Think really hard. 3. Write down the solution. 2011-11-16T21:49:12 lol 2011-11-16T21:50:08 too bad it only works well in a deterministic and nonstochatic enviroment 2011-11-16T21:51:58 *** xathis has quit IRC () 2011-11-16T21:56:04 bmh: What happens when you look at your quad-tree and start and target are in the same square? Do you consider the other squares? 2011-11-16T21:56:08 i just said the same thing.. lol 2011-11-16T21:56:15 deterministic and fully observable* 2011-11-16T21:56:28 i sent an email a while ago to the mailing list 2011-11-16T21:56:38 about possibly getting a plug from ai-class for our contest 2011-11-16T21:56:43 mleise: then just pathfind to it normally ignoring the quad-tree 2011-11-16T21:57:11 no, i mean they are in the same 'hypersquare', like a quarter of the map 2011-11-16T21:58:09 mleise: I'm only proposing a two-level graph 2011-11-16T21:58:38 does that mean that the granularity would be 1x1 and 2x2? 2011-11-16T21:59:46 1x1, sqrt(n) x sqrt(n) for an n x n graph 2011-11-16T22:00:16 *** Antimony has joined #aichallenge 2011-11-16T22:02:12 for a map of 100x100 that means 100 10x10 cells, right? And these are straight forward rectangles on the map? 2011-11-16T22:02:56 mleise: I'm not sure about the point placement strategy (uniform, uniform random, poisson disk sampling, jittered grid) -- but I think it makes sense to construct the cells by BFS 2011-11-16T22:03:22 ok, now i get the picture ;) 2011-11-16T22:03:51 it would be cool to minimize the number of cells on edges. 2011-11-16T22:05:26 You need to have cells consisting only of connected squares, else what would happen in a maze where two points are in the same cell, but not physically connected there. 2011-11-16T22:05:48 if you constructed it by BFS, that isn't possible. 2011-11-16T22:06:05 dito, so no random sampling 2011-11-16T22:07:05 still you may end up with cell A being first and claiming an area that 'carves' out cell C 2011-11-16T22:07:34 so the shortest path inside C for some points may be through that part of A 2011-11-16T22:08:14 I'm not sure about convexity guarantees. 2011-11-16T22:08:58 -> so you need to write a thesis, proof it and come up with an algorithm to guarantee it ;) 2011-11-16T22:10:12 It does exist. If your cells are 1x1 they are all convex. So ... if you go from there you have a solid foundation, hehe 2011-11-16T22:10:29 so.. i forgot about rotational symmetries 2011-11-16T22:10:45 i guess 2 player point symmetry is the same as rotational 2011-11-16T22:10:49 but 4 player is different 2011-11-16T22:11:21 4 player ... wasn't that the axis of evil swastika stuff? 2011-11-16T22:12:40 in other words, are there actually any official maps, that use 4 player rotational symmetry? 2011-11-16T22:12:48 mleise: yeah.. that's the problem 2011-11-16T22:14:19 *** GeorgeSebastian has joined #aichallenge 2011-11-16T22:14:19 *** GeorgeSebastian has joined #aichallenge 2011-11-16T22:14:26 mleise: hmm, apparently i mislabeled it 2011-11-16T22:14:28 lol 2011-11-16T22:15:53 *** Rinum has quit IRC (Ping timeout: 265 seconds) 2011-11-16T22:16:35 At the moment I just want to finish my path finding algorithm, but I am not really motivated for anything on Ants. 2011-11-16T22:16:47 how would you label this map? http://paste.aichallenge.org/xqx7g/ 2011-11-16T22:16:58 it's not rotational 2011-11-16T22:17:05 no it is ... 2011-11-16T22:17:09 mirrored 2011-11-16T22:17:53 yeah, it's point mirror 2011-11-16T22:18:26 I don't think a point mirror exists :D 2011-11-16T22:19:14 To me it looks like one horizontal and one vertical mirror 2011-11-16T22:19:36 Yes, it has two axes of reflective symmetry 2011-11-16T22:24:29 *** wr has joined #aichallenge 2011-11-16T22:24:37 hmmm? 2011-11-16T22:25:04 so I downloaded the ai how do I get a GUI of the play? 2011-11-16T22:25:06 *** Antimony_ has joined #aichallenge 2011-11-16T22:25:41 *** wr has quit IRC (Client Quit) 2011-11-16T22:25:54 *** bmh has quit IRC (Quit: Page closed) 2011-11-16T22:27:04 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-16T22:27:16 *** Antimony_ is now known as Antimony 2011-11-16T22:29:24 I think talking about my path finding idea will motivate me to implement it, but I don't want someone else to try it out before I do ^^ 2011-11-16T22:30:48 if i were to make an irc server, i would have a filter for impacient people 2011-11-16T22:30:55 So I rather listen to others talking about their A* or BFS approaches 2011-11-16T22:31:19 the first few messages you send(if you're new), they'll be buffered, and only after like 2 minutes(if you haven't left) they'll be transmitted to the channel 2011-11-16T22:31:28 that should filter out people 2011-11-16T22:31:36 amstan: no need to redefine irc for this, make an impatience detection algorithm and get a bot to the dirty work lol 2011-11-16T22:31:37 amstan: that's cruel, but effective ;) 2011-11-16T22:31:54 aarossig: yes, but what could a bot do though? 2011-11-16T22:31:56 amstan: nvm then lol 2011-11-16T22:32:07 amstan: you could have a metachannel 2011-11-16T22:32:16 aarossig: by the time the bot gets mad at him because he signed out it's already too late 2011-11-16T22:32:22 amstan: in the words of mcstar.. the realaichallenge 2011-11-16T22:32:43 no.. just have 2 channels 2011-11-16T22:32:51 one for people with answers 2011-11-16T22:32:55 could be any random channel 2011-11-16T22:33:06 and then have a bot that relays the messages between the two 2011-11-16T22:33:15 any messages that come in the random channel are the patient people :D 2011-11-16T22:33:16 lol 2011-11-16T22:33:25 sure, but that's more sketchy than modifying an irc server 2011-11-16T22:33:36 amstan: i suppose 2011-11-16T22:33:45 as if the entire idea isn't sketchy lol 2011-11-16T22:33:56 I like the idea. It alwas strikes me as bizarre when someone joins and leaves 2 minutes after answering a question 2011-11-16T22:34:15 amstan: awesome http://www.shopenzed.com/unearthed-kauri-boxed-mens-pen-xidp233496.html 2011-11-16T22:34:23 nickjohnson: the funny part is that irc is like the absolute fastest to get answers from 2011-11-16T22:34:28 amstan: i feel like your dad would really like that for an xmas gift 2011-11-16T22:34:30 nickjohnson: short of searching yourself 2011-11-16T22:35:15 amstan: i agree, but sometimes... if you're struggling for an hour you just need someone to give you a keyword or even say something totally unrelated that helps get you back on track and solve the issue 2011-11-16T22:35:38 aarossig: yes, but that's no reason to leave after 2 minutes of silence 2011-11-16T22:35:46 agreed 2011-11-16T22:36:33 38 seconds in the last case 2011-11-16T22:36:57 amstan: i want rDNS 2011-11-16T22:37:15 aarossig: huh? 2011-11-16T22:37:33 aarossig: so that my domain resolves for whois and stuff 2011-11-16T22:37:39 amstan* lol 2011-11-16T22:38:12 aarossig: what's the connection with people leaving before they get questions answered? 2011-11-16T22:38:32 amstan: nothing at all, that topic seemed like it was rounding completion 2011-11-16T22:40:49 well, I have _extremely_ basic combat now 2011-11-16T22:41:09 thestinger: cool, what did you do? 2011-11-16T22:41:41 http://paste.aichallenge.org/mEPnX/ 2011-11-16T22:41:56 I just fixed my code for dealing with unmoving enemies 2011-11-16T22:42:04 it's really crappy even for that though 2011-11-16T22:42:46 need to get them to strafe/circle if they aren't going to attack instead of just staying there 2011-11-16T22:43:11 thestinger: have your tried mcstar's tool thing for combat simulations? 2011-11-16T22:43:12 thestinger: that looks pretty good so far 2011-11-16T22:43:29 *** Scryer has joined #aichallenge 2011-11-16T22:43:48 nope, I just implemented a simulation() function based on the battle resolution page 2011-11-16T22:44:06 so I know which ants/enemies will die (assuming enemies don't move) 2011-11-16T22:44:25 How can I tell which of my hills are still alive? Do I have to have an ant near enough to see them? 2011-11-16T22:44:32 Scryer: yes 2011-11-16T22:44:41 Thanks. 2011-11-16T22:44:53 Scryer: if not.. you can know for sure if you just had an ant spawn from one 2011-11-16T22:45:07 I rewrote a lot of my bot so that I do all regular movement and end up with a hash table with ants as keys and directions as values 2011-11-16T22:45:12 and then I change it around for combat 2011-11-16T22:45:53 i should probably eat something more 2011-11-16T22:45:58 Roger - I thought about keeping track of how much food I've collected compared to how many hills should be spawning. Seems like a nuisance. :) 2011-11-16T22:46:56 my cat was very interested in a small chromium window moving across the screen 2011-11-16T22:47:09 amstan: huh? 2011-11-16T22:47:41 thestinger: hey thats cool 2011-11-16T22:47:41 which theme did you use? :p 2011-11-16T22:53:05 *** GeorgeSebastian has quit IRC (Ping timeout: 252 seconds) 2011-11-16T22:54:44 *** TheLinker has quit IRC (Quit: Bye) 2011-11-16T22:55:02 gn guys 2011-11-16T22:55:21 *** Rinum has joined #aichallenge 2011-11-16T22:56:28 looks like my bot is finally improving... http://ants.fluxid.pl/replay.2103 ... still sucks though lol 2011-11-16T22:57:35 I've had two official games since the start of 16 Nov; just noticed that the last guy I played has had five. How is this determined? I haven't been deactivated in that time. 2011-11-16T22:57:56 i just got one game with one other person, so i'm still sub #6000 ranked 2011-11-16T22:58:43 Rinum: haha that is a big pile of ants afraid of each other 2011-11-16T22:59:16 a1k0n: lol and over 1k ants in my hill are too scared to come out 2011-11-16T22:59:22 hilarious 2011-11-16T23:00:27 and i love how there's a HUGE mass surrounding my 4 ants 2011-11-16T23:00:27 5 ants* lol 2011-11-16T23:02:09 and for the 1st time, my bot is actually winning games... wow http://ants.fluxid.pl/replay.2174 2011-11-16T23:03:27 hahah oops. i'm running bots on fluxid with huge amounts of debug output now 2011-11-16T23:03:32 *** u_ has quit IRC (Quit: u_) 2011-11-16T23:03:47 a1k0n: I ended up with a 42GB log file like that... 2011-11-16T23:03:53 whoa 2011-11-16T23:04:16 I was writing _many_ distance maps to a log file each turn 2011-11-16T23:04:24 and I left it on overnight 2011-11-16T23:04:36 damn 2011-11-16T23:04:39 48GB wow lol 2011-11-16T23:04:48 42* lol 2011-11-16T23:05:13 So nice that big disks are cheap these days. :) 2011-11-16T23:05:21 Learn anything from the trace? 2011-11-16T23:05:22 Scryer: not anymore 2011-11-16T23:06:08 Prices went up when I wasn't looking? 2011-11-16T23:12:00 my connection to the TCP server keeps screwing up 2011-11-16T23:12:01 :( 2011-11-16T23:12:42 Scryer: yep, thailand floodings 2011-11-16T23:12:49 Scryer: 180% increase in hd prices 2011-11-16T23:12:53 *** GeorgeSebastian has joined #aichallenge 2011-11-16T23:13:41 *** Raimondi has left #aichallenge 2011-11-16T23:13:49 *** Raimondi has joined #aichallenge 2011-11-16T23:14:59 just noticed pguillory added an 'about me' 2011-11-16T23:15:38 *** Yexo has quit IRC (Ping timeout: 252 seconds) 2011-11-16T23:15:44 Looks like he's reading Sun Tzu. We should worry. 2011-11-16T23:16:28 lol 2011-11-16T23:18:15 *** Yexo has joined #aichallenge 2011-11-16T23:18:30 *** maysam81 has joined #aichallenge 2011-11-16T23:19:28 I guess there actually are ppl with teams http://aichallenge.org/profile.php?user=573 2011-11-16T23:20:11 why is their bio half in german when they're from russia? 2011-11-16T23:20:44 study abroad? 2011-11-16T23:21:06 hi 2011-11-16T23:23:24 no, they aren't german students in moscow, their grammar is too bad. *g* 2011-11-16T23:23:57 maybe they thought waterloo is in germany 2011-11-16T23:24:25 mleise: lolwhat? 2011-11-16T23:27:41 or one of them is part of a german minority in russia and felt like writing some words in the other language 2011-11-16T23:27:45 You mean they're Belgians? That would explain a lot. :) 2011-11-16T23:29:39 *** twymer has quit IRC (Ping timeout: 258 seconds) 2011-11-16T23:31:08 manwe's bot rules 2011-11-16T23:31:13 http://ants.fluxid.pl/replay.2169 2011-11-16T23:32:36 it is awesome http://ants.fluxid.pl/replay.2193 2011-11-16T23:32:47 though I did a sneak attack on his hill there =P 2011-11-16T23:32:58 but then he retaliated x.x 2011-11-16T23:33:00 *** Cowbandit has joined #aichallenge 2011-11-16T23:34:12 how do you debug your bot in java? as in see what values are stored and etc. 2011-11-16T23:34:50 output w/e values to a file 2011-11-16T23:34:56 Rinum: that would have been a timeout on the real servers most definitely ^^ 2011-11-16T23:35:21 lol 2011-11-16T23:35:33 wait so how do you output those values into a file? 2011-11-16T23:35:46 because i've just been setting a breakpoint 2011-11-16T23:35:47 *** replore has quit IRC (Read error: Connection reset by peer) 2011-11-16T23:35:50 and then manually inputting coordinates and stuff 2011-11-16T23:35:51 *** replore__ has joined #aichallenge 2011-11-16T23:36:24 open a file for writing 2011-11-16T23:36:38 and write w/e to that file 2011-11-16T23:37:27 why not use replay the game in the bot with the debugger? 2011-11-16T23:37:35 -use 2011-11-16T23:37:35 well the thing is 2011-11-16T23:37:51 i'm using bluej 2011-11-16T23:37:54 and for some reason 2011-11-16T23:38:06 it has like a 75 line maximum terminal window 2011-11-16T23:38:12 so i can't paste all the bot input int 2011-11-16T23:38:17 in* all at once 2011-11-16T23:39:02 ok, then make your bot load the bot input file. (Add it as a command line parameter) 2011-11-16T23:39:22 o so how does one do that? 2011-11-16T23:39:33 lol sorry i'm a noob 2011-11-16T23:39:36 you have the bot input file? 2011-11-16T23:39:40 yeah 2011-11-16T23:39:52 ok, do you see the main() method? 2011-11-16T23:39:57 in your Java bot? 2011-11-16T23:40:02 ya 2011-11-16T23:40:32 ok the method signature should read something like main(String[] args), right? 2011-11-16T23:40:42 yeah it's that one 2011-11-16T23:40:47 still no strafing/encircling yet but I've improved combat a bit http://paste.aichallenge.org/o73GH/ :) 2011-11-16T23:40:53 and there's only one line it new MyBot().readSystemInput(); 2011-11-16T23:41:17 Cowbandit: ok, now check if args.length is != 0 2011-11-16T23:41:39 have them set to take no losses atm, but that's temporary 2011-11-16T23:41:42 in that case load the file args[0] and read input from there 2011-11-16T23:41:52 oh 2011-11-16T23:42:04 so i just name my file args? 2011-11-16T23:42:07 it would be best, if you added a MyBot().readFileInput(); 2011-11-16T23:42:30 new MyBot().readFileInput(args[0]); 2011-11-16T23:42:38 wait so how do i specify the file that I want to use? 2011-11-16T23:43:07 then when you start the debugger, add "C:\Ants\game_logs\bot0.input" or something like that. 2011-11-16T23:43:32 I *hope* BlueJ allows command-line arguments 2011-11-16T23:43:47 lol hopefully 2011-11-16T23:43:49 thanks for the help 2011-11-16T23:43:55 this should save me a lot of time 2011-11-16T23:44:43 Since file and terminal input are both InputStreams in Java, you should be able to reuse all that code in the bot class. 2011-11-16T23:44:49 *** GeorgeSebastian has quit IRC (Read error: Connection reset by peer) 2011-11-16T23:45:08 kk i'll give it a try 2011-11-16T23:45:15 good luck 2011-11-16T23:50:15 *** Rinum has quit IRC (Quit: Page closed) 2011-11-16T23:50:45 *** Cowbandit has quit IRC (Quit: Page closed) 2011-11-16T23:51:12 *** Scryer has quit IRC (Quit: Page closed) 2011-11-16T23:57:16 *** ChrisH_ has joined #aichallenge