2011-11-15T00:05:08 *** Antimony has joined #aichallenge 2011-11-15T00:05:49 *** aarossig has joined #aichallenge 2011-11-15T00:08:05 *** TheLinker has quit IRC (Quit: Bye) 2011-11-15T00:09:44 *** NightExcessive has joined #aichallenge 2011-11-15T00:12:22 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-15T00:15:23 geez. found another major bug. 2011-11-15T00:16:31 oh? 2011-11-15T00:17:01 I found a hilarious bug with my gradient follower -- if it can't see food it will run into a corner 2011-11-15T00:17:53 yeah, some of my ants in combat are doing random things, literally 2011-11-15T00:18:04 heh, nice. i notied that behavior 2011-11-15T00:18:08 Why did you choose 3x7 for brute force? 2011-11-15T00:18:18 because it was tractable 2011-11-15T00:18:24 also solved 4x7 2011-11-15T00:18:37 just as an experiment, mind. water squares matter a lot. 2011-11-15T00:18:51 must go read to my son 2011-11-15T00:18:54 ah, you're using an empty 4x7! 2011-11-15T00:19:07 a1k0n: Teach your son to read and he'll read to himself ;) 2011-11-15T00:19:54 I realized last night just as I was falling asleep that an optimized distance calculation in my combat failed to take water into account 2011-11-15T00:20:02 gotta fix that 2011-11-15T00:23:02 is anyone actually using a genetic alg? 2011-11-15T00:25:55 *** McLeopold has joined #aichallenge 2011-11-15T00:27:25 bmh: I asked about that and neural network and ppl replied that it wasn't a good solution for this challenge 2011-11-15T00:27:34 grom358: it's a *terrible* idea 2011-11-15T00:28:05 why though? 2011-11-15T00:28:11 amstan: hello, you here? 2011-11-15T00:28:31 grom358: Genetic algorithms tend to be the worst solution to any problem :) 2011-11-15T00:28:51 They become easily stuck at local optima. 2011-11-15T00:31:51 @later tell janzert We should fix the intentional suicide thing by deducting points for players that timeout or crash, any objections? 2011-11-15T00:31:51 McLeopold: Job's done. 2011-11-15T00:32:30 how can u tell a plyer has timeout or crashed though? 2011-11-15T00:32:54 by the fact none of their ants have moved for X turns? 2011-11-15T00:33:00 grom358: we have timers checking for a timeout and a dead process to indicate crashes 2011-11-15T00:33:30 McLeopold: have the worker woes passed? 2011-11-15T00:33:32 it's all in the engine.py code 2011-11-15T00:33:43 bmh: not sure, I've been out for a week 2011-11-15T00:35:05 does that make you the mysterious Jeff? 2011-11-15T00:35:20 Much less mysterious :) 2011-11-15T00:35:25 BenJackson: I don't think so, whois says his name is "Scott" 2011-11-15T00:35:46 *** Antimony has joined #aichallenge 2011-11-15T00:36:31 that genetic algorithm thing reminded me of how I feel in the forums 2011-11-15T00:36:41 BenJackson: stuck? 2011-11-15T00:36:44 oh, you need help with X? I'm torn between helping you with X and telling you X is a terrible idea 2011-11-15T00:36:55 oh, just tell 'em. 2011-11-15T00:37:04 hell no, they might make better bots ;-) 2011-11-15T00:37:10 fair 2011-11-15T00:37:23 but practically speaking everyone probably needs to work through a few ideas to really understand 2011-11-15T00:37:38 heck, my first afternoon I implemented A* in python 2011-11-15T00:37:57 I took two weeks to get diffusion working in Haskell. 2011-11-15T00:37:59 there's probably more long-term benefit in learning how A* works than in winning the contest :) 2011-11-15T00:38:21 Enlightenment occurs when you realize that Dijkstra's algorithm is a special case of A*. 2011-11-15T00:38:53 heck, if all people did was implement pathfinding with BFS, A*, Dijkstra, Bellman-Ford, etc 2011-11-15T00:39:00 they'd learn more than most people do in an intro ai class :) 2011-11-15T00:39:25 well 2011-11-15T00:39:30 different things 2011-11-15T00:39:32 BenJackson: but then they'll increase the supply of competent programmers and depress wages :( 2011-11-15T00:39:33 *** capa has quit IRC () 2011-11-15T00:39:46 I doubt it bmh 2011-11-15T00:39:59 analyst74: I am not serious in the least. 2011-11-15T00:40:07 bmh: mm, it's true I would not have so much stackoverflow rep if I applied my aichallenge forum theory there :) 2011-11-15T00:40:42 it's the amount of imcompetent programmers that depress wages, I think 2011-11-15T00:41:12 that probably includes me. :) 2011-11-15T00:42:57 analyst74: terrible programmers give everyone else an endless supply of tasks. 2011-11-15T00:43:42 I really do need to make my ants care about the hill 2011-11-15T00:43:54 They forage like mad and then someone walks over and kills the hill. 2011-11-15T00:44:47 in a recent game where I got squished there was a moment where I was defending against all 3 other players 2011-11-15T00:44:56 wow 2011-11-15T00:44:59 and some (maybe all) of them missed opportunities to take my hill 2011-11-15T00:45:16 do you handle opponents as individuals or as a mass of ants? 2011-11-15T00:45:19 http://aichallenge.org/visualizer.php?game=94285&user=3473 2011-11-15T00:45:23 analyst74: https://secure.wikimedia.org/wikipedia/en/wiki/Dunning–Kruger_effect :) 2011-11-15T00:45:25 drag the slider to right before white's hill is taken 2011-11-15T00:46:10 BenJackson: rad 2011-11-15T00:46:29 it's kind of interesting how I lost that 2011-11-15T00:46:43 thestinger, that's such a mind blowing theory 2011-11-15T00:46:45 purple and blu happened to wall off such that yellow and orange were focused on me 2011-11-15T00:47:00 *** boegel has joined #aichallenge 2011-11-15T00:47:00 it basically tells you, whatever you feel about your skill level, is wrong 2011-11-15T00:48:15 *** rioy has joined #aichallenge 2011-11-15T00:48:24 hi 2011-11-15T00:48:36 anybody? 2011-11-15T00:48:52 .. 2011-11-15T00:48:53 and btw, strcat, you are such a game disruptor on tcp 2011-11-15T00:48:58 rioy: this is really not the way to join a channel 2011-11-15T00:49:07 oh 2011-11-15T00:49:21 oh...first tine to use this 2011-11-15T00:49:29 analyst74: I've decided my only hope is to be completely aggressive :P, that might change once I get some of this combat stuff done 2011-11-15T00:49:42 rioy: what's up? 2011-11-15T00:50:55 great. trying to figure out hiw to use this .. 2011-11-15T00:51:12 *** boegel has quit IRC (Client Quit) 2011-11-15T00:51:17 rioy: there's lot of people in here, some busy, so if you don't get a response right away don't panick 2011-11-15T00:51:46 @amstan is this a way to speak to someone? 2011-11-15T00:51:47 rioy: An error occured while trying to show the previous error. 2011-11-15T00:52:03 rioy: sure.. except the @ character, the bot will respond to those 2011-11-15T00:52:09 sure! 2011-11-15T00:52:29 thestinger, you're like the reverse sparta, one word, go! 2011-11-15T00:52:44 amstan ....oh so just include the name? 2011-11-15T00:53:01 heh, one of my ants hit the food jackpot in this game: http://ants.fluxid.pl/replay.35743 @ 396 around row 9 col 56 2011-11-15T00:53:04 rioy, that is one convention 2011-11-15T00:53:07 yes, most irc clients have a feature where you get notified if someone mentions your name 2011-11-15T00:53:09 9 food in a little square o_o 2011-11-15T00:53:26 rioy: http://www.irchelp.org/irchelp/ircprimer.html 2011-11-15T00:53:34 *** Palmik has joined #aichallenge 2011-11-15T00:53:44 Finally given up on my principles, and implemented a time-remaining check 2011-11-15T00:53:59 matthewd: I haven't resorted to that yet 2011-11-15T00:54:11 bmh: And I'm in C++ :/ 2011-11-15T00:54:20 matthewd: I'm in Haskell. I'm playing with fire. 2011-11-15T00:54:24 thanks bmh 2011-11-15T00:54:46 *** NightExcessive has quit IRC (Quit: Page closed) 2011-11-15T00:55:00 (that primer is a solid 15 years out of date) 2011-11-15T00:55:01 thanks amstan 2011-11-15T00:55:06 rioy: another thing.. if you have a question, don't ask to ask, just ask 2011-11-15T00:55:14 rioy: other than that, welcome to the channel 2011-11-15T00:55:26 thestinger: wow purple made you pay for the hill right after that 2011-11-15T00:56:17 rioy: sorry for being short earlier, it's just that some people come in, and expect and answer inside a few seconds or else they leave 2011-11-15T00:56:27 amstan: ok! 2011-11-15T00:56:27 I feel like a cruel dictator with an army of expendable soldiers :) 2011-11-15T00:56:50 amstan: im just a newbie here 2011-11-15T00:57:17 amstan gotta spend sometime here~ 2011-11-15T00:57:29 rioy: good behavior on IRC is essential to remaining on IRC 2011-11-15T00:57:31 rioy: are you working on a bot? 2011-11-15T00:57:46 (or trolling) 2011-11-15T00:58:29 amstan yes...just started, trying to implement a*, lol 2011-11-15T00:58:34 my first IRC experience was in 1989 using a client on a VMS system. 22 years later I still suck at IRC 2011-11-15T00:58:39 *** Antimony has quit IRC (Ping timeout: 248 seconds) 2011-11-15T00:58:43 here I am, for example, using freenode's web client 2011-11-15T00:59:02 BenJackson: did you use to telnet into irc servers? as far as i can tell the protocol is designed to be human usable 2011-11-15T00:59:17 BenJackson: the web client is great. I think I've been using IRC for about 11 years now. 2011-11-15T00:59:25 *** Antimony has joined #aichallenge 2011-11-15T00:59:25 hm, no, I think VMS had a real client 2011-11-15T00:59:35 it also had "talk" 2011-11-15T00:59:54 hmm, i should test this 2011-11-15T01:00:05 is it okay to share snippets of code for this challenge? Like I was gonna write an article about BFS and A* in terms of using it in Ants 2011-11-15T01:00:25 grom358: sure, just try to not spoil the challenge of writing a bot 2011-11-15T01:00:36 grom358: so don't offer full working bots, just instructions on how to implement parts 2011-11-15T01:00:58 it would be fun to have a second contest where instead of starting with "starter kits" we start with a library of everyone's best algorithms 2011-11-15T01:02:05 ha -- bring your own glue. It would be like real life. 2011-11-15T01:02:23 *** rioy has quit IRC (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )) 2011-11-15T01:03:02 yeah I not gonna show working bot. Just some pieces on path finding. Upto the reader to integrate it into a bot. 2011-11-15T01:03:18 *** amstan_ has joined #aichallenge 2011-11-15T01:03:31 maybe just even pseudo code at that 2011-11-15T01:03:34 after I wrote a BFS into one algorithm I wrote a C++ BFS iterator 2011-11-15T01:03:40 once you do that you start bfs'ing the shit out of everything 2011-11-15T01:03:42 so, pros and cons 2011-11-15T01:04:49 so.. i joined using my telnet client 2011-11-15T01:04:53 but idk how to send messages, lol 2011-11-15T01:05:05 test 2011-11-15T01:05:07 use /msg 2011-11-15T01:05:14 BenJackson: there's no /msg in telnet 2011-11-15T01:05:18 hmm 2011-11-15T01:05:23 i had to do PRIVMSG #aichallenge test 2011-11-15T01:05:25 for that 2011-11-15T01:05:27 this is fun 2011-11-15T01:05:31 amstan: In telnet no one can hear your scream. 2011-11-15T01:05:45 yes 2011-11-15T01:05:55 PRIVMSG 2011-11-15T01:06:00 haha.. yeah raw IRC.. build your own IRC client lol 2011-11-15T01:06:20 "longer 2011-11-15T01:06:30 for some reason it only sends the first word 2011-11-15T01:06:38 now do FTP! 2011-11-15T01:06:43 BenJackson: did that 2011-11-15T01:07:06 i also did pop and http 2011-11-15T01:07:17 some pop servers are really scum, they kick you when you do typos 2011-11-15T01:07:20 no way to backspace 2011-11-15T01:07:24 if you want to up your game try IMAP 2011-11-15T01:07:25 its nice to learn the nuts and bolts sometimes. I sometimes feel like putting together a basic OS that runs on virtualbox 2011-11-15T01:07:30 sorry, not pop, SMTP 2011-11-15T01:08:12 if you don't break out the IMAP docs there's one really confusing bit: 2011-11-15T01:08:26 every command has to be preceded by a tag (arbitrary string) which the server will use with all replies to that command 2011-11-15T01:08:41 irc is really easy on low level, you can even do help and the server will respond with a list of commands available 2011-11-15T01:08:44 amstan: you need to do privmsg #aichallenge :hi there 2011-11-15T01:08:49 with the colon 2011-11-15T01:08:51 so unlike SMTP where "help" will get you going, IMAP will just bitch at you unti you figure out it's at least 'x help' 2011-11-15T01:09:31 hi there, needs a space before the column as well 2011-11-15T01:09:46 no space after 2011-11-15T01:09:51 right. 2011-11-15T01:09:59 like i had it :) 2011-11-15T01:10:10 my bad 2011-11-15T01:10:13 irc is an odd protocol. 2011-11-15T01:10:20 a1k0n: finished reading "Go the F**k to Sleep"? ;_ 2011-11-15T01:10:23 er, ;) 2011-11-15T01:10:24 this is exausting, i'm out 2011-11-15T01:10:34 *** amstan_ has left #aichallenge ("bye") 2011-11-15T01:10:52 well that was educational, lol 2011-11-15T01:10:54 heh. if only. we're currently at the "you're not thirsty, that's bullshit" stage 2011-11-15T01:11:22 two of my friends have a kid who always stalls at bedtime asking for a snack 2011-11-15T01:11:30 even if he just ate, or just refused to finish dinner 2011-11-15T01:11:38 ..and they give him the snack 2011-11-15T01:11:45 I'm amazed he's not fat 2011-11-15T01:11:45 yeah, there's always some excuse 2011-11-15T01:11:52 oh niec 2011-11-15T01:11:57 I blame child services 2011-11-15T01:13:05 a1k0n: http://tron.aichallenge.org/visualizer.php?game_id=4146300 why did you bot screw up in this game? 2011-11-15T01:14:17 usually stuff like that would not be replicable.. it was limited cpu resources 2011-11-15T01:15:30 I've been in front of a computer for 12 hours. Enough of that 2011-11-15T01:15:36 *** bmh has quit IRC (Quit: Hello reality.) 2011-11-15T01:15:57 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-11-15T01:17:10 sigh.. here i am in a stituation where people want to throw money at me and my goals, yet i can't accept them because of technicalities: http://forums.aichallenge.org/viewtopic.php?f=25&t=1880&start=10 2011-11-15T01:17:56 you could perhaps comment on the technicalities 2011-11-15T01:18:01 we're all wondering 2011-11-15T01:18:06 what i mean by that is that i have to wait for jeff to get the ball rolling 2011-11-15T01:18:27 and jeff is omninonpresent 2011-11-15T01:18:51 a1k0n: you mean the opposite? 2011-11-15T01:19:23 hence the non 2011-11-15T01:19:46 darn wikipedia and autocorrection 2011-11-15T01:19:53 lol 2011-11-15T01:20:07 anyway, i'm going to bed 2011-11-15T01:20:10 omniabsent 2011-11-15T01:20:18 amstan: no! wait! take my m oney! 2011-11-15T01:20:23 jk 2011-11-15T01:20:33 no seriously, take my money 2011-11-15T01:20:49 i should have a meeting with janzert first 2011-11-15T01:20:56 also.. anybody know anything about mcleopold? 2011-11-15T01:21:03 he was here earlier 2011-11-15T01:21:08 I asked if he was the mysterious Jeff 2011-11-15T01:21:09 contestbot: seen mcleopold 2011-11-15T01:21:09 amstan: mcleopold was last seen in #aichallenge 45 minutes and 49 seconds ago: Much less mysterious :) 2011-11-15T01:21:16 aaand that was his response 2011-11-15T01:21:26 I think he told the bot to tell someone something 2011-11-15T01:21:26 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-11-15T01:21:27 he wasn't here for a whole week 2011-11-15T01:21:44 [21:31] @later tell janzert We should fix the intentional suicide thing by deducting points for players that timeout or crash, any objections? 2011-11-15T01:21:52 amstan: he said that, which is how I deduced he was Jeff 2011-11-15T01:22:06 mcleopold!=jeff 2011-11-15T01:22:09 oh 2011-11-15T01:22:09 jeff=j3cameron 2011-11-15T01:22:29 mcleopold's a cool guy that developed a bunch of the code from this contest 2011-11-15T01:22:48 I remember seeing his name while cursing at the symmetry code 2011-11-15T01:22:57 BenJackson: cursing? why? 2011-11-15T01:23:17 I didn't correctly understand what it was trying to do 2011-11-15T01:23:30 BenJackson: need some help with that? i'm working on a new mapgen, so i have a pretty good understanding of how it works 2011-11-15T01:23:32 so trying to figure THAT out just based on what it DID was not easy 2011-11-15T01:23:44 no, I get it now 2011-11-15T01:23:53 map symmetry is not directly related to hill symmetry 2011-11-15T01:23:53 *** ChrisH__ has joined #aichallenge 2011-11-15T01:24:04 it kinda is.. 2011-11-15T01:24:13 well, yes, and get_map_symmetry is how it kinda is 2011-11-15T01:24:15 which took a while to figure out 2011-11-15T01:24:24 look at this: https://github.com/aichallenge/aichallenge/tree/amstan_mapgen/ants/mapgen 2011-11-15T01:24:33 it incorporated all different kinds of symmetries 2011-11-15T01:24:45 btw I did figure out one bit of information the mapgen leaks 2011-11-15T01:25:04 I hadn't figured out yet if it was possible to avoid so I wasn't sure I should rely on it :) 2011-11-15T01:25:07 i have a function for each which a returns a list of points symmetric to an original point 2011-11-15T01:25:25 what info? 2011-11-15T01:25:48 rows and cols must be divisible by #hills 2011-11-15T01:25:56 for some sizes that's useless 2011-11-15T01:25:56 there's no other way 2011-11-15T01:26:03 but the 3, 7 player maps are pretty obvious 2011-11-15T01:26:28 the mapgen should probably work more factors into the size 2011-11-15T01:26:30 don't you get the ammount of players you battle against anyway? 2011-11-15T01:26:35 no 2011-11-15T01:26:46 tricky 2011-11-15T01:26:47 I asked that on the forum, even, because the C++ State gave me the idea it did 2011-11-15T01:27:06 I realized it when I wanted to compute averge hill distance for some heuristic 2011-11-15T01:27:20 yeah, the issue with the current mapgen is that it generates size based on num_players 2011-11-15T01:27:44 I did look at the mapgen and I don't understand exactly how those symmetries work either 2011-11-15T01:27:59 my new mapgen lets you use any sizes, but if you use translational_symmetry maps it'll check the divisibility thing 2011-11-15T01:28:06 but I did figure out the "easiest" way that the game gives up information about symmetry 2011-11-15T01:28:45 and also keep in mind that this is only useful for translational, you can have other symmetries which don't have that problem 2011-11-15T01:29:34 but they could have other problems, like they must be square 2011-11-15T01:29:49 see all the SymmetryExceptions in https://github.com/aichallenge/aichallenge/tree/amstan_mapgen/ants/mapgen/symmetricmap.py 2011-11-15T01:29:51 yeah, I ended up sketching a bunch of tiny maps 2011-11-15T01:30:07 *** JorgeB has joined #aichallenge 2011-11-15T01:30:38 like so many elements of this challenge I finally realized I was going about it the wrong way 2011-11-15T01:31:23 the ants game is a good challenge 2011-11-15T01:31:29 it's simple to describe but there are many subtleties 2011-11-15T01:31:31 *** datachomper has joined #aichallenge 2011-11-15T01:32:11 i like these kinds of things, where it was designed to be simple, but you end up with so many things that the designer didn't even think about 2011-11-15T01:32:26 like minecraft's usage of booster cart glitch 2011-11-15T01:32:34 fuckers broke my mine 2011-11-15T01:32:54 when it was "fixed"? lol 2011-11-15T01:33:16 Yeah, i had a good station with those, it's now all deserted 2011-11-15T01:33:25 the main thing that made me abandon that map was the change in the terrain gen 2011-11-15T01:33:32 yep, same 2011-11-15T01:33:37 there was a nearby ungenerated square and when I generated it I got ugly "cliffs" 2011-11-15T01:33:51 anyway, i should really go to bed, 1:30 am here 2011-11-15T01:33:56 night 2011-11-15T01:33:57 cya 2011-11-15T01:35:51 *** retybok_ has joined #aichallenge 2011-11-15T01:38:06 *** Raimondi has quit IRC (Quit: Get MacIrssi - http://www.sysctl.co.uk/projects/macirssi/) 2011-11-15T01:38:54 *** grom358 has quit IRC (Quit: Ex-Chat) 2011-11-15T01:39:52 *** treeform_ has quit IRC (Remote host closed the connection) 2011-11-15T01:42:26 *** ChrisH__ has quit IRC (Quit: Page closed) 2011-11-15T01:42:54 *** Raimondi has joined #aichallenge 2011-11-15T01:46:17 *** retybok_ has quit IRC (Ping timeout: 245 seconds) 2011-11-15T02:05:35 *** Jak_o_Shadows has quit IRC (Read error: Connection reset by peer) 2011-11-15T02:09:40 *** pyDude has quit IRC (Read error: Connection reset by peer) 2011-11-15T02:12:38 *** Flort has quit IRC (Read error: Operation timed out) 2011-11-15T02:13:25 *** treeform has joined #aichallenge 2011-11-15T02:18:17 *** treeform has quit IRC (Remote host closed the connection) 2011-11-15T02:25:35 *** dr- has joined #aichallenge 2011-11-15T02:28:14 *** Flort has joined #aichallenge 2011-11-15T02:34:34 *** djr_ has quit IRC (Quit: Leaving) 2011-11-15T02:34:57 *** pairofdice has joined #aichallenge 2011-11-15T02:41:41 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 260 seconds) 2011-11-15T02:45:34 *** Cyb has joined #aichallenge 2011-11-15T02:59:25 *** Cyb has quit IRC (Ping timeout: 240 seconds) 2011-11-15T03:07:53 *** ikaros has joined #aichallenge 2011-11-15T03:19:12 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-15T03:28:21 *** kilae has joined #aichallenge 2011-11-15T03:34:55 *** mj41 has joined #aichallenge 2011-11-15T03:38:56 *** sigh has joined #aichallenge 2011-11-15T03:39:32 *** Blkt has joined #aichallenge 2011-11-15T03:40:59 good morning everyone 2011-11-15T03:42:02 good evening 2011-11-15T03:53:08 morning 2011-11-15T04:05:00 *** delt0r__ has joined #aichallenge 2011-11-15T04:05:45 *** delt0r_ has quit IRC (Read error: Operation timed out) 2011-11-15T04:06:55 *** liberforce has joined #aichallenge 2011-11-15T04:07:09 *** McLeopold has quit IRC (Read error: Connection reset by peer) 2011-11-15T04:08:55 *** McLeopold has joined #aichallenge 2011-11-15T04:09:03 *** amstan has quit IRC (Ping timeout: 248 seconds) 2011-11-15T04:09:37 *** iris1 has joined #aichallenge 2011-11-15T04:16:59 *** Raimondi has quit IRC (Read error: Connection reset by peer) 2011-11-15T04:17:04 *** Raimondi has joined #aichallenge 2011-11-15T04:18:04 *** Jak_o_Shadows has joined #aichallenge 2011-11-15T04:18:24 *** Surya has joined #aichallenge 2011-11-15T04:18:56 *** analyst74 has quit IRC (Read error: Connection reset by peer) 2011-11-15T04:31:06 *** analyst74 has joined #aichallenge 2011-11-15T04:37:16 *** Palmik_ has joined #aichallenge 2011-11-15T04:37:28 *** Palmik has quit IRC (Ping timeout: 240 seconds) 2011-11-15T04:39:13 Ahh, new set of videos on #aiclass 2011-11-15T04:39:42 *** Areks has joined #aichallenge 2011-11-15T04:42:24 *** aerique has joined #aichallenge 2011-11-15T04:51:02 *** replore_ has quit IRC (Remote host closed the connection) 2011-11-15T04:57:20 *** mikewintermute has joined #aichallenge 2011-11-15T05:08:08 *** marijnfs has joined #aichallenge 2011-11-15T05:12:28 *** Rufina has joined #aichallenge 2011-11-15T05:13:50 *** mviel_ has joined #aichallenge 2011-11-15T05:13:54 *** grwip has joined #aichallenge 2011-11-15T05:17:45 *** Mankarse has joined #aichallenge 2011-11-15T05:17:50 *** mviel__ has joined #aichallenge 2011-11-15T05:17:56 *** mviel has quit IRC (Ping timeout: 276 seconds) 2011-11-15T05:19:02 *** Rufina has quit IRC (Quit: Page closed) 2011-11-15T05:20:30 *** mviel_ has quit IRC (Ping timeout: 260 seconds) 2011-11-15T05:22:56 Hey guys, I have a problem with the IssueOrder (not always interpreted) details here http://forums.aichallenge.org/viewtopic.php?f=25&t=1902 2011-11-15T05:23:11 Thank you for your attention! 2011-11-15T05:23:21 *** g0llum has joined #aichallenge 2011-11-15T05:26:40 *** Anilm3 has joined #aichallenge 2011-11-15T05:27:00 Hi guys 2011-11-15T05:28:25 *** Piero has joined #aichallenge 2011-11-15T05:28:51 *** Piero is now known as Guest98993 2011-11-15T05:30:31 *** hargrow has joined #aichallenge 2011-11-15T05:31:00 *** Guest98993 has quit IRC (Client Quit) 2011-11-15T05:32:41 *** b0rder_ has quit IRC (Quit: 离开) 2011-11-15T05:34:58 *** zaphus has joined #aichallenge 2011-11-15T05:35:50 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-15T05:38:10 *** HaraKiri has joined #aichallenge 2011-11-15T05:39:00 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-15T05:40:20 did anyone here get the visualiser overlay from the forums to work? 2011-11-15T05:40:40 *** g0llum has joined #aichallenge 2011-11-15T05:41:56 Didn't even find out what I was supposed to dl from git 2011-11-15T05:42:29 the new ants.py and the new /visualozer/ stuff 2011-11-15T05:44:31 I'll try it once I'm working on my bot again 2011-11-15T05:46:52 Harakiri, I just grabbed it - I think it works... although my bot was acting weirdly so it's hard to say! 2011-11-15T05:47:52 HaraKiri: zaphus: could you share a link? 2011-11-15T05:48:09 humm okay. how did you make it work? :D 2011-11-15T05:48:12 http://forums.aichallenge.org/viewtopic.php?f=25&t=1633 2011-11-15T05:51:11 *** Conorach has joined #aichallenge 2011-11-15T05:57:01 All I did was copy the files that you need - which is almost none of them... (checking now) 2011-11-15T05:57:14 in visualiser/data/img there's a new image 2011-11-15T05:57:21 and do your drawing commands actually appear on the visualizer? 2011-11-15T05:58:16 alot of the files in visualizer/js/ were changed 2011-11-15T05:58:16 in visualiser/js/ there are three changed files application, canvaselement and const 2011-11-15T05:58:43 those are the ONLY files I took, replacing the starter tools files of the same names 2011-11-15T05:58:58 and yes, I get drawing (once I got the format correct) 2011-11-15T05:59:05 what language? 2011-11-15T05:59:38 C++, windows 2011-11-15T06:00:09 language shouldnt matter, but it's important to get the commands right! 2011-11-15T06:02:16 theyre not that hard :x 2011-11-15T06:02:19 Note: in the latest version it is off by default, you need to click the new icon 2011-11-15T06:02:44 Harikiri, no - not hard, except if you read the start of the forum rather than the later parts - the format changed 2011-11-15T06:03:35 omg, that was it 2011-11-15T06:03:39 =) 2011-11-15T06:03:47 why would you possible turn that thing off by default haha :D thanks alot 2011-11-15T06:04:35 yeah, off by default is odd - but I must admit after a while I dont want to see it until something odd happens 2011-11-15T06:05:13 yeah maybe 2011-11-15T06:23:30 *** luizribeiro has quit IRC (Ping timeout: 260 seconds) 2011-11-15T06:24:49 *** Mankarse has quit IRC (Quit: Page closed) 2011-11-15T06:26:59 *** ALplus has quit IRC (Ping timeout: 244 seconds) 2011-11-15T06:29:26 *** kaemo has joined #aichallenge 2011-11-15T06:31:06 *** hargrow has quit IRC (Quit: Page closed) 2011-11-15T06:32:25 One in the head makes the dead stay dead! 2011-11-15T06:33:01 AI class is awesome, even though most of it isn't applicable to ants 2011-11-15T06:34:34 *** luizribeiro has joined #aichallenge 2011-11-15T06:35:43 :) 2011-11-15T06:35:59 Japanese classes are pretty awesome too, but what's the point? :p 2011-11-15T06:37:26 Do I ever have a point? 2011-11-15T06:38:15 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-15T06:38:40 :) 2011-11-15T06:39:52 And kanji are pain in the *beep* 2011-11-15T06:40:27 they sure are :/ But nothing you can't overcome with proper practice :] 2011-11-15T06:40:53 we learned them at the rate of 2 a day when I was in Japan, but I don't think i can remember more than 30 or so :/ 2011-11-15T06:40:55 the basic ones. 2011-11-15T06:41:17 I was writing them over and over and over and over :/ 2011-11-15T06:41:25 every free minute :p 2011-11-15T06:41:57 Yeah, they need constant refreshing 2011-11-15T06:43:05 But there's only couple thousand you need to know! 2011-11-15T06:45:35 yeah like 2000 or so to make you considered literate :P Although they say that nowadays the newspapers can be read with knowing only about 800. 2011-11-15T06:53:14 painful truth :) 2011-11-15T07:02:19 *** luizribeiro has quit IRC (Ping timeout: 256 seconds) 2011-11-15T07:03:48 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-11-15T07:05:19 *** patrisk has joined #aichallenge 2011-11-15T07:09:32 *** analyst74 has quit IRC (Ping timeout: 260 seconds) 2011-11-15T07:09:56 *** Kingpin13 has joined #aichallenge 2011-11-15T07:10:56 *** analyst74 has joined #aichallenge 2011-11-15T07:11:19 Anyone using collaborative diffusion who may share a replay? 2011-11-15T07:18:41 Anilm3: lots, but they are mostly painful :) 2011-11-15T07:19:01 let me see! 2011-11-15T07:19:36 Anilm3: http://aichallenge.org/visualizer.php?game=95772&user=10822&turn=162 , http://aichallenge.org/visualizer.php?game=94888&user=10822&turn=41 2011-11-15T07:21:45 Anilm3: oh wait, I overlooked the 'collaborative' part, that's not what I'm doing, sorry :) 2011-11-15T07:22:09 *** xathis has joined #aichallenge 2011-11-15T07:22:33 hahaha 2011-11-15T07:22:54 and what are you doing? 2011-11-15T07:25:05 Anilm3: I'm lazy, so it's simple. I combine different diffusion maps for enemy ants (negative), enemy hills (positive), enemy ants within range N of own hill (positive) and food (positive) , with different decays and starting values. Individual ants maximize the value or, if all equal, move towards the least recently visited square 2011-11-15T07:25:11 wow, it fits in 1 sentence :D 2011-11-15T07:26:57 so that's why your ants always scape from enemy antes even in superiority 2011-11-15T07:27:10 Anilm3: yes, good observation 2011-11-15T07:27:20 *** luizribeiro has joined #aichallenge 2011-11-15T07:27:31 your approach is similar to collaborative diffusion 2011-11-15T07:27:33 Anilm3: the fix is to assign enemy_ants in the vicinity of a larger group of my own ants a positive value instead of a negative one 2011-11-15T07:28:07 Anilm3: this would be very similar to my approach for defending my own hills (also a positive weight for closeby enemy ants) 2011-11-15T07:28:50 or you can take into account your own ants in order counter the decay values 2011-11-15T07:28:55 s/fix/hack/ 2011-11-15T07:29:02 in order to counter* 2011-11-15T07:29:41 Anilm3: yes, that is also possible, but that will make them cluster together more and thus decrease the amount of exploration 2011-11-15T07:30:03 not sure if I understood you correctly though... 2011-11-15T07:30:12 I think you did 2011-11-15T07:31:01 *** TheLinker has joined #aichallenge 2011-11-15T07:31:38 does someone know what kind of approach is xathis using? I looked at some of his replay yesterday and it's really neat, his ants try to control the map, minimizing blind spots in the controlled area, I thought maybe he may be using collab diffusion 2011-11-15T07:32:55 grwip: i've never looked into diffusion, just using a lot of bfs searches 2011-11-15T07:33:12 xathis; hey hi, first time I see you here 2011-11-15T07:34:18 Anilm3: I can't do heavy searching in state-spaces though as I'm using ruby and try to keep it below 100 lines or so :) 2011-11-15T07:34:36 digger3: I see 2011-11-15T07:34:38 xathis: congrats on your bot it's doing pretty well, so not even A* just plain bfs... I guess it doesn't always help to use complex stuff 2011-11-15T07:35:27 grwip: oh i'm using A* as well, but for most of the things i do bfs is enough or even better 2011-11-15T07:35:35 xathis: your idea is to control territory right? I see some of your ants staying around the same place, it looks like they are trying to maximize the visible area 2011-11-15T07:35:42 xathis: oh ok 2011-11-15T07:37:19 grip: in the araes that are safe, i try to have ants on positions so that i can see every tile, all the other ants are sent to explore/fight 2011-11-15T07:37:28 *areas 2011-11-15T07:37:46 good to know you aren't using any strange tactic 2011-11-15T07:38:31 xathis: would you mind saying a few words about how you do battle resolution? no spilling of actual secrets necessary, i would just like to know what the current top bot is using (hand coded logic? alpha/beta? constraint satisfaction? something else entirely?) 2011-11-15T07:39:38 bqf: i'm looking one turn ahead for a group of ants and look at all the possible move states 2011-11-15T07:40:13 bqf: and with group i mean a group of both my ants and enemy ants 2011-11-15T07:41:03 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-11-15T07:41:04 xathis: trully impressive that your only submission is the top 1 2011-11-15T07:41:06 xathis: you can't do that when there are too much of them right? you would have 4^n positions to evaluate... I used a similar approach for battle resolution but switch to monte carlo when there are too much ants... 2011-11-15T07:41:16 xathis: oh wow, from your bot's behavior i was assuming you were doing way more than that. thanks for sharing :) 2011-11-15T07:42:03 Anilm3: thanks, but i took part in the beta phase, so i'Ve had a few submissions that are not displayed 2011-11-15T07:42:35 xathis: oh, not less impressive though :P 2011-11-15T07:42:59 grwip: yeah that's the problem, you can't do larger groups very easy 2011-11-15T07:44:03 *** ikaros has joined #aichallenge 2011-11-15T07:45:25 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 240 seconds) 2011-11-15T07:49:13 *** ALplus has joined #aichallenge 2011-11-15T07:53:34 *** aibot has joined #aichallenge 2011-11-15T07:56:32 *** foRei has joined #aichallenge 2011-11-15T07:59:54 *** marijnfs has quit IRC (Quit: leaving) 2011-11-15T08:09:27 *** aibot has quit IRC (Quit: Page closed) 2011-11-15T08:12:41 *** avdg1 has joined #aichallenge 2011-11-15T08:15:16 *** patrisk has quit IRC (Quit: Leaving) 2011-11-15T08:15:59 *** avdg has quit IRC (Ping timeout: 258 seconds) 2011-11-15T08:20:13 *** ran has joined #aichallenge 2011-11-15T08:20:37 Hi, anyone knows how to easily debug my C# ant ? 2011-11-15T08:24:36 *** Antimony has joined #aichallenge 2011-11-15T08:26:44 *** u_ has joined #aichallenge 2011-11-15T08:26:51 *** zaphus has quit IRC (Quit: Page closed) 2011-11-15T08:27:13 *** u__ has joined #aichallenge 2011-11-15T08:30:19 You can use the new visualizer to ease your pain 2011-11-15T08:30:59 *** u_ has quit IRC (Ping timeout: 244 seconds) 2011-11-15T08:30:59 *** u__ is now known as u_ 2011-11-15T08:31:02 ran: I haven't used it yet, but it looks promising 2011-11-15T08:31:34 What is the new visualizer ? 2011-11-15T08:32:00 http://forums.aichallenge.org/viewtopic.php?f=25&t=1633 2011-11-15T08:34:20 I'll take a look, thanks! 2011-11-15T08:36:07 *** mcstar has joined #aichallenge 2011-11-15T08:37:29 *** Conorach has quit IRC (Ping timeout: 276 seconds) 2011-11-15T08:38:06 *** twymer has joined #aichallenge 2011-11-15T08:39:03 *** Conorach has joined #aichallenge 2011-11-15T08:40:05 *** Antimony has quit IRC (Ping timeout: 276 seconds) 2011-11-15T08:44:42 *** Antimony has joined #aichallenge 2011-11-15T08:49:02 *** ran has quit IRC (Quit: Page closed) 2011-11-15T08:55:39 *** bergmark has joined #aichallenge 2011-11-15T08:57:09 is it possible to get any error or debugging data from the server? have a bot that does really well locally and on tcp but timed out in first 5 moves twice on the game server 2011-11-15T08:58:43 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-15T09:00:45 tcp server has a lot longer default timeout (several seconds) vs the official servers (500ms) 2011-11-15T09:00:57 *** Anilm3 has quit IRC (Ping timeout: 265 seconds) 2011-11-15T09:03:54 *** mikewintermute has joined #aichallenge 2011-11-15T09:06:10 *** Antimony has joined #aichallenge 2011-11-15T09:11:49 *** Areks has quit IRC (Ping timeout: 240 seconds) 2011-11-15T09:17:30 *** zagfai has joined #aichallenge 2011-11-15T09:17:54 : ) 2011-11-15T09:18:05 : ) 2011-11-15T09:21:53 yeah, i figured the default local time was the same as the official server, do you know how to set that? 2011-11-15T09:23:06 --turntime 2345 2011-11-15T09:23:33 It's highly unlikely your computer should have the same turntime as the servers 2011-11-15T09:24:41 just trying to figure out how to roughly reproduce it 2011-11-15T09:25:20 waiting 5 hours between trials isn't really practical, and as far as i can tell from testing my code nothing is particularly slow.. 2011-11-15T09:25:43 Fortunately you can test it locally just fine 2011-11-15T09:26:05 yeah but it's never timed out locally heh 2011-11-15T09:26:33 Look at the exact moment of time-out and figure out what your code is doing 2011-11-15T09:27:10 well my problem is the couple games where it timed out on opening turn or a few in 2011-11-15T09:27:16 when there were only a couple bots and nothing to do 2011-11-15T09:28:55 *** cbad has quit IRC (Read error: Connection reset by peer) 2011-11-15T09:29:09 Java? 2011-11-15T09:29:29 python 2011-11-15T09:29:38 *** cbad has joined #aichallenge 2011-11-15T09:29:38 *** cbad has joined #aichallenge 2011-11-15T09:32:46 *** mleise has joined #aichallenge 2011-11-15T09:34:42 Well, that should already give you a hint. You can always add timing to your code 2011-11-15T09:34:45 pairofdice: did you try to repeat the game (aka setting up a fake bot that replays the moves on top of a simulator repeating the given instructions) 2011-11-15T09:35:00 I don't have the problem ;) 2011-11-15T09:35:14 *** avdg1 is now known as avdg 2011-11-15T09:37:11 *** zagfai has quit IRC (Quit: Page closed) 2011-11-15T09:37:39 pairofdice: yeah, i'm starting to think i have to put a lot more work in on optimization than it may have been worth to use python 2011-11-15T09:38:04 avdg: you can replay from instructions? 2011-11-15T09:38:21 if you create a bot and a simulator for it 2011-11-15T09:38:30 oh 2011-11-15T09:38:30 with open("timing.txt") as test: test.write(endtime - start-time) 2011-11-15T09:38:44 . 2011-11-15T09:39:02 well, it just need to read the files and execute it on the right moment 2011-11-15T09:39:19 *** McLeopold has left #aichallenge 2011-11-15T09:39:43 and the output of the bot can be generated by the given instructions (diff with previous move) 2011-11-15T09:39:49 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-15T09:41:03 (and yes, its a bit of work, but once you have it, it can be very useful for hard core testing, but if you don't need it, just don't code it) 2011-11-15T09:41:57 *** sigh has quit IRC (Remote host closed the connection) 2011-11-15T09:42:37 and ofc make sure your bot uses the your code, or you won't test your code ;-) 2011-11-15T09:44:34 huh 2011-11-15T09:45:53 *** ALplus has quit IRC (Ping timeout: 252 seconds) 2011-11-15T09:46:47 *** Harpyon has joined #aichallenge 2011-11-15T09:52:30 * avdg hopes his bot won't crash the next round on tcp 2011-11-15T09:56:51 so, it crashed 2011-11-15T09:59:17 why don't you just run it locally until it's stable? 2011-11-15T10:01:10 what's the fun in that? 2011-11-15T10:01:11 ;) 2011-11-15T10:01:26 *** amstan has joined #aichallenge 2011-11-15T10:01:27 *** ChanServ sets mode: +o amstan 2011-11-15T10:14:14 qt is nice but i dont get these "layouts" 2011-11-15T10:15:35 *** peyton has joined #aichallenge 2011-11-15T10:16:19 mcstar: think of them like tables in html 2011-11-15T10:16:55 amstan: now im reading a tutorial that hopefully tells me how to apply a streching layout globally 2011-11-15T10:17:03 i.e. to fill the entire space 2011-11-15T10:17:17 amstan: have you done any qt? 2011-11-15T10:17:27 yeah, pyqt too 2011-11-15T10:17:31 but not too much 2011-11-15T10:17:54 friend of mine asked me if its possible to make a tictactoe game in c++ 2011-11-15T10:18:01 of course it is 2011-11-15T10:18:11 but it was more like: can you make a gui for it 2011-11-15T10:18:14 Progresheh 2011-11-15T10:18:20 so i did the gui 2011-11-15T10:18:37 but now i realized im gonna write my battle code in it 2011-11-15T10:18:51 i make a nice gui for my tree-search 2011-11-15T10:19:07 and see what can i replace with heuristics 2011-11-15T10:20:08 amstan: basically, its gonna be a place the ants -> do the battle resolution 2011-11-15T10:20:13 this is the first stage 2011-11-15T10:20:34 do you think others might be interested in this? 2011-11-15T10:20:43 or should i share it at all? 2011-11-15T10:20:54 what? a gui for testing battle code? probably 2011-11-15T10:20:58 yes 2011-11-15T10:20:59 Well, this is a competition :) 2011-11-15T10:21:10 i dont want to share the "battle code" of course 2011-11-15T10:21:24 *** Anilm3 has joined #aichallenge 2011-11-15T10:21:26 but i think this way i can develop mine faster than with the engine 2011-11-15T10:21:29 anyway, ill see 2011-11-15T10:21:32 but isn't this pretty much the same thing as the debugging visualizer? 2011-11-15T10:21:44 whats that? 2011-11-15T10:21:47 hi 2011-11-15T10:21:57 the one that accepts streams? 2011-11-15T10:22:05 the java one? 2011-11-15T10:22:37 mcstar: http://forums.aichallenge.org/viewtopic.php?f=25&t=1633 2011-11-15T10:22:38 im pretty certain that im gonna need this, so im doing it anway 2011-11-15T10:23:20 Feel what the ant feels, be the ant! 2011-11-15T10:23:22 *** liberforce has quit IRC (Ping timeout: 245 seconds) 2011-11-15T10:23:36 amstan: is there code in that forum? a link? 2011-11-15T10:24:03 *** UncleVasya has joined #aichallenge 2011-11-15T10:24:16 https://github.com/j-h-a/aichallenge/tree/vis_overlay 2011-11-15T10:24:31 the guy who made it also frequents this channel 2011-11-15T10:25:27 *** liberforce has joined #aichallenge 2011-11-15T10:25:36 amstan: ok, thx 2011-11-15T10:25:42 still, im gonna write this 2011-11-15T10:27:20 quite nice 2011-11-15T10:27:37 mcstar: what are you working on? 2011-11-15T10:28:09 i just want to do an interactive battle simulator 2011-11-15T10:28:17 just to test my battle code 2011-11-15T10:28:20 cool 2011-11-15T10:28:47 *** treeform has joined #aichallenge 2011-11-15T10:28:54 well, this one that amstan linked is based on the original js visualizer, thats cool 2011-11-15T10:29:31 but i want something to my taste, and more interactive 2011-11-15T10:32:01 With exposions and PhysX support? 2011-11-15T10:32:09 *explosions 2011-11-15T10:32:49 UncleVasya: yes, and it runs on the most powerful graphics engine of the worl 2011-11-15T10:32:51 d 2011-11-15T10:32:58 imagination 2011-11-15T10:33:20 i wish pw would have more flashy things 2011-11-15T10:33:28 like explosions and sounds 2011-11-15T10:33:35 Nooo 2011-11-15T10:33:49 It simple and therefor very nice. 2011-11-15T10:33:57 *it's 2011-11-15T10:34:28 amstan: k, i found it 2011-11-15T10:34:45 *** praveen has joined #aichallenge 2011-11-15T10:34:46 turned out i can right click on the root form and select the appropriate layout 2011-11-15T10:35:01 mcstar: imagination's stimulators aren't cheap :) 2011-11-15T10:35:19 UncleVasya: you grab some 'shrooms 2011-11-15T10:35:27 you just* 2011-11-15T10:35:38 i leave out whole words lol 2011-11-15T10:36:16 mcstar: you make this place happening, I see you all the time :) 2011-11-15T10:36:28 thx 2011-11-15T10:36:42 it is not one person. 2011-11-15T10:37:03 UncleVasya: not one person, what do you mean? 2011-11-15T10:37:08 mcstar is the account for special contest organizers's team 2011-11-15T10:37:23 :) 2011-11-15T10:37:34 praveen: actually im an artificial neural network trained with many personalities 2011-11-15T10:37:59 mcstar: interesting, what is your error rate? 2011-11-15T10:38:26 error? 2011-11-15T10:38:30 humans err 2011-11-15T10:38:34 neural nets dont 2011-11-15T10:38:47 :) 2011-11-15T10:39:04 everything is feature that i exhibit, no bugs whatsoever 2011-11-15T10:39:13 anyway, lets code 2011-11-15T10:39:31 Yeah, start a discussion - exploration? 2011-11-15T10:40:05 just point every ant at the point furthest away from them? :p 2011-11-15T10:40:55 every ant starts at the hill - now what is that point? 2011-11-15T10:41:16 well, first they will seek the closest food of course. 2011-11-15T10:41:19 you have to spread the points in some manner, so they go in different directions 2011-11-15T10:41:25 ah true, 2011-11-15T10:41:42 random direction at start, and follow the wall whenever, unless (some chance) and they move orthagonal? 2011-11-15T10:41:43 :/ 2011-11-15T10:41:49 My ants clearly suck at exploring. 2011-11-15T10:42:09 am just starting to explore. I do gather food alone 2011-11-15T10:42:25 same here. 2011-11-15T10:42:26 we were discussing exploration here when I was in the kindergarden. No let's talk about battle resolution. (because this is the thing I work for at this moment :D) 2011-11-15T10:42:37 *Now 2011-11-15T10:42:41 =] 2011-11-15T10:42:41 lol 2011-11-15T10:42:45 same feelins 2011-11-15T10:43:01 its a really old topic, and everybody does it well 2011-11-15T10:43:03 Unclevasya: Sure, I have some idea on exploration and none on battle resolution 2011-11-15T10:43:35 praveen: anyway, you just discuss it, dont be bothered that we are not interested 2011-11-15T10:43:52 yep, just kidding 2011-11-15T10:45:10 *** amstan_ has joined #aichallenge 2011-11-15T10:45:22 amstan_: hi 2011-11-15T10:45:35 *** amstan_ has quit IRC (Client Quit) 2011-11-15T10:46:00 scizoids have double fun 2011-11-15T10:46:16 *** analyst74 has quit IRC (Quit: HydraIRC -> http://www.hydrairc.com <- Chicks dig it) 2011-11-15T10:46:52 *** Harpyon has quit IRC (Quit: Computer has gone to sleep.) 2011-11-15T10:46:55 i don't: http://aichallenge.org/visualizer.php?game=96037&user=432 2011-11-15T10:47:15 though the solution is obvious 2011-11-15T10:48:31 a1k0n: wha happened? 2011-11-15T10:49:26 a1k0n: experimenting any new algorithm totally? 2011-11-15T10:49:56 aikon is just too good to pay attention to thing like that 2011-11-15T10:51:21 QPainter form the basis of Qt's painting system, Arthur. 2011-11-15T10:52:25 *** cabbage has joined #aichallenge 2011-11-15T10:52:51 *** cabbage is now known as Guest43286 2011-11-15T10:52:58 *** JorgeB has quit IRC (Quit: Computer has gone to sleep.) 2011-11-15T10:53:10 hi there :) 2011-11-15T10:53:51 *** amstan has quit IRC (Ping timeout: 248 seconds) 2011-11-15T10:53:55 @later tell McLeopold nope, I have no objection to getting closer to my original scoring proposal ;) 2011-11-15T10:53:55 janzert: Ready to serve, my lord. 2011-11-15T10:54:04 hi 2011-11-15T10:55:24 *** aerique has quit IRC (Quit: ...) 2011-11-15T10:56:42 only one ant was closest to something to explore at a time 2011-11-15T10:56:56 so all the others wandered aimlessly 2011-11-15T10:57:16 there is a trivial aim that does very well 2011-11-15T10:57:55 actually ive seen that referred as the Ants algorithm 2011-11-15T10:58:02 to* 2011-11-15T10:58:17 a1k0n: you dont use that? 2011-11-15T10:58:53 don't know what you mean but probably not 2011-11-15T11:00:26 i dont want to be too specific, dont want to spoil the fun of others 2011-11-15T11:01:59 *** kilae has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243]) 2011-11-15T11:02:19 *** twymer_ has joined #aichallenge 2011-11-15T11:04:05 *** twymer has quit IRC (Ping timeout: 260 seconds) 2011-11-15T11:06:01 I will have my ants and eat them too 2011-11-15T11:08:38 *** JorgeB has joined #aichallenge 2011-11-15T11:10:27 Does anybody differs enemies? 2011-11-15T11:10:59 I mean into Enemy_bot_1, Enemy_bot_2 2011-11-15T11:12:54 not yet, but perhaps later 2011-11-15T11:12:55 *** McLeopold has joined #aichallenge 2011-11-15T11:13:02 i do when deciding combat outcomes but not otherwise 2011-11-15T11:13:32 janzert: what was your original scoring proposal? 2011-11-15T11:14:07 iirc it was equivalent to the current one, just counted differently 2011-11-15T11:14:38 each hill is 1 point for person controlling the hill at the end of the game, razing the hill transfers control 2011-11-15T11:15:03 *** bmh has joined #aichallenge 2011-11-15T11:15:08 *** bmh has joined #aichallenge 2011-11-15T11:15:48 *** Guest43286 has quit IRC (Ping timeout: 265 seconds) 2011-11-15T11:16:01 the big thing right now is the end bonus doesn't transfer the hill points from one player to another but just adds extra points for the hill 2011-11-15T11:16:42 right, I'm thinking that the -1 should be applied even if the bonus isn't handed out 2011-11-15T11:18:07 well someone should probably be getting points for each hill at the end of the game? Although disappearing points are certainly less likely to cause problems than extra 2011-11-15T11:19:34 I like that a score goes down when losing hills, it give a sense of how a bot is doing 2011-11-15T11:19:36 I suppose I could even see a rational to give timeouts and crashes a zero score though 2011-11-15T11:20:06 The issue raised is that if you timeout, you can make sure you don't lose points 2011-11-15T11:20:14 right 2011-11-15T11:20:20 that's certainly bad 2011-11-15T11:21:07 the end bonus shouldn't ever really be a bonus, but a final transfer of hill control 2011-11-15T11:21:09 okay, I'll make the change so that timeout and crash lose all points for hills, the points for gaining control of those hills will be at a different time, either by razing or as an end bonus 2011-11-15T11:21:39 janzert: agreed, but the visualizer shows how the points were transfered, which is nice 2011-11-15T11:21:40 what I am asking, if a player gets eliminated, should that person earn less points than someone who gets timed out 2011-11-15T11:21:45 *** Antimony has joined #aichallenge 2011-11-15T11:21:50 that should fix the problem with advantage to timing out at least 2011-11-15T11:21:57 *** bmh has quit IRC (Quit: bmh) 2011-11-15T11:21:57 would a bot keep points for hills it razed before it timed out? 2011-11-15T11:21:59 avdg: no, not if they razed a lot of hills 2011-11-15T11:22:10 Yexo: yes 2011-11-15T11:22:31 suppose a bot raised 1 hill and a timed out bot had 3 hills left 2011-11-15T11:22:45 * janzert has to run 2011-11-15T11:22:53 later all 2011-11-15T11:22:57 cya 2011-11-15T11:23:15 s/raised/raided/ 2011-11-15T11:26:04 avdg: 2 points total 2011-11-15T11:30:42 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-15T11:40:04 *** amstan has joined #aichallenge 2011-11-15T11:40:04 *** ChanServ sets mode: +o amstan 2011-11-15T11:42:04 *** dvladim has joined #aichallenge 2011-11-15T11:49:38 *** Larose has joined #aichallenge 2011-11-15T11:52:27 *** datachomper has joined #aichallenge 2011-11-15T11:52:44 How many people are ranked? 2011-11-15T11:53:08 *** Antimony has quit IRC (Ping timeout: 276 seconds) 2011-11-15T11:53:08 *** kaemo has quit IRC (Ping timeout: 276 seconds) 2011-11-15T11:53:13 meh, adding an ant is getting 3ms more expensive for my bot :/ 2011-11-15T11:54:37 *** treeform has quit IRC (Remote host closed the connection) 2011-11-15T11:55:52 *** Harpyon has joined #aichallenge 2011-11-15T11:59:03 *** Antimony has joined #aichallenge 2011-11-15T11:59:42 *** JorgeB has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-11-15T12:00:07 *** JorgeB has joined #aichallenge 2011-11-15T12:07:57 *** delt0r__ has quit IRC (Ping timeout: 245 seconds) 2011-11-15T12:08:00 *** mintsoup has joined #aichallenge 2011-11-15T12:15:59 *** roflmao has joined #aichallenge 2011-11-15T12:16:02 *** twymer_ has quit IRC (Quit: leaving) 2011-11-15T12:17:09 aaah stupid zapping is hard 2011-11-15T12:17:29 *** ikaros has joined #aichallenge 2011-11-15T12:18:14 *** Conorach has quit IRC (Ping timeout: 252 seconds) 2011-11-15T12:18:26 Who is zapping stupids again? 2011-11-15T12:18:39 I am 2011-11-15T12:19:17 I can never get the 2-1 ant ratio down right 2011-11-15T12:19:20 like A or some of the pro bots 2011-11-15T12:19:34 *** liberforce has left #aichallenge 2011-11-15T12:19:35 You don't need a 2-1 ratio just n+1 ratio 2011-11-15T12:19:40 *** Conorach has joined #aichallenge 2011-11-15T12:19:40 yes sorry that 2011-11-15T12:19:49 but i just can't get it down! ah! 2011-11-15T12:20:05 Neither can the top bots a lot of the time 2011-11-15T12:20:44 *** delt0r__ has joined #aichallenge 2011-11-15T12:20:53 either they evade too much or just go suicidal 2011-11-15T12:26:29 *** mj41 has quit IRC (Ping timeout: 252 seconds) 2011-11-15T12:30:50 *** Blkt has quit IRC (Read error: Connection reset by peer) 2011-11-15T12:33:08 *** praveen has quit IRC (Ping timeout: 265 seconds) 2011-11-15T12:40:06 *** Antimony has quit IRC (Ping timeout: 258 seconds) 2011-11-15T12:42:06 *** retybok_ has joined #aichallenge 2011-11-15T12:43:33 *** Antimony has joined #aichallenge 2011-11-15T12:44:11 *** UncleVasya has quit IRC () 2011-11-15T12:51:59 *** retybok_ has quit IRC (Ping timeout: 258 seconds) 2011-11-15T12:56:32 *** Cyb has joined #aichallenge 2011-11-15T12:57:34 *** twymer has joined #aichallenge 2011-11-15T12:57:37 *** treeform has joined #aichallenge 2011-11-15T13:03:13 combat is pretty wonky 2011-11-15T13:03:55 meh, now I know how slow arrays are in js :/ 2011-11-15T13:04:11 sometimes ants attack OUT of range 2011-11-15T13:04:12 it's stupid 2011-11-15T13:06:22 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-11-15T13:06:59 http://i.minus.com/iYb6GfhCHnhRf.png 2011-11-15T13:07:00 *** NightExcessive has joined #aichallenge 2011-11-15T13:07:01 wtf?? 2011-11-15T13:07:20 the idea was that both close in at the same time 2011-11-15T13:07:27 but the attack happened prematurely 2011-11-15T13:07:28 *** Antimony has joined #aichallenge 2011-11-15T13:08:16 roflmao: did you take into account that moves are done before attacking? 2011-11-15T13:08:28 what do you mean? 2011-11-15T13:08:49 you mean 2011-11-15T13:08:50 *** rb_ has quit IRC (Read error: Operation timed out) 2011-11-15T13:08:52 so maybe that bot was trying to get closer 2011-11-15T13:09:00 ah so it attacked in process of moving towards 2011-11-15T13:09:12 the blue circle in front of it? 2011-11-15T13:09:39 yeah, if that ant moved, it could be moving in that blue circle 2011-11-15T13:09:46 hm that's probably it 2011-11-15T13:09:52 although, in theory, both should move in at the same time 2011-11-15T13:09:56 but that's in theory :P 2011-11-15T13:10:13 well, even a computer has to do things 1 at a time 2011-11-15T13:10:27 Yes, but what if the blue moved to the right 2011-11-15T13:10:28 i think it would have been more fun if they could move or shoot but not both 2011-11-15T13:10:29 doing random things parallel is not the way to go 2011-11-15T13:10:40 Then the top orange would be in range but left not 2011-11-15T13:10:49 shooting is done automagically 2011-11-15T13:10:53 pairofdice: blue doesn't move 2011-11-15T13:11:02 it's an extremely constrained and controlled scenario 2011-11-15T13:11:05 blue is a do nothing bot 2011-11-15T13:11:07 Ah 2011-11-15T13:11:48 the blue translucent squares are visualizer 2011-11-15T13:11:50 that's the combat range 2011-11-15T13:11:52 i should make it red 2011-11-15T13:13:42 yeah avdg that was it 2011-11-15T13:13:55 the bottom ant decided to take a path that didn't get it into combat range at the same time as the other one 2011-11-15T13:14:39 problem fixed! 2011-11-15T13:14:41 thanks man :) 2011-11-15T13:17:12 meh, 70 lines added, 6 removed and at least 10 times slower 2011-11-15T13:17:33 thats progress 2011-11-15T13:17:49 *** rb_ has joined #aichallenge 2011-11-15T13:17:53 where is that website where you put up ant replays? 2011-11-15T13:18:03 @paste 2011-11-15T13:18:04 avdg: paste could be http://ants.zeroviz.us/. 2011-11-15T13:18:11 hmm 2011-11-15T13:18:20 what about paste.aichallenge.com 2011-11-15T13:18:32 meh, doesn't work 2011-11-15T13:18:42 http://paste.aichallenge.org/ 2011-11-15T13:18:46 yeah 2011-11-15T13:19:09 wrong tld 2011-11-15T13:19:14 i paste the 2011-11-15T13:19:16 r.replay? 2011-11-15T13:19:18 *** wonklab has left #aichallenge ("Leaving") 2011-11-15T13:19:26 0.replay 2011-11-15T13:19:36 http://paste.aichallenge.org/etepk/ 2011-11-15T13:19:52 seems to work 2011-11-15T13:20:03 I wonder... If I were to put up a TCP "server" (it'd really be a pool of servers) that used an improved protocol, would many people make use of it? :l 2011-11-15T13:21:23 improved protocol? 2011-11-15T13:21:23 if you advertise it well, and its not hard for people to upgrade, why not? 2011-11-15T13:21:27 *** Surya has quit IRC (Quit: Page closed) 2011-11-15T13:21:42 I'd be happy with a tcp server where the ratings actually work 2011-11-15T13:22:09 Garf: Improved protocol by doing something like Kerberos for authentication, and using a worker/manager structure instead of having a single server do all of the work. 2011-11-15T13:22:55 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-15T13:24:25 what's the point of that for a TCP server? 2011-11-15T13:24:34 I don't think any are CPU/client limited 2011-11-15T13:25:04 authentication -> vaguely useful if someones tries to attack you 2011-11-15T13:25:08 Garf: They're network limited, mostly. I have an already existing infrastructure to distribute workers in various location. 2011-11-15T13:25:13 s/location/locations 2011-11-15T13:25:26 really? huh. I count a few kbs per client only 2011-11-15T13:25:40 should handle 1000 clients easily. I dont see that much online anywhere 2011-11-15T13:25:57 Garf: Oh? There's also TCP connection limits (OS/router based) and latencies. 2011-11-15T13:30:15 *** mj41 has joined #aichallenge 2011-11-15T13:30:18 TCP connection limits? 32k or whatever 2011-11-15T13:30:25 latencies? what do you mean by that? 2011-11-15T13:30:45 I'm simply pointing out that IMHO you're solving a problem that doesn't exist. 2011-11-15T13:31:35 hm, what happened to A on fluxid.pl 2011-11-15T13:31:43 i want to fight heem 2011-11-15T13:31:44 Garf: By latencies I mean network latencies. 2011-11-15T13:32:27 how will you solve that without partitioning the playerbase? 2011-11-15T13:33:38 WOOO! Victory over a #94 rank bot 2011-11-15T13:33:55 Garf: I was thinking with a manager/worker scheme I could have clients connect to the manager, which redirects them to a worker nearest them... since latencies between my servers are quite low (and generally much better than the latency between a home connection and my server). And then, of course, the worker relays data back to the manager after a game. 2011-11-15T13:34:25 a1k0n: ask flag_.. but he's not here now 2011-11-15T13:34:26 *** mintsoup has left #aichallenge 2011-11-15T13:34:35 Fluxid: i know. sorry, didn't mean to highlight you. :/ 2011-11-15T13:34:44 also where's your bot? 2011-11-15T13:34:46 a1k0n: no problem 2011-11-15T13:34:56 a1k0n: do you want me to connect it? :) 2011-11-15T13:35:01 NightExcessive: I don't see how this can reduce latency at all. 2011-11-15T13:35:06 one moment 2011-11-15T13:35:14 sure, thanks. 2011-11-15T13:35:40 NightExcessive: unless your routing is so spectacularly better that your c->w->w->c time is better than the current c->s->c time 2011-11-15T13:35:42 Garf: Because network latency from, say, a home connection in Austria to a server in the United States is much higher than the latency from Austria to Germany. 2011-11-15T13:36:22 a1k0n: connected 2011-11-15T13:36:30 NightExcessive: that doesn't help you a thing, unless, as I already stated, you partition the playerbase and favor that the opponent in that game is *also* close to Germany 2011-11-15T13:36:39 NightExcessive: if you don't do that, you'll increase your latency 2011-11-15T13:37:08 or, unless he's trying to make it more fair, so that there are local workers that can enforce the turn limit times but have unlimited turn time between workers 2011-11-15T13:37:22 *** jcarpio has joined #aichallenge 2011-11-15T13:37:24 a1k0n: That's the idea. 2011-11-15T13:38:32 Garf: Essentially, people connect to the worker with the lowest latency to them and that worker managers the turns from that player. At the beginning of a game, a worker would be chosen as the manager of that game- and all other workers would relay data to that worker after their players' turns. 2011-11-15T13:38:49 Garf: It makes turn time limits practical and I think would make for more interesting games. 2011-11-15T13:39:06 i doubt the games would be any more interesting. 2011-11-15T13:39:12 This problem was solved in a far simpeler manner on the chess servers by time stamping the data. 2011-11-15T13:39:13 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-11-15T13:39:17 i don't think anyone is cheating on the turn times right now. 2011-11-15T13:39:37 meh 200ms for 50 bots 2011-11-15T13:40:02 *** Antimony has joined #aichallenge 2011-11-15T13:40:07 besides, it solves a problem that isn't one: with TCP servers you already have hardware inequality, so making it more fair by enforcing time limits gets you nowhere 2011-11-15T13:40:25 and my bot is still not doing anything usefull... 2011-11-15T13:40:45 Garf: Very true. Thanks for allowing me to bounce the idea off of you and providing feedback. :) 2011-11-15T13:40:57 meh, my combat branch takes so long to simulate anything 2011-11-15T13:41:03 My bot has trouble with open maps 2011-11-15T13:41:05 not sure how I can actually turn it into something useful 2011-11-15T13:41:47 NightExcessive: don't let me discourage you. I simply don't think this is the most important issue with the TCP servers right now. Matchmaking and the ratings on them OTOH are completely broken. 2011-11-15T13:42:04 TrueSkill is pretty neat 2011-11-15T13:42:15 *** retybok_ has joined #aichallenge 2011-11-15T13:42:17 pairofdice: if properly working. it isn't on the TCP servers. 2011-11-15T13:42:23 Yeah 2011-11-15T13:42:29 Garf: Well, you brought up things I wouldn't have thought of. I'll, of course, work on things like match making and TrueSkill. 2011-11-15T13:42:55 The TCP servers do seem to have some stability issues 2011-11-15T13:43:03 *** okan_ has joined #aichallenge 2011-11-15T13:43:14 I think there's something wrong with the Sigma calculations 2011-11-15T13:43:15 Oh? 2011-11-15T13:43:40 I think it just has to do with the matchmaking - you pretty much just get put into the first possible game 2011-11-15T13:44:01 http://ants.fluxid.pl/player/roflmao_zap 2011-11-15T13:44:06 err 2011-11-15T13:44:09 Not really thestinger, that shouldn't let Sigma balloon, as every game still bring more info. 2011-11-15T13:44:12 http://ants.fluxid.pl/replay.36557 whyyyyyyyyy did you time out :( 2011-11-15T13:44:33 Right now you can play a few games and see Sigma go from 3.0 -> 3.7 2011-11-15T13:44:48 Even if the results were basically unsurprising. 2011-11-15T13:45:06 http://aichallenge.org/visualizer.php?game=96903 Sweet sweet victory, though my opponent seems to do something weird 2011-11-15T13:45:37 Long streak of losses on random walk maps 2011-11-15T13:45:49 Garf, yeah let's think of a better idea for pairing there 2011-11-15T13:47:39 currently it seems to start a game and just does first comes, first served 2011-11-15T13:48:12 Yeah... Proper match making could get very computationally intensive, though. 2011-11-15T13:48:23 I think you want working ratings first, before you can do matchmaking. 2011-11-15T13:48:34 Without proper ratings, what are you going to base the matchmaking on? :P 2011-11-15T13:48:38 I agree. :P 2011-11-15T13:48:58 Garf, snake bites it's tail ? 2011-11-15T13:49:03 A pair of dice 2011-11-15T13:49:40 i like to blame python for my code being slow, usually it turns out to just be me writing bad code ;) 2011-11-15T13:50:03 twymer: Or Python's intricacies :P 2011-11-15T13:50:03 Did you spot your problem? 2011-11-15T13:50:46 pairofdice: yeah, it's how i handle unexplored territory. which is why if i time out, it's typically very early in the game (when that collection of goals is very large). 2011-11-15T13:51:07 yup, happened to me too 2011-11-15T13:51:18 *** Redgis has joined #aichallenge 2011-11-15T13:51:18 g0llum: I'd wager that you can get way better ratings with the existing data. 2011-11-15T13:51:34 i.e. run a better algorithm over all stored games on the TCP server 2011-11-15T13:51:54 Garf: Any idea how to implement that in Python without making it eat CPU? 2011-11-15T13:52:21 *** mceier has joined #aichallenge 2011-11-15T13:52:27 I think it would be good to see first why TrueSkill (which should work well and be pretty resilient to bad pairing) is having those Sigma issues. 2011-11-15T13:52:54 Maybe this whole thing is a just a minus sign missing somewhere. 2011-11-15T13:52:58 A comma in the wrong place. 2011-11-15T13:53:02 Who knows :P 2011-11-15T13:53:22 *** og01 has quit IRC (Ping timeout: 245 seconds) 2011-11-15T13:53:29 Garf: One thing I noticed is that the script ends the TrueSkill calculation process after a certain time- that could be part of the problem. 2011-11-15T13:53:37 *** og01 has joined #aichallenge 2011-11-15T13:54:05 time, or number of iterations? and in the ratings calc, or in pairing? 2011-11-15T13:54:36 also, on the TCP, people might change their bots without changing the name, so uncertainty should increase over time 2011-11-15T13:54:58 Yes. 2011-11-15T13:55:00 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-11-15T13:55:03 Garf: Iterations. tcpserver.py, line 326 2011-11-15T13:55:29 https://github.com/berak/ants-tcp/blob/master/tcpserver.py#L326 2011-11-15T13:55:58 whois bugnuts? 2011-11-15T13:56:54 he hangs around here a lot 2011-11-15T13:56:56 lurks mostly 2011-11-15T13:57:14 but if you pm him he usually answers 2011-11-15T13:57:39 oh wait 2011-11-15T13:57:59 i was asking about bugnuts on tcp, but just noticed he's here too 2011-11-15T13:59:10 hi guys i have a question , when i am trying test my bot with tools , it s crashes at the first move , actually at the start i doesnt move , but code that i write works on servers , i am writng in C , an i am including it with that "python ""%~dp0sample_bots\C\ant.exe" on the play_one_game.cmd , any idea ? 2011-11-15T13:59:14 *** Antimony has quit IRC (Ping timeout: 276 seconds) 2011-11-15T13:59:18 *** amstan has joined #aichallenge 2011-11-15T13:59:18 *** ChanServ sets mode: +o amstan 2011-11-15T13:59:40 lol 2011-11-15T14:00:10 that string looks pretty malformed to me 2011-11-15T14:00:24 could you paste the full play_one_game cmd ? 2011-11-15T14:00:30 *** dvladim has quit IRC (Ping timeout: 256 seconds) 2011-11-15T14:00:36 @echo off python "%~dp0playgame.py" --engine_seed 42 --player_seed 42 --end_wait=1.25 --verbose --log_dir game_logs --turns 500 --map_file "%~dp0maps\maze\maze_04p_01.map" %* "python ""%~dp0sample_bots\C\ant.exe""" "python ""%~dp0sample_bots\python\LeftyBot.py""" "python ""%~dp0sample_bots\python\HunterBot.py""" "python ""%~dp0sample_bots\python\RandomBot.py""" 2011-11-15T14:00:36 off python %~dp0playgame.py --engine_seed 42 --player_seed 42 --end_wait=1.25 --verbose --log_dir game_logs --turns 500 --map_file %~dp0maps\maze\maze_04p_01.map %* python %~dp0sample_bots\Cnt.exe python %~dp0sample_bots\python\LeftyBot.py python %~dp0sample_bots\python\HunterBot.py python %~dp0sample_bots\python\RandomBot.py 2011-11-15T14:01:21 you only put "python" after bots that are python bots 2011-11-15T14:01:26 orkan_, all your """ got lost there 2011-11-15T14:01:38 LeftyBot and HunterBot are python bots (they end in .py) 2011-11-15T14:01:44 but you do not want to put python before your C bot! 2011-11-15T14:01:51 "python ""%~dp0sample_bots\C\ant.exe" — is wrong 2011-11-15T14:01:55 Why are you trying to run an exe bot through python 2011-11-15T14:02:08 *** mceier has quit IRC (Read error: Connection reset by peer) 2011-11-15T14:02:29 NightExcessive: it's simply detecting if the calculation bailed out, and logging an error if it did. 2011-11-15T14:02:41 *** mceier has joined #aichallenge 2011-11-15T14:02:47 Garf: Ahh. I was just skimming over the code and caught that. :) 2011-11-15T14:04:37 let try it 2011-11-15T14:05:49 NightExcessive, but you are still right, the java impl used there showed some problems in beta with draws, and was discarded for a php version, afaiik 2011-11-15T14:06:15 g0llum: I think I'll stick to the Python version on my server and simply improve it. 2011-11-15T14:06:46 mcstar: how far along are you on the battle visualizer? sounded pretty cool 2011-11-15T14:07:25 I'm experiencing a lot of random timeouts 2011-11-15T14:09:05 twymer: hand not you waked me, pretty far... 2011-11-15T14:09:13 indeed, the official server seems to use PHPSkills 2011-11-15T14:09:30 in api_game_result.php, which I presume gets the results from the workers 2011-11-15T14:11:23 tempting to try to fix this :P 2011-11-15T14:11:37 * Garf looks at the Mu of his bot. 2011-11-15T14:11:43 Temptation over. :P 2011-11-15T14:14:23 *** ztfw has joined #aichallenge 2011-11-15T14:20:23 Garf, if you feel fit to do that PLEASE!!, it's my mu, too, (and of more that 1000 players on fluxid/tcpants/accoun's, NightExcessive's) 2011-11-15T14:20:32 aichallenge: McLeopold epsilon * r6eed29a / ants/ants.py : modified scoring for game, timeouts and crashes immediately remove all hill points for bot - http://git.io/kabuBQ 2011-11-15T14:20:47 *** dendron has joined #aichallenge 2011-11-15T14:21:45 Garf: If you were to put a fix into the master branch of the ants-tcp repo I'd pull it into my fork :) 2011-11-15T14:21:47 man I'm jumping with joy 2011-11-15T14:21:51 Please, can someone enlighten me, can I use QT in challenge? 2011-11-15T14:22:05 isn't that a gui? 2011-11-15T14:22:12 dendron: It's doubtful. Extremely doubtful. 2011-11-15T14:22:43 Thank you. Just STL then. 2011-11-15T14:25:04 *** rb_ has quit IRC (Ping timeout: 240 seconds) 2011-11-15T14:25:45 I got to skill level 40!! 2011-11-15T14:26:20 @later tell janzert the scoring change has been made, can you update the workers? 2011-11-15T14:26:20 McLeopold: Ready to serve. 2011-11-15T14:26:56 McLeopold: Qt is more than just Gui, it's a complete C++ library that covers a wide variaty of topics 2011-11-15T14:27:28 roflmao: congrats :) 2011-11-15T14:27:58 thanks 2011-11-15T14:28:04 *** jcdny is now known as bugnuts2 2011-11-15T14:28:04 on fluxid 2011-11-15T14:28:09 so hoping it's like 70 or so on official 2011-11-15T14:28:14 (that doesn't mean anything) 2011-11-15T14:28:42 roflmao: I'm about 40 on fluxid, near 75 on the official challenge 2011-11-15T14:28:49 double congrats :) 2011-11-15T14:28:51 McLeopold: sure, might be this evening though 2011-11-15T14:28:51 nice! 2011-11-15T14:29:22 *** tobym has joined #aichallenge 2011-11-15T14:30:58 *** cafaro has quit IRC (Ping timeout: 255 seconds) 2011-11-15T14:32:22 is Fluxid around? 2011-11-15T14:32:35 or Accoun? 2011-11-15T14:32:56 or anyone else in charge of tcp servers? 2011-11-15T14:33:29 ok, three workers updating once they're back up I'll do the other three 2011-11-15T14:33:59 McLeopold: here 2011-11-15T14:34:39 McLeopold: Ello, starting one. 2011-11-15T14:35:31 do bots that sometimes issue duplicate orders crash on official? 2011-11-15T14:36:17 Fluxid: can you pull the new ants.py with the scoring changes? 2011-11-15T14:36:46 roflmao: no, the orders should be ignored 2011-11-15T14:37:02 McLeopold: except in the submission test? 2011-11-15T14:37:06 McLeopold: damn, i fucked up my local repo 2011-11-15T14:37:08 or there too? 2011-11-15T14:37:24 janzert: correct, the submission test will fail the submission if it finds a duplicate 2011-11-15T14:37:31 McLeopold: i'll fix it and pull within 1.5h, okay? 2011-11-15T14:37:40 Fluxid: no problem 2011-11-15T14:41:00 McLeopold: scoring changes? 2011-11-15T14:41:21 If your bot crashes or times out, no points for you 2011-11-15T14:41:33 ouch, that's mean 2011-11-15T14:42:15 I was thinking that timing out with a lot of ants wasn't too bad since you usually already have a lot of points 2011-11-15T14:42:17 Fluxid: When you pull that, could you let me know so that I can pull it into my fork? 2011-11-15T14:42:25 I guess I'll have to be more careful :) 2011-11-15T14:43:08 NightExcessive: what did you fork for? What does your tcp server do that others don't? 2011-11-15T14:43:22 retybok_: I'm working on making it a manager/worker model. 2011-11-15T14:43:35 retybok_: Because I don't want the server accepting connections to be running on my web server :) 2011-11-15T14:43:45 nnnnnnnnnnnnnnnnnnnnnnnnnwe uwe used pnnn 2011-11-15T14:43:57 a1k0n: what encoding is this? 2011-11-15T14:44:21 whoa 2011-11-15T14:44:27 i had a very messed up screen session, sorry 2011-11-15T14:44:31 anyways 2011-11-15T14:44:42 we used to use bayeselo on the tcp servers on previous contests, which did a whole-history ranking 2011-11-15T14:45:08 i haven't finished reading scrollback, but did someone find a particular problem causing sigma inflation? 2011-11-15T14:45:27 McLeopold: hey 2011-11-15T14:45:33 find the problem no. observe the issue, yes. 2011-11-15T14:45:52 apparently, it's possible the Java port of TrueSkill is buggy and the PHP one is not 2011-11-15T14:46:08 Garf: What about the Python one? 2011-11-15T14:46:17 no idea 2011-11-15T14:46:44 *** Fandekasp has quit IRC (Ping timeout: 252 seconds) 2011-11-15T14:47:55 NightExcessive, that one had a bias problem, sinking mu over time, if i recall that right 2011-11-15T14:48:12 jezus 2011-11-15T14:48:18 Argh. 2011-11-15T14:48:55 so, getting that php one to work, would be great for the community.. 2011-11-15T14:49:05 the python code is quite compact compared to the PHP/Mono/Java ports 2011-11-15T14:49:07 Getting any of them to work would be great. :l 2011-11-15T14:49:21 I'll be working on the Python one, since it's what I'll be using and is most convenient for me. 2011-11-15T14:49:36 *** dendron has quit IRC (Quit: Page closed) 2011-11-15T14:50:07 *** amstan has quit IRC (Ping timeout: 248 seconds) 2011-11-15T14:50:44 i think the "sigma issue" is mainly a result of there being an increasing number of bots on there, and the bots which are ranked highly haven't played in weeks and hence their sigma is small (cuz when they played there were fewer payers) 2011-11-15T14:50:55 if we could actually match against those players it would settle down 2011-11-15T14:51:01 but of course we can't 2011-11-15T14:51:19 wipe the rankings? 2011-11-15T14:51:32 I have a bot which has now thousands of games. For a while, it mostly played a select group of other bots that was online. It's sigma is >3 2011-11-15T14:51:51 Yeah I think wiping the rankings daily on the tcp server would be a good idea 2011-11-15T14:52:13 what does sigma mean for the rankings? 2011-11-15T14:52:27 rety_book, make that monday morning.. 2011-11-15T14:52:32 because it's buggy, it essentially makes them more random 2011-11-15T14:52:53 wiping the ratings makes little sense IMHO. What would it solve? 2011-11-15T14:53:19 sigma and mu of what? distribution of your trueskill? 2011-11-15T14:53:29 sigma is basically how well the server knows your rank 2011-11-15T14:53:40 Yea 2011-11-15T14:53:55 *** mj41 has quit IRC (Ping timeout: 260 seconds) 2011-11-15T14:54:18 because the rankings are mu - 3*sigma, random sigma means random ranks 2011-11-15T14:55:24 too bad you can't run Bayeselo over multiplayer games 2011-11-15T14:55:34 we should bug Remi about it ;P 2011-11-15T14:55:37 *** Accoun has quit IRC () 2011-11-15T14:56:28 is the sigma and mu calculated over the entire history of your games? 2011-11-15T14:56:29 *** Fandekasp has joined #aichallenge 2011-11-15T14:56:40 NightExcessive: why, it is already there: https://github.com/berak/ants-tcp/commit/ac20bdce60f7944dd4ef295729fc07b1e1e8aa66 2011-11-15T14:56:59 Fluxid: Nevermind :) 2011-11-15T14:58:14 *** analyst74 has joined #aichallenge 2011-11-15T15:00:33 I didn't know the word ilk, but there is actually a very good translation match for me. You would think all these 'easy' words would be the first you learn. 2011-11-15T15:01:00 magiik, no, just over your last mu/sigma and the ranks of your last game 2011-11-15T15:05:26 *** Fandekasp has quit IRC (Ping timeout: 252 seconds) 2011-11-15T15:11:27 *** rb_ has joined #aichallenge 2011-11-15T15:11:35 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-11-15T15:12:00 yeah maybe 2011-11-15T15:12:15 oops (ignore) :D 2011-11-15T15:12:46 *** Accoun has joined #aichallenge 2011-11-15T15:14:44 *** choas has joined #aichallenge 2011-11-15T15:15:05 *** hargrow has joined #aichallenge 2011-11-15T15:15:45 *** AndreyyATC has joined #aichallenge 2011-11-15T15:16:47 *** nn has joined #aichallenge 2011-11-15T15:18:17 guys, i'm gonna do total reboot of tcp server in a minute 2011-11-15T15:20:20 done 2011-11-15T15:20:22 *** kilae has joined #aichallenge 2011-11-15T15:20:26 Fluxid, nice !! 2011-11-15T15:20:54 *** nn has quit IRC (Client Quit) 2011-11-15T15:20:59 if someone needs, i can put up old webserver with old db 2011-11-15T15:21:09 or mebe just post the DB for people? 2011-11-15T15:21:18 sql dump or w/e 2011-11-15T15:21:18 Fluxid: i think reseting tcp is a good idea 2011-11-15T15:21:30 magiik: no way 2011-11-15T15:21:37 magiik: passwds are plaintext 2011-11-15T15:21:46 a reset of the rankings woud have done 2011-11-15T15:21:55 well i meant the rankings data, but if it's all in one table i guess you can't 2011-11-15T15:21:58 what would a reset solve? 2011-11-15T15:22:12 save recent games ? 2011-11-15T15:22:53 can anyone linky me to a description of the skill/ranking system? 2011-11-15T15:23:21 https://research.microsoft.com/en-us/projects/trueskill/details.aspx 2011-11-15T15:23:25 ty 2011-11-15T15:23:34 whathtehell happened on fluxid! 2011-11-15T15:23:39 oh 2011-11-15T15:23:40 i see ;90 2011-11-15T15:23:43 complete reset, fun 2011-11-15T15:24:01 I'm number 3 now! ha 2011-11-15T15:29:33 *** Fandekasp has joined #aichallenge 2011-11-15T15:29:42 :) 2011-11-15T15:29:49 *** luizribeiro has quit IRC (Ping timeout: 240 seconds) 2011-11-15T15:32:10 *** Redgis has quit IRC (Read error: Connection reset by peer) 2011-11-15T15:32:34 *** Redgis has joined #aichallenge 2011-11-15T15:33:11 *** NoxiaZ^ has joined #aichallenge 2011-11-15T15:36:52 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-15T15:37:10 *** g0llum has joined #aichallenge 2011-11-15T15:38:34 Fluxid: already down? 2011-11-15T15:38:54 retybok_: doesn't seem like it 2011-11-15T15:39:28 retybok_,doesn't seem like it 2011-11-15T15:39:59 weird, my bots are suddenly timeing out 2011-11-15T15:40:11 (the smoke_carrot ones) 2011-11-15T15:40:57 i had this shit-weather pase this morning, where my umts-modem would just ... 2011-11-15T15:41:36 but it was working two minutes ago!! And I'm on a rather stable DSL... 2011-11-15T15:42:05 ok, it's working again 2011-11-15T15:42:52 http://ants.fluxid.pl/player/strcat_next weird 2011-11-15T15:43:00 the server doesn't seem to want to rank me at all 2011-11-15T15:43:08 same rety 2011-11-15T15:43:16 mmm, still timing out 2011-11-15T15:43:28 even thoug my bot takes about 10ms to display its moves and "go" 2011-11-15T15:43:30 fluxid occasionally times out 2011-11-15T15:45:03 test other servers, to find out on which side the problem is 2011-11-15T15:45:09 grrr, turns out my wife was uploading pictures :-/ 2011-11-15T15:45:09 hm, i think my client got stuck when you restarted 2011-11-15T15:46:43 *** epicmonkey has joined #aichallenge 2011-11-15T15:46:58 roflmao: more often than before reboot? 2011-11-15T15:48:46 weird, another game and it still hasn't ranked strcat_next at all 2011-11-15T15:50:54 *** Antimony has joined #aichallenge 2011-11-15T15:53:43 *** Anilm3 has quit IRC (Ping timeout: 265 seconds) 2011-11-15T15:54:26 fluxid: yes 2011-11-15T15:57:33 one moment 2011-11-15T15:58:40 *** dr- has quit IRC (Ping timeout: 240 seconds) 2011-11-15T15:59:00 :) 2011-11-15T15:59:18 just so you know, your server is a godsend 2011-11-15T16:00:18 roflmao: removed few lines and restarted 2011-11-15T16:00:32 *** iris1 has quit IRC (Quit: iris1) 2011-11-15T16:00:37 you da man! 2011-11-15T16:00:59 berak tried to set socket timeout globally, this could cause some problems with other threads 2011-11-15T16:01:06 yeah Fluxid, all the flaws it has , thanks for running and caring for that! 2011-11-15T16:01:16 lol, np 2011-11-15T16:01:25 most credit goes to original author anyway 2011-11-15T16:01:49 now all I have to do is make a bot better than yours and everything will be good :P 2011-11-15T16:01:56 you get credit for running it and fixing it in real time :) 2011-11-15T16:02:03 argh, i'm stuck again 2011-11-15T16:03:28 *** aarossig has quit IRC (Ping timeout: 256 seconds) 2011-11-15T16:03:33 *** cafaro has joined #aichallenge 2011-11-15T16:03:33 *** cafaro has joined #aichallenge 2011-11-15T16:05:21 *** aarossig has joined #aichallenge 2011-11-15T16:05:44 *** mj41 has joined #aichallenge 2011-11-15T16:06:17 *** AndreyyATC has quit IRC (Ping timeout: 265 seconds) 2011-11-15T16:09:17 So if I'm understanding TrueSkill properly, best matchmaking is done by choosing people with the closest skill values? 2011-11-15T16:09:54 *** ping-- has quit IRC (Ping timeout: 244 seconds) 2011-11-15T16:09:55 When they have the same probability of winning 2011-11-15T16:10:14 So basically yes 2011-11-15T16:10:30 Prolly closes mu values, not skill values 2011-11-15T16:10:36 Ah. 2011-11-15T16:10:41 Yeah, closest mu 2011-11-15T16:10:48 *** ping-- has joined #aichallenge 2011-11-15T16:13:05 it's an integral over the probability of skill for each player to determine likelihood of a tie 2011-11-15T16:13:18 so it's based on both mu and sigma 2011-11-15T16:13:49 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-15T16:15:41 these huge epic games with a bazillion players drive me nuts 2011-11-15T16:17:41 *** ltriant has joined #aichallenge 2011-11-15T16:19:18 *** Palmik_ has quit IRC (Remote host closed the connection) 2011-11-15T16:20:58 *** kaemo has joined #aichallenge 2011-11-15T16:22:48 Me too, mostly because I lose those 2011-11-15T16:23:37 *** ping-- has quit IRC (Read error: Connection reset by peer) 2011-11-15T16:23:41 *** ping-- has joined #aichallenge 2011-11-15T16:23:55 i often win them depite being eliminated 2011-11-15T16:24:29 I like the fact that you can win while being eliminated. It rewards offensive behavior 2011-11-15T16:24:38 yep 2011-11-15T16:24:44 Otherwise we would see huge buildups and stalemates 2011-11-15T16:24:52 *** amstan has joined #aichallenge 2011-11-15T16:24:52 *** ChanServ sets mode: +o amstan 2011-11-15T16:25:44 *** kaemo has quit IRC (Ping timeout: 252 seconds) 2011-11-15T16:28:27 now i made my bot more conservative, and i lose: http://ants.fluxid.pl/replay.102 2011-11-15T16:29:31 a1k0n: it seems pretty good 2011-11-15T16:29:38 when an ant wants something it gets it 2011-11-15T16:29:44 yeah 2011-11-15T16:30:00 i think the only way to win a giant deathmatch like this is to do symmetry inference really fast 2011-11-15T16:30:17 zerg rush everyone 2011-11-15T16:30:55 a1k0n: you should know the whole map from turn 10 2011-11-15T16:31:06 because at that point you see the other base that is symmetric to your other base 2011-11-15T16:31:07 maaan i timed out?! whaat 2011-11-15T16:31:11 amstan: exactly 2011-11-15T16:31:59 i think it's probably more important at this point for me to add some awareness that if some enemy ant steps on my hill that's bad 2011-11-15T16:32:05 haha 2011-11-15T16:32:17 it's really bad to lose your hill 2011-11-15T16:32:23 a1k0n: it depends, if you have 4 hills it's not that bad 2011-11-15T16:32:26 if you have no hills, gathering food is worthless 2011-11-15T16:32:36 a1k0n: so your bot will be capable of revenge? i think it's a bad idea 2011-11-15T16:32:39 besides denying it to the enemy 2011-11-15T16:32:49 a1k0n: what if one day, you decide not to give it the challenge of compeeting in tcp? 2011-11-15T16:33:04 a1k0n: what if it decides that you're not worthy to be its master anymore 2011-11-15T16:33:24 *** Anilm3 has joined #aichallenge 2011-11-15T16:33:26 *** delt0r__ has quit IRC (Ping timeout: 252 seconds) 2011-11-15T16:33:45 Isn't it possible to download my own code? 2011-11-15T16:35:04 *** ALplus has joined #aichallenge 2011-11-15T16:35:05 Is there a simple way to go about determining map symmetrical axis? Is pattern matching the best way? 2011-11-15T16:36:05 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-11-15T16:36:35 datachomper: if you assume it's translational symmetry 2011-11-15T16:36:41 and if you are aware of all your hills 2011-11-15T16:37:12 the moment you see another hill, you look around it, see which hill(of yours) is symmetrical to 2011-11-15T16:37:16 For some reason my bot (written in Go) takes 40MB of ram when running on a 64 bit machine, and more than 700MB when running on my 32 bit laptop. Does anyone more knowledgeable than I am about Go have an explanation to offer? 2011-11-15T16:37:23 then fill the rest of your map using terrain from around your hills 2011-11-15T16:38:09 retybok_ both same OS and Go version? 2011-11-15T16:38:16 amstan: Ah thanks 2011-11-15T16:38:21 bugnuts: both archlinux with the latest release 2011-11-15T16:38:39 it doesn't really bug me I just find it curious 2011-11-15T16:39:20 *** kilae has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243]) 2011-11-15T16:39:22 weird. you could try dumping MemStats and see if anything stands out 2011-11-15T16:39:38 look at package runtime... 2011-11-15T16:40:43 bugnuts: I'll try when I get a moment 2011-11-15T16:41:32 a1k0n: we should unite against the best player, not bleed eachother to death and wait for him to finish us :) http://ants.fluxid.pl/replay.127 2011-11-15T16:41:35 *** NoxiaZ^ has quit IRC (Ping timeout: 248 seconds) 2011-11-15T16:42:07 So no answer to my question huh! 2011-11-15T16:43:34 Anilm3: you should be more specific 2011-11-15T16:43:45 retybok_: you could try out gccgo on both too to see if it's an 8g bug 2011-11-15T16:43:56 Isn't it possible to download my own code? 2011-11-15T16:44:20 thestinger: that's not a bad idea 2011-11-15T16:44:49 retybok_: totally 2011-11-15T16:45:08 dammit, i keep timing out :( 2011-11-15T16:45:11 a1k0n: I have no idea how to implement strategic intelligence like this, though 2011-11-15T16:45:25 basically impossible since we can't even see each other 2011-11-15T16:45:57 exactly, the fog of war makes it a lot harder 2011-11-15T16:46:20 keeping track of who you see on the borders could probably give you an idea, though 2011-11-15T16:46:29 I think it makes it even more fun! 2011-11-15T16:47:35 amstan: Are all types of symmetry eligible for the tournament maps? 2011-11-15T16:47:44 Anilm3: did you lose your code? 2011-11-15T16:47:47 datachomper: whatever is fair 2011-11-15T16:47:59 datachomper: note that we might change the mapgens, so don't go only by those 2011-11-15T16:48:13 amstan: did you want to talk to me earlier? 2011-11-15T16:48:14 datachomper: actively look for more symmetry types and maybe even give us suggestions if we missed something 2011-11-15T16:48:22 McLeopold: i just wanted to say hi 2011-11-15T16:48:38 amstan: Ok will do 2011-11-15T16:49:05 amstan: so, you gonna split the forums yet? :) 2011-11-15T16:49:21 i'm not sure how i would go to do that 2011-11-15T16:49:45 well, we created an eitherpad, if you just create the new ones, I can move stuff 2011-11-15T16:51:22 *** pairofdice has quit IRC (Quit: Just keep breathing) 2011-11-15T16:52:08 McLeopold: do you still have the link? 2011-11-15T16:52:14 maybe... 2011-11-15T16:53:40 no, have to grep the logs 2011-11-15T16:53:51 keywords? 2011-11-15T16:53:55 etherpad 2011-11-15T16:53:56 mcleopold etherpad 2011-11-15T16:54:03 *** delt0r_ has joined #aichallenge 2011-11-15T16:54:59 http://ietherpad.com/O2cfCoy28r 2011-11-15T16:55:12 yep 2011-11-15T16:55:29 ok.. so what's the diff with the current setup? 2011-11-15T16:55:32 all those new forums? 2011-11-15T16:55:45 just the bottom 4 2011-11-15T16:55:54 so, 3 new forums 2011-11-15T16:56:00 3? 2011-11-15T16:56:07 2 2011-11-15T16:56:22 we have 2 spies, lol 2011-11-15T16:56:24 Starter Packages and Tools -- and -- Rules Questions 2011-11-15T16:56:28 had* 2011-11-15T16:57:09 McLeopold: can we rename it to something else? rules will make a ton of people ask if they can write to files 2011-11-15T16:57:18 Fluxid: did you change the maps/food spawn rate? 2011-11-15T16:57:20 well, that is a rules question 2011-11-15T16:57:29 yes, but that shouldn't be in ants then 2011-11-15T16:57:30 more bots are timing out now, and there are a shitton of ants in play 2011-11-15T16:57:36 amstan: we could make a sticky 2011-11-15T16:57:47 Game Rules Questions? 2011-11-15T16:58:20 but then where to general contest questions go? 2011-11-15T16:58:43 like? 2011-11-15T16:58:50 writing to files 2011-11-15T16:59:02 right now I think they are in suggestions? 2011-11-15T16:59:21 well.. they're pretty rare, and if they exist they're answered quickly 2011-11-15T16:59:27 i don't think we need a forum for that 2011-11-15T16:59:39 so, Suggestions (Contest Questions welcome here)? 2011-11-15T16:59:52 retybok_: hah our last game was just insane 2011-11-15T17:00:06 i would like to avoid debates on contest rules 2011-11-15T17:00:08 956 ants in the hive? 2011-11-15T17:00:14 a1k0n: link? 2011-11-15T17:00:18 http://ants.fluxid.pl/replay.134 2011-11-15T17:00:54 wow, I end up with a huge backlog of ants 2011-11-15T17:01:23 *** hargrow has quit IRC (Ping timeout: 265 seconds) 2011-11-15T17:02:08 thestinger: er, you're strcat, right 2011-11-15T17:02:13 yeah 2011-11-15T17:02:34 then who is smoke_carrot 2011-11-15T17:02:42 McLeopold: what's the point of bot development then? 2011-11-15T17:03:07 stuff I don't need to pay attention to as an admin 2011-11-15T17:03:14 but who's going to post in there? 2011-11-15T17:03:30 people who want to talk to each other via the forums 2011-11-15T17:03:37 nobody's going to post some code they made and ask why it doesn't work 2011-11-15T17:04:12 amstan: I think they will 2011-11-15T17:04:32 I see your point, but it's a mess right now, so we need extra containers until it's cleaned up 2011-11-15T17:04:37 delt0r_: well, it's just that nobody has posted such a thing yet that i distinctivelly remember 2011-11-15T17:04:40 Maybe it can go away later 2011-11-15T17:05:08 a1k0n: smoke_carrot is me 2011-11-15T17:05:23 retybok_: oh ok. i suspected as much. didn't you have a different name before? 2011-11-15T17:05:34 Here's a non-admin dev post: http://forums.aichallenge.org/viewtopic.php?f=25&t=1663 2011-11-15T17:05:38 on the main site I'm "pierrejoel" 2011-11-15T17:06:00 anyways yeah that game was pretty insane 2011-11-15T17:06:15 McLeopold: this would go to the starter packages forums if it was up to me, it's more relevant to the users in there 2011-11-15T17:06:18 I think my timeout was due to me trying to watch my bot output in screen 2011-11-15T17:06:20 McLeopold: yet admins wouldn't care about it 2011-11-15T17:06:37 Fluxid: What about more maps? 2011-11-15T17:07:26 retybok_: that can be very expensive... 2011-11-15T17:07:56 delt0r_: especially when you have 200 ants and are logging like crazy :) 2011-11-15T17:09:09 amstan: okay, maybe it goes away, but I at least need the other ones to get started on cleanup 2011-11-15T17:09:39 *** Redgis has quit IRC (Quit: ... mains libres) 2011-11-15T17:10:04 <_flag> Okay, who made B on the tcp server? 2011-11-15T17:10:18 lol 2011-11-15T17:10:22 <_flag> That was going to be my next version :( 2011-11-15T17:10:29 i was thinking of doing that 2011-11-15T17:10:35 *** Jak_o_Shadows has joined #aichallenge 2011-11-15T17:10:42 I better make C really quick... 2011-11-15T17:10:57 my ants are retarded 2011-11-15T17:11:07 McLeopold: so you want the game rules discussions forum? 2011-11-15T17:11:11 the walk right past undefended hills 2011-11-15T17:11:12 <_flag> Also, are you guys trying to fix the ridiculously high sigmas on tcp with the reset? 2011-11-15T17:11:13 yes 2011-11-15T17:11:17 run away from food 2011-11-15T17:11:26 delt0r_: yet you just won :) 2011-11-15T17:11:39 well i do ok on a few maps 2011-11-15T17:11:39 McLeopold: i broke the forums visualizer when i moved the forums 2011-11-15T17:11:51 _flag: how the hell do you have 49 games already? are you running a cluster of bots? 2011-11-15T17:11:57 McLeopold: i can't load the games off the main site anymore because of cross site scripting preventions 2011-11-15T17:11:58 delt0r_: at least I killed you :) 2011-11-15T17:11:58 retybok_: I have a very simple bot 2011-11-15T17:12:03 <_flag> a1k0n: I'm running 2 bots at once 2011-11-15T17:12:05 *** Cyb has quit IRC () 2011-11-15T17:12:05 :( 2011-11-15T17:12:10 <_flag> And my games usually end pretty fast 2011-11-15T17:12:11 delt0r_: are you the ruby guy? 2011-11-15T17:12:12 McLeopold: any ideas? 2011-11-15T17:12:13 changing it to a influence map style now 2011-11-15T17:12:17 a1k0n: he also just murders us so damn fast :P 2011-11-15T17:12:20 retybok_: java 2011-11-15T17:12:21 <_flag> Want me to turn one off? 2011-11-15T17:12:25 i keep getting embroiled in huge games 2011-11-15T17:12:30 coffee all the way down 2011-11-15T17:12:30 no, it's fine 2011-11-15T17:12:44 amstan: just copy the visualizer.js into the forums domain 2011-11-15T17:13:00 McLeopold: visualizer code works, it just that it can't load the replay 2011-11-15T17:13:14 oh 2011-11-15T17:13:35 load the js from the main site? 2011-11-15T17:13:48 I'm not up on my XSS rules 2011-11-15T17:14:26 McLeopold: so, i'm loading both the .js and .replay from the main site 2011-11-15T17:14:32 .js loads, .replay doesn't 2011-11-15T17:14:35 retybok_: I am watching the games live ;) 2011-11-15T17:14:49 just took out green --yea 2011-11-15T17:14:50 it would be really awesome if the tcp server showed live games 2011-11-15T17:14:57 think they where timed out 2011-11-15T17:15:13 a1k0n: you can watch from your bots view with my visulaizer... 2011-11-15T17:15:18 McLeopold: so what's going in the ant game rules forum? 2011-11-15T17:15:26 oh really? 2011-11-15T17:15:37 yea 2011-11-15T17:15:42 you just wrap your bot with it? 2011-11-15T17:15:47 its just a filter to your bot 2011-11-15T17:15:51 awesome 2011-11-15T17:15:58 so rather than "bot command 2011-11-15T17:16:03 amstan: rules about the game, like scoring, movement and such 2011-11-15T17:16:22 its now "java -jar AntWatcher2000.jar "bot command" " 2011-11-15T17:16:31 post it on the forums 2011-11-15T17:16:32 McLeopold: i can only find 2 topics already about that 2011-11-15T17:16:48 so far* 2011-11-15T17:17:09 I know I've seen more. But even if it is a low volume, it is nice to have them in one place 2011-11-15T17:17:47 k 2011-11-15T17:17:48 it's up 2011-11-15T17:17:53 post *is* on the forums 2011-11-15T17:17:59 not it 2011-11-15T17:18:26 amstan: so, you renamed Dev to Starter, so I can't separate out the stuff 2011-11-15T17:19:01 thestinger: omg you bastard haha 2011-11-15T17:19:02 McLeopold: originally i meant that to be a developer forums, but it seems like most of the action is on github 2011-11-15T17:19:30 also what the fuck, i timed out? no i didn't 2011-11-15T17:19:34 arrrggghh 2011-11-15T17:19:51 ok, but we need a way to separate admin questions from questions to and from contestants 2011-11-15T17:20:00 a1k0n: what did I do? :P 2011-11-15T17:20:04 http://ants.fluxid.pl/replay.160 2011-11-15T17:20:10 killed me with your last three ants or so 2011-11-15T17:20:54 McLeopold: but i disagree, if admins are supposed to watch the starter pack forum, they can watch that 2011-11-15T17:21:15 McLeopold: if someone posts a third party tool, there should be no reason not to put it in the same forums, from a users perspective 2011-11-15T17:21:52 you want users to find information easily, it's not meant to be a ticketing system for us on user's issues 2011-11-15T17:22:05 amstan: I'm concerned about missing a question asking for an official admin answer, because there is too much stuff to go through 2011-11-15T17:22:18 also, a ton of the questions asked don't require admins to answer them, including stuff about starter packages 2011-11-15T17:23:24 *** Antimony has joined #aichallenge 2011-11-15T17:25:09 thestinger: http://ants.fluxid.pl/replay.178 <- How did I win that? :) 2011-11-15T17:30:11 *** peyton has quit IRC (Quit: peyton) 2011-11-15T17:30:13 McLeopold: oh yeah, have you seen this? http://tron.aichallenge.org http://planetwars.aichallenge.org 2011-11-15T17:32:18 amstan: yep, nice work, is the sql compressed? 2011-11-15T17:32:42 McLeopold: tron is, planetwars will be soon 2011-11-15T17:32:53 but planetwars games table only compressed like 0.01% 2011-11-15T17:33:44 good night everybody 2011-11-15T17:34:11 *** retybok_ has left #aichallenge 2011-11-15T17:34:29 *** ztfw has quit IRC (Ping timeout: 252 seconds) 2011-11-15T17:34:55 *** peyton has joined #aichallenge 2011-11-15T17:35:41 *** NightExcessive has quit IRC (Quit: Page closed) 2011-11-15T17:35:56 *** NightExcessive has joined #aichallenge 2011-11-15T17:39:54 *** xathis has quit IRC () 2011-11-15T17:40:02 *** bergmark has quit IRC (Ping timeout: 245 seconds) 2011-11-15T17:42:23 *** revell has joined #aichallenge 2011-11-15T17:42:36 hi all 2011-11-15T17:42:57 has anyone said anything since retybok_ left? 2011-11-15T17:43:32 can anyone help me configure the "tutorial.sh" script so no food is dropped on map 2011-11-15T17:43:38 can anyone help me configure the "tutorial.sh" script so no food is dropped on map? 2011-11-15T17:44:02 adding "--food_rate 0,0" doesn't seem to work 2011-11-15T17:44:05 --food none instead of whatever it's set to now 2011-11-15T17:44:12 and get rid of --scenario 2011-11-15T17:44:47 food rate will probably work too, but --scenario places stuff on the map at the start 2011-11-15T17:45:18 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-15T17:45:25 awesome! 2011-11-15T17:45:29 thanks! ^^ 2011-11-15T17:48:28 *** revell has quit IRC (Quit: Page closed) 2011-11-15T17:52:11 *** HaraKiri has quit IRC () 2011-11-15T17:52:15 *** peyton has quit IRC (Quit: peyton) 2011-11-15T17:55:02 *** twymer has quit IRC (Ping timeout: 245 seconds) 2011-11-15T17:57:20 *** peyton has joined #aichallenge 2011-11-15T18:04:45 *** ikaros has joined #aichallenge 2011-11-15T18:06:07 *** luizribeiro has joined #aichallenge 2011-11-15T18:12:20 100% influence or potential map version running now 2011-11-15T18:12:35 no direct bfs or path finding... 2011-11-15T18:12:51 delt0r_: replay? 2011-11-15T18:13:01 still playing... 2011-11-15T18:13:10 don't know how well it will do tbh 2011-11-15T18:13:31 but i am tring simple things to see what works... rather than winning 2011-11-15T18:13:45 shame i can only enter once at the end ;) 2011-11-15T18:14:01 :( 2011-11-15T18:14:08 got a bug 2011-11-15T18:14:22 my ants are walking over themselfs ... damit 2011-11-15T18:14:30 that hurts 2011-11-15T18:15:52 *** foRei has quit IRC (Quit: Bye) 2011-11-15T18:16:21 and potentials are propagating through water... not cool. Stupid bugs 2011-11-15T18:16:41 hehehe 2011-11-15T18:17:02 That's how it goes all the time right? You make some cool improvements, and suddenly your ants are failing so hard... 2011-11-15T18:17:06 then you go hunting for bugs 2011-11-15T18:17:13 and fine tune the settings 2011-11-15T18:17:26 and in the end it is working better (at least, that's what we all hope) :) 2011-11-15T18:17:41 yes, my new improved version always loses to my old one 2011-11-15T18:17:53 :) 2011-11-15T18:20:27 I do testing whenever I fix a significant "bug" now, sometimes they are really features 2011-11-15T18:22:31 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-15T18:22:45 :D 2011-11-15T18:22:47 haha 2011-11-15T18:23:25 *** mj41 has quit IRC (Ping timeout: 240 seconds) 2011-11-15T18:25:25 *** twymer has joined #aichallenge 2011-11-15T18:26:34 *** bergmark has joined #aichallenge 2011-11-15T18:29:30 delt0r_: any progress? 2011-11-15T18:31:52 *** epicmonkey has quit IRC (Ping timeout: 240 seconds) 2011-11-15T18:33:53 well if more bugs count as progress 2011-11-15T18:33:55 then yes 2011-11-15T18:34:06 apparently they do 2011-11-15T18:34:10 Anilm3: are you interested in the results 2011-11-15T18:34:19 since i cannot beat my buggy bot with my debugged bot 2011-11-15T18:34:21 a1k0n: then i am the progress master 2011-11-15T18:38:26 aichallenge: Janzert epsilon * r931e4fb / (8 files in 5 dirs): 2011-11-15T18:38:26 aichallenge: Merge pull request #367 from jackpal/epsilon 2011-11-15T18:38:26 aichallenge: Add starter pack and server support for the Dart language. - http://git.io/zNEMjQ 2011-11-15T18:38:33 *** grom358 has joined #aichallenge 2011-11-15T18:38:37 *** lavalamp has joined #aichallenge 2011-11-15T18:42:39 aichallenge: janzert epsilon * r0f8755f / setup/retrieve_languages.py : Add dart to retrieve_languages script - http://git.io/yh15bQ 2011-11-15T18:42:43 delt0r_: I am interested 2011-11-15T18:42:59 In how it performs 2011-11-15T18:45:31 mine suck big time 2011-11-15T18:45:47 s 2011-11-15T18:46:42 they say qt is easy, but still, im a bit confused 2011-11-15T18:48:01 mcstar: but it is 2011-11-15T18:48:06 mcstar: what's the problem? 2011-11-15T18:48:31 nothing, just that ive never done this really before 2011-11-15T18:48:39 remember, im not a CS guy 2011-11-15T18:48:54 mcstar: do you know python? 2011-11-15T18:49:19 my GUI experience is some windows guis, that i dont remember anymore, and an opengl madelbrot viewer 2011-11-15T18:49:30 yes 2011-11-15T18:49:39 not like a guru, but i know it to get along 2011-11-15T18:49:42 you could try pyqt, it's much easier to test and stuff 2011-11-15T18:49:46 but thx, i dont want pyqt now 2011-11-15T18:50:02 yeah, im sure cause you can develop dynamically 2011-11-15T18:50:09 but im sticking to c++ 2011-11-15T18:50:09 also.. a lot of people seem to recomend qtcreator 2011-11-15T18:50:14 i need the experience 2011-11-15T18:50:20 i used that 2011-11-15T18:50:36 but it doesnt support full c++11 syntax 2011-11-15T18:50:46 so i switched to kdevelop,. which is even better 2011-11-15T18:50:52 mcstar: you mean qt the widget tool kit? 2011-11-15T18:50:57 yes 2011-11-15T18:51:03 lol, clang_complete has been annoying me due to c++11 stuff 2011-11-15T18:51:08 all gui toolkits sux 2011-11-15T18:51:09 ALL 2011-11-15T18:51:11 amstan: btw i dont really need it, im not designing the gui, im programmin git 2011-11-15T18:51:31 it doesn't like the double right angle brackets for templates and makes vim underline them in red 2011-11-15T18:51:38 mcstar: it could be useful for giving you ideas what on what kind of layout/widgets you want to have 2011-11-15T18:52:07 yep, right, but i find it more elegant to actually call functions than to rely on an xml 2011-11-15T18:52:11 but this is personal 2011-11-15T18:52:42 so my point is that qt is good(seems to be) but overwhelming at the same time 2011-11-15T18:52:51 ah, i did a bit of raw X too 2011-11-15T18:52:59 did a plotter for a project 2011-11-15T18:53:20 i'm not saying to use the stuff generated by qtcreator, just use it as documentation and rapid sketching of what you want 2011-11-15T18:53:21 *** Anilm3 has quit IRC (Quit: Page closed) 2011-11-15T18:53:45 amstan: k, i understand you, but this gui will be quite simple 2011-11-15T18:54:03 *** okan_ has quit IRC (Quit: Page closed) 2011-11-15T18:54:16 delt0r_: which ones did you try? 2011-11-15T18:54:53 I really hate gtk/qt/tk/whatever 2011-11-15T18:54:59 haha 2011-11-15T18:55:03 when I do GUI crap I just hack it together in html/javascript 2011-11-15T18:55:06 i like opinionated people 2011-11-15T18:55:28 thestinger: so you prefer a browser interface? 2011-11-15T18:55:30 python/C++ just aren't good at dealing with asynchronous stuff like a GUI imo 2011-11-15T18:55:46 thestinger: that probably gives you horrible usability 2011-11-15T18:56:03 amstan: well, I've only made GUI stuff for myself :P 2011-11-15T18:56:04 what about tabs, tooltips, speed, memory usage? 2011-11-15T18:56:24 you can do tabs/tooltips with javascript 2011-11-15T18:56:41 yeah, but you have to code a lot more to get to the same result 2011-11-15T18:56:58 theres some new java thing javafx, whats that? 2011-11-15T18:57:08 mcstar: many over the years 2011-11-15T18:57:37 delt0r_: but now nowadays right? since you are into research again... 2011-11-15T18:57:39 not matter what... its more or less hard work to anything rite. 2011-11-15T18:57:44 but not* 2011-11-15T18:57:54 well now i stick with java... 2011-11-15T18:58:03 no need to use n different gui toolkits 2011-11-15T18:58:17 yes i still need to write them from time to time 2011-11-15T18:58:21 that program you showed me didnt even have a gui right? 2011-11-15T18:58:23 it was cli 2011-11-15T18:58:39 yes... but there is a gui for teaching with it 2011-11-15T18:58:44 k 2011-11-15T18:58:50 I will be adding more too 2011-11-15T18:59:13 i cant stand swing 2011-11-15T18:59:19 amstan: gnome 3 is basically just piles of css/javascript, so I guess you're right about the horrible usability thing 2011-11-15T18:59:20 the problem is that a lot of users will not care that it deletes there HDD, but that the font is not the same as app X 2011-11-15T18:59:22 i opened netbeans, it was horrible 2011-11-15T19:00:17 that kind of stuff is slow in all gui toolkits... every font every aa setting its a royal PITA 2011-11-15T19:00:24 i see many people write "there" while they mean "their" why is that: 2011-11-15T19:00:25 ? 2011-11-15T19:00:26 mcstar: why --from the look? 2011-11-15T19:00:35 i mean their 2011-11-15T19:00:38 ment 2011-11-15T19:00:40 whatever 2011-11-15T19:00:40 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-15T19:00:43 :D 2011-11-15T19:00:47 k, not just you 2011-11-15T19:00:57 i see it on forums all arounf 2011-11-15T19:00:59 gramma you know what 2011-11-15T19:01:10 *** mceier has quit IRC (Quit: leaving) 2011-11-15T19:01:13 delt0r_: yes, from the look, looked horrible, bad fonts 2011-11-15T19:01:17 shitty performance 2011-11-15T19:01:34 i couldnt even get it to work wo a hack though 2011-11-15T19:01:36 the font setting are easily changed... 2011-11-15T19:01:50 but yea eclipse starts with better defaults 2011-11-15T19:01:53 i have a tiling wm, awesome 2011-11-15T19:02:04 and it doesnt set a root wm something or other 2011-11-15T19:02:20 but the standard doesnt require it to, and java gui is not conformant 2011-11-15T19:02:30 so almost all my java windows showed up empty 2011-11-15T19:02:49 *** datachomper has joined #aichallenge 2011-11-15T19:03:00 however you made my point perfectly--users care about shiny... not serious bugs... so no matter how good the app, you get 10000 comments on why its crap becuase the font is not the right size and a button is not round 2011-11-15T19:03:05 i was surprised that eclipse is gtk 2011-11-15T19:03:33 delt0r_: ill upload a screenie 2011-11-15T19:03:48 i am about to go to bed 2011-11-15T19:04:48 *** Harpyon has quit IRC (Quit: Computer has gone to sleep.) 2011-11-15T19:05:25 http://imgur.com/NK349 2011-11-15T19:05:30 ok 2011-11-15T19:05:39 well, im perfectly happy with this gui 2011-11-15T19:06:15 i dont want to get rid of the toolbar, cause i can easily select functions from there 2011-11-15T19:07:11 but i totally love kde apps, but i dont use the de anymore 2011-11-15T19:07:18 mcstar: my point is that NO gui is really conforming... to whatever standard... not really.. you the gui designer has to do it. 2011-11-15T19:07:21 maybe if i had a better vga.... 2011-11-15T19:07:50 delt0r_: you misunderstood me: i was talking about the ICCM standard 2011-11-15T19:07:56 i think it is iccm 2011-11-15T19:08:04 *** Antimony has joined #aichallenge 2011-11-15T19:08:05 EWMH is the other window manage one 2011-11-15T19:08:10 window manager* 2011-11-15T19:08:24 yeah maybe that 2011-11-15T19:08:30 why do i remember iccm? 2011-11-15T19:08:33 awesome has good support for all of the standard (except for a few new things) 2011-11-15T19:08:54 mcstar: there's another standard called something like that 2011-11-15T19:09:02 i noticed 2 bugs in awesome 2011-11-15T19:09:14 first, the clock is 1 minute behind the system clock 2011-11-15T19:09:30 the widget only updates once a minute I think 2011-11-15T19:09:35 and theres a glitch with floating windows 2011-11-15T19:09:45 I really don't like how awesome does the status bar 2011-11-15T19:10:16 *** Conorach has quit IRC (Ping timeout: 240 seconds) 2011-11-15T19:10:20 i chose awesome over xmonad, because it needed less setup, and has goo mouse support 2011-11-15T19:10:29 g 2011-11-15T19:10:37 I recently switched from awesome -> i3 2011-11-15T19:10:50 yeah, tried that one too 2011-11-15T19:11:00 maybe later ill revisit 2011-11-15T19:11:33 it's nice, but it takes some getting used to the paradigm it uses for windows (a tree with vertical/horizontal splits and containers, instead of layouts) 2011-11-15T19:11:40 the really appealing thing about i3 is that its homepage feature Deborah Morgan 2011-11-15T19:11:58 gah.. what happened on fluxid?? 2011-11-15T19:12:11 replays are gone 2011-11-15T19:12:24 was just looking at one 2011-11-15T19:12:47 working fine here 2011-11-15T19:13:12 * delt0r_ really goes to bed this time 2011-11-15T19:14:41 It looks like fluxid has been wiped clean 2011-11-15T19:14:51 yeah, it was 2011-11-15T19:15:12 any reason for it? 2011-11-15T19:15:32 I think to clear out all the bots that don't play anymore 2011-11-15T19:15:44 yeah rankings were messed up. 2011-11-15T19:19:15 hmm, how hard would it be to switch from js to c++11? 2011-11-15T19:20:26 *** Rinum has joined #aichallenge 2011-11-15T19:21:49 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-15T19:26:53 amstan: i see qt prefers dynamically allocated elements 2011-11-15T19:26:56 widgets 2011-11-15T19:27:08 there are a lot of pointer members 2011-11-15T19:27:20 but i dont see any destrcutors in the examples 2011-11-15T19:27:32 so i guess qt somehow frees those resources? 2011-11-15T19:27:51 *** datachomper has joined #aichallenge 2011-11-15T19:29:26 *** Antimony has quit IRC (Ping timeout: 276 seconds) 2011-11-15T19:31:51 *** Harpyon has joined #aichallenge 2011-11-15T19:32:21 amstan, or another admin here? 2011-11-15T19:32:33 ikaros: yes? 2011-11-15T19:32:49 luas standard lib "timer" only has second resolution 2011-11-15T19:33:12 i wanted to know if its possible to install liblua-socket? 2011-11-15T19:33:31 because it offers a millisecond resolution timer.. or even more exact 2011-11-15T19:33:32 ikaros: most likely 2011-11-15T19:33:51 that would be cool.. otherwise i probably had a problem :) 2011-11-15T19:33:56 apparently, the sane version of strcat is way better than the insane one 2011-11-15T19:34:10 ikaros: are you familiar with github? 2011-11-15T19:34:17 kind of 2011-11-15T19:34:18 why? 2011-11-15T19:34:24 and python? 2011-11-15T19:34:35 not so much 2011-11-15T19:34:42 *** Antimony has joined #aichallenge 2011-11-15T19:34:44 depends on the task :O 2011-11-15T19:35:22 ikaros: the installation of the worker depends on this file: https://github.com/aichallenge/aichallenge/blob/epsilon/setup/worker_setup.py 2011-11-15T19:35:45 it would go the fastest if you were to just make the change yourself and submit a pull request 2011-11-15T19:36:52 okay i will look into this tomorrow since its very late here 2011-11-15T19:36:57 but looks easy 2011-11-15T19:37:06 thx .. if theres a problem ill ask again :) 2011-11-15T19:39:12 g'night 2011-11-15T19:39:17 ikaros: bye 2011-11-15T19:39:20 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-15T19:42:53 amstan: Why OpenJDK 6 instead of OpenJDK 7... or better yet, Sun Java 7? 2011-11-15T19:43:37 NightExcessive: because what's the most easily installed with a few characters in that file 2011-11-15T19:44:13 amstan: OpenJDK 7 can be installed by replacing the 6 with a 7 IIRC... and Sun Java 7 just has to be extracted and copied to a different folder. 2011-11-15T19:44:30 NightExcessive: can you guarantee that it won't break present submissions? 2011-11-15T19:44:35 java is the most used language right now 2011-11-15T19:44:38 amstan: The reason I ask is: Sun Java 7 has *huge* performance improvements. 2011-11-15T19:44:50 *** Israfel has quit IRC (Ping timeout: 258 seconds) 2011-11-15T19:44:58 amstan: I don't know enough about Java to guarantee anything myself. 2011-11-15T19:45:23 NightExcessive: if you're really for this, you could submit a post in the language support forums 2011-11-15T19:45:27 see what other people thin 2011-11-15T19:45:29 think* 2011-11-15T19:45:37 amstan: Alrighty. I'll go have a look. 2011-11-15T19:46:21 amstan: From the small amount of looking I did before I said anything, it looks like Sun Java 7 is completely backwards compatible with Java 6. 2011-11-15T19:46:53 NightExcessive: you're more than welcome to push for this, you just have to convince janzert that nothing bad will happen 2011-11-15T19:47:03 and that it's easy to do 2011-11-15T19:47:12 amstan: Alrighty. :) 2011-11-15T19:48:52 *** analyst74 has quit IRC (Quit: HydraIRC -> http://www.hydrairc.com <- It'll be on slashdot one day...) 2011-11-15T19:49:35 *** Rinum has quit IRC (Ping timeout: 265 seconds) 2011-11-15T19:53:26 *** McLeopold has left #aichallenge 2011-11-15T19:56:15 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-15T19:58:36 *** Rinum has joined #aichallenge 2011-11-15T20:02:06 *** mikewintermute has joined #aichallenge 2011-11-15T20:02:18 delt0r_: well.. your pile-of-ants-under-a-hill defense seems to hvae worked: http://ants.fluxid.pl/replay.319 2011-11-15T20:02:41 *** mcstar has quit IRC (Quit: WeeChat 0.3.6) 2011-11-15T20:06:19 a1k0n: Until he runs out of ants under the hill :) 2011-11-15T20:06:38 which seems to be after the turn limit 2011-11-15T20:06:46 Yeah. 2011-11-15T20:06:47 i need to do some tuning to park some ants outside the hill 2011-11-15T20:06:58 they don't seem to realize that moving onto the hill is certain death 2011-11-15T20:07:45 hahaha 2011-11-15T20:08:13 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-11-15T20:08:16 and the rest of my ants.. i have no idea 2011-11-15T20:08:22 this thing is really buggy now 2011-11-15T20:08:26 that's why it's working better 2011-11-15T20:09:09 *** userjjb has joined #aichallenge 2011-11-15T20:11:36 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-15T20:14:14 Trying to figure out the Go libraries confuses me... why must making a sha256 hash be so complicated? 2011-11-15T20:14:28 :p "ants not razing hills" 2011-11-15T20:15:10 NightExcessive: it's not :P 2011-11-15T20:15:30 NightExcessive: they don't really have much howto stuff or examples in the docs yet 2011-11-15T20:15:33 thestinger: Oh? 2011-11-15T20:15:53 thestinger: I figured out sha256.New() and now I'm searching through the docs to find the Digest thing 2011-11-15T20:16:34 md5sum := md5.New(); md5sum.Write(bytes) 2011-11-15T20:16:54 md5sum.Sum() then gets you the hash 2011-11-15T20:17:15 thestinger: Ahhh. Thanks! :) 2011-11-15T20:17:18 if you want it in hex you have to use Sprintf or whatever 2011-11-15T20:18:09 combat w/ CD is really hard x.x whereas food gathering with CD is simple 2011-11-15T20:18:12 fmt.Sprintf("%x", md5sum.Sum()) would get you the e691414cadf3cc226ad078a43630e948 style string 2011-11-15T20:18:25 thestinger: Is that "PHP style?" 2011-11-15T20:18:32 thestinger: Sorry, not too big into crypto. 2011-11-15T20:18:41 dunno 2011-11-15T20:18:53 the libraries I've used call it the hex digest I think 2011-11-15T20:18:55 * avdg wonders where that line of php is written 2011-11-15T20:20:29 I probably shouldn't have done my TCP server in a language that I barely know :v 2011-11-15T20:20:40 Go is very nice for servers though 2011-11-15T20:20:49 Mhm. That's why I used it. 2011-11-15T20:20:52 I guess stackless python would be nice too 2011-11-15T20:20:55 Things like netchan are soooo useful. 2011-11-15T20:20:57 I can't stand the event-based stuff anymore 2011-11-15T20:21:00 *** delt0r_ has joined #aichallenge 2011-11-15T20:21:01 I don't like Python... at all. 2011-11-15T20:23:05 *** Antimony has joined #aichallenge 2011-11-15T20:24:38 i guess i'm not doing combat w/ CD afterall... too complicated 2011-11-15T20:24:43 *** Harpyon has quit IRC (Quit: Computer has gone to sleep.) 2011-11-15T20:25:30 *** Jak_o_Shadows has quit IRC (Read error: Connection reset by peer) 2011-11-15T20:26:20 *** b0rder_ has joined #aichallenge 2011-11-15T20:28:29 http://forums.aichallenge.org/viewtopic.php?f=24&t=1914 2011-11-15T20:29:32 *** choas_ has joined #aichallenge 2011-11-15T20:29:41 overloaded operators ftw 2011-11-15T20:30:03 why does he wants to compare objects by converting them to strings first? 2011-11-15T20:30:14 no idea lol 2011-11-15T20:30:22 I don't even know how to respond to that. 2011-11-15T20:30:25 lol 2011-11-15T20:31:08 i know i should be replying to that.. but like.. 2011-11-15T20:31:23 ... 2011-11-15T20:31:27 amstan: should I? 2011-11-15T20:31:34 *** Jak_o_Shadows has joined #aichallenge 2011-11-15T20:31:36 aarossig: if you can be nice about it 2011-11-15T20:31:47 suggest to overload the = operator 2011-11-15T20:31:48 I can.. I guess I better make an account on he forums lol 2011-11-15T20:31:54 the* 2011-11-15T20:31:56 aarossig: you have one, you're a mod 2011-11-15T20:32:06 thing is.. I don't know VB so i'll probably be wrong on this stuff 2011-11-15T20:32:13 amstan: I noticed a security issue. Want me to send you it here or in a PM? 2011-11-15T20:32:24 NightExcessive: pm's good 2011-11-15T20:33:08 *** choas has quit IRC (Ping timeout: 276 seconds) 2011-11-15T20:33:47 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-11-15T20:33:51 amstan: I ordered 10 monitors for 5 dual-screen setups, the monitors are VGA and the motherboard is DVI-D 2011-11-15T20:33:58 amstan: fffffffuuuuuuuuuuuu 2011-11-15T20:34:58 i'll leave you to reply to this VB post.. I don't remember my account credentials and I really have no reputable name on the forums anyway... 2011-11-15T20:35:28 you can probably rename it 2011-11-15T20:38:04 contestbot: seen janzert 2011-11-15T20:38:05 amstan: janzert was last seen in #aichallenge 6 hours and 56 seconds ago: or there too? 2011-11-15T20:40:16 oh man.. cannot unsee: tc pans.com 2011-11-15T20:45:49 need something? 2011-11-15T20:46:07 amstan ? 2011-11-15T20:50:31 time to rewrite all my movement code again... combat screws up everything xD 2011-11-15T20:51:20 *** yoden has joined #aichallenge 2011-11-15T20:51:48 janzert: hmm, what happened to workers 91-80, we're only using workers under id 80 2011-11-15T20:52:57 one ec2 zone is bring a worker up for a short time then shutting it down right now 2011-11-15T20:57:16 *** treeform has quit IRC (Read error: Operation timed out) 2011-11-15T20:59:12 rats. my bugs are still no match for _flag 2011-11-15T21:02:06 *** djr_ has joined #aichallenge 2011-11-15T21:02:35 what is the game rate on profile page on aichallenge? 2011-11-15T21:02:55 under Recent Submissions 2011-11-15T21:03:13 *** analyst74 has joined #aichallenge 2011-11-15T21:03:28 grom358: see bottom: http://aichallenge.org/server_stats.php 2011-11-15T21:04:14 amstan: should sure I follow 2011-11-15T21:04:24 amstan: not sure I follow* 2011-11-15T21:04:37 grom358: it's games per minute being played for the contest 2011-11-15T21:04:39 *** Antimony has joined #aichallenge 2011-11-15T21:04:53 the higher that, the more chances that you're going to play more often 2011-11-15T21:05:02 amstan: yes but how does that relate to the Game rate column in recent submissions on my profile page? 2011-11-15T21:05:26 like when it says 333 what does that mean 2011-11-15T21:05:42 oh 2011-11-15T21:05:45 i'm not sure about that 2011-11-15T21:06:03 that game rate is average minutes between games for that submission 2011-11-15T21:06:41 if you hover your mouse over the column name it will give a description 2011-11-15T21:06:43 janzert: why is that useful for contestants? 2011-11-15T21:06:53 * janzert shrugs 2011-11-15T21:06:56 janzert: thanks 2011-11-15T21:07:13 I didn't put it there I just know what it is ;) 2011-11-15T21:08:25 is turntime measured from when the engine says "turn #" and starts giving you input or when it says "go" indicating it's finished? 2011-11-15T21:08:50 u_: basically from when you start receiving input 2011-11-15T21:11:04 *** Larose has quit IRC (Ping timeout: 240 seconds) 2011-11-15T21:11:22 *** Larose has joined #aichallenge 2011-11-15T21:12:54 *** Israfel has joined #aichallenge 2011-11-15T21:13:24 apparently dragging text into terminal takes 100ms... 2011-11-15T21:13:30 i hope 2011-11-15T21:13:31 :l 2011-11-15T21:13:46 u_: why would you hope such a slow thing? 2011-11-15T21:14:18 either that or my bot is taking 100ms to read ~100 ant's worth of data 2011-11-15T21:15:40 i changed it to just ignore every line and count the ms until "go," and it still takes about 100 ms 2011-11-15T21:15:57 *** grwip has quit IRC (Quit: Leaving) 2011-11-15T21:16:34 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-11-15T21:16:41 so either line <- getLine ; if "go" `isPrefixOf` line takes a millisecond in haskell or that. 2011-11-15T21:18:44 Patience pays! http://ants.fluxid.pl/replay.328 2011-11-15T21:19:09 around turn 220, a lone red ant at (58, 22) 2011-11-15T21:19:56 http://aichallenge.org/visualizer.php?game=97772 purple is quite a character 2011-11-15T21:21:57 yeah.. it has to be the way i fed it the test input. my bot should be receiving and taking to heart plenty of "w r c"'s in actual games, and it's plenty snappy there 2011-11-15T21:22:37 nice maze game dominance 2011-11-15T21:22:44 *** datachomper has joined #aichallenge 2011-11-15T21:23:03 you are using some sort of ML for this? 2011-11-15T21:23:08 it surprises me how purple wasted ants by moving them back and forth in the beginning yet still one 2011-11-15T21:23:28 haskell 2011-11-15T21:24:09 yyyyeah i give it ~50 times more test data and it still takes 100 ms 2011-11-15T21:24:14 :D 2011-11-15T21:29:11 fluxid must have been reset since my bot isn't ranked 350th anymore ;-) 2011-11-15T21:29:18 *** GeorgeSebastian has joined #aichallenge 2011-11-15T21:29:36 it just got a whole lot more addictive 2011-11-15T21:31:17 *** datachomper has joined #aichallenge 2011-11-15T21:37:03 *** roflmao has quit IRC (Ping timeout: 248 seconds) 2011-11-15T21:38:21 *** treeform has joined #aichallenge 2011-11-15T21:43:01 *** Nealefelaen has joined #aichallenge 2011-11-15T21:45:45 *** mleise has quit IRC (Ping timeout: 260 seconds) 2011-11-15T21:47:32 *** Fandekasp has quit IRC (Ping timeout: 245 seconds) 2011-11-15T21:47:44 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-15T21:49:28 *** treeform has quit IRC (Remote host closed the connection) 2011-11-15T21:52:45 Hum. Making a distributed TCP server set in Go is harder than I thought. :x 2011-11-15T21:52:59 I'm sure that implementing TrueSkill in Go won't quite be easy, either. 2011-11-15T21:53:25 *** Fandekasp has joined #aichallenge 2011-11-15T21:53:48 *** Rinum has quit IRC (Ping timeout: 265 seconds) 2011-11-15T21:58:27 *** JorgeB has quit IRC (Quit: Computer has gone to sleep.) 2011-11-15T22:01:53 NightExcessive: make it call external programs for that 2011-11-15T22:02:07 amstan: My workers don't have the RAM to run Java 2011-11-15T22:02:30 yep, java's a bitch to support 2011-11-15T22:02:37 So I'll probably just port the Java library to Go. 2011-11-15T22:02:53 isn't the library javascript or php or something? 2011-11-15T22:02:59 php i think, the latest one we used 2011-11-15T22:04:29 I know one of them in the Git repo is PHP and the other is Java... and I can't remember what the third is. Python? 2011-11-15T22:04:34 yes 2011-11-15T22:04:46 And I remember people in here were saying the PHP and Python ones didn't work properly. 2011-11-15T22:06:12 Plus I like the idea that my workers have absolutely no requirements for software- since Go is statically linked. 2011-11-15T22:06:26 I simply compiled it on one Debian 6 64-bit machine and then distribute it to the others. 2011-11-15T22:06:35 s/compiled/compile/ 2011-11-15T22:07:04 what are you building anyway? workers? for what? 2011-11-15T22:07:08 ... with the view range/radius (square root of 77, ~8.77), why can't the ants see 9 squares away? 2011-11-15T22:07:27 because 9 squared is 81 which is more than 77 2011-11-15T22:07:42 so you always round down? 2011-11-15T22:07:48 there's no rounding 2011-11-15T22:07:52 that's why it's squared 2011-11-15T22:08:11 if you're square rooting, you're doing it wrong 2011-11-15T22:08:27 compare everything while squared, so it's integers 2011-11-15T22:08:43 mm, no, I have it as: if(x^2)+(y^2)>r^2 2011-11-15T22:08:47 amstan: I'm making a TCP server, but more distributed- partially for fun, partially because I don't want to run arbitrary servers on my web hosting server and partially because I think it might make interesting results. 2011-11-15T22:09:14 but why do you need a distributed tcp server? 2011-11-15T22:09:31 you can easily run a ton of users one one small server 2011-11-15T22:09:55 amstan: I don't *need* one :P 2011-11-15T22:10:00 amstan: But it's something to do. 2011-11-15T22:10:07 make a bot? 2011-11-15T22:10:34 amstan: Already did- then got fed up because it wouldn't work. 2011-11-15T22:12:12 contestbot: later tell mcleopold you have spam on paste.aichallenge.org, any ideas on a fix? 2011-11-15T22:12:12 amstan: Yep. 2011-11-15T22:12:51 amstan: add a question about the ants game :) 2011-11-15T22:13:08 thestinger: hmm? 2011-11-15T22:13:09 amstan: cppreference has something where they ask a C++ question before you can edit, and the Arch Linux wiki does something similar 2011-11-15T22:13:11 cheers amstan, not using the square root makes it make much more sense =) 2011-11-15T22:13:26 so you just need to ask something anyone doing this contest would know 2011-11-15T22:13:54 keeps out the persistent human spammers in addition to bots 2011-11-15T22:13:57 thestinger: that could fix it i guess 2011-11-15T22:14:12 temporarly, unless they're interested 2011-11-15T22:14:19 persistent human spammers? how? 2011-11-15T22:15:18 well, I guess it depends on how persistent they are :P 2011-11-15T22:15:49 the Arch Wiki has a question that you basically answer by copypasting it into a *nix command line, and it got rid of all the human spammers 2011-11-15T22:16:29 not sure about what could be used for this 2011-11-15T22:17:26 I guess you could just require an account (maybe with an uploaded bot if they go the spammers just start registering, they'll only go so far) 2011-11-15T22:20:07 I managed to add 150 lines of code "refactoring"... 2011-11-15T22:23:21 I seem to have written some new code with a bunch of obvious bugs that may still be better than the old version I happen to be testing against 2011-11-15T22:24:32 I think my new code fixed a bunch of bugs, but it appears to be inferior 2011-11-15T22:24:59 maybe the map I'm testing on is just rigged against the first player... 2011-11-15T22:26:20 *** Larose has quit IRC (Read error: Connection reset by peer) 2011-11-15T22:32:10 *** replore has joined #aichallenge 2011-11-15T22:33:12 *** RyokuHasu has joined #aichallenge 2011-11-15T22:33:48 *** roflmao has joined #aichallenge 2011-11-15T22:34:05 http://ants.fluxid.pl/replay.517 um 2011-11-15T22:35:00 I don't think I've ever had 900 ants 2011-11-15T22:35:38 yes, this is the reason why I'm building static defenses into my AI =D 2011-11-15T22:36:45 blue's cloggin' the tubes 2011-11-15T22:37:11 roflmao: Your bot is very nice at combat. 2011-11-15T22:37:20 not really but thanks :) 2011-11-15T22:37:22 i like your movement style roflmao 2011-11-15T22:37:42 and goddammit, i thought i was gonna be the first one with a buddy system 2011-11-15T22:38:15 I'm just trying to rip it off of A 2011-11-15T22:38:16 ha 2011-11-15T22:38:28 rip it off A? 2011-11-15T22:38:38 A is a player who has a incredible combat system 2011-11-15T22:38:39 just amazing 2011-11-15T22:38:47 top of leaderbords on fluxid 2011-11-15T22:39:02 A too good :( 2011-11-15T22:39:21 http://ants.fluxid.pl/replay.535 2011-11-15T22:39:29 I have no idea how he does it so he never loses an engagement 2011-11-15T22:39:33 in the sense that when his bots zap 2011-11-15T22:39:38 he always zaps more bot than the enemy 2011-11-15T22:39:40 so he keeps his! 2011-11-15T22:40:12 but thanks for the compliments all 2011-11-15T22:41:36 roflmao: He seems to be using some kind of flocking technique. 2011-11-15T22:42:04 And a really good one, at that. 2011-11-15T22:42:12 could be but I don't think that's what's making him the best player 2011-11-15T22:42:20 he simply has by far the best combat system developed 2011-11-15T22:42:26 look at the small individual engagements 2011-11-15T22:42:35 he always sets it up so he attacks with n+1 ants 2011-11-15T22:42:44 only move to a square if any square that can attack that square is being attacked by at least two of your ants 2011-11-15T22:42:45 Mh. Good point. 2011-11-15T22:43:08 any ant^ instead of any square^ 2011-11-15T22:43:19 enemy ant^ =P 2011-11-15T22:43:35 Sometimes the fact that Go errors if you don't use an import is *really* frustrating during debugging. 2011-11-15T22:44:24 *** Rinum has joined #aichallenge 2011-11-15T22:44:38 *** Larose has joined #aichallenge 2011-11-15T22:44:42 that's a good idea neale 2011-11-15T22:45:07 it seems to be the base rule that would give that behaviour? 2011-11-15T22:45:19 man, a week later and still no idea how to handle combat 2011-11-15T22:45:27 and once I've gotten my bot up to the attacking stage I'm going to try to implement it =D 2011-11-15T22:45:33 neale makes snse 2011-11-15T22:45:58 roflmao: that's sort of what I'm trying to implement atm 2011-11-15T22:46:31 only move to a square if any square that can attack the square you're moving to 2011-11-15T22:46:32 The only thing hard about that to implement is predicting the enemy- especially if it's RandomBot... 2011-11-15T22:46:39 is being attacked by MORE of your ants than your enemies ants? 2011-11-15T22:46:42 I edited all of my movement code, so all movement choices are now made at the end of the turn, and then combat stuff changes/gets rid of moves 2011-11-15T22:46:55 thestinger: does it work? 2011-11-15T22:47:25 well, it's no different than what I had before atm (strcat_sane) 2011-11-15T22:47:40 but I had to refactor it all like that so I can actually implement what I want >.< 2011-11-15T22:47:45 heh 2011-11-15T22:47:53 I think I need to rebuild my turn from the ground up 2011-11-15T22:47:56 I'll do that next iteration 2011-11-15T22:48:07 I like your idea of modifying moves based on combat 2011-11-15T22:48:31 *** userjjb has quit IRC (Quit: Page closed) 2011-11-15T22:50:38 hey, here's an idea 2011-11-15T22:50:45 let's all team up and make a super bot 2011-11-15T22:50:47 that'll beat ant 2011-11-15T22:50:57 A* 2011-11-15T22:51:25 I don't think he can be beaten :P 2011-11-15T22:51:28 :( 2011-11-15T22:51:39 I just wish he'd tell us WHAT he is using? 2011-11-15T22:51:41 huy! 2011-11-15T22:51:44 just the concept 2011-11-15T22:51:50 well, at combat 2011-11-15T22:51:50 does anyone have a replay of him against a crashed/timedout bot on a narrow map? =) 2011-11-15T22:52:18 if he can't deal with static ants I might be able to beat him with my idea 2011-11-15T22:52:25 strcat_sane can actually handle static enemies with proper combat now 2011-11-15T22:52:35 but the code is almost useless against moving ants 2011-11-15T22:52:57 I'm going to have different classes of ants, so both moving and static 2011-11-15T22:53:07 Is it bad that a Go program that prints "Hello world!\n" is 1.1MB? :v 2011-11-15T22:53:09 defender ants just hold strategic locations =) 2011-11-15T22:53:11 thestinger: YES, you really need prediction for your combat to be flawless 2011-11-15T22:53:16 which just makes things very hard 2011-11-15T22:53:17 NightExcessive: lol =D 2011-11-15T22:53:22 NightExcessive: run 'strip -s' on it 2011-11-15T22:53:33 well, it's big because it's statically linked 2011-11-15T22:53:44 hey, here's an idea! 2011-11-15T22:53:45 thestinger: Is that included in Go? (I'm running on a Windows laptop with the weeklies) 2011-11-15T22:53:59 treat combat radius as one square bigger than it is in reality 2011-11-15T22:54:07 or two 2011-11-15T22:54:12 by increasing it by one you plan for any enemy moves 2011-11-15T22:54:14 roflmao: yeah, I'm doing that with my diffusion maps =) 2011-11-15T22:54:16 roflmao: Tw- darn you a1k0n! 2011-11-15T22:54:18 *** dvladim has joined #aichallenge 2011-11-15T22:54:21 why two? 2011-11-15T22:54:47 NightExcessive: well, it's a *nix program but I'm sure you can strip debugging symbols on windows 2011-11-15T22:54:59 anyway, the binary is that big b/c it's statically linked to the Go runtime 2011-11-15T22:55:02 thestinger: Probably takes too much time. It just gave me a good smile. 2011-11-15T22:55:10 you can run it on a computer w/o Go installed, just with the binary 2011-11-15T22:55:21 thestinger: Yeah, I figured. That's why I'm using it- compile it once and then run it on any system with the same architecture/OS. 2011-11-15T22:55:22 *** Larose has quit IRC (Read error: Connection reset by peer) 2011-11-15T22:55:26 a C program compiled with 'gcc -static' is similar in size with glibc 2011-11-15T22:55:59 sooo I have nothing but eating, combat and exploration in this test version 2011-11-15T22:56:10 currently it's at 150 ants to 15 2011-11-15T22:56:18 where 15 is a few versions back on my main line 2011-11-15T22:56:27 *** TheLinker has quit IRC (Quit: Bye) 2011-11-15T22:56:27 so it's not attacking, just milling around 2011-11-15T22:56:37 NightExcessive: yup, 759K for a C hello world compiled with -static -O2 2011-11-15T22:56:50 strip -s gets that down to 682K 2011-11-15T22:57:14 *** NightExcessive_ has joined #aichallenge 2011-11-15T22:57:32 The web client crashes when you use /list... :l 2011-11-15T22:58:17 go hello world compiled with 6g/6l is 1.2MB (down to 782K with strip -s) 2011-11-15T22:58:24 so it's in the same ballpark :) 2011-11-15T22:58:37 thestinger: 1.1MB on Windows amd64 2011-11-15T22:58:44 :P 2011-11-15T22:59:12 well, it might be the MiB v MB thing 2011-11-15T22:59:16 I bet they're the same size 2011-11-15T22:59:22 well, almost the same size 2011-11-15T22:59:24 Ahhh. Good point. 2011-11-15T22:59:42 You never know if Windows has its units abbreviations right. :P 2011-11-15T23:00:01 *** NightExcessive has quit IRC (Ping timeout: 265 seconds) 2011-11-15T23:00:35 *** NightExcessive_ is now known as NightExcessive 2011-11-15T23:01:09 NightExcessive: gccgo does dynamic linking by default, and it's 22K (16K stripped) :) 2011-11-15T23:01:38 thestinger: Does it have a static linking option? Aand doesn't it have not-so-great support for goroutines? 2011-11-15T23:02:03 well, you can do any of the gcc stuff (so -static should work) - and yeah, I think it spawns a thread for each goroutine 2011-11-15T23:02:37 I heard channels weren't perfect either in gccgo, but I'm loving Go so far. 2011-11-15T23:02:44 I think it might be fixed, but they have to wait for a whole gcc release cycle to get the new version out 2011-11-15T23:03:13 What's the goroutine option for 6g? -maxcpus or something? 2011-11-15T23:03:29 Yay Go! it is teh awesome for sure. 2011-11-15T23:03:29 you can import the runtime package and set GOMAXPROCS 2011-11-15T23:03:47 the goroutines get multiplexed over the number of threads GOMAXPROCS is set to (1 by default) 2011-11-15T23:03:54 Ahh. Is there any way to grab the current number of (physical) cores? 2011-11-15T23:04:05 *** djr__ has joined #aichallenge 2011-11-15T23:04:09 I'm not sure, it might actually make your code slower 2011-11-15T23:04:27 *** djr__ has quit IRC (Read error: Connection reset by peer) 2011-11-15T23:04:37 goroutines/channels won't require locking since they're concurrent (context switches), but not parallel 2011-11-15T23:04:41 Why's that? Context switching? 2011-11-15T23:04:48 as soon as you have multiple threads, you get lots of locking 2011-11-15T23:05:02 Yeah... I suppose it really just depends on the application. 2011-11-15T23:07:29 I think they've actually sped up channels a lot, so it might be better than when I last tried 2011-11-15T23:08:52 thestinger: Is it possible to compile/link for Windows from Linux? (By setting GOOS) 2011-11-15T23:09:42 I haven't tried that, but you can definitely cross-compile with the standard Go compilers, at least across architectures 2011-11-15T23:10:28 ... wine and cygwin? 2011-11-15T23:11:03 Nealefelaen: ah, it looks like it should work 2011-11-15T23:11:18 you're basically always cross-compiling 2011-11-15T23:11:33 *** yoden has quit IRC (Quit: Leaving.) 2011-11-15T23:12:33 http://golang.org/doc/install.html#environment 2011-11-15T23:13:08 Welp... I'm off for a bit while I upgrade my desktop to openSUSE 12.1 and reformat this laptop to openSUSE. :) 2011-11-15T23:13:19 *** dvladim has quit IRC (Read error: Operation timed out) 2011-11-15T23:13:20 *** NightExcessive has quit IRC (Quit: Ciao) 2011-11-15T23:14:24 *** Larose has joined #aichallenge 2011-11-15T23:24:51 *** Larose has quit IRC (Read error: Connection reset by peer) 2011-11-15T23:30:22 woot =D finally figured out how to get the cartesian points of the perimiter of a circle so that it matches how the engine sees it 2011-11-15T23:32:20 haha i'm in 1st on fluxid.pl however briefly, cuz A timed out in a game with me 2011-11-15T23:32:32 =D congrats 2011-11-15T23:32:47 god my bot is so broken too 2011-11-15T23:33:04 mine just crashed at turn 6... 2011-11-15T23:35:25 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-15T23:36:14 *** bmh has joined #aichallenge 2011-11-15T23:36:14 *** bmh has joined #aichallenge 2011-11-15T23:36:18 hi 2011-11-15T23:36:28 hey 2011-11-15T23:36:33 hello... =) 2011-11-15T23:36:36 a1k0n: we're giving up on diffusion. 2011-11-15T23:36:41 bmh: heh 2011-11-15T23:36:46 bmh: http://ants.fluxid.pl/ranking 2011-11-15T23:36:56 (fortuitous timeouts from A) 2011-11-15T23:37:01 I knew I should have never told you to implement a Gibbs Sampler. 2011-11-15T23:37:08 *** roflmao has quit IRC (Quit: Leaving.) 2011-11-15T23:37:10 sssssssssssssshh! 2011-11-15T23:37:24 I'm trying to persuade Eric to implement a sort-of-TSP solver for the early game. 2011-11-15T23:37:46 a1k0n: your bot still does stupid things. 2011-11-15T23:37:54 yes, it's completely broken right now 2011-11-15T23:38:03 i just killed it to try to figure out wtf is going on 2011-11-15T23:38:25 once i have enough ants (or something) it totally breaks down on collecting food 2011-11-15T23:38:55 do you trade ants to stop people from seeing your hill? 2011-11-15T23:39:29 i might. i trade ants more and more urgently as enemies approach my hill 2011-11-15T23:40:25 though it tends to just commit suicide a lot 2011-11-15T23:41:23 once I can formulate the early game foraging problem I think I'll be able to solve it fairly close to optimally 2011-11-15T23:42:22 wtf? http://ants.fluxid.pl/player/sashaSochka 2011-11-15T23:42:34 tcp rankings are definitely broken 2011-11-15T23:42:54 hah, glory is fleeting 2011-11-15T23:45:40 yay. new bug 2011-11-15T23:46:22 http://ants.fluxid.pl/replay.616 sometimes my ants just decide to chill for a few turns... 2011-11-15T23:48:35 at the opening? 2011-11-15T23:48:45 yeah, I think it's a bug in my foraging code 2011-11-15T23:48:57 that branch doesn't have defense or anything that would make them stay still 2011-11-15T23:49:14 ugh.. weird bug 2011-11-15T23:49:39 yay figured it out, foraging bug 2011-11-15T23:49:52 s/figured it out/figured out one of the causes/ 2011-11-15T23:50:39 *** kingemer has quit IRC (Read error: Connection reset by peer) 2011-11-15T23:51:06 *** u_ has quit IRC (Quit: u_) 2011-11-15T23:54:59 a1k0n: have you implemented APSP? 2011-11-15T23:56:01 (I should have payed closer attention when I was taking Planar Graph Optimization Algorithms) 2011-11-15T23:56:08 payed paid 2011-11-15T23:56:54 er. not that i know of. african people's socialist party? 2011-11-15T23:57:01 all pairs shortest path 2011-11-15T23:57:14 oh. yes, i have implemented floyd-warshall in the past, on a much smaller graph 2011-11-15T23:57:57 H^3 * W^3 seems like a lot of storage/time 2011-11-15T23:58:20 the graph is planar, manhattan and the nodes are of degree 4 at the worst 2011-11-15T23:58:56 yes... but there are still H*W graphs of size H*W to generate 2011-11-15T23:59:03 so i guess that's ^2 space, ^3 time 2011-11-15T23:59:31 *** Antimony has joined #aichallenge 2011-11-15T23:59:36 you can be lazy and just solve for a subset of the points you're interested in 2011-11-15T23:59:49 yes you could. if you're doing that, then you should just use A*.