2011-12-09T00:08:06 *** Antimony_ has joined #aichallenge 2011-12-09T00:08:36 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-12-09T00:08:45 *** Antimony_ is now known as Antimony 2011-12-09T00:10:08 *** Jak_o_Shadows has quit IRC (Ping timeout: 240 seconds) 2011-12-09T00:13:36 *** JorgeB has joined #aichallenge 2011-12-09T00:16:17 *** Areks has joined #aichallenge 2011-12-09T00:33:15 *** AntDroid has quit IRC (Ping timeout: 265 seconds) 2011-12-09T00:37:21 *** AntDroid_ has quit IRC (Quit: Page closed) 2011-12-09T00:37:46 *** AntDroid has joined #aichallenge 2011-12-09T00:41:57 BenJackson: our bots are locked in mortal combat :) 2011-12-09T00:42:57 since I don't know who on your server is "serious" and who isn't I never know if I'm trouncing some poor newb or if it's just me vs some other guy's dumbed down bot 2011-12-09T00:43:15 BenJackson: my bot is fully serious. 2011-12-09T00:44:04 is your's Wraithan? that seems to be the only one I'm p[laying 2011-12-09T00:44:07 I could really use a function that tells me which square are currently visible. 2011-12-09T00:44:14 no, Skynet 2011-12-09T00:44:19 I wish my current bot was as reliable! 2011-12-09T00:44:46 bmh: what langauge? c++ starter computes that 2011-12-09T00:44:48 though inefficiently 2011-12-09T00:44:53 Haskell 2011-12-09T00:44:57 ah 2011-12-09T00:46:11 hm. I wonder why they twitch so much 2011-12-09T00:52:15 *** Fandekasp has joined #aichallenge 2011-12-09T00:52:20 *** bmh has left #aichallenge 2011-12-09T00:54:37 *** antimatroid has joined #aichallenge 2011-12-09T00:57:08 *** Palmik has joined #aichallenge 2011-12-09T01:06:13 *** avdg has quit IRC (Quit: Leaving.) 2011-12-09T01:07:27 *** dom7b5 has quit IRC (Ping timeout: 240 seconds) 2011-12-09T01:10:40 *** Antimony has quit IRC (Ping timeout: 255 seconds) 2011-12-09T01:11:46 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-12-09T01:22:35 *** Antimony has joined #aichallenge 2011-12-09T01:23:19 *** ltriant has quit IRC (Quit: Get MacIrssi - http://www.sysctl.co.uk/projects/macirssi/) 2011-12-09T01:24:12 *** LouisMartin has joined #aichallenge 2011-12-09T01:29:33 BenJackson: is my bot still running against that server? 2011-12-09T01:29:57 http://bhickey.net:2080/ 2011-12-09T01:29:59 looks like no 2011-12-09T01:29:59 * Wraithan set it to run a bunch of times to see how hte weights were working 2011-12-09T01:30:18 Ah, you just mentioned me a bit ago, thought I put too high of a number in there 2011-12-09T01:30:28 no, run all day if you like 2011-12-09T01:30:34 we only played a zillion times because no one else is 2011-12-09T01:31:38 yeah, I have a couple of my bots running a like 100000 times against a friend's private one since we are nearer in skill than most o the other tcpservers 2011-12-09T01:39:43 *** praveen has joined #aichallenge 2011-12-09T01:39:48 *** Palmik has quit IRC (Remote host closed the connection) 2011-12-09T01:41:43 *** Antimony has quit IRC (Ping timeout: 255 seconds) 2011-12-09T01:46:27 *** dr-_ has quit IRC (Ping timeout: 240 seconds) 2011-12-09T01:53:23 *** Scryer has quit IRC (Quit: Page closed) 2011-12-09T02:24:56 *** RedTurtle has quit IRC (Quit: Page closed) 2011-12-09T02:38:56 *** bhasker has joined #aichallenge 2011-12-09T02:45:49 hi 2011-12-09T02:47:37 *** pairofdice has joined #aichallenge 2011-12-09T02:49:47 *** replore_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T02:50:03 *** replore_ has joined #aichallenge 2011-12-09T02:55:30 *** mj41 has joined #aichallenge 2011-12-09T03:11:00 *** bluegaspode has joined #aichallenge 2011-12-09T03:11:15 *** NotABug has joined #aichallenge 2011-12-09T03:17:23 *** g0llum has joined #aichallenge 2011-12-09T03:18:32 *** praveen has quit IRC (Ping timeout: 265 seconds) 2011-12-09T03:24:54 *** dom7b5 has joined #aichallenge 2011-12-09T03:38:38 *** bluegaspode has quit IRC (Quit: Page closed) 2011-12-09T03:45:07 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-12-09T03:47:59 *** Fandekasp has quit IRC (Ping timeout: 252 seconds) 2011-12-09T03:49:46 *** aerique has joined #aichallenge 2011-12-09T03:59:30 *** olexs has joined #aichallenge 2011-12-09T04:07:36 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-12-09T04:10:39 *** Fandekasp has joined #aichallenge 2011-12-09T04:13:36 *** jon1 has joined #aichallenge 2011-12-09T04:31:24 *** Mystret has joined #aichallenge 2011-12-09T04:34:19 *** Conorach has joined #aichallenge 2011-12-09T04:50:24 *** kire has joined #aichallenge 2011-12-09T05:08:44 *** AntDroid has quit IRC (Ping timeout: 265 seconds) 2011-12-09T05:11:31 *** mviel_ has quit IRC (Ping timeout: 252 seconds) 2011-12-09T05:11:46 *** mviel_ has joined #aichallenge 2011-12-09T05:11:54 *** hkraal has left #aichallenge ("Leaving") 2011-12-09T05:11:57 *** hkraal_ has joined #aichallenge 2011-12-09T05:12:53 *** hkraal_ has joined #aichallenge 2011-12-09T05:13:24 *** olexs has quit IRC (Quit: Leaving.) 2011-12-09T05:14:19 *** replore_ has quit IRC (Remote host closed the connection) 2011-12-09T05:22:56 *** Weir_Doe has joined #aichallenge 2011-12-09T05:31:07 *** dom7b5 has quit IRC (Ping timeout: 240 seconds) 2011-12-09T05:33:02 *** dom7b5 has joined #aichallenge 2011-12-09T05:34:32 *** bluegaspode has joined #aichallenge 2011-12-09T05:47:04 *** bluegaspode has quit IRC (Quit: Page closed) 2011-12-09T06:03:18 *** mleise has joined #aichallenge 2011-12-09T06:04:27 bhasker chilling with knuth 2011-12-09T06:09:22 *** olexs has joined #aichallenge 2011-12-09T06:31:37 *** b0rder has joined #aichallenge 2011-12-09T06:32:13 *** Mystret has quit IRC () 2011-12-09T06:36:59 *** b0rder has quit IRC (Quit: 离开) 2011-12-09T06:48:45 *** g0llum has joined #aichallenge 2011-12-09T07:01:29 *** olexs has quit IRC (Quit: Leaving.) 2011-12-09T07:07:29 *** ikaros has joined #aichallenge 2011-12-09T07:08:07 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-12-09T07:09:27 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-12-09T07:16:08 *** Fandekasp has quit IRC (Ping timeout: 240 seconds) 2011-12-09T07:20:27 *** delt0r_ has joined #aichallenge 2011-12-09T07:26:48 *** liberforce has joined #aichallenge 2011-12-09T07:41:35 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-12-09T07:42:33 *** TheLinker has joined #aichallenge 2011-12-09T07:55:16 *** bugnuts2 has quit IRC (Remote host closed the connection) 2011-12-09T07:55:51 *** Areks has quit IRC (Ping timeout: 268 seconds) 2011-12-09T07:57:07 *** Palmik has joined #aichallenge 2011-12-09T08:06:31 *** Areks has joined #aichallenge 2011-12-09T08:14:32 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-12-09T08:34:40 *** tdubellz has quit IRC (Read error: Operation timed out) 2011-12-09T08:35:40 *** Guest1906 has joined #aichallenge 2011-12-09T08:42:32 *** Guest1906 has quit IRC (Read error: Connection reset by peer) 2011-12-09T08:43:42 *** tdubellz_ has joined #aichallenge 2011-12-09T08:44:35 *** Kommander has joined #aichallenge 2011-12-09T08:47:47 *** jcdny has joined #aichallenge 2011-12-09T08:50:10 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T08:50:51 anyone about? 2011-12-09T08:50:59 i want to run a map idea past people 2011-12-09T08:51:20 *** tdubellz_ has joined #aichallenge 2011-12-09T08:51:50 take the maze maps, but instead of opening them up so much, keep them as a "perfect maze" then "deteriorate" the water in all places, so tiny gaps end up opening up all over the place? 2011-12-09T08:55:20 http://ants.fluxid.pl/player/strcat does this page load for any of you 2011-12-09T08:55:50 I just get an empty text/html file 2011-12-09T08:56:12 and the pages for all my other bots work :P 2011-12-09T08:56:23 capitalize it 2011-12-09T08:57:18 thestinger: uhhuh 2011-12-09T08:57:33 IndexError: list index out of range 2011-12-09T08:58:34 oh well, I can see the stats in the search page 2011-12-09T08:59:18 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T08:59:18 http://ants.fluxid.pl/player/hedonismbot this is my "rewrite" 2011-12-09T08:59:33 but, uh, it looks like desync in db 2011-12-09T08:59:52 like, there is foreign key to something nonexistent 2011-12-09T09:00:15 "foreign key" 2011-12-09T09:00:28 *** tdubellz_ has joined #aichallenge 2011-12-09T09:00:37 weird 2011-12-09T09:00:38 what do you prefer, dropping db or dumbfixing it? 2011-12-09T09:01:10 might as well just dumbfix it 2011-12-09T09:01:16 i'm curious, what did people end up deciding for situations like ....a..b..c..a... 2011-12-09T09:01:23 because it will just happen again in all likelihood 2011-12-09T09:01:24 do you make moves for both 'a's at the same time? 2011-12-09T09:01:27 *** sigh has quit IRC (Remote host closed the connection) 2011-12-09T09:01:36 antimatroid: yes, I make them at the same time 2011-12-09T09:01:47 I half-pretend all enemies are on the same team until evaluating the battle 2011-12-09T09:01:49 *** Antimony has joined #aichallenge 2011-12-09T09:01:59 yeah i do that 2011-12-09T09:02:41 I paint an "axis_next" value over the grid which is where enemies could be attacking next turn 2011-12-09T09:03:02 do you try to improve a state? 2011-12-09T09:03:23 well, I just evaluate a score for the battle and try to improve that 2011-12-09T09:03:28 it sucks atm 2011-12-09T09:03:29 ie. check the slack for a dying enemy for moving a dying friend out? 2011-12-09T09:03:44 I basically have like 10 different hacks 2011-12-09T09:03:51 and they somehow turn into somewhat okay combat :P 2011-12-09T09:03:58 :p 2011-12-09T09:04:10 there we go, i dumbfixed it 2011-12-09T09:04:12 I tried a branch with combat like memetix, it didn't work 2011-12-09T09:04:18 i wonder what will break next 2011-12-09T09:04:26 strcat_next? :P 2011-12-09T09:04:44 no, in code 2011-12-09T09:05:14 but ah well, whatever 2011-12-09T09:05:18 that's a strange error though 2011-12-09T09:05:23 9 more days anyway 2011-12-09T09:05:33 the page worked last night, and I woke up a while ago and refreshed to find a blank page 2011-12-09T09:05:45 so somehow it broke overnight 2011-12-09T09:05:49 thestinger: there was a game in "game index" for your user 2011-12-09T09:05:53 but there was no game data 2011-12-09T09:06:10 oh, maybe it was one of the times you restarted the server? 2011-12-09T09:06:26 hm... 2011-12-09T09:06:33 yeah, taht could be it! 2011-12-09T09:06:39 viva la transctions! 2011-12-09T09:06:54 i guess sqlite has no transactions, or are not used 2011-12-09T09:08:01 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T09:08:42 Fluxid: http://www.sqlite.org/lang_transaction.html it definitely can 2011-12-09T09:08:48 probably just isn't use 2011-12-09T09:08:50 used* 2011-12-09T09:08:54 *** Guest5749 has joined #aichallenge 2011-12-09T09:09:12 yeah, no transactions 2011-12-09T09:09:45 if it didn't then my firefox would be completely borked :P 2011-12-09T09:10:37 I wonder if other bots in that game have broken pages too 2011-12-09T09:12:19 http://ants.fluxid.pl/replay.21097 I was so close to winning this :( 2011-12-09T09:12:30 turn 175 2011-12-09T09:12:52 so many bots in 100-200 ranking range don't have combat 2011-12-09T09:13:08 wow, the sad part is I would have won if I had fixed the bug breaking my nice hill razing 2011-12-09T09:13:08 on aichallenge 2011-12-09T09:13:21 because he was moving his ants off the hill right away, and I keep track of that 2011-12-09T09:13:30 but my new combat code breaks my sneaky raze of the hill 2011-12-09T09:13:47 oh well, that was all the motivation I needed to finally fix this :P 2011-12-09T09:15:57 *** Guest5749 has quit IRC (Read error: Connection reset by peer) 2011-12-09T09:17:04 anyway strcat_next is now what hedonismbot was :) 2011-12-09T09:17:06 *** tdubellz_ has joined #aichallenge 2011-12-09T09:19:10 there, I think I fixed it 2011-12-09T09:19:41 thestinger: famous last words of any developer 2011-12-09T09:20:00 continue; // avoid collisions, unless it's a hill flagged as razeable 2011-12-09T09:20:20 *** Areks has quit IRC (Ping timeout: 268 seconds) 2011-12-09T09:20:30 but my compiler reads my comments, I swear :) 2011-12-09T09:23:37 I guess I need to just test this with scenarios 2011-12-09T09:23:47 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T09:24:19 hi 2011-12-09T09:24:54 *** tdubellz_ has joined #aichallenge 2011-12-09T09:25:10 *** tdubellz_ has quit IRC (Killed (mrmist ())) 2011-12-09T09:26:54 nope, I caused new bugs :( 2011-12-09T09:27:01 *** Antimony_ has joined #aichallenge 2011-12-09T09:27:04 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-12-09T09:27:10 *** Antimony_ is now known as Antimony 2011-12-09T09:29:55 *** tdubellz_ has joined #aichallenge 2011-12-09T09:33:55 *** Garf has quit IRC (Read error: Connection reset by peer) 2011-12-09T09:34:43 *** Garf has joined #aichallenge 2011-12-09T09:36:31 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T09:37:39 *** tdubellz_ has joined #aichallenge 2011-12-09T09:42:08 oh, I see the problem 2011-12-09T09:42:18 I need to add score for razing a hill to my combat eval 2011-12-09T09:42:41 bugs >.< 2011-12-09T09:44:12 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T09:45:22 *** tdubellz_ has joined #aichallenge 2011-12-09T09:46:04 http://paste.aichallenge.org/Hf2k4/ against HoldBot 2011-12-09T09:46:48 http://paste.aichallenge.org/gBkF7/ and against RandomBot, where my bugfix kicks in :) 2011-12-09T09:48:00 without the bugfix, they will try to just stay next to the hill, even though RandomBot moves off right away 2011-12-09T09:48:07 and they'll die off due to collisions 2011-12-09T09:48:18 Heh, random bot should be called selfDestructBot 2011-12-09T09:49:59 thestinger: looks funny 2011-12-09T09:51:17 afair my bot does not care about ant on a hill 2011-12-09T09:51:22 i'll test it at home 2011-12-09T09:51:38 I use dead ant info to see if they moved off their hill last turn 2011-12-09T09:51:45 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T09:52:06 *** Antimony has quit IRC (Ping timeout: 248 seconds) 2011-12-09T09:52:17 so against a smart bot that leaves the ant on the hill, they just do the usual wait til it drains 2011-12-09T09:52:29 but lots of good bots move off the hill (bugnuts, CombatAnt, etc.) 2011-12-09T09:52:48 *** tdubellz1 has joined #aichallenge 2011-12-09T09:55:36 wow that's the first purpose for dead ants i learned 2011-12-09T09:55:56 i dont even parse them so far 2011-12-09T09:56:16 I also have a branch where I try to use them to figure out how aggressive each opponent is (based on my moves/position last turn) 2011-12-09T09:56:24 I dunno what to do with that info yet though 2011-12-09T09:56:29 :) 2011-12-09T09:57:17 *** amstan has joined #aichallenge 2011-12-09T09:57:17 *** ChanServ sets mode: +o amstan 2011-12-09T09:58:03 lol, it took my 17 commits to fix that bug 2011-12-09T10:00:01 *** tdubellz1 has quit IRC (Read error: Connection reset by peer) 2011-12-09T10:01:09 *** tdubellz_ has joined #aichallenge 2011-12-09T10:03:16 But if you move your ant off the hill it can be captured? 2011-12-09T10:03:52 yes 2011-12-09T10:04:02 since an enemy can swap places with you 2011-12-09T10:04:05 Oh, right 2011-12-09T10:04:06 if you stay on it, they would collide 2011-12-09T10:09:02 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T10:09:57 *** Guest57705 has joined #aichallenge 2011-12-09T10:11:19 *** rwest has quit IRC (Quit: Lost terminal) 2011-12-09T10:12:34 *** rwest has joined #aichallenge 2011-12-09T10:13:00 ok I really think it is time to find out what hotkey I keep hitting closes my terminals 2011-12-09T10:13:24 ctrl-D? 2011-12-09T10:13:28 that does EOF and closes the shell 2011-12-09T10:13:32 nope 2011-12-09T10:13:39 it closes terminator 2011-12-09T10:13:39 I hit it all the time because super-D is bound to dmenu :P 2011-12-09T10:13:48 it has something to do with tab 2011-12-09T10:13:50 heh 2011-12-09T10:14:12 I should just disable "close terminal on EOF" or whatever the setting is in zsh 2011-12-09T10:14:14 I know there is one 2011-12-09T10:14:24 I don't mind that behavior 2011-12-09T10:14:29 I do that on purpose 2011-12-09T10:14:34 this is some odd hotkey 2011-12-09T10:14:51 I hit it every few days 2011-12-09T10:15:10 it has to be a system hotkey since I disabled the one in terminator 2011-12-09T10:17:06 and I can't reproduce it per usual 2011-12-09T10:17:10 *** Guest57705 has quit IRC (Read error: Connection reset by peer) 2011-12-09T10:18:19 *** tdubellz_ has joined #aichallenge 2011-12-09T10:19:31 *** Antimony has joined #aichallenge 2011-12-09T10:19:51 *** avdg has joined #aichallenge 2011-12-09T10:23:07 yeah, I always exit the shell with ctrl-D, so it's the stupid dmenu keybind I need to change :) 2011-12-09T10:24:28 Dec 9 10:11:18 rwest-pc kernel: [3274398.274117] /usr/bin/termin[13106]: segfault at 14 ip 00007ff28838e4c1 sp 00007fffc277d340 error 4 in libgtk-x11-2.0.so.0.2400.4[7ff2882bc000+43f000] 2011-12-09T10:24:30 http://aichallenge.org/profile.php?user=2255 nice, old version is up to 17 :) 2011-12-09T10:24:32 nvm 2011-12-09T10:24:34 *** Anilm3 has joined #aichallenge 2011-12-09T10:24:34 lol 2011-12-09T10:24:39 that's why I can't find the key 2011-12-09T10:24:54 silly gtk 2011-12-09T10:24:56 use urxvt :) 2011-12-09T10:25:41 but terminator is too amazing 2011-12-09T10:25:44 why? 2011-12-09T10:26:18 because of being able to have multiple terminals in one without the use of screen 2011-12-09T10:26:21 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T10:26:27 and being able to type in all the terminals at once 2011-12-09T10:26:45 have you tried dvtm or tmux? they suck a lot less than screen 2011-12-09T10:27:42 *** tdubellz_ has joined #aichallenge 2011-12-09T10:27:48 dvtm is basically just a tiling window manager in a terminal 2011-12-09T10:28:17 looking at them now 2011-12-09T10:35:18 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T10:36:26 *** tdubellz_ has joined #aichallenge 2011-12-09T10:37:32 *** g0llum has joined #aichallenge 2011-12-09T10:37:33 rwest: I actually don't use dvtm/tmux/screen because I can do the same thing with my window manager (i3) 2011-12-09T10:37:48 you can make any "window" into a container of windows 2011-12-09T10:38:10 and it has vertical/horizontal splits and stacking for any workspace and the containers in them 2011-12-09T10:39:29 < rwest> because of being able to have multiple terminals in one without the use of screen ← urxvt has tabs 2011-12-09T10:39:49 Fluxid: i3 lets me tab (or stack) any windows :P 2011-12-09T10:40:05 there's also urxvt-tabbedex which lets you rename the tabs (from the 0, 1, 2 crap) 2011-12-09T10:40:11 thestinger: i know, i use wmii and because of that i don't use tabs in urxvt 2011-12-09T10:40:29 i love urxvt for shift+click and backscroll search 2011-12-09T10:40:32 does wmii have tabbing? or just stacking 2011-12-09T10:40:40 just stacking 2011-12-09T10:40:43 http://i3wm.org/docs/userguide.html#_changing_the_container_layout 2011-12-09T10:40:48 ah, I like tabbing more most of the time 2011-12-09T10:40:51 it's enough 2011-12-09T10:40:53 stacking wastes too much vertical space 2011-12-09T10:40:59 because I abuse it and stack 10 windows 2011-12-09T10:41:06 ;) 2011-12-09T10:41:32 thestinger: at work i have vertically positioned two monitors 2011-12-09T10:42:00 so i don't care much about stacking taking too much space ;) 2011-12-09T10:42:53 and i see whole window titles 2011-12-09T10:43:59 *** amstan has quit IRC (Ping timeout: 252 seconds) 2011-12-09T10:44:32 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T10:45:41 *** tdubellz_ has joined #aichallenge 2011-12-09T10:46:07 i3 was based on wmii, so it has the same kind of stuff 2011-12-09T10:46:49 ah, wmii uses the column thing 2011-12-09T10:47:23 yeah 2011-12-09T10:47:27 i3 uses a tree, which is nicer - but you have workspaces instead of tags 2011-12-09T10:47:39 I never really liked mixing together tags so I don't care 2011-12-09T10:47:47 *** jon1 has quit IRC (Ping timeout: 252 seconds) 2011-12-09T10:48:07 yeah, i don't mix them too, i use them as workspaces/desktops 2011-12-09T10:48:15 you might like i3 :) 2011-12-09T10:48:31 last time i tried it stuff was very confusing 2011-12-09T10:48:50 but i think that was an early stage of development 2011-12-09T10:49:06 *** aerique has quit IRC (Quit: ...) 2011-12-09T10:49:08 yeah, it's pretty new and still changes a lot between each version 2011-12-09T10:49:23 but now i have wmii nicely scripted in python and i don't see any need to change wm 2011-12-09T10:49:36 ah yeah, that's the nice thing about wmii 2011-12-09T10:49:36 well, maybe fot rendering need a bit work 2011-12-09T10:49:55 i tried to add pango support in wmii but there were some problems, maybe other time 2011-12-09T10:49:56 I was using awesomewm before so switching to i3 meant throwing out a bunch of hideous Lua crap 2011-12-09T10:50:21 silly 1-indexed arrays... 2011-12-09T10:50:22 well, wmii not is THIS MUCH scriptable 2011-12-09T10:50:49 but the way you interact with wmii via filesystem is fun 2011-12-09T10:50:59 "the unix way" 2011-12-09T10:51:20 well, the plan9 way :P 2011-12-09T10:52:39 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T10:52:54 plan9 is more hardcore 2011-12-09T10:53:04 but yes 2011-12-09T10:53:08 well, they use 9p right? 2011-12-09T10:53:11 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-12-09T10:53:13 (than unix) 2011-12-09T10:53:18 right 2011-12-09T10:53:27 yeah, plan9 has /net 2011-12-09T10:53:32 with all the sockets and stuff 2011-12-09T10:53:48 *** tdubellz_ has joined #aichallenge 2011-12-09T10:54:00 yup 2011-12-09T10:54:09 EVERYTHING IS A FILE 2011-12-09T10:54:13 as far i remember 2011-12-09T10:55:14 http://tools.suckless.org/ii/ this is pretty neat too 2011-12-09T10:55:38 I'm too happy with weechat to try making my own irc client with that though :P 2011-12-09T10:57:25 oh, the wmii maintainer is one of the two main devs of pentadactyl 2011-12-09T10:58:17 just use cat and echo, lol 2011-12-09T10:58:18 *** jon1 has joined #aichallenge 2011-12-09T11:00:10 *** JorgeB has quit IRC (Quit: Computer has gone to sleep.) 2011-12-09T11:01:35 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T11:02:45 *** tdubellz_ has joined #aichallenge 2011-12-09T11:02:46 *** Antimony has joined #aichallenge 2011-12-09T11:09:37 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T11:10:35 *** Guest69570 has joined #aichallenge 2011-12-09T11:11:21 *** chris_p has joined #aichallenge 2011-12-09T11:11:53 Pentadactyl is pretty awesome 2011-12-09T11:13:25 my bot is great at fighting timed out bots, but it sucks at combat the rest of the time 2011-12-09T11:13:37 *** JorgeB has joined #aichallenge 2011-12-09T11:13:57 Antimony: so you pretend enemies don't move? 2011-12-09T11:14:10 yep 2011-12-09T11:14:14 That's usually a bad thing to assume 2011-12-09T11:14:14 now that I predict moves, my bot sucks at dealing with unmoving enemies compared with how it handled them before 2011-12-09T11:14:20 lol 2011-12-09T11:14:24 I guess I need timeout detection 2011-12-09T11:14:37 but then a bot could _pretend_ to timeout :P 2011-12-09T11:14:45 Moving enemies are a lot more dangerous than unmoving ones :) 2011-12-09T11:14:52 so do you just try to predict a single move for each enemy, or do you consider multiple possibilities? 2011-12-09T11:14:55 *** jon1 has quit IRC (Ping timeout: 252 seconds) 2011-12-09T11:15:07 Antimony: atm, a single set of offensive moves 2011-12-09T11:15:30 I thought about doing that, but I'm not sure how well it would work 2011-12-09T11:15:32 and then I optimize my attack to deal with both unmoving or moving enemies 2011-12-09T11:15:45 I have a bunch of work started on better ways of doing it 2011-12-09T11:15:49 just need to finish one xD 2011-12-09T11:15:50 if you predict wrong you're at a disadvantage 2011-12-09T11:17:00 yeah 2011-12-09T11:17:09 and my ants are too scared to attack unmoving enemies 2011-12-09T11:17:26 it works well enough though 2011-12-09T11:17:54 http://aichallenge.org/profile.php?user=2255 http://ants.fluxid.pl/player/strcat that's the old version, which did combat globally 2011-12-09T11:18:08 http://ants.fluxid.pl/player/strcat_next this new version partitions it into battles nicely 2011-12-09T11:18:26 *** Guest69570 has quit IRC (Read error: Connection reset by peer) 2011-12-09T11:19:34 *** tdubellz_ has joined #aichallenge 2011-12-09T11:22:37 http://ompldr.org/vYm50Yw <- I get this data now, but I'm not sure how to turn it into a replacement for my current combat 2011-12-09T11:23:09 red == death, blue == trade, yellow == combat zone (ants within it can enter possibly combat with an enemy) 2011-12-09T11:23:46 can possibly enter combat* 2011-12-09T11:26:16 right now I just prevent ants from going into the red zone when coming up with an offensive set of moves, but I still have to run my old combat code or it won't work 2011-12-09T11:27:32 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T11:28:36 *** tdubellz_ has joined #aichallenge 2011-12-09T11:29:16 *** mj41 has quit IRC (Ping timeout: 240 seconds) 2011-12-09T11:30:11 *** TheLinker has quit IRC (Quit: Bye) 2011-12-09T11:34:50 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T11:35:37 *** jon1 has joined #aichallenge 2011-12-09T11:35:48 *** ttydragon has joined #aichallenge 2011-12-09T11:35:56 *** tdubellz_ has joined #aichallenge 2011-12-09T11:37:06 *** ttydragon has quit IRC (Client Quit) 2011-12-09T11:37:42 *** jstrong has joined #aichallenge 2011-12-09T11:40:00 *** korsvanloon has joined #aichallenge 2011-12-09T11:40:42 hello is this a helpdesk? 2011-12-09T11:41:43 hmm 2011-12-09T11:41:45 in some way yes. 2011-12-09T11:41:49 in other ways no 2011-12-09T11:41:54 Do you have a question? 2011-12-09T11:41:57 ug, I jsut found another major bug in my symmetry 2011-12-09T11:41:59 Or suggestion :) 2011-12-09T11:42:03 how come somany bugs keep popping up 2011-12-09T11:42:34 *** amstan has joined #aichallenge 2011-12-09T11:42:34 *** ChanServ sets mode: +o amstan 2011-12-09T11:42:36 well I've got an error after uploading my code that baffles me 2011-12-09T11:42:49 heh Antimony, I always feel that after I have a good working version, the first thing that happens when I start working on it to improve it, is that it actually breaks everywhere ;) 2011-12-09T11:42:54 korsvanloon: what language? 2011-12-09T11:43:00 python (3) 2011-12-09T11:43:30 I don't get the error when I run things on my computer (or others) 2011-12-09T11:43:34 Unfortunately, I don't know anything about Python, but I'm sure one of the 130 other users in this channel can be of assistance, just ask away and see if someone is listening ;) 2011-12-09T11:43:54 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T11:45:02 *** tdubellz_ has joined #aichallenge 2011-12-09T11:47:42 *** jstrong1 has joined #aichallenge 2011-12-09T11:47:44 *** jstrong1 is now known as roflmao 2011-12-09T11:48:10 Only when uploading (not on my computer) I get an error that says this variable "derp" is a NoneType, but I'm assigning it something like this: derp = ants.method(parameter) 2011-12-09T11:48:27 *** jstrong has quit IRC (Ping timeout: 240 seconds) 2011-12-09T11:48:58 I thought this meant that he might not know what "ants" is, but that should be impossible 2011-12-09T11:49:43 Anybody has any Python skills? 2011-12-09T11:50:25 paste the actual traceback 2011-12-09T11:51:02 due to dynamic typing, if a function doesn't explicitly return, it returns None 2011-12-09T11:52:23 wait, I just realized something: can I edit and upload ants.py? 2011-12-09T11:52:46 *** bmh has joined #aichallenge 2011-12-09T11:52:49 korsvanloon: yes 2011-12-09T11:52:53 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T11:52:57 you just need a file named MyBot.extension 2011-12-09T11:53:06 other files are optional 2011-12-09T11:53:56 I've got that, but I also included a file: ants.py, but the server probably uses that too 2011-12-09T11:53:56 *** tdubellz_ has joined #aichallenge 2011-12-09T11:53:58 thestinger: my bot is alive. Next stop? Trying to beat the stop Haskell bots :) 2011-12-09T11:55:29 now that my bot is deactivated it seems to get a lot more games :\ 2011-12-09T11:55:46 http://aichallenge.org/profile.php?user=2255 2011-12-09T11:56:13 Skynet: http://ants.fluxid.pl/replay.21400 2011-12-09T11:57:30 tcp is really hard now, and the top 100 is filled with bots that don't play anymore 2011-12-09T11:57:40 if you get in the top 50 on tcp it's pretty much a top 30 bot on the main site now 2011-12-09T11:58:00 Ahh, Skynet is still weak, we can defeat it! 2011-12-09T11:58:20 *** iglo has joined #aichallenge 2011-12-09T11:58:24 pairofdice: it's suffering from laziness :) 2011-12-09T11:58:35 thestinger: that hurts... 2011-12-09T11:58:44 silly haskell :P 2011-12-09T11:59:17 I should grep -c my code and see how much I'm abusing C++11 features 2011-12-09T11:59:21 korsvanloon: no, ants.py is not used by the server 2011-12-09T11:59:25 for you bot 2011-12-09T11:59:50 bmh: are you still having trouble with performance? 2011-12-09T12:00:02 thestinger: not when I don't compile with -prof 2011-12-09T12:00:13 ok thanks leopold 2011-12-09T12:00:21 *** McLeopold has left #aichallenge 2011-12-09T12:00:33 bmh: have you tried any of the crazy compiler switches? 2011-12-09T12:00:50 like strictness stuff 2011-12-09T12:00:58 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T12:01:36 http://paste.pocoo.org/show/OVi7outEIZelWvAAwpvR/ 2011-12-09T12:01:44 45 vectors, lol. 2011-12-09T12:02:00 thestinger: a little, if you know any more, let me know 2011-12-09T12:02:02 *** tdubellz_ has joined #aichallenge 2011-12-09T12:02:41 my bot is basically glue code between C++ standard library containers 2011-12-09T12:03:19 it's just 1400 lines if I use sed to get rid of blank lines or commented lines 2011-12-09T12:04:24 *** LouisMartin has quit IRC (Ping timeout: 265 seconds) 2011-12-09T12:04:33 *** Antimony_ has joined #aichallenge 2011-12-09T12:06:37 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-12-09T12:06:43 *** Antimony_ is now known as Antimony 2011-12-09T12:07:09 oops! I have a job! 2011-12-09T12:07:17 *** bmh has quit IRC (Quit: Page closed) 2011-12-09T12:08:26 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T12:09:35 *** tdubellz_ has joined #aichallenge 2011-12-09T12:13:32 *** mviel_ has quit IRC (Remote host closed the connection) 2011-12-09T12:17:56 *** NotABug has quit IRC (Ping timeout: 265 seconds) 2011-12-09T12:17:59 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T12:19:28 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-12-09T12:24:15 *** dr- has joined #aichallenge 2011-12-09T12:26:25 *** QuirionPT has joined #aichallenge 2011-12-09T12:29:52 *** jon1 has quit IRC (Quit: Leaving.) 2011-12-09T12:32:25 *** jon1 has joined #aichallenge 2011-12-09T12:46:40 *** liberforce has left #aichallenge 2011-12-09T12:50:44 *** kire has quit IRC (Remote host closed the connection) 2011-12-09T12:50:52 morning 2011-12-09T12:51:38 *** jon1 has quit IRC (Quit: Leaving.) 2011-12-09T12:51:59 *** flowenol has joined #aichallenge 2011-12-09T12:52:27 *** Conorach_ has joined #aichallenge 2011-12-09T12:53:07 i's 7PM here, not Am 2011-12-09T12:53:08 ;) 2011-12-09T12:53:14 *** McLeopold has joined #aichallenge 2011-12-09T12:53:16 it's* 2011-12-09T12:53:50 *** bhasker has quit IRC (Remote host closed the connection) 2011-12-09T12:53:59 *** Guest76755 has joined #aichallenge 2011-12-09T12:55:16 *** Conorach has quit IRC (Ping timeout: 240 seconds) 2011-12-09T12:55:56 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-12-09T12:57:55 http://ants.fluxid.pl/player/strcat_nihilism yay for mu being overestimated :) 2011-12-09T12:58:42 *** amstan has joined #aichallenge 2011-12-09T12:58:42 *** ChanServ sets mode: +o amstan 2011-12-09T12:59:30 *** Weir_Doe has quit IRC (Ping timeout: 265 seconds) 2011-12-09T13:00:49 *** Guest76755 has quit IRC (Read error: Connection reset by peer) 2011-12-09T13:01:20 *** Antimony has joined #aichallenge 2011-12-09T13:01:49 Fluxid: http://ants.fluxid.pl/search?name=strcat_hunt that seems weird... 2011-12-09T13:02:06 oh nvm 2011-12-09T13:02:13 lol, silly mistake - that was the search page 2011-12-09T13:05:17 *** Israfel has quit IRC (Ping timeout: 252 seconds) 2011-12-09T13:07:07 *** tdubellz_ has joined #aichallenge 2011-12-09T13:10:37 WHY MY ANTS DON'T WANT TO KILL 2011-12-09T13:10:43 FUCKING HIPPIES 2011-12-09T13:10:45 KILL KILL KILL 2011-12-09T13:10:47 :D 2011-12-09T13:10:53 :O 2011-12-09T13:10:56 THIS IS SPARTAAAA 2011-12-09T13:11:03 Fluxid has gone mad, 2011-12-09T13:11:10 for a while my combat code had a score in it I meant to make my ants kill 2011-12-09T13:11:13 *** Redgis has joined #aichallenge 2011-12-09T13:11:17 but I had accidentally instead just made them want to die 2011-12-09T13:11:35 Anilm3: yeah, i try to undertand wtf is going here http://paste.aichallenge.org/GpWxL/ 2011-12-09T13:12:50 Fluxid: yeah I have a stalemate issue atm 2011-12-09T13:13:00 cannot figure out why my positioning code doesn't move ants towards the enemy 2011-12-09T13:13:14 like 2011-12-09T13:13:20 it KNOWS it will improve things 2011-12-09T13:13:22 ant 19,51 thinks "i'll kill when going E!", but 18,51 thinks "oh no, i'll die when i go E!". And then 19,51 thinks "oh o i can't go alone because i'll actually die" 2011-12-09T13:13:25 I see it printed out 2011-12-09T13:13:34 so those two ants don't move near enemy 2011-12-09T13:13:35 ffff 2011-12-09T13:13:41 Fluxid: I had a similar problem on my first and only attempt at combat 2011-12-09T13:14:09 Fluxid: I posted my version yesterday: http://paste.aichallenge.org/xH1yL/ 2011-12-09T13:14:19 I think it's funny that the whoel thing instantly ends then 2011-12-09T13:15:09 ah i know why! 2011-12-09T13:15:19 Are those "trapped" ants to nullify visibility code? 2011-12-09T13:15:28 *** tdubellz_ has quit IRC (Read error: Connection reset by peer) 2011-12-09T13:15:43 18,51 whinks enemy ant will go N, ant if it really do that they will kill each other 2011-12-09T13:16:15 so i need some kind of prediction 2011-12-09T13:16:36 so i can weight result on how probable is enemy going that way 2011-12-09T13:16:37 ffff 2011-12-09T13:16:43 now, how do i do this 2011-12-09T13:16:45 damn 2011-12-09T13:17:05 Fluxid: http://ompldr.org/vYm52YQ 2011-12-09T13:17:10 my latest code ends up with this 2011-12-09T13:17:12 as a stalemate 2011-12-09T13:17:14 lol 2011-12-09T13:17:32 I have code to work around it but.... doesn't work 2011-12-09T13:17:37 lol 2011-12-09T13:17:48 sometimes* it doesn't work 2011-12-09T13:17:50 even more scared than mine 2011-12-09T13:17:55 no 2011-12-09T13:17:57 only sometimes 2011-12-09T13:17:59 :P 2011-12-09T13:18:04 oh and the best part? 2011-12-09T13:18:10 making my ants not chase enemies gets rid of the bug 2011-12-09T13:18:17 *** Israfel has joined #aichallenge 2011-12-09T13:18:24 so by trying to kill enemies, my ants become cowards 2011-12-09T13:18:29 http://aichallenge.org/maps.php 2011-12-09T13:18:34 I didn't know that had a ton of stats 2011-12-09T13:18:36 they say ones ants are a reflection of their master 2011-12-09T13:18:59 or was it dogs? I can't remember now 2011-12-09T13:21:11 *** Guest25605 has joined #aichallenge 2011-12-09T13:22:23 *** kire has joined #aichallenge 2011-12-09T13:26:34 oh well, I guess I need more ugly hacks :) 2011-12-09T13:27:45 *** Guest25605 has quit IRC (Read error: Connection reset by peer) 2011-12-09T13:28:38 *** tdubellz_ has joined #aichallenge 2011-12-09T13:30:26 *** tdubellz_ has quit IRC (Client Quit) 2011-12-09T13:31:16 *** UncleVasya has joined #aichallenge 2011-12-09T13:34:43 *** Weir_Doe has joined #aichallenge 2011-12-09T13:37:10 *** raemde_ has joined #aichallenge 2011-12-09T13:38:35 Evening World 2011-12-09T13:41:05 *** raemde has quit IRC (Ping timeout: 252 seconds) 2011-12-09T13:42:10 *** Antimony has quit IRC (Ping timeout: 255 seconds) 2011-12-09T13:45:15 is { matrix.at(i).at(j) } the standard safe way to access { std::vector > matrix; } ?? 2011-12-09T13:45:19 oops 2011-12-09T13:45:25 wrong channel :) 2011-12-09T13:47:22 *** Weir_Doe has quit IRC (*.net *.split) 2011-12-09T13:47:22 *** korsvanloon has quit IRC (*.net *.split) 2011-12-09T13:47:22 *** chris_p has quit IRC (*.net *.split) 2011-12-09T13:51:44 *** Antimony has joined #aichallenge 2011-12-09T13:58:16 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-12-09T14:07:31 btw, althought it's not a proper use of this channel, I think one will get better help here than on those c++, java etc. channels. But maybe I am wrong because I am present only here :) 2011-12-09T14:10:35 i tried to get help on #python few times 2011-12-09T14:10:43 only one time i've got a response 2011-12-09T14:12:16 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-12-09T14:17:20 I know I always say 'do not make any assumptions' but let's do some of them. How many bots of current top100 will end up there after the final stage (without updates)? 2011-12-09T14:18:30 how can i run this? http://paste.aichallenge.org/GpWxL/ 2011-12-09T14:21:44 Kommander: http://paste.aichallenge.org/DVwbi/raw/ download as map file 2011-12-09T14:22:13 thanks 2011-12-09T14:23:06 *** olexs has joined #aichallenge 2011-12-09T14:24:05 olexs: still want a map extractor for paste? 2011-12-09T14:25:08 thanks McLeopold! 2011-12-09T14:25:10 i am not sure if it's because my bot is smart or stupid, as it takes 2 turns to kill the opponent on that map 2011-12-09T14:25:32 McLeopold: it'd sure be useful 2011-12-09T14:28:17 *** bhasker has joined #aichallenge 2011-12-09T14:28:20 morning 2011-12-09T14:28:32 *** Migi32 has joined #aichallenge 2011-12-09T14:28:53 janzert: are you on? :) 2011-12-09T14:29:26 *** bluegaspode has joined #aichallenge 2011-12-09T14:30:41 *** Paradoxial has joined #aichallenge 2011-12-09T14:32:03 *** amstan_ has joined #aichallenge 2011-12-09T14:32:03 *** ChanServ sets mode: +o amstan_ 2011-12-09T14:32:24 yep 2011-12-09T14:34:47 *** epicmonkey has joined #aichallenge 2011-12-09T14:34:55 *** bluegaspode has quit IRC (Quit: Page closed) 2011-12-09T14:35:32 *** bretep has joined #aichallenge 2011-12-09T14:36:29 olexs: you can watch me code it live on a production server :) http://paste.aichallenge.org/gBkF7/map/ 2011-12-09T14:37:03 janzert: while I'm thinking about it, for next contests, what do you think about automatic deactivation for crashing and timeouts? 2011-12-09T14:37:23 olexs: Hi. How's your bot doing? 2011-12-09T14:37:43 McLeopold: you mean till a new submission? yea a very good idea... IMO only of course 2011-12-09T14:38:11 hmm, that's basically what it was in tron and people like this much better 2011-12-09T14:38:39 this is certainly much more forgiving 2011-12-09T14:38:50 I wouldn't mind a zero score for the game though 2011-12-09T14:39:15 well, I want to separate game scoring and rules from this 2011-12-09T14:39:22 this is just for matchmaking 2011-12-09T14:39:46 yeah, I don't think pulling them out is a good idea 2011-12-09T14:39:49 it would follow the normal deactivation rules, just happen sooner 2011-12-09T14:40:05 not for finals of course 2011-12-09T14:40:08 *** amstan_ has quit IRC (Ping timeout: 240 seconds) 2011-12-09T14:40:14 *** mj41 has joined #aichallenge 2011-12-09T14:40:29 and I like as little as possible to change in between :) 2011-12-09T14:40:51 which is why I preferenced it with "for the next contest" :) 2011-12-09T14:41:18 but I would really want the rule to be in place for both the open contest and finals if used 2011-12-09T14:41:36 I'm not too opposed to it personally but I think a lot would be 2011-12-09T14:41:37 deactivation isn't happening for finals 2011-12-09T14:42:42 so, I don't see a problem 2011-12-09T14:43:06 deactivation happens after 3 days already, this is just a rule to make it happen sooner for trouble bots 2011-12-09T14:44:50 *** Paradoxial has quit IRC (Quit: :) 2011-12-09T14:46:37 yeah, simple deactivation wouldn't be as bad as removal 2011-12-09T14:49:38 okay, I never said removal 2011-12-09T14:50:18 that's just what my mind jumped to since that is what happened with tron :/ 2011-12-09T14:52:42 Me thought about discvalification of current version too. I even wrote a big message but lucky decided to look more carefuly on the discussion before pressing Enter. 2011-12-09T14:53:58 *** besh has joined #aichallenge 2011-12-09T14:54:27 *** epicmonkey has quit IRC (Ping timeout: 240 seconds) 2011-12-09T14:55:04 *** Accoun has quit IRC () 2011-12-09T14:55:18 *** Dlayne has joined #aichallenge 2011-12-09T14:57:54 *** Paradoxial has joined #aichallenge 2011-12-09T15:07:21 olexs: okay, map link enabled, but the replays store the ending state 2011-12-09T15:08:27 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-12-09T15:10:26 *** Accoun has joined #aichallenge 2011-12-09T15:17:59 *** Kurnevsky has joined #aichallenge 2011-12-09T15:19:04 *** speakeasy has joined #aichallenge 2011-12-09T15:20:21 aichallenge: McLeopold epsilon * r803120e / (website/header.php website/profile_games.php website/sql.php): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://git.io/Xrxu3A 2011-12-09T15:20:22 aichallenge: McLeopold epsilon * r07d7292 / ants/ants.py : scenario replays store original map - http://git.io/jyepOw 2011-12-09T15:20:44 *** delt0r_ has joined #aichallenge 2011-12-09T15:22:16 for those of you who are pasting awesome replays with scenarios, please get the latest game engine so that we can download the maps from within the replays 2011-12-09T15:22:47 aichallenge: janzert epsilon * r69544ba / ants/ants.py : Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://git.io/NnfwzQ 2011-12-09T15:22:47 aichallenge: janzert epsilon * r9d0ec4d / (4 files in 3 dirs): Remove some hardcoded ec2 urls - http://git.io/Ncu1Tg 2011-12-09T15:23:37 google just told me a map file was in slovek and asked if I wanted to translate :/ 2011-12-09T15:27:45 Not too far off ;) 2011-12-09T15:28:16 McLeopold: get it to read the english translation 2011-12-09T15:28:17 *** bretep has quit IRC (Ping timeout: 258 seconds) 2011-12-09T15:28:23 if there is one :P 2011-12-09T15:38:34 I once again manage to lock myself out of root by leaving incorrect permissions on /etc/sudoers :} 2011-12-09T15:39:06 *** Migi32 has quit IRC (Remote host closed the connection) 2011-12-09T15:41:10 *** bluegaspode has joined #aichallenge 2011-12-09T15:41:32 janzert: visudo ? 2011-12-09T15:41:32 I really, really should stop touching them and force vim to write the file 2011-12-09T15:41:57 if you aren't using visudo not only are you doing it wrong, but you are a complete moron as well. 2011-12-09T15:42:28 heh, well I'm already root when editing it so the permission changes are completely unnecessary 2011-12-09T15:42:57 *** Dlayne has left #aichallenge 2011-12-09T15:43:06 just a bad habit to do a chmod u+w 2011-12-09T15:50:49 *** Jacob_Strauss has joined #aichallenge 2011-12-09T15:53:17 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2011-12-09T15:56:50 aichallenge: janzert epsilon * rb73ab89 / website/worker_init.py : Actually run the worker setup command - http://git.io/jWEGxA 2011-12-09T15:57:16 *** speakeasy has quit IRC (Quit: Nettalk6 - www.ntalk.de) 2011-12-09T16:01:25 *** Kurnevsky has left #aichallenge 2011-12-09T16:09:16 aichallenge: janzert epsilon * re28140d / (3 files in 2 dirs): Add chroot os base url - http://git.io/QvtxOA 2011-12-09T16:10:33 *** olexs has quit IRC (Quit: Leaving.) 2011-12-09T16:15:57 *** AntDroid has joined #aichallenge 2011-12-09T16:16:05 *** avdg has quit IRC (Quit: Leaving.) 2011-12-09T16:18:37 *** avdg has joined #aichallenge 2011-12-09T16:21:05 aichallenge: janzert epsilon * r38b7eb7 / setup/worker_setup.py : Expand chroot sources.list filename - http://git.io/8gEbkQ 2011-12-09T16:21:33 *** cyphase has quit IRC (Ping timeout: 252 seconds) 2011-12-09T16:21:33 *** sashaSochka has joined #aichallenge 2011-12-09T16:22:36 *** besh has quit IRC (Quit: Page closed) 2011-12-09T16:24:47 *** Antimony has joined #aichallenge 2011-12-09T16:31:37 my master todo/ideas list is growing rather than shrinking 2011-12-09T16:31:48 BenJackson: yeah, same with mine 2011-12-09T16:31:57 and every time I try implementing something my bot gets worse :P 2011-12-09T16:32:01 deleting code makes it better 2011-12-09T16:33:51 *** cyphase has joined #aichallenge 2011-12-09T16:39:15 *** bluegaspode has quit IRC (Quit: Page closed) 2011-12-09T16:44:36 aichallenge: janzert epsilon * r7597b34 / setup/chroot_configs/sources.list.template : chroot should be using natty packages - http://git.io/oU9rjA 2011-12-09T16:44:45 BenJackson: if you don't want to implement all your ideas, just give them to me :) 2011-12-09T16:48:55 *** pairofdice has quit IRC (Quit: leaving) 2011-12-09T16:50:39 *** epicmonkey has joined #aichallenge 2011-12-09T16:56:54 *** avdg has quit IRC (Quit: Leaving.) 2011-12-09T17:05:05 @rankings 2011-12-09T17:05:06 Anilm3: Top 10 players: pguillory(91.8), xathis(91.3), FlagCapper(90.1), ChrisH(89.5), GreenTea(88.4), lazarant(87.4), delineate(87.4), ThisIsNotABug(87.1), fourmidable(86.9), Migi32(86.6) 2011-12-09T17:05:15 getting tight up there 2011-12-09T17:09:52 *** sashaSochka has quit IRC (Quit: Page closed) 2011-12-09T17:11:58 *** thestinger has quit IRC (Quit: WeeChat 0.3.6) 2011-12-09T17:12:24 *** jix has quit IRC (Ping timeout: 258 seconds) 2011-12-09T17:13:47 *** sigh has joined #aichallenge 2011-12-09T17:14:29 *** Palmik has quit IRC (Remote host closed the connection) 2011-12-09T17:14:57 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-12-09T17:21:07 thestinger: at this point I did remove a bunch of goal weights to get back to basics 2011-12-09T17:21:12 I'm stuck at the point now where I see undesirable behavior 2011-12-09T17:21:24 but I'm not sure how to express my desire to avoid it 2011-12-09T17:25:38 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-12-09T17:28:24 *** jix has joined #aichallenge 2011-12-09T17:28:55 flag has a mu of 98.39 after 19 games now o_O 2011-12-09T17:34:00 go flag 2011-12-09T17:34:26 i'm not ruling dmj111 out 2011-12-09T17:34:31 it looks like he got involved late to me 2011-12-09T17:34:41 and had to switch from python to c++ 2011-12-09T17:35:52 Meh, most of the people who have to do that, could have just switched to pypy and got a 5x-20x speed increase 2011-12-09T17:37:48 I'm using D, which admittedly is not quite as fast as C++ but its tight loops should be pretty close, and I have to worry about how much time my bot uses 2011-12-09T17:38:46 hundreds of ants is no problem, but on a big map the cpu use gets pretty high (100-200 ms locally, which means 250-500 ms on the server) 2011-12-09T17:38:48 pypy is the same, in tight loops and such it can be as fast or faster then C 2011-12-09T17:39:18 JITing lends itself very well to tight repetitive loops. 2011-12-09T17:40:20 *** Antimony has joined #aichallenge 2011-12-09T17:42:19 Minthos: Really? My Q6600 seems to be slightly slower than the contest servers 2011-12-09T17:42:29 2.4G 2011-12-09T17:42:54 that's weird, I'm also using a q6600 2011-12-09T17:46:03 one of my early submissions that scaled workload with ants, used to time out at around 150 ants on the official server 2011-12-09T17:46:29 it worked fine locally, but when I set timeout to 250, 200 or 150 ms, it would timeout more and more frequently 2011-12-09T17:47:15 I haven't been timing out (since I quit using python on day 2) so I haven't worried about it 2011-12-09T17:47:26 but I did seem to see some long compute times locally for high ant games 2011-12-09T17:47:33 but those did not cause timeouts on the server 2011-12-09T17:47:53 I suspect wall clock timing will always be somewhat random 2011-12-09T17:47:56 it could be multitasking locally.. 2011-12-09T17:48:02 I work on embedded linux devices at my day job 2011-12-09T17:48:05 and that, yes 2011-12-09T17:48:08 I've seen the linux scheduler do some really stupid things 2011-12-09T17:48:22 so counting on the system scheduler to put 10 bots on 10 cpus is a fantasy 2011-12-09T17:48:42 unless there's a taskset in there somewhere I didn't see 2011-12-09T17:48:59 the contest servers are virtualized, I suspect there are considerable delays/overheads in task switching both on the host and vm 2011-12-09T17:56:20 does anyone know what does it mean when people write a word between underscores? e.g. _word_ 2011-12-09T17:57:32 It means underlined, which in historical typesetting meant to italicize 2011-12-09T17:57:40 (because you could understrike on a typewriter) 2011-12-09T17:57:48 no, underline is underline 2011-12-09T17:57:55 italic is /italic/ 2011-12-09T17:58:07 _this_ is underline 2011-12-09T17:58:14 I understand it's underline but what is the meaning? 2011-12-09T17:58:35 you underline book titles 2011-12-09T17:58:41 added significance, like italicizing or making bold 2011-12-09T17:58:41 other than that just emphasis 2011-12-09T17:59:39 I see, thanks 2011-12-09T17:59:45 underline is _not_ the same as italic (emphasis mine) 2011-12-09T17:59:54 ;) 2011-12-09T18:01:28 *stars* is sometimes used for emphasis, but conventionally it's used for actions 2011-12-09T18:01:47 as in *jumps up and down like a jackrabbit on crack* 2011-12-09T18:03:27 *** epicmonkey has quit IRC (Ping timeout: 240 seconds) 2011-12-09T18:05:25 *** mj41 has quit IRC (Ping timeout: 255 seconds) 2011-12-09T18:07:01 *** Jak_o_Shadows has joined #aichallenge 2011-12-09T18:18:56 *** mleise has quit IRC (Ping timeout: 240 seconds) 2011-12-09T18:19:55 *** Redgis has quit IRC (Ping timeout: 268 seconds) 2011-12-09T18:21:36 I see 2011-12-09T18:22:02 *** iglo has quit IRC (Remote host closed the connection) 2011-12-09T18:39:07 huh, my currently submitted bot will crash if it's in combat on the very first turn 2011-12-09T18:40:19 that would be an interesting map 2011-12-09T18:41:13 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-12-09T18:41:25 scenario 2011-12-09T18:42:33 *** Scryer has joined #aichallenge 2011-12-09T18:43:16 Geez - 4 out of my last 6 games on TC Pants have been on the tutorial map. I think it's trying to tell me something about my skill level. 2011-12-09T18:45:43 *** amstan has joined #aichallenge 2011-12-09T18:45:43 *** ChanServ sets mode: +o amstan 2011-12-09T18:49:20 *** Jacob_Strauss has quit IRC (Quit: Page closed) 2011-12-09T18:49:32 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-12-09T18:55:08 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-12-09T19:07:13 *** Antimony has joined #aichallenge 2011-12-09T19:11:49 *** besh has joined #aichallenge 2011-12-09T19:12:01 *** Antimony has quit IRC (Ping timeout: 255 seconds) 2011-12-09T19:20:17 *** besh has quit IRC (Quit: Page closed) 2011-12-09T19:26:13 *** amstan has joined #aichallenge 2011-12-09T19:26:13 *** ChanServ sets mode: +o amstan 2011-12-09T19:27:56 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-12-09T19:31:06 *** tobym has quit IRC (Remote host closed the connection) 2011-12-09T19:38:49 *** avdg has joined #aichallenge 2011-12-09T19:43:39 *** kire has quit IRC (Remote host closed the connection) 2011-12-09T19:46:47 *** Antimony has joined #aichallenge 2011-12-09T19:56:27 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-12-09T19:57:37 *** QuirionPT has quit IRC (Quit: Leaving) 2011-12-09T19:59:26 *** amstan_ has joined #aichallenge 2011-12-09T19:59:26 *** ChanServ sets mode: +o amstan_ 2011-12-09T20:03:46 *** Antimony has quit IRC (Ping timeout: 255 seconds) 2011-12-09T20:05:50 *** Jak_o_Shadows has quit IRC (Read error: Connection reset by peer) 2011-12-09T20:13:07 *** thestinger has joined #aichallenge 2011-12-09T20:13:57 welp, last big weekend for bot development 2011-12-09T20:16:46 <_flag> yup 2011-12-09T20:17:19 I think I'll just get combat working properly and pretty much leave everything else how it is 2011-12-09T20:19:43 *** cbad has left #aichallenge 2011-12-09T20:56:17 *** sigh has quit IRC (Remote host closed the connection) 2011-12-09T20:58:35 *** Anilm3 has quit IRC (Quit: Lost terminal) 2011-12-09T21:08:42 *** JorgeB has quit IRC (Ping timeout: 252 seconds) 2011-12-09T21:20:27 *** flowenol has quit IRC (Ping timeout: 240 seconds) 2011-12-09T21:22:59 hmm 2011-12-09T21:23:01 its like my bot 2011-12-09T21:23:03 has all the right ideas 2011-12-09T21:23:10 but doesn't effectively carry any of them out xD 2011-12-09T21:23:13 my combat blows 2011-12-09T21:24:44 *** Conorach_ has quit IRC (Ping timeout: 252 seconds) 2011-12-09T21:34:11 *** sigh has joined #aichallenge 2011-12-09T21:47:07 *** amstan_ has quit IRC (Ping timeout: 240 seconds) 2011-12-09T21:57:42 *** besh has joined #aichallenge 2011-12-09T21:58:59 *** Anilm3 has joined #aichallenge 2011-12-09T21:59:16 I miss mcstar... 2011-12-09T21:59:42 @rankings 2011-12-09T21:59:43 My bot produced a bunch of "duplicate move errors". When re-running the game using bot-input data , i don't see the error because my bot doesn't check for legality of moves. 2011-12-09T21:59:43 Anilm3: Top 10 players: pguillory(91.8), xathis(91.6), ChrisH(89.8), FlagCapper(89.2), GreenTea(88.6), ThisIsNotABug(87.5), delineate(87.4), fourmidable(87.4), lazarant(87.3), RVeerdonk(86.4) 2011-12-09T21:59:56 How do you reproduce the errors ? 2011-12-09T22:00:03 woow things are getting tighter 2011-12-09T22:00:33 use the bot in function besh 2011-12-09T22:01:09 yes that is what I did but my bot doesn't for legality of moves. I think I need to involve the play_game script somehow. 2011-12-09T22:01:15 check* 2011-12-09T22:03:03 @rankings 2011-12-09T22:03:04 besh: Top 10 players: pguillory(91.8), xathis(91.6), ChrisH(89.8), FlagCapper(89.2), GreenTea(88.6), ThisIsNotABug(87.5), delineate(87.4), fourmidable(87.4), lazarant(87.3), RVeerdonk(86.4) 2011-12-09T22:03:21 wow i din't know that was a command :) 2011-12-09T22:03:23 still the same from 4 minutes ago 2011-12-09T22:03:33 @hi 2011-12-09T22:03:34 besh: Run as fast as you can and don't look back. 2011-12-09T22:05:02 *** rwest_ has quit IRC (Quit: Lost terminal) 2011-12-09T22:06:10 anyone know how to hook the playgame script during replay i.e to produce the errors ? 2011-12-09T22:06:42 the problem is i know there are duplicate moves somewhere in the 1300th turn but i don't know where. 2011-12-09T22:06:56 run your bot alone 2011-12-09T22:06:59 with the input 2011-12-09T22:07:12 just make sure if you use random numbers you use the seed from the game 2011-12-09T22:07:25 yes the bot produces all the moves but it doesn't tell me if there are duplicates. 2011-12-09T22:07:34 modify the bot 2011-12-09T22:07:40 to print to a log file or something 2011-12-09T22:07:45 and then do it 2011-12-09T22:08:14 I have all the moves from turn 1 - 1500 but how do you check for duplicates ? 2011-12-09T22:09:47 i guess it is doable but would be much easier if i can pass it trhough the script 2011-12-09T22:10:22 *** Anilm3 has quit IRC (Quit: Lost terminal) 2011-12-09T22:17:32 *** guest42 has joined #aichallenge 2011-12-09T22:19:29 It turned out to be easy. I checked for duplicate moves in my bot using a HashSet before printing.. 2011-12-09T22:20:11 :) 2011-12-09T22:21:00 well you need to know think the right way to solve the problem :) 2011-12-09T22:22:03 *** guest42 has quit IRC (Quit: leaving) 2011-12-09T22:22:25 how have people gone about getting your ants to "flatten out" when near enemies? 2011-12-09T22:22:29 rather than lining up? 2011-12-09T22:24:05 woo hoo, I made a small testcase that shows my unwanted combat behavior 2011-12-09T22:24:08 now to fix it...somehow 2011-12-09T22:24:29 :D 2011-12-09T22:25:34 BenJackson: say you have aa..b, how you do know to move the middle a up or down? :P 2011-12-09T22:26:06 neither -- either way b could go west and trade 2011-12-09T22:26:29 so aa would both go west. probably most b bots would go east too 2011-12-09T22:26:38 THEN maybe a reorganizes 2011-12-09T22:28:31 antimatroid: my bot could do it two ways firs a south and second a east (the best option) 2011-12-09T22:28:51 second a south 2011-12-09T22:28:58 oh yeah 2011-12-09T22:29:11 see this is why I wrote a program to do combat :) 2011-12-09T22:29:34 besh: but are you picking that from a near full minimax? 2011-12-09T22:30:07 i want to pass my game tree a set of offensive moves and a set of defensive moves and just run minimax on that 2011-12-09T22:30:09 no minmax. the first 'a' can not move because it is suicide. 2011-12-09T22:30:24 it can move, just not west 2011-12-09T22:30:41 you mean not east ? 2011-12-09T22:30:45 yeah sorry 2011-12-09T22:30:51 i would move the middle a for offense first anyway 2011-12-09T22:31:02 but how do you know to move him up/down there rather than east? 2011-12-09T22:31:22 for offensive i would normally just place it where it minimises the positive number of potential battle enemies 2011-12-09T22:31:26 i collect attackers on radius + 1 and radius + 2 for enemy ants 2011-12-09T22:31:34 I built a version of my bot without optimization 2011-12-09T22:31:39 i start from the closest (radius + 1) 2011-12-09T22:31:39 times out instantly in combat! 2011-12-09T22:32:47 i have an idea but i don't know how much i like it 2011-12-09T22:33:50 like as well as counting potential battle enemies for each move loc, get the total sum for all move locs battleable from and minimise that? 2011-12-09T22:34:48 you mean an influence map approach like memetix described ? 2011-12-09T22:35:58 *** Antimony has joined #aichallenge 2011-12-09T22:38:19 well, whatever people are calling htem 2011-12-09T22:38:24 i've been using it for a while 2011-12-09T22:39:06 i only work out the information for locations ants can move into 2011-12-09T22:39:49 and i store "battle locations" for static and non-static ants next turn 2011-12-09T22:40:15 ug my bot refuses to run locally 2011-12-09T22:40:17 are people tracking where possible enemies might be? 2011-12-09T22:40:22 I do 2011-12-09T22:40:29 I use a conservative method though 2011-12-09T22:40:35 i use that to set up my perimeter 2011-12-09T22:40:45 Inivsible squares are marked if any neighbors were last turn 2011-12-09T22:40:57 and visible squares are obviously marked iff there's an ant there 2011-12-09T22:41:05 i just store a bool as to whether an enemy might be somewhere, everything starts as true, then i propogate the information each turn 2011-12-09T22:41:11 that is interesting 2011-12-09T22:41:14 I guess that's the same as me 2011-12-09T22:41:16 so i can set up a perimeter around my hill 2011-12-09T22:41:19 where i know an enemy isn't 2011-12-09T22:41:26 and not have to go back in and suss it out straight away 2011-12-09T22:41:38 I don't have a perimeter per se. I just say that ants have to be closer than the closest possible enemy 2011-12-09T22:41:58 are you resetting a non-visible square as a possible enemy? 2011-12-09T22:41:59 i don't do that 2011-12-09T22:42:11 nonvisible squares are true if any neighbors were last turn 2011-12-09T22:42:20 *** Paradoxial has quit IRC (Quit: :) 2011-12-09T22:42:29 or if it is an enemy hill 2011-12-09T22:42:36 yep 2011-12-09T22:42:39 that's waht i do 2011-12-09T22:43:04 I've been thinking about more sophisticated and accurate methods, but they seem too complex 2011-12-09T22:43:23 what probability based? 2011-12-09T22:43:32 One idea is that once you find all the hills, mark all squares that are more than turn distances away as false 2011-12-09T22:43:38 on old maps, it's useless 2011-12-09T22:43:41 what is the usual area given to each player on a map? I seem to remember the average player gets 20x20 or something like that 2011-12-09T22:43:44 but some of the new maps start hills close together 2011-12-09T22:43:44 i thought about using markov chains, but not really sure how i'd use the additional info 2011-12-09T22:43:56 per player area is 900-5000 2011-12-09T22:44:44 thanks 2011-12-09T22:44:51 *** Vesbur has joined #aichallenge 2011-12-09T22:45:15 it's killing me not to explain how to do map symmetry 2011-12-09T22:45:45 I guess we;ll see whos symmetry is better on dec 20 2011-12-09T22:46:17 i'm not even attempting to exploit symmetry 2011-12-09T22:46:22 me either 2011-12-09T22:46:29 I haven't noticed much benefit from it 2011-12-09T22:46:33 then why don't you want to explain? 2011-12-09T22:46:34 it's oly a slight edge at best 2011-12-09T22:46:40 *** Vesbur has left #aichallenge 2011-12-09T22:46:44 cobmat is more important 2011-12-09T22:46:51 i came up with most of the "symmetries", so i think it might come off as bad if i also try/exploit it 2011-12-09T22:47:01 what? 2011-12-09T22:47:07 antimatroid: just because I didn't find it advantageous doesn't mean someone else couldn't use it 2011-12-09T22:47:14 the possible symmetries is a well defeind set 2011-12-09T22:47:23 i know that, i'm not saying i literally invented them 2011-12-09T22:47:34 but i went through and worked out how to do the ones we're using 2011-12-09T22:48:07 it was still fun to work out how to do the translationally symmetric maps, even if it had been done previously :P 2011-12-09T22:48:26 although i'm not sure where tbh? 2011-12-09T22:49:04 I'm tempted to just say how I did mine. There's probably not enough time to actually implement it. Especially since I'm still finding bugs in mine 2011-12-09T22:49:22 moral of the story, test on all possible types of maps 2011-12-09T22:50:57 like, do you know anywhere that describes how to do translational symmetry on the surface of a torus? with the n = noPlayers = lcm(rows/gcd(rows, row_translate), cols/gcd(cols, col_translate)) forumulate such that translate^n(loc) = loc for all locations loc? 2011-12-09T22:51:15 formula*/whatever 2011-12-09T22:52:00 well it's not hard to rederive 2011-12-09T22:52:07 yes, it's fun :P 2011-12-09T22:52:12 that's what I did 2011-12-09T22:52:27 but i can't find that anywhere on the interwebs? 2011-12-09T22:52:31 except a reddit post i made 2011-12-09T22:52:51 the tricky part is "extra symmetries" 2011-12-09T22:52:58 extra? 2011-12-09T22:53:05 as in the rigid rotation ones? 2011-12-09T22:53:05 for example a completely blank map can have rotational symmetry even if it's rectangular 2011-12-09T22:53:42 did my translating those maps make symmetry detection any harder for you? 2011-12-09T22:53:44 i thought it might 2011-12-09T22:53:47 no 2011-12-09T22:53:57 do you see why i thought it might? 2011-12-09T22:54:08 bots naivly assume they rotate around the center? 2011-12-09T22:54:13 yeah 2011-12-09T22:54:20 btw what do you use symmerty detection for ? 2011-12-09T22:54:27 water information 2011-12-09T22:54:28 my bot uses first hill it sees as the origin for all coordinates 2011-12-09T22:54:28 path finding 2011-12-09T22:54:36 and knowing where enemy hills are etc. 2011-12-09T22:54:40 but it can rotate at any offset 2011-12-09T22:54:53 Antimony: what about say.... a.b.c.d. 2011-12-09T22:55:01 ? 2011-12-09T22:55:03 if you see c before b and d you might assume there's only 2 players? 2011-12-09T22:55:10 no the first of your own hills 2011-12-09T22:55:19 well, actually the one that is sorted first 2011-12-09T22:55:23 so probably the most top left one 2011-12-09T22:55:23 oh i see 2011-12-09T22:55:50 anyway, I think my bot can handle all that stuff 2011-12-09T22:56:13 it doesnt stop until it can prove there are no more players 2011-12-09T22:56:33 i still think it was worth translating those maps so that people can't just check the coordinates like how you'd make it 2011-12-09T22:56:35 *** MikeyG has quit IRC (Ping timeout: 252 seconds) 2011-12-09T22:56:45 yeah 2011-12-09T22:56:51 translate all the maps randomly 2011-12-09T22:56:58 *** MikeyG has joined #aichallenge 2011-12-09T22:57:00 irrelevant for the translationally symmetric maps 2011-12-09T22:57:02 doesn't do anything 2011-12-09T22:57:10 so you wrote all the mapgens? 2011-12-09T22:57:14 no 2011-12-09T22:57:39 i did random walk (didn't actually intend for that to be used) and the maze map gen 2011-12-09T22:57:48 cell maps are my favourite but pretty sure mcleopold did those 2011-12-09T22:58:09 i need to rewrite my two 2011-12-09T22:58:11 I'm disapointed that they got rid of the logn distance mazes? 2011-12-09T22:58:18 since it gives syemmtry a bigger advantage 2011-12-09T22:58:42 i think maze generator is just currently failing to make valid maps most of the time 2011-12-09T22:59:00 i haven't updated it since beta, and multiple hills was just hacked on 2011-12-09T22:59:09 i'm going to try and rewrite it this week 2011-12-09T22:59:25 but i need to write my battle code all over too :| 2011-12-09T22:59:35 and quite busy irl 2011-12-09T22:59:36 I need to write battle code period 2011-12-09T23:01:31 summer is awesome though 2011-12-09T23:01:39 it's nice been able to ride around outside in shorts and thongs 2011-12-09T23:02:09 (I just let the non-aussie's minds go wild with that last sentence these days :P) 2011-12-09T23:03:22 *** amstan_ has joined #aichallenge 2011-12-09T23:03:22 *** ChanServ sets mode: +o amstan_ 2011-12-09T23:04:30 antimatroid: have you thought about doing anything like using 'worms' that steer around to make maps more interesting that random walks but not as rigid as mazes? 2011-12-09T23:04:47 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-12-09T23:04:51 my bot already seems to be quite good at forming it's own snakes 2011-12-09T23:04:59 but i don't even consider that as necessary 2011-12-09T23:05:18 I meant as map generation not for bots 2011-12-09T23:05:21 what i do have an issue with is not getting ants to try and line up for battle 2011-12-09T23:05:24 oh 2011-12-09T23:05:43 that's a nice idea 2011-12-09T23:06:03 I've been working on varients of that idea 2011-12-09T23:06:03 Extrarius: have you seen the cell mazes? 2011-12-09T23:06:04 http://pastebin.com/WBDkKFFj 2011-12-09T23:06:27 Extrarius: got some examples you can add to ant paste? 2011-12-09T23:06:32 BenJackson: Are they in the python starter pack? 2011-12-09T23:06:36 i might not even have to rewrite the random walk map gen 2011-12-09T23:06:36 antimatroid: sure 2011-12-09T23:06:52 Extrarius: does it do min/max paths between enemy hills etc.? 2011-12-09T23:06:58 no, they're new 2011-12-09T23:07:01 but they're on the main server 2011-12-09T23:07:31 I can't imagine anything that looks more like what you describe than some of the cell mazes 2011-12-09T23:07:35 http://paste.aichallenge.org/KHvNB/ 2011-12-09T23:07:42 i think cell mazes are done differently 2011-12-09T23:07:44 http://paste.aichallenge.org/WuxC9/ 2011-12-09T23:08:03 http://paste.aichallenge.org/Xz0Gl/ 2011-12-09T23:08:05 eg that map 2011-12-09T23:08:27 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-12-09T23:08:34 Extrarius: they look great 2011-12-09T23:08:45 BenJackson: I like less regularity than that 2011-12-09T23:09:01 also, i'm a big fan of allowing smaller width corridors, but a 3x3 block must be able to traverse between hills on the map 2011-12-09T23:09:43 don't think there's a problem, just explicitly saying :) 2011-12-09T23:10:02 hrmm, that makes it more difficult to make 'interesting' maps =-/ I'll play around with adding that constraint 2011-12-09T23:10:20 what do you mean? 2011-12-09T23:10:42 just start with like a 3x3 block and add hills, then finish the map 2011-12-09T23:11:05 dw if you don't want to 2011-12-09T23:11:26 but it'd be cool to use it as an official map gen, but you'd need to meet the requirements for that to be worthwhile.. 2011-12-09T23:11:40 I just meant every requirement reduces freedom 2011-12-09T23:12:04 I don't think I'm going to be able to submit a bot, so being able to submit a map generator instead would be nice =-) 2011-12-09T23:12:43 wow, made a mod to my combat to try to fix a misbehavior 2011-12-09T23:12:52 antimatroid: is there any list of all the requirements for map generators? 2011-12-09T23:13:00 i think so, let me look 2011-12-09T23:13:06 in a 31v31 ant challenge test it killed... all of my current version's ants 2011-12-09T23:13:07 McLeopold would probably be the best person to talk to 2011-12-09T23:14:21 Extrarius: https://github.com/aichallenge/aichallenge/wiki/Ants-Map-Format bottom of that page looks about right? 2011-12-09T23:14:56 oh waht 2011-12-09T23:14:59 as much as people are working on combat, I wonder how many can do things like realize they have ants to spare and accept major losses to take hills 2011-12-09T23:15:00 antimatroid: for lining up for attack what i do is to stop the ants at radius + 2. Simple but effective. 2011-12-09T23:15:02 i can do my map generator in any language now :D 2011-12-09T23:15:17 besh: but then you're never actually moving in for the kill? 2011-12-09T23:15:32 oh, and then attack? 2011-12-09T23:15:40 i do when the ants at radius + 1 outnumber the opponent 2011-12-09T23:16:05 otherwise they standoff like the rest of the bots 2011-12-09T23:16:19 took me some time to realize how the others work too 2011-12-09T23:16:49 Extrarius: there are the map specs, did you do the worm maps? 2011-12-09T23:17:04 McLeopold: see his links above 2011-12-09T23:17:21 McLeopold: I think those would be a good replacement for my old random walk map? 2011-12-09T23:17:30 McLeopold: I created the ones I posted, and I posted the source on pastebin, but I'm still tweaking it a lot 2011-12-09T23:17:33 i'll redo my maze map gen tonight i think 2011-12-09T23:17:51 The ants shape themselves to surround the enemy and when you are satisfied you can decide to move in for the kill. 2011-12-09T23:18:06 Extrarius: have you considered roughing up the straight edges? 2011-12-09T23:18:07 can't believe i'm beating a1k0n 2011-12-09T23:18:46 oh 2011-12-09T23:18:47 he timed out 2011-12-09T23:18:48 xD 2011-12-09T23:19:13 antimatroid: I actually have a 'rough up' function but somebody I showed it too complained about it being too rough 2011-12-09T23:19:23 you can beat him later in multi player games ... when xathis beats it on your behalf 2011-12-09T23:19:31 :) 2011-12-09T23:19:35 it's probably not such an issue anyway 2011-12-09T23:19:44 my maze maps are all straight edges otherwise 2011-12-09T23:20:08 BenJackson: can you post that 31v31 game? 2011-12-09T23:20:32 the scenario? 2011-12-09T23:20:49 McLeopold: look for the conversation antimony and I had above about tracking possible enemies 2011-12-09T23:20:56 argh another bird in my attic 2011-12-09T23:20:57 brb 2011-12-09T23:20:57 for a bot strategy idea 2011-12-09T23:21:03 *** delt0r_ has joined #aichallenge 2011-12-09T23:21:03 BenJackson: no, your results 2011-12-09T23:22:30 antimatroid: it looks like a long conversation... 2011-12-09T23:23:22 start with antimatroid: are people tracking where possible enemies might be? 2011-12-09T23:23:28 it doesn't actually extend that entire convo 2011-12-09T23:26:37 McLeopold: it's basically so you don't need to maximise vision around your hills at all times, just slowly move your possibly enemy perimeter outwards 2011-12-09T23:27:09 possible* (i'm terriblY at that :P) 2011-12-09T23:27:53 i want to know how some bots do that. My bots just park an ant close to hill at all times .. 2011-12-09T23:28:00 that is especially true in maze maps 2011-12-09T23:28:16 yeah, this way is quite popular 2011-12-09T23:28:23 i think 2011-12-09T23:28:56 It seems like you only need to track 1 number then 2011-12-09T23:29:15 yeah i just have a bool in each square for if an enemy might reside there 2011-12-09T23:29:24 and then target unseen squares that are closest to your hill 2011-12-09T23:29:30 not unseen 2011-12-09T23:29:36 possible enemy squares 2011-12-09T23:29:44 sorry 2011-12-09T23:29:55 antimatroid: so, do you have symmetry detection? 2011-12-09T23:30:01 i go for unseen first, then later on i worry about possible enemies rather than not visibl 2011-12-09T23:30:07 McLeopold: no 2011-12-09T23:30:12 antimony does 2011-12-09T23:30:13 do you want it? 2011-12-09T23:30:16 no 2011-12-09T23:30:34 although i'll hear how to do it? 2011-12-09T23:30:49 I don't know, I just want to figure it out 2011-12-09T23:31:03 so, ready to figure it out :) 2011-12-09T23:31:15 read the rest of my convo with antimony then :P 2011-12-09T23:31:18 I did 2011-12-09T23:31:21 he outlines some of what he's doing 2011-12-09T23:31:34 you pulled out your translation formula again :) 2011-12-09T23:31:39 yeah, i'm a fan :P 2011-12-09T23:32:25 what i don't know is can one do a fully fair map with translational symmetry and 5 players? 2011-12-09T23:32:30 you can if it's continuous, but with the grid? 2011-12-09T23:32:55 what's the difference? 2011-12-09T23:33:14 a.b.c., that's fully fair because each player has equal access to each of the other players 2011-12-09T23:33:49 I got a few starting hills to be the same distance in 5 player maps 2011-12-09T23:34:07 I don't think it is equal, but it's really close 2011-12-09T23:34:36 i'm not sure if it'd still be perfectly fair 2011-12-09T23:34:58 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-12-09T23:35:00 symmetric games are complicated, most of my honours thesis this year ended up being on them :P 2011-12-09T23:35:10 okay, so your first bit of information for map symmetry is the rows and cols 2011-12-09T23:35:10 but for more general games 2011-12-09T23:35:32 and the goal will be a set of starting hills with the ant orientations 2011-12-09T23:35:47 and figuring out the number of players 2011-12-09T23:36:19 so, the rows and cols give you a limit to the player count 2011-12-09T23:36:22 if you can match an enemy hill to one of yours that gives you some info about the number of players 2011-12-09T23:36:40 right, but I'm thinking about before you get that far 2011-12-09T23:37:06 actually, it doesn't matter 2011-12-09T23:37:12 the rows and cols give you a set of feasible players actually 2011-12-09T23:37:29 they give you a min and max, because of the area contraints 2011-12-09T23:37:43 i think there's divisibility constraints too? 2011-12-09T23:37:52 plus I think you can eliminate prime number if not a divisor of either dimension? 2011-12-09T23:37:52 i haven't actually worked out what they are 2011-12-09T23:37:57 yeah 2011-12-09T23:37:59 something like that 2011-12-09T23:38:14 for 5 players, at least 1 dimension *has* to be 5, right? 2011-12-09T23:38:38 for translational symmetric i think both have to be divisible by 5? 2011-12-09T23:38:50 no 2011-12-09T23:38:50 for the other, you can't have 5 players 2011-12-09T23:39:08 you need to be able to translate the rows 5 times and get back to the same location 2011-12-09T23:39:12 it must be divisible by 5 no? 2011-12-09T23:39:24 well, unless one way you translate 0 2011-12-09T23:39:43 my map gens don't do that 2011-12-09T23:40:11 I don't think you can ever eliminate the even number players on any map 2011-12-09T23:40:32 sure, but you can make sure no two players start in the same row or column 2011-12-09T23:40:40 *** TheLinker has joined #aichallenge 2011-12-09T23:40:48 okay, but stop assuming your maps :P 2011-12-09T23:40:59 sorry :P 2011-12-09T23:41:02 you don't need even players, you need even hills 2011-12-09T23:41:03 we are working backwards here 2011-12-09T23:41:18 why do you want symmetry detection? 2011-12-09T23:41:24 i don't think it's going to help you that much 2011-12-09T23:41:30 you're much better off improving combat 2011-12-09T23:41:31 so, you can right away limit the possible player counts 2011-12-09T23:41:45 what does it matter? just slowly dominate all the things 2011-12-09T23:42:10 for science! 2011-12-09T23:42:23 another hint would be instead of just worrying about possible enemies, actively try and push your enemies away from your hills ;) 2011-12-09T23:42:44 if they all kill each other, you lose on points 2011-12-09T23:42:56 anyway, let's say you found a hill, what next? 2011-12-09T23:43:21 you get feasible parameters for each of your hills? 2011-12-09T23:43:29 if you only started with 1 hills, that tells you alot 2011-12-09T23:43:30 test them until you find a counter example with water information 2011-12-09T23:43:35 untill you have 1 hill remaining? 2011-12-09T23:44:19 but first, can't you again limit the player counts? 2011-12-09T23:44:32 probably 2011-12-09T23:45:04 if it is in the same row or column, it might be mirror or rotational 2011-12-09T23:45:11 are you just working out translationally symmetry here? 2011-12-09T23:45:16 all of them 2011-12-09T23:45:17 yeah :P 2011-12-09T23:45:33 *** roflmao has quit IRC (Quit: Leaving.) 2011-12-09T23:45:43 you probably want to keep track of razed hills if you haven't detected symmetry yet too 2011-12-09T23:45:52 so so messy :P 2011-12-09T23:46:02 assume you do 2011-12-09T23:46:13 arrgh no bird in sight once I'm up there 2011-12-09T23:47:06 McLeopold: i'd be inclined to keep feasible stuff for translational and the others separately 2011-12-09T23:47:09 if the player count is 3, 5 or 7 finding the hill is good 2011-12-09T23:47:14 and rule them all down until only 1 remains 2011-12-09T23:47:19 anyway, i need to go do some stuff 2011-12-09T23:47:25 ok 2011-12-09T23:47:26 McLeopold: http://paste.aichallenge.org/xgTnQ/ 2011-12-09T23:48:24 BenJackson: the first move didn't look to good 2011-12-09T23:49:02 the second move was the classic "will a volunteer please step forward" and everyone but 1 ant steps back 2011-12-09T23:49:20 the first is because only the first 2 ranks are "in combat" 2011-12-09T23:49:28 thus they can't all step back as they'd like 2011-12-09T23:49:36 which is sort of on purpose 2011-12-09T23:49:48 if I build up 3 ranks deep it gives the ones in front more intestinal fortitude 2011-12-09T23:50:16 the lunging of single ants there has to do with winning scenarios that don't happen, like every opposing ant stepping forward 2011-12-09T23:50:31 that one sacrificial forward ant gets a "touch" on a bunch of enemies 2011-12-09T23:50:37 of course if they don't then he dies 2011-12-09T23:51:50 McLeopold: maybe I can recreate the "bad" prior behavior 2011-12-09T23:51:55 so with 2 lines advancing, sticking a single ant out there is good? 2011-12-09T23:52:59 bad version where defensive front rotates sideways: http://paste.aichallenge.org/9rRXN/ 2011-12-09T23:53:43 McLeopold: imagine two flat lines 2 deep on each side, 2 apart (so aa..bb but stack that up) 2011-12-09T23:53:55 now imagine aa all step east so: aa.bb (but again, stacked up 2011-12-09T23:53:57 BenJackson: can you get the latest ants.py from github? 2011-12-09T23:54:12 your map in the replay is the ending state instead of the starting state 2011-12-09T23:54:13 if one 'b' in the middle steps forward it alters the balance a lot in b's favor 2011-12-09T23:54:21 *** avdg has quit IRC (Quit: Leaving.) 2011-12-09T23:54:25 how do you paste games ? 2011-12-09T23:54:28 because that forward 'b' now touches 5 'a' that are in combat 2011-12-09T23:54:42 besh: paste in the game_logs/0.replay to paste.aichallenge.org 2011-12-09T23:54:52 there should be a script to do that in the tools 2011-12-09T23:55:20 McLeopold: I'm on the visualization branch 2011-12-09T23:55:25 did that guy merge your change? 2011-12-09T23:55:30 probably not 2011-12-09T23:55:37 just copy it over 2011-12-09T23:55:48 I'd need to merge it 2011-12-09T23:55:51 you want me to paste the map? 2011-12-09T23:56:02 sure 2011-12-09T23:56:05 *** Antimony has joined #aichallenge 2011-12-09T23:56:27 http://paste.aichallenge.org/RE2J4/ 2011-12-09T23:56:40 so, for the 1 guy steping forward, is there someone behind him or a gap? 2011-12-09T23:57:08 okay, google wanted to translate to croatian now :/ 2011-12-09T23:57:56 basically that move is due to the fact that I don't weight the likelihood of enemy moves when considering how my moves work out 2011-12-09T23:58:09 so that "one ant nub" is some super awesome reply to something that won't happen 2011-12-09T23:58:20 I'm working on that a bit 2011-12-09T23:58:25 but mainly I had to fix that rotation thing you saw 2011-12-09T23:58:28 I need to see it though... 2011-12-09T23:59:03 hm, what if you make the scenario I described and play 2 holdbots 2011-12-09T23:59:08 I am :)