2011-11-25T00:04:16 *** ltriant has joined #aichallenge 2011-11-25T00:05:51 *** delt0r_ has joined #aichallenge 2011-11-25T00:16:22 *** smiley1983 has joined #aichallenge 2011-11-25T00:16:34 *** cyphase has quit IRC (Ping timeout: 244 seconds) 2011-11-25T00:21:48 *** peyton has joined #aichallenge 2011-11-25T00:24:11 *** Areks has joined #aichallenge 2011-11-25T00:28:22 *** mleise has quit IRC (Quit: Leaving.) 2011-11-25T00:33:03 *** ThatSnail has quit IRC (Ping timeout: 248 seconds) 2011-11-25T00:37:13 *** Euphony has joined #aichallenge 2011-11-25T00:44:22 ok wtf 2011-11-25T00:44:32 I timed out on tcp despite returning 4500 ms early 2011-11-25T00:45:11 *** peyton has quit IRC (Quit: peyton) 2011-11-25T00:50:14 *** Palmik has joined #aichallenge 2011-11-25T00:53:46 *** antimatroid has quit IRC (Ping timeout: 244 seconds) 2011-11-25T01:03:27 *** antimatroid has joined #aichallenge 2011-11-25T01:05:05 *** replore has quit IRC (Remote host closed the connection) 2011-11-25T01:17:03 *** ltriant has quit IRC (Quit: Get MacIrssi - http://www.sysctl.co.uk/projects/macirssi/) 2011-11-25T01:18:59 is anyone here? 2011-11-25T01:25:49 yes 2011-11-25T01:29:11 *** praveen has joined #aichallenge 2011-11-25T01:32:24 *** Accoun has quit IRC () 2011-11-25T01:33:29 *** amstan_ has joined #aichallenge 2011-11-25T01:33:29 *** ChanServ sets mode: +o amstan_ 2011-11-25T01:33:29 *** amstan has quit IRC (Read error: Connection reset by peer) 2011-11-25T01:34:56 *** cyphase has joined #aichallenge 2011-11-25T01:36:38 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-25T01:38:55 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-11-25T01:39:59 *** Antimony has joined #aichallenge 2011-11-25T01:47:28 Has anyone gotten the tcp scripts to work? 2011-11-25T01:51:23 *** Accoun has joined #aichallenge 2011-11-25T01:52:50 ... 2011-11-25T01:53:02 *** Fandekasp has joined #aichallenge 2011-11-25T01:58:43 *** grc has joined #aichallenge 2011-11-25T02:03:06 which scripts? :P 2011-11-25T02:04:50 *** replore has joined #aichallenge 2011-11-25T02:06:00 *** roflmao has quit IRC (Quit: Leaving.) 2011-11-25T02:11:20 *** replor___ has joined #aichallenge 2011-11-25T02:11:27 *** replore has quit IRC (Read error: Connection reset by peer) 2011-11-25T02:24:05 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-25T02:30:19 *** sigh has joined #aichallenge 2011-11-25T02:34:11 *** HaraKiri has joined #aichallenge 2011-11-25T02:54:09 *** johnsonav has joined #aichallenge 2011-11-25T03:00:00 *** kaemo has quit IRC (Ping timeout: 244 seconds) 2011-11-25T03:01:00 *** jnwhiteh has quit IRC (*.net *.split) 2011-11-25T03:01:10 *** aerique has joined #aichallenge 2011-11-25T03:06:07 *** speakeasy has joined #aichallenge 2011-11-25T03:06:18 *** jnwhiteh has joined #aichallenge 2011-11-25T03:06:35 *** killian has joined #aichallenge 2011-11-25T03:07:20 *** kire has joined #aichallenge 2011-11-25T03:07:48 *** kaemo has joined #aichallenge 2011-11-25T03:08:15 *** epicmonkey has joined #aichallenge 2011-11-25T03:08:53 *** treeform has quit IRC (Remote host closed the connection) 2011-11-25T03:12:29 Antimony: sure I'm on the tcp servers all the time 2011-11-25T03:17:16 *** Conorach has joined #aichallenge 2011-11-25T03:19:39 *** earphone has joined #aichallenge 2011-11-25T03:22:22 *** smiley1983 has quit IRC (Remote host closed the connection) 2011-11-25T03:23:11 *** speakeasy has quit IRC (Ping timeout: 248 seconds) 2011-11-25T03:23:56 *** smiley1983 has joined #aichallenge 2011-11-25T03:26:27 *** Rav3nsW0rd has quit IRC (Quit: Leaving) 2011-11-25T03:29:30 *** earphone has quit IRC (Quit: Page closed) 2011-11-25T03:34:28 *** speakeasy has joined #aichallenge 2011-11-25T03:35:08 *** antimatroid has quit IRC (Ping timeout: 244 seconds) 2011-11-25T03:36:20 *** Blkt has joined #aichallenge 2011-11-25T03:39:23 *** antimatroid has joined #aichallenge 2011-11-25T03:44:26 *** mj41 has joined #aichallenge 2011-11-25T03:50:06 *** speakeasy has quit IRC (Ping timeout: 240 seconds) 2011-11-25T03:51:43 http://ants.fluxid.pl/replay.14787 yay, sometimes my bot works. 2011-11-25T03:51:56 other times it just kills itself ASAP 2011-11-25T03:51:59 *** grwip has joined #aichallenge 2011-11-25T03:52:57 I think I've finally fixed stalemates but I managed to break all my new combat code... 2011-11-25T03:53:49 *** replor___ has quit IRC (Ping timeout: 240 seconds) 2011-11-25T03:54:43 good morning everyone 2011-11-25T03:56:24 *** praveen has quit IRC (Ping timeout: 265 seconds) 2011-11-25T03:57:52 *** jon1 has joined #aichallenge 2011-11-25T04:02:04 *** smiley1983 has quit IRC (Ping timeout: 260 seconds) 2011-11-25T04:02:20 *** smiley1983 has joined #aichallenge 2011-11-25T04:02:56 *** Palmik has quit IRC (Remote host closed the connection) 2011-11-25T04:03:17 *** Palmik has joined #aichallenge 2011-11-25T04:09:08 *** smiley1983 has quit IRC (Remote host closed the connection) 2011-11-25T04:09:31 *** poka has joined #aichallenge 2011-11-25T04:11:13 *** kire has quit IRC (Read error: Connection reset by peer) 2011-11-25T04:13:33 *** smiley1983 has joined #aichallenge 2011-11-25T04:17:45 *** poka has quit IRC (Quit: Page closed) 2011-11-25T04:19:21 *** HaraKiri has quit IRC () 2011-11-25T04:23:08 *** smiley1983 has quit IRC (Ping timeout: 252 seconds) 2011-11-25T04:33:51 *** dommod has joined #aichallenge 2011-11-25T04:44:00 *** antimatroid has quit IRC (Ping timeout: 276 seconds) 2011-11-25T04:44:17 *** dommod has quit IRC (Quit: Page closed) 2011-11-25T04:44:56 *** smiley1983 has joined #aichallenge 2011-11-25T04:47:05 *** praveen has joined #aichallenge 2011-11-25T05:24:13 *** avdg1 has joined #aichallenge 2011-11-25T05:24:45 *** smiley1983 has quit IRC (Ping timeout: 260 seconds) 2011-11-25T05:27:05 *** avdg has quit IRC (Ping timeout: 244 seconds) 2011-11-25T05:36:38 *** smiley1983 has joined #aichallenge 2011-11-25T05:47:36 *** b0rder has quit IRC (Quit: 离开) 2011-11-25T05:50:01 *** ikaros has joined #aichallenge 2011-11-25T05:51:00 *** g0llum has joined #aichallenge 2011-11-25T06:07:38 *** praveen has quit IRC (Quit: Page closed) 2011-11-25T06:21:02 *** mviel__ has joined #aichallenge 2011-11-25T06:24:37 *** mviel_ has quit IRC (Ping timeout: 240 seconds) 2011-11-25T06:29:11 *** praveen has joined #aichallenge 2011-11-25T06:29:13 *** sigh has quit IRC (Remote host closed the connection) 2011-11-25T06:34:14 *** epicmonkey has quit IRC (Ping timeout: 245 seconds) 2011-11-25T06:47:53 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-25T06:48:53 *** smiley1983 has quit IRC (Ping timeout: 252 seconds) 2011-11-25T06:49:08 *** smiley1983 has joined #aichallenge 2011-11-25T06:51:22 *** Euphony has quit IRC (Quit: Page closed) 2011-11-25T06:51:54 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-11-25T06:59:34 *** liberforce has joined #aichallenge 2011-11-25T07:12:54 *** Euphony has joined #aichallenge 2011-11-25T07:14:28 *** praveen_ has joined #aichallenge 2011-11-25T07:16:07 *** sigh has joined #aichallenge 2011-11-25T07:20:34 *** praveen_ has quit IRC (Quit: Page closed) 2011-11-25T07:23:59 *** Yexo has quit IRC (Quit: Leaving) 2011-11-25T07:32:00 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-11-25T07:43:31 *** antimatroid has joined #aichallenge 2011-11-25T07:52:06 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-11-25T07:52:53 *** foRei has joined #aichallenge 2011-11-25T07:57:02 *** HaraKiri has joined #aichallenge 2011-11-25T07:57:44 midnight 2011-11-25T07:57:44 time to start programming 2011-11-25T07:58:34 *** pairofdice has joined #aichallenge 2011-11-25T08:00:21 i still think that priority queue thing is a pita 2011-11-25T08:05:21 *** delt0r_ has joined #aichallenge 2011-11-25T08:05:53 write your own priority queue then, where you offer stronger guarantees 2011-11-25T08:06:08 they are extremely simple to implement 2011-11-25T08:12:12 meh 2011-11-25T08:12:18 it's unideal but not suboptimal 2011-11-25T08:21:04 3 negations in one sentence 2011-11-25T08:21:26 that's how i roll :P 2011-11-25T08:21:35 *** avdg has joined #aichallenge 2011-11-25T08:22:50 *** avdg2 has joined #aichallenge 2011-11-25T08:23:25 *** avdg1 has quit IRC (Ping timeout: 240 seconds) 2011-11-25T08:25:49 *** avdg has quit IRC (Ping timeout: 240 seconds) 2011-11-25T08:30:14 suboptimal isn't a negation :f 2011-11-25T08:30:31 Well... 2011-11-25T08:31:13 hmm good point 2011-11-25T08:35:55 Close enough I guess 2011-11-25T08:41:36 it's quite messy to collect all targets with a single bfs 2011-11-25T08:41:45 well, a set of targets 2011-11-25T08:52:37 *** Areks has quit IRC (Ping timeout: 240 seconds) 2011-11-25T08:57:22 *** Euphony has quit IRC (Quit: Page closed) 2011-11-25T09:03:48 *** praveen has quit IRC (Ping timeout: 265 seconds) 2011-11-25T09:04:00 How is it messy 2011-11-25T09:11:29 *** Antimony has joined #aichallenge 2011-11-25T09:15:13 *** ikaros has joined #aichallenge 2011-11-25T09:15:45 *** amstan_ has quit IRC (Ping timeout: 260 seconds) 2011-11-25T09:22:06 http://aichallenge.org/visualizer.php?game=126987&user=757 this hurt 2011-11-25T09:22:13 my bot is so bad... 2011-11-25T09:23:35 you need to implement anticlustring 2011-11-25T09:27:13 Well it was quite good, until it wasn't. 2011-11-25T09:28:11 i always loose huge amounts of ants from simply rushing forward 2011-11-25T09:29:25 my pseudocombat fails horribly. if not this i could win this game 2011-11-25T09:29:31 wow, what a game 2011-11-25T09:29:55 xoroshiy is so averse to combat he lets you come almost into his hive before engaging your ants 2011-11-25T09:30:19 he does not engage 2011-11-25T09:30:26 he keeps a wall of three ants 2011-11-25T09:30:38 and i smash a row og ants right into it 2011-11-25T09:30:40 Try to advance widely, not as a snake. I think it should help. 2011-11-25T09:31:07 his ants could not move at all 2011-11-25T09:32:24 *** ikaros has quit IRC (Read error: Connection reset by peer) 2011-11-25T09:33:54 *** smiley1983 has quit IRC (Ping timeout: 276 seconds) 2011-11-25T09:35:17 *** smiley1983 has joined #aichallenge 2011-11-25T09:44:35 *** cbad has quit IRC (Read error: Connection reset by peer) 2011-11-25T09:44:57 *** cbad has joined #aichallenge 2011-11-25T09:44:57 *** cbad has joined #aichallenge 2011-11-25T09:52:35 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-11-25T09:55:30 *** ikaros has joined #aichallenge 2011-11-25T10:02:03 *** Harpyon has joined #aichallenge 2011-11-25T10:02:53 *** Harpyon_ has joined #aichallenge 2011-11-25T10:07:07 *** peyton has joined #aichallenge 2011-11-25T10:07:27 *** Antimony has joined #aichallenge 2011-11-25T10:08:22 *** ThatSnail has joined #aichallenge 2011-11-25T10:08:54 *** UncleVasya has joined #aichallenge 2011-11-25T10:12:00 *** iglo has joined #aichallenge 2011-11-25T10:12:43 Fluxid: I'm just adding +3 instead of +1 when coming from a square with an ant in breadth-first searches, so they don't form lines anymore 2011-11-25T10:21:06 *** amstan has joined #aichallenge 2011-11-25T10:21:06 *** ChanServ sets mode: +o amstan 2011-11-25T10:21:42 contestbot: seen janzert 2011-11-25T10:21:42 amstan: janzert was last seen in #aichallenge 11 hours, 9 minutes, and 1 second ago: @later tell amstan yeah disabling the account should be good there. 2011-11-25T10:21:56 contestbot: later tell janzert apparently jeff has new funds 2011-11-25T10:21:56 amstan: I think that worked... 2011-11-25T10:24:11 thestinger: i don't think i understand 2011-11-25T10:26:22 well, when you do your BFS for movement, you do +1 to the old value for each tile outwards from the target, right? 2011-11-25T10:27:04 *** Anilm3 has joined #aichallenge 2011-11-25T10:27:40 *** Boumbles has joined #aichallenge 2011-11-25T10:27:41 hi 2011-11-25T10:29:20 Fluxid: http://pastie.org/2919790 2011-11-25T10:30:32 *** peyton has quit IRC (Quit: peyton) 2011-11-25T10:31:01 it ends up making them pathfind around each other and take alternate routes if the short way is clogged with ants 2011-11-25T10:31:06 atm they spread out too far though 2011-11-25T10:31:42 *** peyton has joined #aichallenge 2011-11-25T10:35:18 *** Antimony_ has joined #aichallenge 2011-11-25T10:36:57 *** Antimony has quit IRC (Ping timeout: 276 seconds) 2011-11-25T10:36:58 *** Antimony_ is now known as Antimony 2011-11-25T10:37:03 ah, this way 2011-11-25T10:37:08 yeah, i thought of it 2011-11-25T10:37:42 i'm going to do this after i rewrite pathfinding in C ;) 2011-11-25T10:41:34 *** Antimony_ has joined #aichallenge 2011-11-25T10:42:25 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-11-25T10:43:59 http://ants.fluxid.pl/map/mc_test_map_04p_02.map ... huh? 2011-11-25T10:45:45 thestinger: mcleopold's test map 2011-11-25T10:45:58 thestinger: did you see the games? awesome 2011-11-25T10:46:33 *** Antimony_ has quit IRC (Ping timeout: 244 seconds) 2011-11-25T10:46:34 nope, didn't see them 2011-11-25T10:46:37 hm, those shouldn't be in game anyway 2011-11-25T10:46:56 http://ants.fluxid.pl/replay.13612 2011-11-25T10:47:12 do people avoid collisions with path finding? 2011-11-25T10:47:27 ie. don't let two ants paths intersect on the same location at the same turn? 2011-11-25T10:47:30 *** ikaros has quit IRC (Read error: Connection reset by peer) 2011-11-25T10:47:30 not just the next turn? 2011-11-25T10:48:19 antimatroid: my design doesn't really allow such things, since i don't remember paths per ant 2011-11-25T10:48:27 antimatroid: yes 2011-11-25T10:48:38 fluxid: neither, but i'm thinking maybe i should 2011-11-25T10:48:46 I have a hash table for avoiding this 2011-11-25T10:49:06 antimatroid: I linked to an example of what I do now http://pastie.org/2919790 - the individual ants just move in a direction that brings them closer to a target 2011-11-25T10:49:10 (once I do the planning for an ant, I update the hash table accordingly) 2011-11-25T10:49:16 antimatroid: you mean avoiding collisions n-turns ahead? 2011-11-25T10:49:25 so they do pathfind around each other a bit, but not really 2011-11-25T10:49:26 Fluxid: yeah 2011-11-25T10:49:30 (and once an ant gives up on it's plan or dies, I update it too) 2011-11-25T10:49:41 like set > to mark when a location is taken 2011-11-25T10:49:57 but i don't actually generate the paths after i finish path finding 2011-11-25T10:50:06 i just store the initial search direction and target 2011-11-25T10:50:42 *** ikaros has joined #aichallenge 2011-11-25T10:50:55 Fluxid: those cell_maze maps were neat, now I know my pathfinding does work 2011-11-25T10:51:07 it's just combat/fleeing that sucks (which I knew) 2011-11-25T10:51:25 http://ants.fluxid.pl/replay.15173 2011-11-25T10:51:29 I mean foraging, not pathfinding 2011-11-25T10:51:34 what do you guys think? is it worth avoiding collisions after next turn? 2011-11-25T10:52:25 I think it doesn't hurt to do that. it's just a few more lines of code. :) 2011-11-25T10:52:27 I think it's worth weighting those tiles, so movement through them costs more 2011-11-25T10:52:34 i'm thinking more time wise :P 2011-11-25T10:52:44 thestinger: that sounds horrible :P 2011-11-25T10:52:46 thestinger: draskbott behaves funny here, lol 2011-11-25T10:53:11 it's not costly in terms of computational complexity too. 2011-11-25T10:53:21 well i have to generate paths 2011-11-25T10:53:48 luizribeiro: i have more important problems with my bot now, so maybe later 2011-11-25T10:55:24 *** AntDroid has joined #aichallenge 2011-11-25T10:59:04 i have to write messy functions to pass them around like variables in c++ don't i? 2011-11-25T10:59:07 is that horribly slow? 2011-11-25T10:59:21 i want to generalise my path finding stuff, but it requires throwing around functions 2011-11-25T11:00:12 C++11 has lambdas 2011-11-25T11:00:26 what are lambdas? 2011-11-25T11:00:36 anonymous functions you can pass around 2011-11-25T11:00:49 instead of just function pointers and functors 2011-11-25T11:01:49 I think the old C++ way was to pass around functors (structs with operator() overloaded) 2011-11-25T11:02:17 so gross 2011-11-25T11:02:31 that's what std::less, std::hash and std::greater are 2011-11-25T11:02:33 how have others generalised path finding? 2011-11-25T11:02:51 lambdas i take it for you? 2011-11-25T11:02:56 well, not yet :P 2011-11-25T11:03:07 haven't figured out the C++ syntax for them, but that's how I'll do it 2011-11-25T11:03:25 what a pita :P 2011-11-25T11:03:27 just need to pass in a lambda for stuff like the end condition 2011-11-25T11:03:53 i need to pass it in for what to do to the state when it finds a target and is updating information 2011-11-25T11:04:10 i'm making it so it collects a set of targets in one bfs sweep 2011-11-25T11:04:19 reusing ants that have already been moved if i tell it to 2011-11-25T11:04:21 yeah, you could stick that in a lambda or functor and pass it in 2011-11-25T11:04:35 functors are really weird though 2011-11-25T11:04:45 I think python actually has the same thing... never used it though 2011-11-25T11:05:58 http://docs.python.org/reference/datamodel.html#object.__call__ 2011-11-25T11:06:02 yeah, python has that too :\ 2011-11-25T11:10:12 meh, C++ has the worst syntax for everything 2011-11-25T11:11:28 ++ operator is pretty nice :) 2011-11-25T11:11:38 auto lambda = []int(int i)->int{return i * 2;}; 2011-11-25T11:11:42 meh. 2011-11-25T11:12:34 *** dom7b5 has quit IRC (Ping timeout: 245 seconds) 2011-11-25T11:12:44 haskell syntax: \i -> 2 * i 2011-11-25T11:13:20 thestinger: Do you have a bot on the official sever? 2011-11-25T11:13:20 *** BenJackson has quit IRC (Ping timeout: 265 seconds) 2011-11-25T11:13:33 UncleVasya: yeah, strcat (that version is ancient though) 2011-11-25T11:13:54 ok 2011-11-25T11:14:03 *** peyton has quit IRC (Quit: peyton) 2011-11-25T11:15:47 wow thestinger ancient but still on position 32! 2011-11-25T11:16:08 it has no combat 2011-11-25T11:16:16 and the foraging sucks 2011-11-25T11:16:22 it just gets early points from rushing 2011-11-25T11:16:49 *** peyton has joined #aichallenge 2011-11-25T11:17:53 on PlanetWars early rush was effective too. 2011-11-25T11:18:09 *** kire has joined #aichallenge 2011-11-25T11:18:55 *** sigh has quit IRC (Remote host closed the connection) 2011-11-25T11:19:15 *** aerique has quit IRC (Quit: ...) 2011-11-25T11:20:42 aichallenge: janzert epsilon * r7646779 / website/server_stats.php : Display lower server game rates - http://git.io/9icS1g 2011-11-25T11:21:16 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2011-11-25T11:21:19 thestinger: what do you mean by rushing? 2011-11-25T11:22:17 *** Surya has joined #aichallenge 2011-11-25T11:22:19 aichallenge: janzert epsilon * ra8049a7 / website/server_stats.php : Really display lower server game rates - http://git.io/VCL_bA 2011-11-25T11:22:22 it explores in a big group (which is pretty bad for finding food), finds an enemy hill and kills it 2011-11-25T11:22:35 so it can take out a few enemy hills early on before people have many ants 2011-11-25T11:22:46 wow that's cool 2011-11-25T11:23:06 but that version has no defense and it sucks at foraging (it isn't good at spreading out to find food) 2011-11-25T11:23:10 so it doesn't really last long 2011-11-25T11:23:31 *** peyton has quit IRC (Quit: peyton) 2011-11-25T11:23:40 it's an interesting approach though 2011-11-25T11:25:33 my bot still does something similar 2011-11-25T11:25:47 but it's a lot better after the first 100 turns 2011-11-25T11:25:54 *** peyton has joined #aichallenge 2011-11-25T11:26:13 it also walks quite nice 2011-11-25T11:26:32 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-25T11:26:39 are you looking at the version on the site or tcp server? 2011-11-25T11:26:46 site 2011-11-25T11:27:23 http://ants.fluxid.pl/replay.11443 2011-11-25T11:27:29 movement is different now 2011-11-25T11:27:40 they used to just spread out when there were multiple equally good paths 2011-11-25T11:28:00 movement is really the only thing I have working well 2011-11-25T11:28:07 maybe foraging too now that I worked on it a bit 2011-11-25T11:28:56 it's quite nice, you should upload it to the site 2011-11-25T11:29:19 can't atm 2011-11-25T11:29:26 I have combat stuff that takes too much time 2011-11-25T11:29:32 oh I see 2011-11-25T11:29:37 and it doesn't really work 2011-11-25T11:30:06 it works when enemies don't move :) 2011-11-25T11:30:13 haha 2011-11-25T11:30:26 for some reason the code for dealing with enemies that move takes a bunch of time, and seems to accomplish nothing 2011-11-25T11:30:36 after another 50 bug fixes it should sort of work 2011-11-25T11:31:06 I had to recode my bot from scratch in order to repair bugs and clean up a little bit the code 2011-11-25T11:34:22 I really need to clean up my code 2011-11-25T11:34:31 *** mviel_ has joined #aichallenge 2011-11-25T11:35:04 up to 1453 lines now without my new combat code (which is ~100 more lines) 2011-11-25T11:36:06 *** Boumbles has quit IRC (Quit: Page closed) 2011-11-25T11:37:59 *** mviel__ has quit IRC (Ping timeout: 252 seconds) 2011-11-25T11:38:09 *** Anilm3 has quit IRC (Quit: Lost terminal) 2011-11-25T11:41:18 *** mj41 has quit IRC (Ping timeout: 240 seconds) 2011-11-25T11:42:37 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-11-25T11:42:53 *** jon1 has quit IRC (Quit: Leaving.) 2011-11-25T12:00:36 aichallenge: janzert epsilon * rb8b5334 / website/server_stats.php : Switch to minutes per game for individual servers - http://git.io/9enb1w 2011-11-25T12:02:00 *** ikaros has joined #aichallenge 2011-11-25T12:02:55 aichallenge: janzert epsilon * rf6a991a / website/server_stats.php : Remove extraneous call - http://git.io/09P29Q 2011-11-25T12:04:18 aichallenge: janzert epsilon * rdef7f42 / website/server_stats.php : Drop server error percentage - http://git.io/wqf0Ew 2011-11-25T12:05:14 *** kapoc has joined #aichallenge 2011-11-25T12:12:25 *** dom7b5 has joined #aichallenge 2011-11-25T12:13:52 http://en.cppreference.com/w/cpp/utility/functional/function nice, this is what I needed 2011-11-25T12:14:03 higher order function time 2011-11-25T12:14:22 *** smiley1983 has quit IRC (Ping timeout: 260 seconds) 2011-11-25T12:15:00 aichallenge: janzert epsilon * rc862f94 / website/server_stats.php : Fix server rate styling - http://git.io/LER2ig 2011-11-25T12:15:43 *** mviel_ has quit IRC (Quit: Leaving) 2011-11-25T12:19:07 what is EPM anyway? 2011-11-25T12:19:28 I don't really know anymore :/ 2011-11-25T12:19:46 also it's kind of funny how it does all the cell maze maps for a while, then it does all the other ones 2011-11-25T12:19:49 it used to be number of timeouts and crashes but it's obviously not anymore 2011-11-25T12:20:22 it's because it is trying to keep the same number of games per map per day 2011-11-25T12:20:29 i'm also suspicious of my last game, where pguillory and i both timed out 2011-11-25T12:20:36 my bot should basically never ever timeout now 2011-11-25T12:20:58 *** smiley1983 has joined #aichallenge 2011-11-25T12:21:08 hmm :( 2011-11-25T12:21:10 yeah, i know, and they were all introduced at once 2011-11-25T12:21:18 right 2011-11-25T12:21:45 over the next few days they'll probably spread out 2011-11-25T12:22:11 damn, so i can't have a function pointer to a function for an object? 2011-11-25T12:22:19 an instance of an object* 2011-11-25T12:22:32 you mean an object method? 2011-11-25T12:23:12 antimatroid: http://www.parashift.com/c++-faq-lite/pointers-to-members.html 2011-11-25T12:23:17 I'm reading that atm 2011-11-25T12:23:19 so.. i assume it's showing me the first 1000 characters of my stdout here when i time out? 2011-11-25T12:23:28 wouldn't the *last* 1000 characters make much more sense? 2011-11-25T12:23:31 well i have say void State::updateStuff(), then I want for a State state, void *foo() and foo = &state.updateStuff? 2011-11-25T12:23:35 er, stderr, not stdout 2011-11-25T12:23:46 a1k0n: first and last 2011-11-25T12:24:00 i only have the first 2011-11-25T12:24:25 I think the blank line followed by the line with '..' is the split? 2011-11-25T12:24:37 oh, i see 2011-11-25T12:25:44 weird. it didn't print out the line that appears right after the map that takes a couple microseconds to evaluate 2011-11-25T12:26:07 thestinger: http://codepad.org/6HFjgLqh 2011-11-25T12:26:11 that's what i'm trying to get working 2011-11-25T12:26:19 or anyone else :P 2011-11-25T12:26:58 otherwise i might just have to have void updateStuff(State &state) outside State? 2011-11-25T12:28:51 I think I'm just going to make a lambda which takes a reference to State as an argument 2011-11-25T12:29:03 and pass that to methods 2011-11-25T12:29:13 I'm pretty much trying to figure out the same thing 2011-11-25T12:29:18 function pointers are really ugly 2011-11-25T12:29:53 std::function<> and lambdas are supposed to make it nice, but I can't find any good documentation 2011-11-25T12:30:14 *** smiley1983 has quit IRC (Ping timeout: 244 seconds) 2011-11-25T12:32:22 void method(std::function f); 2011-11-25T12:32:38 and then I'll pass it a lambda which alters a reference to State 2011-11-25T12:33:13 *** Surya has quit IRC (Quit: Page closed) 2011-11-25T12:33:16 *** liberforce has left #aichallenge 2011-11-25T12:35:38 thestinger: http://codepad.org/TAoAyA1R that worked, although kinda gross 2011-11-25T12:35:53 *** Blkt has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2011-11-25T12:36:25 antimatroid: if you use a lambda, you can make the function a local variable so it isn't as ugly 2011-11-25T12:36:30 cell_maze_p03_03 is just too big. 2011-11-25T12:36:39 i'm not using c++11 though :P 2011-11-25T12:37:15 *** smiley1983 has joined #aichallenge 2011-11-25T12:37:36 why not? 2011-11-25T12:39:12 nice 2011-11-25T12:39:23 auto f = [](State &s)->void{ s.test = true; }; 2011-11-25T12:39:25 f(*this); 2011-11-25T12:39:27 that works 2011-11-25T12:39:29 and I can pass it around 2011-11-25T12:39:40 void update_radius_info(std::function f, int radius); 2011-11-25T12:39:43 like that 2011-11-25T12:39:54 obfuscated by design 2011-11-25T12:39:57 awesome 2011-11-25T12:40:09 not really :\ 2011-11-25T12:40:16 it's just a really verbose lambda syntax 2011-11-25T12:41:42 Prelude> map (\i -> i + 1) [1..10] 2011-11-25T12:41:44 [2,3,4,5,6,7,8,9,10,11] 2011-11-25T12:41:48 it's just a really ugly version of that 2011-11-25T12:43:04 *** fnord has joined #aichallenge 2011-11-25T12:43:30 *** fnord is now known as Guest73993 2011-11-25T12:46:25 *** nplus has quit IRC (Quit: ChatZilla 0.9.87-rdmsoft [XULRunner 1.9.0.17/2009122204]) 2011-11-25T12:47:27 http://pastie.org/2920308 that works :) 2011-11-25T12:51:19 *** smiley1983 has quit IRC (Ping timeout: 245 seconds) 2011-11-25T12:57:52 *** mcstar has joined #aichallenge 2011-11-25T13:00:38 1 files changed, 48 insertions(+), 247 deletions(-) 2011-11-25T13:00:43 yay for higher order functions :) 2011-11-25T13:02:11 also cut down on the size of the binary so it's actually doing what it should be 2011-11-25T13:04:39 *** dom7b5 has quit IRC (Ping timeout: 245 seconds) 2011-11-25T13:05:38 *** epicmonkey has joined #aichallenge 2011-11-25T13:06:14 *** LouisMartin has joined #aichallenge 2011-11-25T13:09:19 *** dom7b5 has joined #aichallenge 2011-11-25T13:09:40 how do you copy an object when you have a reference? 2011-11-25T13:10:11 * avdg2 wonders if clone would do the job 2011-11-25T13:10:29 ie. if you had an input parameter int &ip and you wanted int d=??ip so that it wasn't just a reference? 2011-11-25T13:11:21 antimatroid: your doing it in c++ right? 2011-11-25T13:11:31 yeah 2011-11-25T13:11:40 http://en.wikipedia.org/wiki/Cloning_(programming) 2011-11-25T13:12:55 seems it just does it for me anyway 2011-11-25T13:13:00 nevermind my silly question :P 2011-11-25T13:13:12 I think it was a good question 2011-11-25T13:13:30 most people won't notice that they have a reference instead of a copy 2011-11-25T13:13:48 which is a big logical beginner mistake 2011-11-25T13:14:00 int d=ip; copies ip, wont it? 2011-11-25T13:14:09 i've made those mistakes enough times to know the difference :P 2011-11-25T13:14:21 mcstar: yep 2011-11-25T13:14:23 *** treeform has joined #aichallenge 2011-11-25T13:14:27 scalar values are copies 2011-11-25T13:14:47 assignment is the copy operaiton 2011-11-25T13:14:48 objects and arrays are mostly refered 2011-11-25T13:14:51 as far as i know 2011-11-25T13:14:53 i.e. = 2011-11-25T13:15:07 still the same if I have like State &state, then State newState = state? 2011-11-25T13:15:13 so for objects and arrays you mostly copy the reference 2011-11-25T13:15:13 yes 2011-11-25T13:15:13 i assume so, but making sure.. 2011-11-25T13:15:36 antimatroid: but you have to take care for that copy to make sense 2011-11-25T13:15:47 (dynamic allocations...) 2011-11-25T13:15:58 yeah 2011-11-25T13:16:17 *** Antimony has joined #aichallenge 2011-11-25T13:16:24 mcstar: i'm generalising my collection searches 2011-11-25T13:16:34 function pointers are a bit messy :\ 2011-11-25T13:16:53 std::function myfun; 2011-11-25T13:17:09 i use std::function everywhere 2011-11-25T13:17:14 no need to use function pointers 2011-11-25T13:17:19 Woohoo! I just beat up starter bots again. A couple more days and I'll have a meaningful ranking 2011-11-25T13:18:16 mcstar: that's not with plain c++ though right? 2011-11-25T13:18:19 *** smiley1983 has joined #aichallenge 2011-11-25T13:18:26 its stl 2011-11-25T13:18:32 i think its in c++03 too 2011-11-25T13:18:35 but check it 2011-11-25T13:19:15 antimatroid: c++11 2011-11-25T13:19:17 my A* is not working ... 2011-11-25T13:19:28 mcstar: how do i get that? :P 2011-11-25T13:19:30 antimatroid: but its very convenient 2011-11-25T13:19:34 *** JorgeB has joined #aichallenge 2011-11-25T13:19:39 yeah, i think you've convinced me :P 2011-11-25T13:19:42 #include 2011-11-25T13:20:05 for c++03 you need boost function 2011-11-25T13:20:20 *** Anilm3 has joined #aichallenge 2011-11-25T13:20:42 Antimony: what did you do with boost? 2011-11-25T13:20:59 a lot of parts of boost were added to the standard in c++11 2011-11-25T13:21:04 but if you're using c++03, you need boost 2011-11-25T13:21:12 function, unordered_set, shareD_ptr, etc. 2011-11-25T13:21:14 Antimony: you said you made a script or something 2011-11-25T13:21:18 mcstar: it's saying function is not a member of std :\ 2011-11-25T13:21:19 oh yeah that 2011-11-25T13:21:21 did it work? 2011-11-25T13:21:31 wouldn't i need to download something for 11? 2011-11-25T13:21:33 Antimony: where is it? 2011-11-25T13:21:35 Hi 2011-11-25T13:21:38 on the forums 2011-11-25T13:21:43 antimatroid: --std=c++0x 2011-11-25T13:21:48 just one - 2011-11-25T13:21:53 2 is ok to 2011-11-25T13:21:55 o 2011-11-25T13:21:58 mcstar: that means nothing to me? :\ 2011-11-25T13:22:02 http://forums.aichallenge.org/viewtopic.php?f=25&t=1944 2011-11-25T13:22:06 antimatroid: compile with that flag 2011-11-25T13:22:16 CFLAG+=--std=c++0x 2011-11-25T13:22:21 CFLAGS+=--std=c++0x 2011-11-25T13:22:25 CXXFLAGS+=--std=c++0x 2011-11-25T13:22:29 depending what us set 2011-11-25T13:22:39 mcstar: yeah, I'm using std::function and lambdas now 2011-11-25T13:22:41 so nice 2011-11-25T13:24:22 *** Guest73993 is now known as fnord42 2011-11-25T13:24:28 i'm retarded, where anyone know where i can set that stuff in code::blocks? 2011-11-25T13:25:00 got to project properties 2011-11-25T13:25:07 *** JorgeB has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-11-25T13:25:07 wiat 2011-11-25T13:25:12 project build options 2011-11-25T13:25:14 *** JorgeB has joined #aichallenge 2011-11-25T13:25:40 Go to Project->Build Options and it's under compile flags 2011-11-25T13:26:31 antimatroid: what codeblocks? 2011-11-25T13:26:36 put it into the makefile 2011-11-25T13:26:41 it's an IDE 2011-11-25T13:26:45 i know 2011-11-25T13:26:49 i just pretend i dont 2011-11-25T13:26:54 he knows i know 2011-11-25T13:27:14 i even tried it 2011-11-25T13:27:18 hat eit 2011-11-25T13:27:21 clang_complete has finally decided to kill itself now that I'm using lambdas 2011-11-25T13:27:34 *** dom7b5 has quit IRC (Ping timeout: 245 seconds) 2011-11-25T13:27:45 Antimony: got it, thanks :) 2011-11-25T13:27:56 also, you have too many letters from my username in your username :P 2011-11-25T13:28:06 its annyoing 2011-11-25T13:28:09 it is 2011-11-25T13:28:25 do myhills contains my hills even if I'm out of visual range? 2011-11-25T13:28:36 in the C++ starter? no 2011-11-25T13:28:49 you have to see the hill to know if it's still there 2011-11-25T13:29:05 thestinger: ok, thanks 2011-11-25T13:30:47 that changes everything :( 2011-11-25T13:31:35 how do i make the server recognise c++11 when i upload later? 2011-11-25T13:31:48 switch the extension to .cpp and the compiler script uses -std=c++0x 2011-11-25T13:31:49 Anilm3: it's really easy to make it do it 2011-11-25T13:32:06 the extension on MyBot.xx is all that matters iirc 2011-11-25T13:32:49 thestinger: the server compiler uses my Makefile? 2011-11-25T13:32:55 Anilm3: no 2011-11-25T13:33:32 https://github.com/aichallenge/aichallenge/blob/epsilon/worker/compiler.py 2011-11-25T13:33:48 Anilm3: add an oldHills, oldFood etc. vector to state, then read input, update vision, for any old food if the location isn't visible re add it like it was given as input info otherwise it was removed since last seen (or hill razed since last seen etc.), then once you've updated the information copy the current food, hills etc. into the old info vectors 2011-11-25T13:34:17 for C, it compiles all the .c files to .o, then links them together (without link-time optimization sadly, I forget if gcc 4.5 has it though) 2011-11-25T13:34:30 C++ is the same 2011-11-25T13:34:32 if you're clever you should also be able to track where your ants spawn to see if a hill of yours out of view has been razed because an ant didn't spawn there 2011-11-25T13:34:54 *** ztfw has joined #aichallenge 2011-11-25T13:34:56 thestinger: cheers :) 2011-11-25T13:35:07 antimatroid: thanks 2011-11-25T13:35:42 *** smiley1983 has quit IRC (Ping timeout: 240 seconds) 2011-11-25T13:35:43 i actually use sets for those kinds of things 2011-11-25T13:35:44 do you do any kind of enemy movements prediction? 2011-11-25T13:35:54 and have like moveable and unmoved ants sets for my ants 2011-11-25T13:35:59 only for combat 2011-11-25T13:36:13 they could use march=native on the c/c++ compiler 2011-11-25T13:36:20 you could presumably "collect food" etc for them and try to cut them off i guess 2011-11-25T13:36:21 it sometimes helps 2011-11-25T13:36:23 that could be cool 2011-11-25T13:36:44 making the assumption that they don't collect food they can get first no matter what 2011-11-25T13:37:33 if you're going to this much trouble you probably want to hope you're going to do quite well, so it's usually a safe assumption your competition will be reasonably smart about what they do 2011-11-25T13:38:18 I just assume all enemies are stationary 2011-11-25T13:38:22 it works occasionally 2011-11-25T13:38:46 thestinger: i don't get what you're using lambdas for? 2011-11-25T13:38:52 care to make me a simple code pad example? :P 2011-11-25T13:38:59 sure, sec 2011-11-25T13:39:09 std::function seems to do all i need, but might as well learn while i'm at it :P 2011-11-25T13:39:28 well, have you used anonymous functions before? 2011-11-25T13:39:30 in other langs 2011-11-25T13:39:33 *** smiley1983 has joined #aichallenge 2011-11-25T13:39:43 i use lambdas anywhere when i want to geenralize something 2011-11-25T13:40:07 and i need to provide access to the this pointer for example 2011-11-25T13:40:19 i can share data with different objects 2011-11-25T13:40:39 nope :P 2011-11-25T13:40:42 wait a moment, have multihill maps been removed? 2011-11-25T13:40:43 antimatroid: lambdas can capture variables from the calling environment 2011-11-25T13:40:46 c++ is pretty much the only language i've properly used 2011-11-25T13:41:45 ah, not calling, but where i define that labda 2011-11-25T13:41:55 yeah 2011-11-25T13:42:01 you can pass stuff to them where they get called 2011-11-25T13:42:01 *** teste has joined #aichallenge 2011-11-25T13:42:02 i dont know what environment that is 2011-11-25T13:42:09 *** teste has quit IRC (Client Quit) 2011-11-25T13:42:09 so you can do that anyway 2011-11-25T13:42:23 its a controlled bridge 2011-11-25T13:42:26 i like it 2011-11-25T13:42:34 antimatroid: basically, you can define a function as a local variable 2011-11-25T13:43:03 and you can have it capture state from the function you define it in 2011-11-25T13:43:18 that too, i can set up a function variable, but after some citerion holds, i can change that function varianle to another funciton 2011-11-25T13:43:30 so i dont neeed to check selection variables all the time 2011-11-25T13:43:35 its neater 2011-11-25T13:44:07 *** avdg2 has quit IRC (Remote host closed the connection) 2011-11-25T13:45:43 antimatroid: the syntax is just [](int a, int b)->int{ return a + b; } 2011-11-25T13:45:52 and you can use auto if you want to assign it to a variable 2011-11-25T13:46:00 you dont need -> 2011-11-25T13:46:05 so 'auto f = lambda' or whatever 2011-11-25T13:46:13 since auto does local type inference in C++11 2011-11-25T13:46:14 * antimatroid has no idea what you two are on about 2011-11-25T13:46:23 antimatroid: trying to help you 2011-11-25T13:46:28 :P 2011-11-25T13:46:30 other than that 2011-11-25T13:46:39 thestinger: you dont have to provide return type 2011-11-25T13:46:57 so, 'auto add = [](int a, int b){ return a + b; }' ? 2011-11-25T13:47:03 yes 2011-11-25T13:47:24 is this so i could have a function for like state.aFunction? 2011-11-25T13:47:27 antimatroid: alright, so try 'auto add = [](int a, int b){ return a + b; }' and then call it as a function 2011-11-25T13:47:47 the first nice thing about it is that you have a local nested function 2011-11-25T13:48:21 so you don't need to jump somewhere else to see the definition for something you only use in one place 2011-11-25T13:48:25 auto addFn3=[this, &rc](int r, int c){ 2011-11-25T13:48:39 function addFn=addFn1; 2011-11-25T13:48:42 function addFn=addFn3; 2011-11-25T13:49:04 *** avdg has joined #aichallenge 2011-11-25T13:49:38 i'm confused :S 2011-11-25T13:49:39 nevermind 2011-11-25T13:49:49 antimatroid: and you can change the definition of the function, since it's just a local variable 2011-11-25T13:50:19 and you can give it variables from the current function to start it off with some start or whatever 2011-11-25T13:50:41 antimatroid: [] captures the variables 2011-11-25T13:50:49 &rc is captured by reference 2011-11-25T13:51:01 rc is defined where addFn, addFn3 is 2011-11-25T13:51:07 () is the parameter list 2011-11-25T13:51:13 just like normal funcitons 2011-11-25T13:51:19 *** Antimony has quit IRC (Ping timeout: 245 seconds) 2011-11-25T13:52:02 function addFn=addFn1; 2011-11-25T13:52:07 this coulbe also: 2011-11-25T13:52:13 auto addFn=addFn1; 2011-11-25T13:52:33 antimatroid: auto just infers the type for you 2011-11-25T13:53:14 so instead of something like 'std::vector::const_iterator it' you can just write 'auto it' 2011-11-25T13:53:15 function == float(*)(int, int) 2011-11-25T13:54:19 I put off figuring out C++11 lambdas until today, and I regret it 2011-11-25T13:54:26 for some reason I thought they would be horrible 2011-11-25T13:54:45 thestinger: i started to use them the moment i heard about them 2011-11-25T13:54:47 in my eyes it looks horrible atm 2011-11-25T13:55:07 but maybe one day :-) 2011-11-25T13:55:18 mcstar: I've used function pointers before in C, and I thought it would as horrible as that xD 2011-11-25T13:55:43 lambdas as function pointers? 2011-11-25T13:55:57 they are different concepts 2011-11-25T13:56:00 yeah 2011-11-25T13:56:05 of course you know that 2011-11-25T13:56:07 I didn't think they would be proper lambdas 2011-11-25T13:56:36 after seeing what a 'functor' was in C++ I lost hope for lambdas and just never looked up any docs 2011-11-25T13:56:36 i didnt have time to theorize lambda's abilities 2011-11-25T13:56:54 huh 2011-11-25T13:56:56 functor 2011-11-25T13:56:58 lol 2011-11-25T13:57:11 i guess you know they arent really functors 2011-11-25T13:57:17 some languages have functors 2011-11-25T13:57:27 yeah, they are friggen structs with operator() overloaded what the hell 2011-11-25T13:57:28 *** jcdny has joined #aichallenge 2011-11-25T13:57:29 *** dom7b5 has joined #aichallenge 2011-11-25T13:57:36 in c++ a functor is just higher order s 2011-11-25T13:57:37 bs 2011-11-25T13:57:38 I really don't understand what the purpose is tbh 2011-11-25T13:57:51 syntax 2011-11-25T13:58:03 im mean sytactic sugar 2011-11-25T13:58:17 oh 2011-11-25T13:58:28 it's a way of making the ugliness of function pointers go away? 2011-11-25T13:58:35 no 2011-11-25T13:58:37 i think 2011-11-25T13:58:45 you just spare a .memberfunction 2011-11-25T13:58:46 *** Antimony has joined #aichallenge 2011-11-25T13:58:51 well, std::less, std::hash, std::greater, etc. are functors 2011-11-25T13:58:51 myclass() 2011-11-25T13:58:56 instead of 2011-11-25T13:58:56 rather than functions 2011-11-25T13:59:01 myclass.call() 2011-11-25T13:59:10 yeah, python lets you define __call__() for the same thing 2011-11-25T13:59:37 but I think the idiom for functors in C++ is a struct with a single method (the operator() overload) 2011-11-25T13:59:51 that's what std::less, etc. are 2011-11-25T14:00:08 *** bugnuts has quit IRC (Ping timeout: 244 seconds) 2011-11-25T14:00:11 the optional third argument to std::sort is a functor 2011-11-25T14:00:13 i read somewhere that in c++ functors are functions that take function arguemnts 2011-11-25T14:00:31 *** smiley1983 has quit IRC (Ping timeout: 248 seconds) 2011-11-25T14:00:51 int i = add42(8); // and "call" it 2011-11-25T14:00:53 if you know that what you described are functors ill believe you 2011-11-25T14:00:54 what 2011-11-25T14:00:58 weird paste 2011-11-25T14:01:13 *** smiley1983 has joined #aichallenge 2011-11-25T14:01:17 http://stackoverflow.com/questions/356950/c-functors-and-their-uses 2011-11-25T14:01:28 nvm, apparently any class with operator() overloaded is a functor 2011-11-25T14:01:53 lame 2011-11-25T14:02:11 I think it's just to get rid of the ugly function pointer syntax, before they had std::function 2011-11-25T14:02:19 and it's also a way to avoid 'evil' static local variables 2011-11-25T14:02:27 same thing though, different syntax 2011-11-25T14:02:33 c++ doesnt have closures 2011-11-25T14:02:38 thats why they needed that 2011-11-25T14:02:47 ah 2011-11-25T14:03:26 have you had to make any functors? 2011-11-25T14:03:36 (setf myfun (let ((var 2)) (lambda (x) (+ x var)))) 2011-11-25T14:03:49 now myfun has state 2011-11-25T14:03:55 its a proper closure in cl 2011-11-25T14:03:58 *** rexou has quit IRC (Read error: Operation timed out) 2011-11-25T14:04:09 lambda captures var 2011-11-25T14:04:21 ah 2011-11-25T14:04:34 i needed to add std::hash> 2011-11-25T14:04:37 apparently you aren't supposed to use C++ functors anymore 2011-11-25T14:04:46 why? 2011-11-25T14:04:47 except when interfacing with the STL ofc 2011-11-25T14:05:34 Antimony: because there are lambdas and a nice way to pass around functions now (keep in mind I just read this on stackoverflow from a few commenters :P) 2011-11-25T14:05:47 anyway, for that hash i needed to overload operator(), but i never did it vefore 2011-11-25T14:06:00 yeah, I had to do that for std::hash too 2011-11-25T14:06:16 static hash ull_hasher; 2011-11-25T14:06:20 i did speed tests 2011-11-25T14:06:34 unfortunately, my version of gcc doesn't support lambdas 2011-11-25T14:06:37 theres no speed-improvement to use a static member here 2011-11-25T14:06:38 I'm stuc with ugly code 2011-11-25T14:06:55 i initilize ull_hasher outside: 2011-11-25T14:07:06 std::hash std::hash::ull_hasher; 2011-11-25T14:07:24 oh, i broke my hashfunction just to see what will happen 2011-11-25T14:07:36 i got a ~ 100x slowdown 2011-11-25T14:07:43 but results were still correct 2011-11-25T14:07:43 *** kire has quit IRC (Remote host closed the connection) 2011-11-25T14:08:05 (not that it shouldnt be, but i checked) 2011-11-25T14:08:29 you can check the number of collisions 2011-11-25T14:08:33 it handles them fine 2011-11-25T14:08:40 yep 2011-11-25T14:08:51 it was collided every time 2011-11-25T14:09:06 *** rexou has joined #aichallenge 2011-11-25T14:13:34 *** Anilm3 has quit IRC (Quit: Lost terminal) 2011-11-25T14:15:41 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-25T14:17:07 hey, xathis's skill is coming down 2011-11-25T14:18:54 *** smiley1983 has quit IRC (Ping timeout: 240 seconds) 2011-11-25T14:19:24 *** smiley1983 has joined #aichallenge 2011-11-25T14:19:41 *** BenJackson has joined #aichallenge 2011-11-25T14:19:56 @seen a1k0n 2011-11-25T14:19:56 BenJackson: a1k0n was last seen in #aichallenge 2 minutes and 48 seconds ago: hey, xathis's skill is coming down 2011-11-25T14:20:05 a1k0n: http://ants.fluxid.pl/replay.14820 quite a battle 2011-11-25T14:20:35 nice. 2011-11-25T14:20:45 bummer about the turn limit, i would have gotten you in the end. :) 2011-11-25T14:20:58 I'm sure I would have turned it around ;-) 2011-11-25T14:21:37 jesus my bot sacrifices a lot of ants needlessly 2011-11-25T14:23:13 I'd say I was winning that one in combat 2011-11-25T14:23:24 but I failed to leave any explorers out there 2011-11-25T14:23:38 yeah 2011-11-25T14:23:53 i just made up for my suicide tactics by fast collection 2011-11-25T14:25:01 i have two version of my bot, one of which should theoretically be much better in that situation, but i'm not sure whether it plays better overall 2011-11-25T14:26:33 *** smiley1983 has quit IRC (Remote host closed the connection) 2011-11-25T14:26:49 *** smiley1983 has joined #aichallenge 2011-11-25T14:28:24 *** HaraKiri has quit IRC () 2011-11-25T14:29:17 yeah, that's just it 2011-11-25T14:29:50 it's the same old thing I keep reminding myself of: 2011-11-25T14:29:57 it's hard to write a program to do something if you don't know what you want 2011-11-25T14:30:10 *** liberforce has joined #aichallenge 2011-11-25T14:30:20 so I know I want "more explorers" but it's hard to draw a line in that game and say where the bot went wrong 2011-11-25T14:30:25 and exactly what it would do differently 2011-11-25T14:30:29 right 2011-11-25T14:30:55 well, i know exactly what to do, but it's hard for me to evaluate different versions of my bot by playing them against each other 2011-11-25T14:31:14 since they're so similar, they tend to not engage hills and just pick off just-formed battle groups 2011-11-25T14:31:59 *** smiley1983 has quit IRC (Ping timeout: 248 seconds) 2011-11-25T14:32:02 so it's just a map full of lone ants collecting food and killing each other until the turn limit 2011-11-25T14:34:42 *** pairofdice has quit IRC (Quit: leaving) 2011-11-25T14:38:32 http://fluxid.pl/misc/2011-11-25-203342_669x694_scrot.png :( 2011-11-25T14:39:19 Fluxid: ? 2011-11-25T14:40:03 where did you get that map ? 2011-11-25T14:40:12 your shell script bot didn't kill holdbot? 2011-11-25T14:40:18 i made it myself 2011-11-25T14:40:18 g0llum: they're just text files, you can make them in an editor 2011-11-25T14:40:20 a1k0n: right 2011-11-25T14:40:32 a1k0n: and it los more ants that holdbot while attacking 2011-11-25T14:40:33 lol 2011-11-25T14:40:38 k,k 2011-11-25T14:40:46 i'm like 2011-11-25T14:40:48 is that up on paste somewhere? 2011-11-25T14:40:54 you... seriously wrote a bot in bourne shell? 2011-11-25T14:40:59 *** mj41 has joined #aichallenge 2011-11-25T14:41:17 a1k0n: well, no, it's just a wrapper for python bot 2011-11-25T14:41:21 heh, ok, i thought so 2011-11-25T14:41:23 it redirects logging and stuff 2011-11-25T14:42:05 what's worse, it depends on seed 2011-11-25T14:42:15 with different seed it killed holdbot 2011-11-25T14:42:20 hahah my bot loses to holdbot too 2011-11-25T14:42:59 *** Matt_ has joined #aichallenge 2011-11-25T14:43:01 hmm, that's depressing 2011-11-25T14:43:25 *** Matt_ is now known as Guest41292 2011-11-25T14:43:32 *** Guest41292 is now known as MattBradley 2011-11-25T14:43:32 i kind of predict that the enemy isn't stationary 2011-11-25T14:43:33 http://paste.aichallenge.org/KzRFO/ here's the map 2011-11-25T14:44:21 http://paste.aichallenge.org/DgWCM/ and here's fourth game 2011-11-25T14:44:37 hold bot only lost *one* ant :| 2011-11-25T14:46:18 http://paste.aichallenge.org/CHRCm/ cool, someone uses two ants to kill single ants 2011-11-25T14:46:45 that scares me :( 2011-11-25T14:46:57 my ants are going to need a buddy system 2011-11-25T14:47:02 works really well: http://paste.aichallenge.org/P7HPs/ 2011-11-25T14:47:14 i'm scared too 2011-11-25T14:47:17 wow 2011-11-25T14:47:19 that is really good 2011-11-25T14:47:55 well, whoever that is just won the contest 2011-11-25T14:47:58 xD 2011-11-25T14:48:16 he actually moves his ants to kill... it's far from my capabilites 2011-11-25T14:48:20 hopefully it's really slow 2011-11-25T14:48:28 http://paste.aichallenge.org/lbvap/ 2011-11-25T14:48:39 he'll be killing my ants. one. by. one. 2011-11-25T14:49:06 BenJackson: cool 2011-11-25T14:49:09 *** smiley1983 has joined #aichallenge 2011-11-25T14:49:23 *** smiley1983 has joined #aichallenge 2011-11-25T14:49:45 my bot is stupid, but still takes the hill: http://paste.aichallenge.org/DIijW/ 2011-11-25T14:50:23 I accidentallyr an that with a turn limit the first time 2011-11-25T14:50:27 well, my bot succeeds about half the time 2011-11-25T14:50:51 I thought my bot was going to sneak around on the left 2011-11-25T14:51:08 *** choas has joined #aichallenge 2011-11-25T14:51:09 Fluxid: that waas my paste 2011-11-25T14:51:11 i didn't know that weight of 0.05 of randomness is so powerfull in such simple case 2011-11-25T14:51:16 Fluxid: well of all the things I may do wrong I'm pretty sure a stationary bot is not going to defeat me ;-) 2011-11-25T14:51:25 mcstar: you mean using two ants to kill? 2011-11-25T14:51:29 yep 2011-11-25T14:51:32 you see the bug too 2011-11-25T14:51:34 mcstar: man, that's really cool 2011-11-25T14:51:38 the first pair isnt supposed to die 2011-11-25T14:51:47 not really, i'm too amazed 2011-11-25T14:51:47 ;) 2011-11-25T14:52:15 heh, if i comment out the thing where it predicts enemy moves: http://paste.aichallenge.org/wDdqr/ 2011-11-25T14:52:29 mcstar: any keyword on what you do? 2011-11-25T14:52:54 Those "two ant kills" aren't that impressive 2011-11-25T14:52:55 Fluxid: the other one is mine too :D 2011-11-25T14:53:03 yeah, thats simple 2011-11-25T14:53:04 my bot would do that too against randombot 2011-11-25T14:53:23 BenJackson: i'm impressed because i can't do that.. i think 2011-11-25T14:53:24 if the enemy dodges you basically can't kill them unless you run them into an obstacle 2011-11-25T14:54:12 and then you actually have to chase them down 2011-11-25T14:54:35 *** Accoun has quit IRC () 2011-11-25T14:54:40 i'm reworking my combat right now 2011-11-25T14:54:52 i wonder how it'll work 2011-11-25T14:54:53 *** yyTH has joined #aichallenge 2011-11-25T14:55:02 combat, or rather interaction with enemy 2011-11-25T14:55:51 *** peyton has quit IRC (Quit: peyton) 2011-11-25T14:59:48 *** peyton has joined #aichallenge 2011-11-25T15:00:42 *** mtmcannon has joined #aichallenge 2011-11-25T15:06:09 http://fluxid.pl/misc/2011-11-25-210118_681x693_scrot.png this is BAD 2011-11-25T15:06:40 * BenJackson LAUGHS 2011-11-25T15:07:29 but this is test run after changes... 2011-11-25T15:08:55 hi 2011-11-25T15:09:06 what's keeping me from resubmitting on the official server is that I fixed combat but it turned out my bot sort of relied on the blundering semi-suicidal combat quality 2011-11-25T15:09:18 so I need to make other improvements to regain parity 2011-11-25T15:10:03 *** Accoun has joined #aichallenge 2011-11-25T15:10:50 *** cafaro has quit IRC (Quit: farewell) 2011-11-25T15:14:15 *** JorgeB has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-11-25T15:14:23 *** yyTH has left #aichallenge ("Sto andando via") 2011-11-25T15:15:59 *** HaraKiri has joined #aichallenge 2011-11-25T15:19:21 *** hacklash has joined #aichallenge 2011-11-25T15:20:12 you mean party? 2011-11-25T15:22:13 well, it sucks 2011-11-25T15:22:25 in fluxid's situation my ants dont move 2011-11-25T15:22:35 they stand in front of the enemy ants 2011-11-25T15:23:20 have any of you considered some asynchronous message passing between ants? 2011-11-25T15:23:33 so as to coordinate them for common goals? 2011-11-25T15:23:45 "hey buddy, lets move there ok? im going" 2011-11-25T15:23:54 "damn, dont maan, im going there" 2011-11-25T15:24:05 "ok, then i find another objective :(" 2011-11-25T15:24:08 Did you watch my replay? 2011-11-25T15:24:16 you can see bits where everyone shuffles in place 2011-11-25T15:24:22 *** Antimony has joined #aichallenge 2011-11-25T15:24:27 you see it in other bots too 2011-11-25T15:24:36 i consider shuffling bad 2011-11-25T15:24:38 something of a hint about what's goign on 2011-11-25T15:24:52 2 ants changing places? 2011-11-25T15:24:53 Has anyone here done symmetry detection yet? 2011-11-25T15:25:08 Antimony: I implemented a proof of concept but I'm not using it 2011-11-25T15:25:28 I got stuck when I realized that it's psosible to have half square reflections 2011-11-25T15:25:34 *** Anilm3 has joined #aichallenge 2011-11-25T15:25:47 BenJackson: im asking about the AMP, which would replace nested function calls 2011-11-25T15:25:49 Hi 2011-11-25T15:26:01 Have you gotten the tcp client to work? 2011-11-25T15:26:19 sure 2011-11-25T15:26:28 it more of a technicality im asking about, not tactics 2011-11-25T15:26:41 are you on windows? 2011-11-25T15:26:50 not for that 2011-11-25T15:27:42 *** amstan_ has joined #aichallenge 2011-11-25T15:27:42 *** ChanServ sets mode: +o amstan_ 2011-11-25T15:27:54 I'm on Windows, but when I try to use tcp, it randomly crashes a lot 2011-11-25T15:28:06 with random os errors 2011-11-25T15:28:15 C++ bot? 2011-11-25T15:28:22 yeah 2011-11-25T15:28:24 the C++ starter kit crashes on empty input 2011-11-25T15:28:30 how so? 2011-11-25T15:28:41 it dosen't initialize rows/cols 2011-11-25T15:28:52 it has no code whatsoever to handle the case of no input 2011-11-25T15:28:56 oh 2011-11-25T15:29:09 well that can't explain why it would crash on say turn 112 2011-11-25T15:29:09 I never noticed it until I used the tcpclient 2011-11-25T15:29:13 no 2011-11-25T15:29:28 BenJackson: weird, mine has never crashed 2011-11-25T15:29:39 what's an easy way to cause the problem? 2011-11-25T15:29:44 crashing on turn 112 is probably a bug in your code stimulated by input you get from complex opponents and not your local testing 2011-11-25T15:29:55 ./MyBot < /dev/null 2011-11-25T15:30:15 it just outputs 'go' 2011-11-25T15:30:16 but why would it say windows error in the client then? 2011-11-25T15:30:19 and valgrind is clean 2011-11-25T15:30:23 shouldn't it just say my bot crashed? 2011-11-25T15:30:33 I haven't really touched the code for input 2011-11-25T15:30:44 also, if you don't pass input the bot doesn't crash, it just waits for input 2011-11-25T15:32:05 thestinger: State doesn't initializes rows/cols and then allocates grid iirc 2011-11-25T15:32:15 so if it happens those are 0 anyway you wouldn't see it 2011-11-25T15:32:39 that happens if you get passed invalid input, not no input at all 2011-11-25T15:32:45 with no input it just waits 2011-11-25T15:32:55 I mean no input + EOF 2011-11-25T15:32:58 BenJackson: oh, that's true 2011-11-25T15:33:18 maybe I should put in some code to verify that the input is valid 2011-11-25T15:33:30 but that's not the problem here 2011-11-25T15:33:55 weird, it still doesn't crash if I init rows to -1 2011-11-25T15:34:15 there, got it to crash 2011-11-25T15:34:17 -2 works 2011-11-25T15:34:19 xD 2011-11-25T15:35:32 anyway the tcpclient would time out and close the pipe to terminate the bot and crash it, but I don't think it's giong to matter for normal contest purposes 2011-11-25T15:42:28 *** liberforce has quit IRC (Quit: Leaving.) 2011-11-25T15:42:40 *** peyton has quit IRC (Quit: peyton) 2011-11-25T15:43:08 somehow some of my ants refuse to liste and go close to enemy with, for example, -738 points (the more the higher propability to go this way, usually its between 0 and 5) 2011-11-25T15:43:35 listen, closer, it's* 2011-11-25T15:43:45 of going* 2011-11-25T15:45:27 *** peyton has joined #aichallenge 2011-11-25T15:51:50 I fucked up a score function the other day 2011-11-25T15:51:56 oh man, i've reduced my python bot to an absolute bare minimum, and there are still random slowdowns of 20-60ms. i'm this close to giving up :( 2011-11-25T15:52:01 I could see from the visualizer it was doing the wrong thing 2011-11-25T15:52:03 but the code looked fine 2011-11-25T15:52:06 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-11-25T15:52:09 so I deleted it and wrote it again 2011-11-25T15:52:10 worked. 2011-11-25T15:52:25 *** peyton has quit IRC (Quit: peyton) 2011-11-25T15:53:03 I had some problems too, and after 2hrs testing I realized that I just copied some functions but forgot to modify them 2011-11-25T15:53:18 garbage collection maybe 2011-11-25T15:53:43 oh, worked you said 2011-11-25T15:53:58 i disabled gc, all file output. tried 3 machines, linux, windows, eacah with 32 and 64bit, with and without psycho and/or pypy 2011-11-25T15:54:30 *** Redgis has joined #aichallenge 2011-11-25T15:54:32 ah, found a problem 2011-11-25T15:55:17 Anilm3: oh I'm not using python 2011-11-25T15:55:20 I started with python and found it waaay too slow 2011-11-25T15:56:08 *** kire has joined #aichallenge 2011-11-25T15:56:23 *** smiley1983 has quit IRC (Ping timeout: 244 seconds) 2011-11-25T15:56:36 HaraKiri: no idea, i didn't measure time to detect such slowdowns.. 2011-11-25T15:57:04 sure. i'm just raging into the wild, ignore me :x 2011-11-25T15:57:45 but maybe someone else knows 2011-11-25T15:57:58 that's the issue with interpreted languages, you don't really know what's going on 2011-11-25T15:58:21 garbage collection isn't really a problem with cpython 2011-11-25T15:58:23 HaraKiri: how do you detect that there is slowdown? 2011-11-25T15:58:31 yes but it's really just some readings in a bool array. how can that possible not be in the amount of time every time? it is infuriating 2011-11-25T15:58:39 a lot of timers all over the program Fluxid 2011-11-25T15:59:13 refcounting is used for everything, I think GC just deals with circular references 2011-11-25T15:59:14 HaraKiri: you mean you check if function is taking constant time, but it does not? 2011-11-25T15:59:44 do you measure real time or cpu time? 2011-11-25T15:59:58 i've reduced my program to ~20 lines or so, there are no fuctions :D. it's ~1000 array readings. most of the time it's ~1ms, but some rounds it takes 80ms 2011-11-25T16:00:11 O_o 2011-11-25T16:00:32 reduced for tracking down taht bug of course, not the "total" bot itself 2011-11-25T16:00:51 no idea 2011-11-25T16:01:33 i have enough optimising and measuring at my job so, sorry, i give up immediately 2011-11-25T16:01:43 but rewriting in C++ at this stage is so much work :(. haha ok 2011-11-25T16:02:52 *** fnord42 has quit IRC (Ping timeout: 265 seconds) 2011-11-25T16:03:20 *** mtmcannon has quit IRC (Ping timeout: 265 seconds) 2011-11-25T16:03:21 *** MattBradley has quit IRC (Ping timeout: 265 seconds) 2011-11-25T16:03:56 Anilm3: same problem in pypy, which is a JIT. just by the way, to defend python :D 2011-11-25T16:04:48 *** BenJackson has quit IRC (Ping timeout: 265 seconds) 2011-11-25T16:04:48 *** LouisMartin has quit IRC (Ping timeout: 265 seconds) 2011-11-25T16:04:48 *** AntDroid has quit IRC (Ping timeout: 265 seconds) 2011-11-25T16:05:06 *** delt0r_ has joined #aichallenge 2011-11-25T16:05:11 *** kurti has joined #aichallenge 2011-11-25T16:07:31 *** besh has joined #aichallenge 2011-11-25T16:09:59 HaraKiri: rewriting on C++ would give you a lot of advantages 2011-11-25T16:10:16 well it would be faster of course. but that's basically it :D 2011-11-25T16:10:26 it's a good reason 2011-11-25T16:10:39 Hara' dont feel too bad about the slowdowns 2011-11-25T16:10:44 i agree! but it's not as nice to write as python 2011-11-25T16:10:48 Even in C++, I've seen random slowdowns of over 30ms 2011-11-25T16:10:53 how can that be? 2011-11-25T16:10:56 it's just OS stuff outside of your control 2011-11-25T16:11:03 CPU caches are evil 2011-11-25T16:11:06 i write in python... and c extension for speedups 2011-11-25T16:11:06 they hate me 2011-11-25T16:11:21 Question: Are the locations of all of my hills sent at turn 0 ? 2011-11-25T16:11:27 yes 2011-11-25T16:11:33 besh: they should 2011-11-25T16:11:36 The only solution I've found is to return extra early 2011-11-25T16:11:46 it's just so frustrating because by whole bot only takes ~20ms at the starting phase of the game, and then i see shit like 80ms at turn 2 2011-11-25T16:12:16 Ok. so I can pre-calculate some things for the visisble area around the hill. 2011-11-25T16:12:23 besh: I don't think so 2011-11-25T16:12:30 I think they are sent on 1 2011-11-25T16:12:41 turn 0 corresponds to configuration 2011-11-25T16:12:42 yes 2011-11-25T16:12:51 besh, your visible area will change 2011-11-25T16:13:27 I also want all the waters around the hills within the visible radius 2011-11-25T16:14:12 *** iglo has quit IRC (Remote host closed the connection) 2011-11-25T16:14:18 *** Rav3nsW0rd has joined #aichallenge 2011-11-25T16:14:35 Do I get one ant per hill ? If that is the case , then that is all I need. 2011-11-25T16:15:51 yeah, you start with an ant on every hill 2011-11-25T16:16:17 Ok. So all the water around the hill should be sent at turn 1. 2011-11-25T16:16:49 besh, but notice, that all maps in the contest curretly have only 1 hill per player 2011-11-25T16:17:25 *** Palmik has quit IRC (Remote host closed the connection) 2011-11-25T16:17:56 *** BenJackson has joined #aichallenge 2011-11-25T16:18:10 the multi hill ones seem to be gone 2011-11-25T16:18:57 http://fluxid.pl/misc/2011-11-25-221354_704x699_scrot.png duh, it didn't do any move... it's too scared 2011-11-25T16:18:58 I think so. But they may be included in the final tournament 2011-11-25T16:19:05 but at least i didn;t loose any ant 2011-11-25T16:19:06 *** peyton has joined #aichallenge 2011-11-25T16:19:21 yea, prepare for the wurst 2011-11-25T16:24:18 *** JorgeB has joined #aichallenge 2011-11-25T16:29:16 wurstli 2011-11-25T16:35:49 *** kaemo has quit IRC (Ping timeout: 244 seconds) 2011-11-25T16:41:00 don't bite into anything you see on irc, there might be a hook in it 2011-11-25T16:41:15 http://fluxid.pl/misc/2011-11-25-223607_682x695_scrot.png look at that group of ants... scared of ONE enemy ant. awesome 2011-11-25T16:41:22 *** smiley1983 has joined #aichallenge 2011-11-25T16:42:20 in some degree it is a success 2011-11-25T16:42:35 just lots of tweaks are needed 2011-11-25T16:43:01 Is BenJackson still here? 2011-11-25T16:43:12 I am 2011-11-25T16:43:37 What do you think of the new version of my bot? http://ants.fluxid.pl/player/Parasprites 2011-11-25T16:43:55 Well I guess it dies too fast on tcp to show much difference 2011-11-25T16:45:29 you seem to explore and gather fine 2011-11-25T16:45:29 Antimony: it's hard for me to say what do i think of it, but let me test it with my current bot version 2011-11-25T16:45:36 you've got to jump the hurdle of combat 2011-11-25T16:45:57 combat is so messy. there should be a pacificst mode :) 2011-11-25T16:46:04 *** peyton has quit IRC (Quit: peyton) 2011-11-25T16:47:11 my bot is kind of pacifist 2011-11-25T16:47:15 in a way, this tries to be a multiplayer game, but it isn't. it is just a blown up 2 player 2011-11-25T16:47:33 The problem is that I implemented some rudimentary combat but it doesn't work very well 2011-11-25T16:47:45 It assumes all enemies are stationary, so it's helpless against mobile enemies 2011-11-25T16:48:51 *** dom7b5_ has joined #aichallenge 2011-11-25T16:51:04 you are battling water then 2011-11-25T16:51:08 *** smiley1983 has quit IRC (Ping timeout: 252 seconds) 2011-11-25T16:51:21 Hmm, I really need to figure out a way to prevent diagonal ants from getting stuck 2011-11-25T16:51:36 I should have razed purples hill here. http://ants.fluxid.pl/replay.15601 2011-11-25T16:52:34 *** dom7b5 has quit IRC (Ping timeout: 245 seconds) 2011-11-25T16:55:47 my bot is terrible, and i have no idea why it ranks so well 2011-11-25T16:56:27 *** Anilm3_ has joined #aichallenge 2011-11-25T16:56:43 *** Anilm3 has quit IRC (Quit: Lost terminal) 2011-11-25T16:56:55 low sigma? 2011-11-25T16:57:09 *** Anilm3 has joined #aichallenge 2011-11-25T17:01:26 *** mcstar has quit IRC (Quit: WeeChat 0.3.6) 2011-11-25T17:01:46 *** smiley1983 has joined #aichallenge 2011-11-25T17:02:33 *** Anilm3 has quit IRC (Ping timeout: 252 seconds) 2011-11-25T17:05:59 ahh the fun part 2011-11-25T17:06:07 where I do git add -p and split my huge mass of changes into meaninful blocks 2011-11-25T17:06:11 if I can figure out what they do :) 2011-11-25T17:09:28 *** valentin has joined #aichallenge 2011-11-25T17:10:57 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-11-25T17:16:11 *** mleise has joined #aichallenge 2011-11-25T17:16:38 *** epicmonkey has quit IRC (Ping timeout: 244 seconds) 2011-11-25T17:23:13 @later tell amstan great, if he can get me the new amount I'll put it to use :) 2011-11-25T17:23:13 janzert: As you wish. 2011-11-25T17:24:14 *** kapoc has quit IRC (Ping timeout: 245 seconds) 2011-11-25T17:26:02 *** Anilm3 has joined #aichallenge 2011-11-25T17:34:58 *** HaraKiri has quit IRC () 2011-11-25T17:39:01 Fluxid: pleae don't delete records until sunday xD 2011-11-25T17:39:13 lol, ok 2011-11-25T17:39:19 :P 2011-11-25T17:39:36 are you planning on adding new features like tcpants? 2011-11-25T17:42:54 Fluxid: beep 2011-11-25T17:43:47 wow, I never realized how terrible State::updateVisionInformation was 2011-11-25T17:44:25 Anilm3: nope 2011-11-25T17:44:31 yeah, I profiled it today and it took almost 40% of the time 2011-11-25T17:45:03 have you managed to optimize it? 2011-11-25T17:45:07 I'm about to! 2011-11-25T17:45:58 *** Antimony has joined #aichallenge 2011-11-25T17:46:25 I wonder how many peopel worry about doing a BFS for every ant without realizing it's doing one for every ant already 2011-11-25T17:47:00 haha 2011-11-25T17:47:08 what... 2011-11-25T17:47:16 I didn't manage to spead it up 2011-11-25T17:47:25 speed* 2011-11-25T17:47:26 what updateVisionInformation does? 2011-11-25T17:47:51 it puts every tile on viewradius to visible 2011-11-25T17:47:52 because i don't use any starter package 2011-11-25T17:47:52 it does a BFS from every ant and computes the eulidean distance at every visit 2011-11-25T17:48:05 oh boy 2011-11-25T17:48:09 as a means of finding the vision of all ants 2011-11-25T17:48:13 it's essentially a BFS 2011-11-25T17:48:18 std::vector > visited(rows, std::vector(cols, 0)); 2011-11-25T17:48:22 that's inside the "for all ants" loop 2011-11-25T17:48:48 * BenJackson head asplode 2011-11-25T17:48:57 i... don't understand this line, but whatever 2011-11-25T17:49:10 that's making a 2-dimensional vector of bool 2011-11-25T17:49:20 first it allocates one row worth of "false" 2011-11-25T17:49:36 then for every row it copies that row of false to a new one, duplicating it rows times 2011-11-25T17:49:57 such that visited[r][c] is a bool (initially false) 2011-11-25T17:50:00 it uses it for the bfs 2011-11-25T17:50:32 do you really use 0 in place of false? 2011-11-25T17:50:43 this code does 2011-11-25T17:50:51 terrible terrible terrible 2011-11-25T17:50:53 the C++ starter bot is not the greatest C++ 2011-11-25T17:51:04 true 2011-11-25T17:51:07 for(int a=0; a<(int) myAnts.size(); a++) 2011-11-25T17:51:18 the C++ starter bot is terrible 2011-11-25T17:51:19 forget iterators, it doesn't even use unsigned ints ;-) 2011-11-25T17:51:42 my bot has consumed the starter bot like a parasite 2011-11-25T17:51:45 by the way, is there any way to submit my fixed version of the timer if I don't have Git? 2011-11-25T17:51:46 you can see vestiges of the original 2011-11-25T17:51:57 learn git, it's worth it 2011-11-25T17:51:59 *** AntDroid has joined #aichallenge 2011-11-25T17:52:19 Antimony: are you talking about the C++ timer? 2011-11-25T17:52:21 using 0 instead of false leads to great bugs like 2011-11-25T17:52:32 bool* some_flag 2011-11-25T17:52:41 how do people predict enemy positions in next turn? 2011-11-25T17:52:43 some_flag = 0; 2011-11-25T17:52:50 (gg) 2011-11-25T17:53:04 any idea how to disable symmetry checking for maps? i can't use a non symmetric map for testing. 2011-11-25T17:53:05 the spurious semicolons in the C++ start drive me nuts too 2011-11-25T17:53:12 AntDroid: --food none 2011-11-25T17:53:32 BenJackson: that disables symmetry? 2011-11-25T17:53:36 that should just be implied by --scenario because everone is confused by it 2011-11-25T17:53:44 Yes, the symmetry requirement is for the food placement system 2011-11-25T17:53:51 oh 2011-11-25T17:54:33 *** mleise has quit IRC (Quit: Leaving.) 2011-11-25T17:55:16 no answer 2011-11-25T17:55:22 ohh, BenJackson, thanks for pointing that out 2011-11-25T17:55:26 oh well, i'll look on forums 2011-11-25T17:55:49 Fluxid: ? 2011-11-25T17:55:57 oh 2011-11-25T17:56:15 *** Antimony has quit IRC (Ping timeout: 248 seconds) 2011-11-25T17:56:17 well, I *don't* predict them (except in one tiny case) 2011-11-25T17:56:27 I figure they could do anything 2011-11-25T17:56:50 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-25T17:57:29 BenJackson: thanks 2011-11-25T17:58:03 who the hell uses a pointer to bool anyway? 2011-11-25T17:58:17 * BenJackson 2011-11-25T18:02:59 *** choas has quit IRC (Ping timeout: 245 seconds) 2011-11-25T18:03:05 Fluxid: you could assume they will do the best movement possible 2011-11-25T18:03:27 and do some sort of minimax 2011-11-25T18:03:35 1 level 2011-11-25T18:03:36 well I just grepped my folder of personal projects (about 50 KSLOC of cpp) and not a single bool pointer declared 2011-11-25T18:04:05 I always thought I'd do minmax for combat 2011-11-25T18:04:16 then I realized the game is more like RPS than tic-tac-toe 2011-11-25T18:04:26 *** SlashV has joined #aichallenge 2011-11-25T18:04:51 BenJackson: just to decide what to do next it can help 2011-11-25T18:05:13 I'm not using it btw so I can't really say 2011-11-25T18:09:40 *** Jak_o_Shadows has joined #aichallenge 2011-11-25T18:09:54 BenJackson: the fact that it uses std::vector for that is pretty bad 2011-11-25T18:10:21 because it's a bunch of bitfields, not bools 2011-11-25T18:10:42 true, can't take a pointer to those at all 2011-11-25T18:10:51 they really shouldn't have done that 2011-11-25T18:11:09 well you can just make it std::vector 2011-11-25T18:11:10 silly though 2011-11-25T18:11:24 it's much slower to access a bool vector 2011-11-25T18:11:47 updateVisionInformation or whatever it's called was fast enough so I didn't bother changing it much 2011-11-25T18:12:04 In that function, that's the least of your concerns actually 2011-11-25T18:12:20 thestinger: I profiled it and it took almost 40% of the total computation time 2011-11-25T18:12:28 you want slow, try using the visualizer to verify your vision replacement works! 2011-11-25T18:13:17 *** hacklash has left #aichallenge 2011-11-25T18:13:26 Anilm3: 40% of the computation time of your bot? 2011-11-25T18:13:32 yes 2011-11-25T18:14:40 how long exactly? it takes a fraction on a ms for me now 2011-11-25T18:14:45 it's going to scale with ant count 2011-11-25T18:15:01 and cost more on larger maps 2011-11-25T18:15:17 thestinger: I had up to 200 ants on that game 2011-11-25T18:15:30 *** Rav3nsW0rd has quit IRC (Read error: Connection reset by peer) 2011-11-25T18:16:00 *** Rav3nsW0rd has joined #aichallenge 2011-11-25T18:16:34 thestinger: I can't tell, I don't remember 2011-11-25T18:16:50 *** NoxiaZ^ has joined #aichallenge 2011-11-25T18:17:05 a scenario with 200 ants would be a good test 2011-11-25T18:17:13 ok, I found some code with bool pointers in it - I grepped /usr/src, about 800 KSLOC of c, and found 56 occurrences 2011-11-25T18:17:36 Minthos: what is KS? 2011-11-25T18:17:47 kilo source 2011-11-25T18:18:30 thanks 2011-11-25T18:19:01 do the other starter bots use a better matrix method or something? 2011-11-25T18:19:31 ah, so the python bot precalculates it then sets it around each ant 2011-11-25T18:21:03 similar to the python game engine 2011-11-25T18:21:27 even if you just iterated over the rectangle and tested sqrt every time it'd probably beat the stock C++ impl 2011-11-25T18:26:49 the java starter bot does it the same way as python 2011-11-25T18:30:27 *** treeform has quit IRC (Remote host closed the connection) 2011-11-25T18:31:13 *** treeform_ has joined #aichallenge 2011-11-25T18:31:49 I don't know what everyone else does, but a simple dumb loop written in D for all my ants takes about 1 ms with 50 ants locally 2011-11-25T18:32:14 if the loop is written in python, I could see how that might explode to a signifcant processing load 2011-11-25T18:33:53 there's actually a more subtle bug in the C++ implementation 2011-11-25T18:34:18 now all you C++ starter kit users can obsess over that ;-) 2011-11-25T18:35:13 and I can sit here and say "sounds like typical C++" and look forward to lots of C++ in my next job :) 2011-11-25T18:35:20 *** smiley1983 has quit IRC (Ping timeout: 258 seconds) 2011-11-25T18:36:30 luckily for my mental health I'm already pretty fucked up so I doubt C++ can make it worse :p 2011-11-25T18:36:56 *** cyphase has quit IRC (Read error: Operation timed out) 2011-11-25T18:37:05 my day job is embedded C++ 2011-11-25T18:39:10 BenJackson: what bug are you talking about? 2011-11-25T18:39:35 no no, you need to obsess over it rather than working on your bot ;-) 2011-11-25T18:39:48 it doesn't caclculate vision on water correctly 2011-11-25T18:40:14 on the updatevisioninformation? 2011-11-25T18:40:15 huh, how does water change vision? 2011-11-25T18:40:54 *** mtmcannon has joined #aichallenge 2011-11-25T18:41:51 it never resets water squares because nothing interesting happens on them 2011-11-25T18:41:55 but one of the things is vision 2011-11-25T18:41:58 so they are isVisible forever 2011-11-25T18:42:23 you'd have to be doing something that would care to notice 2011-11-25T18:42:35 but it was obvious when I used the visualizer to check my vision algorithm against the replay tool 2011-11-25T18:43:06 oh, I've fixed that 2011-11-25T18:43:21 me too 2011-11-25T18:43:30 I think update vision rewrite did speed up my bot considerably 2011-11-25T18:43:36 I noticed because it wasn't resetting enemy attack range counts to 0 2011-11-25T18:43:40 although to make that claim I should make a test scenario 2011-11-25T18:43:50 so my bot was thinking there were imaginary battles everywhere 2011-11-25T18:43:51 oh, doh. i was wondering wth was going on with water tiles being visible all the time 2011-11-25T18:43:55 took a while to debug that 2011-11-25T18:44:06 arr I knew I should have left you in the dark! 2011-11-25T18:44:10 now a1k0n will take the next step 2011-11-25T18:44:14 I didn't realize it caused a vision bug too :P 2011-11-25T18:44:24 ..i was just trying to draw a cool colored map 2011-11-25T18:44:38 ‘__gnu_cxx::__enable_if::__type {aka double} 2011-11-25T18:44:40 I kind of want to download all the games on a map and make a heatmap for ants and battles and stuff 2011-11-25T18:44:40 what 2011-11-25T18:44:57 enable_if is a template 2011-11-25T18:45:32 enable_if::__type is type if bool is true 2011-11-25T18:45:32 *** mj41 has quit IRC (Ping timeout: 252 seconds) 2011-11-25T18:45:49 http://www.boost.org/doc/libs/1_48_0/libs/utility/enable_if.html 2011-11-25T18:46:02 it's probably in std::tr1 too 2011-11-25T18:46:07 oh, yeah 2011-11-25T18:46:11 and C++11 ofc 2011-11-25T18:47:31 I'm glad my next job will pay enough to finance generous amouns of alcohol, cause I'll need it 2011-11-25T18:48:00 convoluted stuff like enable_if is mostly useful to people writing more libraries 2011-11-25T18:49:09 ok shutting up and going to bed now, just in case any future employers are reading this and taking me too seriously :) 2011-11-25T18:49:13 nn :) 2011-11-25T18:49:53 *** cyphase has joined #aichallenge 2011-11-25T18:50:14 bye 2011-11-25T18:50:22 dream with ants 2011-11-25T18:50:57 BenJackson: alright, I do it with vision offsets like the py/java bots now 2011-11-25T18:51:03 time to benchmark that 2011-11-25T18:51:09 I don't really care enough to really optimize it 2011-11-25T18:51:27 *** grwip has quit IRC (Quit: Leaving) 2011-11-25T18:52:35 *** Fandekasp has quit IRC (Ping timeout: 258 seconds) 2011-11-25T18:53:20 offsets: 0.253916 2011-11-25T18:53:22 bfs: 2.83313 2011-11-25T18:53:24 with 100 ants 2011-11-25T18:53:52 and I already switched the BFS to use euclidean_distance2 and viewdistance2, and std::vector 2011-11-25T18:53:59 which sped it up 2x 2011-11-25T18:54:06 so yeah, I can see that it was a problem now 2011-11-25T18:54:08 xD 2011-11-25T18:54:55 now I just need to code a test to make sure they really do the same thing and this case is closed 2011-11-25T18:55:05 *** ztfw has quit IRC (Remote host closed the connection) 2011-11-25T18:56:46 *** Cyndre has quit IRC (Read error: Connection reset by peer) 2011-11-25T18:56:59 wow the visions offsets seem nice 2011-11-25T18:57:05 *** Cyndre has joined #aichallenge 2011-11-25T18:57:06 quite better than the BFS 2011-11-25T18:57:10 *** smiley1983 has joined #aichallenge 2011-11-25T18:57:18 yeah, and it scales much better 2011-11-25T18:58:07 you can basically just copy the python code, I'm sure it can be optimized a lot more though 2011-11-25T18:58:58 *** cowbandit has joined #aichallenge 2011-11-25T18:59:01 I'll leave that task for tomorrow, I think I'm going to bed now 2011-11-25T18:59:07 hello 2011-11-25T18:59:11 thanks for the tips 2011-11-25T19:01:02 *** Anilm3 has quit IRC (Quit: Lost terminal) 2011-11-25T19:01:18 *** dr- has quit IRC (Ping timeout: 240 seconds) 2011-11-25T19:02:06 *** cbad has quit IRC (Ping timeout: 240 seconds) 2011-11-25T19:02:30 *** welterde has quit IRC (Ping timeout: 240 seconds) 2011-11-25T19:02:33 *** cbad has joined #aichallenge 2011-11-25T19:02:33 *** cbad has joined #aichallenge 2011-11-25T19:03:15 *** dr- has joined #aichallenge 2011-11-25T19:06:54 *** AlliedEnvy has quit IRC (Ping timeout: 240 seconds) 2011-11-25T19:09:32 *** AlliedEnvy has joined #aichallenge 2011-11-25T19:10:06 *** jix has quit IRC (Ping timeout: 240 seconds) 2011-11-25T19:10:54 *** kire has quit IRC (Ping timeout: 240 seconds) 2011-11-25T19:11:05 *** jix has joined #aichallenge 2011-11-25T19:11:37 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-25T19:12:03 *** foRei has quit IRC (Read error: Connection reset by peer) 2011-11-25T19:12:51 *** kire has joined #aichallenge 2011-11-25T19:14:09 Does anyone know how to bunch your ants very close together? Like when my ants are traveling to a destination in a line, they seem to always be one space apart from another ant in front. 2011-11-25T19:15:56 BenJackson: oh, I think that vision bug would have been fixed when I switched to last_seen instead of visible 2011-11-25T19:16:00 which is why I never noticed it 2011-11-25T19:16:03 *** Antimony has joined #aichallenge 2011-11-25T19:16:22 *** ikaros has joined #aichallenge 2011-11-25T19:16:35 ah yeah 2011-11-25T19:16:47 what are vision offsets? 2011-11-25T19:16:58 cowbandit: do you have code to avoid having one ant step where another ant already is? 2011-11-25T19:17:06 mtmcannon: https://github.com/aichallenge/aichallenge/blob/epsilon/ants/dist/starter_bots/python/ants.py#L224 2011-11-25T19:17:08 yeah 2011-11-25T19:17:12 it was from the tutorial 2011-11-25T19:17:13 oh 2011-11-25T19:17:14 look how python calculates which tiles are visible 2011-11-25T19:17:17 i see what you're saying 2011-11-25T19:17:17 cowbandit: ok now think what happens when you iterate that 2011-11-25T19:17:23 it makes a list of vision offsets, then paints those around each ant basically 2011-11-25T19:17:28 Ok, thanks! 2011-11-25T19:17:29 cowbandit: oh I see. thanks 2011-11-25T19:18:17 much faster than the BFS the C++ bot does 2011-11-25T19:18:26 and I was lazy so I did the same thing for attackradius 2011-11-25T19:18:27 er, thanks to thestinger that is 2011-11-25T19:18:33 lol 2011-11-25T19:18:53 I just made the function generic and used it all over, so my whole bot got faster by overcoming that laziness 2011-11-25T19:19:15 I might have finally found the right knob to install in my bot 2011-11-25T19:19:22 a whole new class of knob I had not exploited 2011-11-25T19:19:25 muahaha 2011-11-25T19:19:32 knob? 2011-11-25T19:20:14 tunable parameter 2011-11-25T19:20:23 *** TheLinker has joined #aichallenge 2011-11-25T19:20:46 *** welterde has joined #aichallenge 2011-11-25T19:20:48 ah 2011-11-25T19:21:06 I'm trying to figure out how to do symmetry 2011-11-25T19:21:10 I've given up on combat for now 2011-11-25T19:21:23 I'm just going to optimize everything now, too lazy to think 2011-11-25T19:21:38 *** treeform_ has quit IRC (Remote host closed the connection) 2011-11-25T19:21:39 thestinger: I do the same thing 2011-11-25T19:21:41 maybe I'll have some great idea if I just stop trying to come up with one :) 2011-11-25T19:21:45 no new ideas? clean up and do checkins 2011-11-25T19:21:47 optimize old code 2011-11-25T19:21:49 yeah 2011-11-25T19:21:52 I like doing that 2011-11-25T19:21:59 when I code new stuff it's just like code vomit 2011-11-25T19:22:02 add some new metadata I detect about the map 2011-11-25T19:22:04 needs to be polished up a lot later 2011-11-25T19:22:20 there's an awesome trick to symmetry 2011-11-25T19:22:21 dead ants are awesome :) 2011-11-25T19:22:29 which I look forward to making people forehead-slap about after the contest 2011-11-25T19:22:44 what trick> 2011-11-25T19:22:47 lol 2011-11-25T19:22:53 ...after... the contest! 2011-11-25T19:22:54 My bot completely ignores dead ants 2011-11-25T19:22:56 you'll have to wait til he gets 1st place 2011-11-25T19:22:56 lol 2011-11-25T19:22:58 *** JorgeB has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-11-25T19:23:04 yeah, I deleted the dead ant code 2011-11-25T19:23:13 but I haven't done much with # enemy detection or the like 2011-11-25T19:23:14 *** JorgeB has joined #aichallenge 2011-11-25T19:23:22 BenJackson: how do you handle hill razing? do you always try to wait until the hill drains of enemies? 2011-11-25T19:23:34 I should try that 2011-11-25T19:23:46 I'm not sure 2011-11-25T19:23:48 I didn't used to check 2011-11-25T19:23:54 ah 2011-11-25T19:23:57 I use dead ants for that 2011-11-25T19:23:59 I don't have any hill razing code in my bot. I just pretend that hills are food 2011-11-25T19:24:02 I think now there's some heuristic about seeing it empty before I charge in 2011-11-25T19:24:06 to figure out if they're staying on the hill or moving off 2011-11-25T19:24:10 *** Conorach has quit IRC (Ping timeout: 260 seconds) 2011-11-25T19:24:32 worst case scenario is that someone counters my exact code and it's like I just didn't handle that little thing at all 2011-11-25T19:24:36 I tried to make my ants more defensive 2011-11-25T19:24:44 then I ran a test game and saw my lone defender run away 2011-11-25T19:24:50 lol 2011-11-25T19:24:53 and then I realized it was right -- all it could have done was suicide, which is still bad 2011-11-25T19:25:00 well 2011-11-25T19:25:02 maybe 2011-11-25T19:25:07 in that exact case 2011-11-25T19:25:11 maybe the enemy would have taken an extra turn to attack 2011-11-25T19:25:14 based on their combat code 2011-11-25T19:25:17 and you might get an extra ant 2011-11-25T19:25:30 in this case it was my lone ant and a wall of 3 bearing down, but yes 2011-11-25T19:25:46 time for more lambdas :) 2011-11-25T19:25:57 you're using boost::lambda? 2011-11-25T19:26:01 what do I need to checkout to submit my starter bot changes? 2011-11-25T19:26:07 BenJackson: I'm using C++11 2011-11-25T19:26:14 lambdas and std::function<> are standard 2011-11-25T19:26:18 they're awesome 2011-11-25T19:26:19 antimony: I think it depends on the starter 2011-11-25T19:26:25 C++ 2011-11-25T19:26:31 *** xathis has joined #aichallenge 2011-11-25T19:26:38 some are in the contest repo, but that repo has some submodules (sub git repositories) for some of the languages 2011-11-25T19:27:28 they actually let you do functional stuff in C++ 2011-11-25T19:28:12 https://github.com/aichallenge/aichallenge/wiki/_access 2011-11-25T19:28:32 derp 2011-11-25T19:28:33 https://github.com/aichallenge/aichallenge 2011-11-25T19:29:12 under ants/dist/starter_bots/cpp 2011-11-25T19:29:59 ok, I've got git isntalled on an ubutntu vm 2011-11-25T19:30:04 now I have to learn how to use it 2011-11-25T19:30:49 *misuse 2011-11-25T19:31:14 http://ftp.newartisans.com/pub/git.from.bottom.up.pdf 2011-11-25T19:31:16 *** Extrarius has quit IRC (Read error: Connection reset by peer) 2011-11-25T19:31:22 that starts from the internals, which seems strange 2011-11-25T19:31:32 but really I found git much easier to understand when I knew how it worked 2011-11-25T19:31:34 *** Extrarius has joined #aichallenge 2011-11-25T19:31:48 the official docs are good 2011-11-25T19:31:52 the internals are very simple really 2011-11-25T19:31:59 whereas there are a zillion commands 2011-11-25T19:32:00 http://book.git-scm.com/3_getting_a_git_repository.html 2011-11-25T19:32:03 starting there basically 2011-11-25T19:32:15 once you know the internals then it's much easier to reason about what the commands do 2011-11-25T19:32:26 well "community book" (official enough for me) 2011-11-25T19:32:28 or realize a command to do "X" must exist because it's obviously useful 2011-11-25T19:33:18 it doesn't like me 2011-11-25T19:33:25 Permission denied (publickey) remote host hang up 2011-11-25T19:33:43 that's a url, not a git repo 2011-11-25T19:33:52 it shows you the url on that page, sorry 2011-11-25T19:34:31 ops quick question: I want to iterate through each of my ants to perform moves, but what if ant1's move depended on ant100's projected move? Would I have to iterate through my ants multiple times to check for the projected moves? Or is there a more efficient way to do this 2011-11-25T19:35:24 I just wrote a constraint solver 2011-11-25T19:35:37 it assigns moves to all ants at once 2011-11-25T19:35:49 same here pretty much 2011-11-25T19:35:51 cowbandit, you can implement a push method on your ants 2011-11-25T19:35:53 oh ok 2011-11-25T19:36:06 although early on I just used an algorithm which happened to yield "compatible" resutls for all ants 2011-11-25T19:36:07 so one tells the other that it wanna take its place 2011-11-25T19:36:24 oh, I have to tell GitHub my public key 2011-11-25T19:36:25 and the other tries to move first 2011-11-25T19:36:32 you can do it recursively 2011-11-25T19:36:36 oh i see what you mean 2011-11-25T19:36:43 oi thanks 2011-11-25T19:36:48 just avoid cycles :) 2011-11-25T19:38:01 * BenJackson fires off 20 iterations on fluxid and goes to hit the treadmill 2011-11-25T19:38:18 only 20? 2011-11-25T19:38:20 Why not do infinite? 2011-11-25T19:38:34 might as well be reasonable about how many replays I'll watch 2011-11-25T19:38:41 I run an old version full time for other people to practice on 2011-11-25T19:39:07 lol speaking of which, do you know if running a ton of iterations will hurt your computer? 2011-11-25T19:39:08 holy crap I must be playing a slow bot 2011-11-25T19:39:13 because sometimes when i'm on the tcp server 2011-11-25T19:39:18 with 1 ant turns are taking 1s+ 2011-11-25T19:39:27 and i have like 200+ ants, my computer beeps every time there's a turn 2011-11-25T19:39:53 shouldn't hurt your computer 2011-11-25T19:40:00 it's like a clicking sound 2011-11-25T19:40:07 my old desktop is really screwed up and I can _hear_ CPU load 2011-11-25T19:40:14 my bot is running infinitely 2011-11-25T19:40:16 it makes a high pitched noise 2011-11-25T19:40:18 so it can keep you company 2011-11-25T19:40:19 the thermal diodes in modern CPUs are so sensitive that the CPU fan can ramp up almost instantly under load 2011-11-25T19:40:21 and bump up your score 2011-11-25T19:40:30 it's probably something about to explode in the PSU :\ 2011-11-25T19:40:31 lol ok, so i'm good then? 2011-11-25T19:41:12 Antimony: The best part about resubmitting to aichallenge was krushing bots until I got close to my final place 2011-11-25T19:41:44 it's a nice change of pace from close matches on fluxid 2011-11-25T19:42:17 Woohoo, I sucessfully sshed to github 2011-11-25T19:42:21 now to set up git 2011-11-25T19:42:37 *** Redgis has quit IRC (Ping timeout: 240 seconds) 2011-11-25T19:42:54 *** xathis has quit IRC (Ping timeout: 240 seconds) 2011-11-25T19:43:37 my C++ bots have lasted 20, 40 and 87 games on aichallenge so far 2011-11-25T19:43:54 given my ever-descreasing submission rate this may end up being my last one :) 2011-11-25T19:44:25 I really want to see this fluxid replay now because it's nearly turn 250 and my bot has been down to 1 ant multiple times and still must have hills 2011-11-25T19:45:42 oh no, vs A_experimental 2011-11-25T19:45:44 that can't be good 2011-11-25T19:45:58 I don't think it's as good as A 2011-11-25T19:46:55 ah, I misread, I just lost my hill early 2011-11-25T19:47:02 these 10 player maps are stupid 2011-11-25T19:47:23 I like those new maps 2011-11-25T19:47:46 http://ants.fluxid.pl/replay.15144 2011-11-25T19:47:56 *** Ikol has joined #aichallenge 2011-11-25T19:48:02 basically another player suicided into me early 2011-11-25T19:48:31 the random_walk ones suck imo 2011-11-25T19:48:39 the big ones with only a few players are fine 2011-11-25T19:48:47 but most of them are crowded with the hills way too close 2011-11-25T19:49:16 I like how blue broke you out, then kicked your ass 2011-11-25T19:49:30 xD 2011-11-25T19:49:40 that's a bug in my defense code 2011-11-25T19:49:40 ug, more random timeouts 2011-11-25T19:49:45 the mapgens really need code to detect stalemate chokepoints like that 2011-11-25T19:49:50 it's so depressing getting 0 every single game 2011-11-25T19:49:52 they stop foraging because my limit on defenders doesn't work 2011-11-25T19:49:57 mabye all fixed-width halls are, though, given the combat rules 2011-11-25T19:50:08 no 2011-11-25T19:50:12 I'm mostly fine with chokepoints now 2011-11-25T19:50:14 according to my calculations 2011-11-25T19:50:23 any chokepoint of width at least 3 can't be a stalemate 2011-11-25T19:50:29 I'm just going to have to adjust how the 'score' I use for combat is calculated 2011-11-25T19:50:38 it's impossible to defend it without casualitties 2011-11-25T19:50:59 *** AntDroid has quit IRC (Ping timeout: 265 seconds) 2011-11-25T19:51:00 Antimony: there are width 1 chokepoints in some of the random walks :) 2011-11-25T19:51:01 *** Rav3nsW0rd has quit IRC (Ping timeout: 240 seconds) 2011-11-25T19:51:05 I know 2011-11-25T19:51:10 those are pretty open maps so it doesn't really matter 2011-11-25T19:51:13 I think I was the first person to bring attention to it 2011-11-25T19:51:25 there's one map that can be walled off almost compeltely 2011-11-25T19:51:36 *** bmh has joined #aichallenge 2011-11-25T19:56:01 a1k0n: re:your fight code -- do your ants communicate with each other, or examine their neighbors? 2011-11-25T19:57:59 *** goffrie has quit IRC (Quit: Konversation terminated!) 2011-11-25T19:58:06 *** McLeopold_ has joined #aichallenge 2011-11-25T19:58:22 *** goffrie has joined #aichallenge 2011-11-25T19:59:14 *** AntDroid has joined #aichallenge 2011-11-25T19:59:51 *** JorgeB has quit IRC (Ping timeout: 276 seconds) 2011-11-25T20:03:00 janzert: ping 2011-11-25T20:03:49 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-25T20:05:52 McLeopold_: pong! 2011-11-25T20:05:58 hi bmh 2011-11-25T20:10:18 *** newant has joined #aichallenge 2011-11-25T20:11:06 *** kire has quit IRC (Remote host closed the connection) 2011-11-25T20:18:28 *** Harpyon_ has quit IRC (Quit: Computer has gone to sleep.) 2011-11-25T20:19:32 *** smiley1983 has quit IRC (Ping timeout: 244 seconds) 2011-11-25T20:20:19 *** besh has quit IRC (Quit: Page closed) 2011-11-25T20:20:47 *** smiley1983 has joined #aichallenge 2011-11-25T20:23:18 bmh: they can see each others moves 2011-11-25T20:23:40 *** Harpyon has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-11-25T20:24:04 guys total newbie first timer needs help with c starter kit implementations 2011-11-25T20:24:26 a1k0n: I was thinking of trying a polya's urn appraoch where each ant pushes its move into the urns of nearby ants 2011-11-25T20:24:42 though I'm not sure if that's strictly less efficient than other approaches 2011-11-25T20:24:50 *** NoxiaZ^ has quit IRC () 2011-11-25T20:25:09 i don't follow 2011-11-25T20:25:14 hey anybody doin it in c here? 2011-11-25T20:25:41 newant: c++, but what's your question? 2011-11-25T20:26:26 no i need a c kind of guide for getting started coz i dont understand python could u guide me a little 2011-11-25T20:26:27 my nearby_enemies and nearby_friends functions are so much faster now :) 2011-11-25T20:27:29 bmh: i guess you're assuming nearby ants will want to go the same direction, envisioning a sort of "wall of ants"? 2011-11-25T20:27:36 a1k0n: start with a uniform distribution for each ant, run minimax/maximin for each ant and push the resulting move to the nearby ants (if you decide to move West, nearby ants become more likely to move West) 2011-11-25T20:28:52 that might be worth trying, hm 2011-11-25T20:29:17 I don't think it will converge. 2011-11-25T20:30:31 alk0n can you b kind enough to direct me to a guide in c 2011-11-25T20:32:31 still working on my pathfinder. It just topped 100 lines, which is about enough to write a window manager. 2011-11-25T20:32:33 for.. the tutorial of this contest? 2011-11-25T20:32:48 yes 2011-11-25T20:32:48 hooray! 2011-11-25T20:33:05 Of all the games on tcp, my bot finally didn't get last place! 2011-11-25T20:33:30 i don't think such a thing exists. 2011-11-25T20:33:44 you'll have to read the java one and try to figure out how the C starter pack does it 2011-11-25T20:34:02 by the way, I just submitted by improved c++ timer 2011-11-25T20:34:10 hopefully it will be included in the starter pack 2011-11-25T20:34:34 ok anything for c++ maybi can draw parallels 2011-11-25T20:37:31 anybody please i really want to do this but for lack of knowledge or python and experience i am finding it tough 2011-11-25T20:37:56 if anybody could help i would really b grateful 2011-11-25T20:38:26 *** sigh has joined #aichallenge 2011-11-25T20:38:50 newant: have you thought of taking the opportunity to learn python? 2011-11-25T20:39:13 it only takes a couple minutes to get started with python 2011-11-25T20:39:14 it's really easy 2011-11-25T20:40:19 but i know c so maybe if i could just get a push then i could do well.... i can think in c more or less .... learning python would take time and i have certain constraints 2011-11-25T20:40:55 thestinger: http://ants.fluxid.pl/replay.15800 2011-11-25T20:41:03 there's your answer: I apparently charge in like crazy :) 2011-11-25T20:42:58 its a massive pink blob! 2011-11-25T20:43:24 *** Antimony has quit IRC (Ping timeout: 245 seconds) 2011-11-25T20:44:48 BenJackson: I enjoy how your ant swarm twitches every few turns 2011-11-25T20:45:00 yay, programming time 2011-11-25T20:45:10 almost 12 hours earlier than yesterdays starting time 2011-11-25T20:45:53 bmh: it was kind of seething there 2011-11-25T20:46:34 BenJackson: whoa. He had 100 ants stored up! I didn't notice 2011-11-25T20:46:51 neither did my bot! 2011-11-25T20:49:44 hmmm, it's a bit messy to make sure you don't walk an ant onto an acjacent food square 2011-11-25T20:50:34 I had to fix a few bugs with that 2011-11-25T20:50:40 *** Ikol is now known as Rav3nSw0rd 2011-11-25T20:50:41 since you don't mind after the first "move" whether it walks over food but you do care if a second adjcant ant then tries to pathfind over a square where an ant has to stay still to collect it 2011-11-25T20:50:50 yeah, i just realised last night my logic isn't sound 2011-11-25T20:51:02 you also have to factor it into combat :) 2011-11-25T20:51:08 yeah that's easy 2011-11-25T20:51:22 i'm just going to say collected food for a bot is -- to deaths :P 2011-11-25T20:51:36 then my battle code will do the rest 2011-11-25T20:51:46 are you doing it in c? 2011-11-25T20:52:02 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-25T20:52:16 *** Surya has joined #aichallenge 2011-11-25T20:52:48 *** JorgeB has joined #aichallenge 2011-11-25T20:52:54 BenJackson: so your ants push each other or something now? 2011-11-25T20:53:04 push? 2011-11-25T20:54:18 I think I'm just tired xD 2011-11-25T20:54:37 no, i do a sort of separate bfs from all ants simultaneously with a single search queue, then when i move an ant i possibly re use it and add in search information for it (and its availability) at it's path destination location and if the ant was previously unmoved for each of it's non-moved into but searched adjacent squares i add moves for any adjcaent unmoved ants into those locations 2011-11-25T20:54:49 that last bit is how i avoid collisions 2011-11-25T20:55:00 (answering for myself :P) 2011-11-25T20:55:20 and i do that to collect a set of targets 2011-11-25T20:55:46 remove pretty much all the 's 2011-11-25T20:56:13 BenJackson: is yours much like that? 2011-11-25T20:56:27 that's my generalised collect targets with bfs 2011-11-25T20:56:51 i may write one for a* too but when the number of targets is large i don't trust my heuristic function to not be too slow 2011-11-25T20:57:27 similar 2011-11-25T20:58:05 oh crud 2011-11-25T20:58:08 thestinger: oh you don't meant adjacent 2011-11-25T20:58:14 antimatroid: you aren't just using a manhattan heuristic? 2011-11-25T20:58:18 thestinger: you could think of it like that 2011-11-25T20:58:27 bmh: h(x) = min manhatten distance from x to a target 2011-11-25T20:58:44 Oh, by slow you mean "will expand too many nodes" 2011-11-25T20:58:46 if the number of targets is large and you're calling that for every location you search i'm not sure if that's worse than just using bfs? 2011-11-25T20:58:58 thestinger: they don't literally push (like a gravity simulation) but they maximize a condition that causes them to spread out 2011-11-25T20:59:14 which is basically what I've been trying to achive since I submitted V7 to aichallenge 2011-11-25T20:59:21 no, i mean for each search location you have to calculate the manhatten distance to each of the targets 2011-11-25T20:59:30 but most of my attempts got me rapidly snuffed out because my code kind of relied on travelling en masse 2011-11-25T20:59:32 if the number of targets is large surely that would slow things down quite a bit? 2011-11-25T21:00:00 antimatroid: I've been slaving away at my hierarhical A* pathfinder... which hopefully I'll finish one of these days 2011-11-25T21:00:10 hierarchical? 2011-11-25T21:00:15 I'm concerned that my pre-processor won't be fast enough, and I need to add incremental updates to it 2011-11-25T21:00:20 are any of you using multi target a*? 2011-11-25T21:00:52 i "was", but it's really broken 2011-11-25T21:01:02 multi target bfs 2011-11-25T21:01:05 hence i'm generalising my pathfinding stuff so it only needs to be right in one place 2011-11-25T21:01:19 yeah that's what i'm working on now, but it might be neat to have multi target a* too 2011-11-25T21:01:25 I don't do any A* 2011-11-25T21:01:30 I implemented it for my early python effort 2011-11-25T21:02:54 i just realised i don't think i need path information for ants to be separate 2011-11-25T21:03:58 *** newant has quit IRC (Ping timeout: 265 seconds) 2011-11-25T21:04:14 *** fpcfan has quit IRC (Ping timeout: 245 seconds) 2011-11-25T21:05:09 is anyone estimating the portion of the map controlled by your opponents? It seems to me that if you have low 'map share' you should more aggresively trade in order to break out 2011-11-25T21:05:35 BenJackson: do you store separate path information numbers for each ant? or just for the whole search? 2011-11-25T21:06:01 i think my reuse of ants requires me to have separate path info for each ant, but i'm not positive :\ 2011-11-25T21:06:17 bmh: no but it's fairly easy to do 2011-11-25T21:06:38 you probably want to be super offensive if you have lots more ants and less territory than an enemy 2011-11-25T21:07:54 aughh this game makes me cringe: http://ants.fluxid.pl/replay.12515 2011-11-25T21:09:12 cowbandit: why are you parking ants around your hill like that? 2011-11-25T21:09:40 sure, it will stop a wandering ant, but it won't do jack against a reasonable opponent 2011-11-25T21:09:49 yeah that's a good point 2011-11-25T21:10:04 do you think it'd be better to just have 1 ant parked? 2011-11-25T21:10:15 or are static defenses like bad in general? 2011-11-25T21:11:18 static defenses are as good as useless 2011-11-25T21:11:44 cowbandit: parked ants are a bad idea 2011-11-25T21:11:57 blah 2011-11-25T21:12:12 if that location is selected for spawning food it will wait to spawn it there until you move your ant 2011-11-25T21:12:25 ah 2011-11-25T21:12:28 it'll even queue them up and wait to spawn them 2011-11-25T21:12:41 i guess i was hesitant patrol groups though 2011-11-25T21:12:47 i just don't know how to implement them 2011-11-25T21:13:35 I'm gonna go to sleep now, cya 2011-11-25T21:13:40 that was the most fair thing i could come up with for symmetric food spawning to work 2011-11-25T21:14:04 it means your ants can't potentially influence where food will spawn elsewhere on the map 2011-11-25T21:14:13 only prevent food from spawning adjacent to you 2011-11-25T21:14:26 until you move, so you better move before an enemy comes in :P 2011-11-25T21:22:22 *** fpcfan has joined #aichallenge 2011-11-25T21:24:54 *** treeform has joined #aichallenge 2011-11-25T21:26:59 http://pastebin.com/KE9uzc5v 2011-11-25T21:27:08 there's a food collection test example for people who re use ants 2011-11-25T21:27:36 well, you want more rows to get rid of the wrapping thing 2011-11-25T21:27:49 [18:05] BenJackson: do you store separate path information numbers for each ant? or just for the whole search? 2011-11-25T21:28:02 just the whole search for that 2011-11-25T21:28:20 BenJackson: do you reuse ants for food collection? and do you see the problem i'm highlighting with the example linked above? 2011-11-25T21:28:23 assume no wrapping 2011-11-25T21:28:37 atm my bot doesn't do that properly 2011-11-25T21:28:47 it would use the left ant to collect both food items 2011-11-25T21:29:20 cowbandit: your cringeworthy game is essentially the same problem my current submission has 2011-11-25T21:29:39 *** cyphase has quit IRC (Ping timeout: 245 seconds) 2011-11-25T21:29:40 lol yeah it's frustrating 2011-11-25T21:29:41 against lesser opponents the "dive on the ant hill" thing works 2011-11-25T21:29:49 but against better opponents you need to win the map, then win the hill 2011-11-25T21:29:57 yeah 2011-11-25T21:29:58 in a 10 player ffa you need to do both at once :) 2011-11-25T21:30:09 lol yea 2011-11-25T21:30:10 i've been thinking 2011-11-25T21:30:17 do you think a strategy of attrition would be better? 2011-11-25T21:30:23 like what if i just focused on scouting and getting food 2011-11-25T21:30:29 anyone looked at my example? would your bots use one or both ants to collect the food if you're reusing ants? 2011-11-25T21:30:30 and not really care about hills 2011-11-25T21:30:33 *** bmh has quit IRC (Ping timeout: 265 seconds) 2011-11-25T21:30:58 I don't know what "reusing" is but for a long time now I'd send one bot for each food 2011-11-25T21:31:00 well, bad example, place some water below the right a :P 2011-11-25T21:31:18 another challenge would be to put a 3rd food next to one of the others 2011-11-25T21:31:55 if i'm reusing ants, when i make a move i'll mark when the ant will be available to move and where and continue using that ant for collecting targets found 2011-11-25T21:32:15 basically i do that when collecting any targets i know i will get 2011-11-25T21:32:27 and that the ant will be alive afterwards 2011-11-25T21:34:33 http://ants.fluxid.pl/replay.15859 2011-11-25T21:34:40 cowbandit: click about 2/3rds through that 2011-11-25T21:34:52 where yellow and orange have similar numbers and are down to 1v1 2011-11-25T21:35:23 then click on the fog of war for each 2011-11-25T21:35:28 *** Antimony has joined #aichallenge 2011-11-25T21:35:31 sorry, yellow and purple 2011-11-25T21:35:45 anyway, note that they are equally matched at that instant, but purple sees more map 2011-11-25T21:35:50 oh snap 2011-11-25T21:35:50 and then look at the trend 2011-11-25T21:36:01 when I look at replays that's the sort of thing I do 2011-11-25T21:36:12 look at the start, look at the points where ants get an advantage 2011-11-25T21:36:22 try to see why the tide turns 2011-11-25T21:36:37 ah that's really helpful 2011-11-25T21:37:00 it's sometimes tough though 2011-11-25T21:37:10 like i feel in a lot of games 2011-11-25T21:37:15 the tide turns due to a stroke of luck 2011-11-25T21:37:55 or like the advantage changes slowly 2011-11-25T21:39:09 but gaining map control really seems extremely important 2011-11-25T21:41:59 *** besh has joined #aichallenge 2011-11-25T21:42:38 *** McLeopold_ has quit IRC (Ping timeout: 265 seconds) 2011-11-25T21:46:29 *** jstrong has joined #aichallenge 2011-11-25T21:51:23 virut 2011-11-25T21:51:26 oh winblows how i hate you 2011-11-25T21:52:13 jstrong: join the club 2011-11-25T21:52:28 got a stupid patching virus on my winblows partition 2011-11-25T21:52:48 BenJackson: hey thanks for the help today, gonna go sleep now. 2011-11-25T21:52:56 unfortunately since my hackintosh mobo died I'm dependant on windows for some adobe tools (got photoshop working on wine but it runs too slow for normal usage) 2011-11-25T21:52:57 *** cowbandit has quit IRC (Quit: Page closed) 2011-11-25T21:53:06 I don't know where I left my adobe photoshop CD 2011-11-25T21:53:23 so I can't just reformat. I'm going to have to commando squad instead of nuke :( 2011-11-25T21:54:39 stupid windows allowing root access to any random EXE that's willing to give it a blowjob 2011-11-25T21:54:56 Holy cow! I actually won a game on TCP somehow! 2011-11-25T21:55:32 anyone finds pigullory vs xathis games very exicting? http://aichallenge.org/visualizer.php?game=128653&user=4513 2011-11-25T21:56:16 besh: that is intense! 2011-11-25T21:56:18 wow that's the definition of a frontline 2011-11-25T21:56:41 it is like their ants have lives and ours are so so 2011-11-25T21:57:07 how do they even do that ? Influence maps ? 2011-11-25T21:57:12 that looks exactly like a real battle would look 2011-11-25T21:57:29 antimatroid: check it out: http://aichallenge.org/visualizer.php?game=128653&user=4513 2011-11-25T21:57:55 contestbot: later tell mcleopold congrats at making such awesome game rules, this looks almost like a real fight: http://aichallenge.org/visualizer.php?game=128653&user=4513 2011-11-25T21:57:55 amstan_: OK 2011-11-25T22:03:53 wow, the thin red line 2011-11-25T22:03:59 I wonder what combat algorithms they use 2011-11-25T22:08:26 *** suissidle has joined #aichallenge 2011-11-25T22:09:16 *** suissidle has left #aichallenge 2011-11-25T22:09:38 *** rajanaresh has joined #aichallenge 2011-11-25T22:10:15 but no actual ants were harmed 2011-11-25T22:12:15 when a territory is breached, they send reinforcements immediately. 2011-11-25T22:12:55 i am guessing they have some kind of influence maps. but what do i know 2011-11-25T22:15:32 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-11-25T22:25:38 @later tell antimatroid can you review https://github.com/aichallenge/aichallenge/pull/380 please 2011-11-25T22:25:38 janzert: Job's done. 2011-11-25T22:27:20 *** amstan_ has quit IRC (Ping timeout: 260 seconds) 2011-11-25T22:28:38 I think i have managed to get my bot play like pguillory vs xathis !? 2011-11-25T22:29:04 Instead of going into the terriroty, go there and "pause" at the frontier. 2011-11-25T22:29:28 4 vs 2 for the new version. 2011-11-25T22:29:54 *** cyphase has joined #aichallenge 2011-11-25T22:29:57 *** McLeopold_ has joined #aichallenge 2011-11-25T22:32:00 *** smiley1983 has quit IRC (Ping timeout: 260 seconds) 2011-11-25T22:32:20 *** smiley1983 has joined #aichallenge 2011-11-25T22:50:13 combat combat combat 2011-11-25T22:50:16 you so difficult 2011-11-25T22:50:19 yeah 2011-11-25T22:50:25 that's my bots main weakness 2011-11-25T22:50:26 easy when targets don't move 2011-11-25T22:50:31 it has great gathering 2011-11-25T22:51:12 what's your skill lvl on tcp 2011-11-25T22:52:20 (just curious to see how far you got without combat 2011-11-25T22:53:09 not very good 2011-11-25T22:53:16 I always get trashed by the top bots on tcp 2011-11-25T22:53:29 http://ants.fluxid.pl/player/Parasprites 2011-11-25T22:53:48 37 dang you must have great gathering 2011-11-25T22:54:00 *** Rav3nSw0rd has quit IRC (Quit: Leaving) 2011-11-25T22:54:30 *** Rav3nsW0rd has joined #aichallenge 2011-11-25T22:59:17 *** gcflymoto has joined #aichallenge 2011-11-25T23:10:06 *** besh has quit IRC (Quit: Page closed) 2011-11-25T23:17:16 *** treeform_ has joined #aichallenge 2011-11-25T23:19:59 *** treeform has quit IRC (Ping timeout: 248 seconds) 2011-11-25T23:22:16 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-11-25T23:23:10 *** gcflymoto has quit IRC (Ping timeout: 264 seconds) 2011-11-25T23:24:37 *** Surya has quit IRC (Ping timeout: 265 seconds) 2011-11-25T23:24:37 *** AntDroid has quit IRC (Ping timeout: 265 seconds) 2011-11-25T23:24:37 *** mtmcannon has quit IRC (Ping timeout: 265 seconds) 2011-11-25T23:27:29 here's a question, normally you might have say const Object &varName for an input parameter, when defining a function >, how do you make Object& a constant reference? 2011-11-25T23:29:09 janzert: can i just trust that the guy did it alright? :P 2011-11-25T23:29:54 the windows timer always sucked, but the other one (which would be used on the server) is solid 2011-11-25T23:35:38 "const Location&" would that be right? for being a constant reference to a Location? 2011-11-25T23:35:41 it compiles 2011-11-25T23:36:30 *** Rav3nsW0rd has quit IRC (Quit: Have to fix symlinks on my unix systems :( and for some reason I'm inspired to do so now of all times) 2011-11-25T23:40:16 *** avdg has quit IRC (Quit: Leaving.) 2011-11-25T23:40:34 huh fascinating 2011-11-25T23:40:39 this new bot has a very interesting combat pattern 2011-11-25T23:40:52 ... it like clearly tries to zap but doesn't seem to seal the deal 2011-11-25T23:42:25 *** avdg has joined #aichallenge 2011-11-25T23:43:01 *** Antdroid has joined #aichallenge 2011-11-25T23:43:29 http://ants.fluxid.pl/replay.16005 2011-11-25T23:43:32 i'm friedman 2011-11-25T23:43:42 also -- my ant kind of gets scared off my own hill ,gotta fix that lol 2011-11-25T23:43:48 but weird .. i won on ants but got 0 score 2011-11-25T23:44:12 *** TheLinker has quit IRC (Quit: Bye) 2011-11-25T23:46:37 *** avdg has quit IRC (Ping timeout: 240 seconds) 2011-11-25T23:52:37 *** rb_ has quit IRC (Read error: Operation timed out) 2011-11-25T23:52:49 *** rb_ has joined #aichallenge 2011-11-25T23:53:12 *** delt0r_ has quit IRC (Ping timeout: 276 seconds) 2011-11-25T23:54:51 *** Antimony has joined #aichallenge 2011-11-25T23:56:31 *** McLeopold_ has quit IRC (Ping timeout: 265 seconds) 2011-11-25T23:57:00 *** amstan_ has joined #aichallenge 2011-11-25T23:57:00 *** ChanServ sets mode: +o amstan_