2011-11-18T00:04:00 *** twymer has quit IRC (Ping timeout: 260 seconds) 2011-11-18T00:05:51 bah, stanford are doing an online game theory course, but it's like at hte typical introductory level for first years :\ 2011-11-18T00:06:10 *** TheLinker has quit IRC (Quit: Bye) 2011-11-18T00:06:58 *** u_ has quit IRC (Quit: u_) 2011-11-18T00:19:36 *** analyst74 has left #aichallenge 2011-11-18T00:20:55 *** Areks has joined #aichallenge 2011-11-18T00:22:27 *** treeform has joined #aichallenge 2011-11-18T00:23:48 *** modafinil has quit IRC (Read error: Operation timed out) 2011-11-18T00:25:38 *** smiley1983 has joined #aichallenge 2011-11-18T00:26:38 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-11-18T00:27:12 *** djr_ has joined #aichallenge 2011-11-18T00:30:01 *** smiley1983 has quit IRC (Ping timeout: 258 seconds) 2011-11-18T00:30:29 *** Chris_0076 has joined #aichallenge 2011-11-18T00:36:11 evening 2011-11-18T00:38:08 *** modafinil has joined #aichallenge 2011-11-18T00:38:47 afternoon 2011-11-18T00:41:30 *** callahan has quit IRC (Remote host closed the connection) 2011-11-18T00:44:58 *** Nbooo has quit IRC (Ping timeout: 258 seconds) 2011-11-18T00:46:03 *** Nbooo has joined #aichallenge 2011-11-18T00:54:35 whatsup 2011-11-18T00:56:55 *** smiley1983 has joined #aichallenge 2011-11-18T01:00:38 aichallenge: Janzert epsilon * r3ea455b / (2 files in 2 dirs): 2011-11-18T01:00:38 aichallenge: Merge pull request #372 from jackpal/epsilon 2011-11-18T01:00:38 aichallenge: Update dart starter package - http://git.io/5-oZAw 2011-11-18T01:02:25 *** Garf has joined #aichallenge 2011-11-18T01:03:21 *** Accoun has quit IRC () 2011-11-18T01:04:26 *** smiley1983 has quit IRC (Remote host closed the connection) 2011-11-18T01:05:11 *** ltriant has quit IRC (Quit: Get MacIrssi - http://www.sysctl.co.uk/projects/macirssi/) 2011-11-18T01:10:51 *** k has joined #aichallenge 2011-11-18T01:11:17 *** k is now known as Guest30343 2011-11-18T01:11:46 fluxid is dead? 2011-11-18T01:12:23 *** Guest30343 has quit IRC (Client Quit) 2011-11-18T01:12:50 Garf: works for me 2011-11-18T01:12:59 http://ants.fluxid.pl/player/strcat_rewrite 2011-11-18T01:13:09 *** smiley1983 has joined #aichallenge 2011-11-18T01:17:03 timeout for both my bots on the first turn :-/ 2011-11-18T01:17:23 my bot is playing a game right now 2011-11-18T01:17:47 it might be a network issue on your end I guess? 2011-11-18T01:17:50 nope 2011-11-18T01:18:22 *** smiley1983 has quit IRC (Ping timeout: 245 seconds) 2011-11-18T01:19:07 looks like the tcpclient didnt understand it was connected, and tried to log back in 2011-11-18T01:19:08 *** antimatroid has quit IRC (Quit: Leaving.) 2011-11-18T01:19:15 *** antimatroid1 has joined #aichallenge 2011-11-18T01:19:23 which got a "is already playing a game" error 2011-11-18T01:23:16 avdg: ping? 2011-11-18T01:23:20 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-18T01:25:18 anyone here have a bot on the official server that 1, prints something to stderr and 2, has timed out or crashed in a game? 2011-11-18T01:25:50 *** Accoun has joined #aichallenge 2011-11-18T01:26:37 *** antimatroid1 has quit IRC (Ping timeout: 240 seconds) 2011-11-18T01:30:53 *** Antimony has joined #aichallenge 2011-11-18T01:31:02 *** s_seletskiy has joined #aichallenge 2011-11-18T01:31:35 aichallenge: janzert epsilon * r000e7e2 / website/visualizer.php : Fix user name displayed to admin for game errors - http://git.io/uPHvvQ 2011-11-18T01:31:50 *** s_seletskiy has quit IRC (Client Quit) 2011-11-18T01:32:47 aichallenge: janzert epsilon * r0daea7a / website/sql.php : And corresponding sql for last commit - http://git.io/MpR9cA 2011-11-18T01:32:54 *** seletskiy has joined #aichallenge 2011-11-18T01:35:18 and I still broke the visualizer :/ 2011-11-18T01:39:24 aichallenge: janzert epsilon * r6ab45e5 / website/visualizer.php : Add missing semicolon - http://git.io/bvHT1Q 2011-11-18T01:39:47 *** Antimony has quit IRC (Ping timeout: 258 seconds) 2011-11-18T01:46:19 aichallenge: janzert epsilon * raad55a6 / website/sql.php : Now remove the extra comma - http://git.io/FOGfKA 2011-11-18T01:50:11 argh 2011-11-18T01:50:28 *** Palmik has joined #aichallenge 2011-11-18T01:50:44 aichallenge: janzert epsilon * re955fd3 / website/sql.php : and the extra equal sign - http://git.io/qAIIVw 2011-11-18T01:53:07 *** TwistedLogic has joined #aichallenge 2011-11-18T01:56:47 *** antimatroid has joined #aichallenge 2011-11-18T01:57:08 *** dr- has joined #aichallenge 2011-11-18T01:57:10 *** TwistedLogic has quit IRC (Client Quit) 2011-11-18T01:59:23 *** praveen has joined #aichallenge 2011-11-18T02:06:58 *** bmh has joined #aichallenge 2011-11-18T02:07:30 *** bmh has joined #aichallenge 2011-11-18T02:08:13 a1k0n: good to see you're no longer #1 on fluxid :) 2011-11-18T02:10:54 *** antimatroid has quit IRC (Quit: Leaving.) 2011-11-18T02:11:10 *** antimatroid has joined #aichallenge 2011-11-18T02:12:26 *** Flort has joined #aichallenge 2011-11-18T02:12:38 *** Fandekasp has joined #aichallenge 2011-11-18T02:15:41 *** Mooloo has quit IRC (Ping timeout: 252 seconds) 2011-11-18T02:16:51 *** bmh has quit IRC (Quit: bmh) 2011-11-18T02:16:55 *** retybok has quit IRC (Ping timeout: 260 seconds) 2011-11-18T02:17:50 Hi everyone! Is there a way to download my own code from profile? 2011-11-18T02:24:04 *** agentcobra has joined #aichallenge 2011-11-18T02:33:48 *** jasox has joined #aichallenge 2011-11-18T02:34:53 *** sigh has joined #aichallenge 2011-11-18T02:35:35 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-11-18T02:40:49 *** Nealefelaen has joined #aichallenge 2011-11-18T02:40:54 *** HaraKiri has joined #aichallenge 2011-11-18T02:43:00 HaraKiri: a couple of days ago you asked about food decision logic? And I said I had it so that each ant weighs different strengths? 2011-11-18T02:43:41 it turns out it is sub-optimal, it can lead to ants moving away from food one tile away in some circumstances 2011-11-18T02:44:42 oh ok, good to know 2011-11-18T02:44:48 =) 2011-11-18T02:47:22 *** praveen has quit IRC (Ping timeout: 265 seconds) 2011-11-18T02:52:25 *** agentcobra has quit IRC (Quit: Quitte) 2011-11-18T02:54:20 *** epicmonkey has joined #aichallenge 2011-11-18T02:59:31 *** andar has quit IRC (Read error: Operation timed out) 2011-11-18T03:01:41 updated my bot to work with new rules 2011-11-18T03:01:47 now to try it out on some tcp server 2011-11-18T03:03:45 which tcp server are people using? 2011-11-18T03:03:57 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-18T03:04:34 anyone? 2011-11-18T03:04:36 *** aerique has joined #aichallenge 2011-11-18T03:09:21 bhasker: probably http://ants.fluxid.pl/ 2011-11-18T03:09:29 hmm thanks 2011-11-18T03:09:41 i tried tcpants.com but don't see too many bots on it 2011-11-18T03:09:42 bhasker: note: IRC users are expected to reply slowly (as in: hours) 2011-11-18T03:10:59 this channel is more active than that:) 2011-11-18T03:14:40 *** krishna has quit IRC (Ping timeout: 240 seconds) 2011-11-18T03:15:01 *** Jak_o_Shadows1 has joined #aichallenge 2011-11-18T03:18:07 *** Jak_o_Shadows has quit IRC (Ping timeout: 260 seconds) 2011-11-18T03:18:08 these games take too long:-\ 2011-11-18T03:18:12 even on tcp server 2011-11-18T03:18:20 1000 turn limit is far too long 2011-11-18T03:22:12 bhasker: It will not take 1000 turns if your bot can eliminate others 2011-11-18T03:22:38 yea i know its just that most bots don't seem to go in for the kill like my bot is surely losing this game it has like 1 ant left 2011-11-18T03:22:43 but stupid enemy isn't kiling it 2011-11-18T03:22:54 so its been hanging around for like the last 150 turns :-\ 2011-11-18T03:23:11 bhasker: Yeah faced it. No other way - wait to die 2011-11-18T03:24:22 *** Palmik has quit IRC (Remote host closed the connection) 2011-11-18T03:24:46 i hate the fact that they changed the rules just before going live:( 2011-11-18T03:34:49 *** Blkt has joined #aichallenge 2011-11-18T03:40:46 hokay, time to start writing a bot :) 2011-11-18T03:41:09 aichallenge: janzert epsilon * r9bdce12 / website/visualizer.php : Check for that game error query actually has a result - http://git.io/2A3aMA 2011-11-18T03:41:51 *** Jak_o_Shadows1 has quit IRC (Read error: Connection reset by peer) 2011-11-18T03:42:12 *** djr_ has quit IRC (Quit: Leaving) 2011-11-18T03:44:26 Garf: don't kill me! 2011-11-18T03:44:39 i was only sleeping! 2011-11-18T03:45:42 Yexo: http://fluxid.pl/game_3636.input http://fluxid.pl/game_3636.output http://fluxid.pl/game_3636.errors 2011-11-18T03:46:30 HaraKiri: found out the flaw in the logic I was using, if you want food as the highest priority then you use extra data (other priorities) only to decide ties between distances to food =) no more stupid behaviour 2011-11-18T03:47:36 cool 2011-11-18T03:47:58 i really found that the overlay visualizer thing helps alot for decision debugging 2011-11-18T03:48:52 overlay visualizer? 2011-11-18T03:49:38 http://forums.aichallenge.org/viewtopic.php?f=25&t=1633 2011-11-18T03:50:02 cheers =) I've been dumping ascii to a file =D 2011-11-18T03:50:21 *** ping-- has left #aichallenge 2011-11-18T03:50:41 aww you shouldnt :D. http://alexshestakov.com/JetFly/Images/8024a1be57ad43a3ad79725b68b665bf.png <-- example of what you can do 2011-11-18T03:51:04 I was actually considering writing my own visualizer, lol 2011-11-18T03:52:37 just adding arrows from ants to target is 1 line code extra and helps a lot 2011-11-18T03:52:37 looks very useful, much better than reading it from my ascii maps 2011-11-18T03:54:13 *** smiley1983 has joined #aichallenge 2011-11-18T03:54:18 *** mj41 has joined #aichallenge 2011-11-18T03:56:32 do people just use play_one_game_live when testing locally? 2011-11-18T03:57:31 I've made a couple of different scripts, and if there's a specific bug I want to reproduce I've made some specific maps too 2011-11-18T03:58:22 *** smiley1983 has quit IRC (Ping timeout: 245 seconds) 2011-11-18T03:59:17 *** hkraal has quit IRC (Quit: Leaving) 2011-11-18T04:03:35 *** g0llum has joined #aichallenge 2011-11-18T04:04:28 *** Jak_o_Shadows has joined #aichallenge 2011-11-18T04:08:45 *** pairofdice has joined #aichallenge 2011-11-18T04:09:18 *** treeform has quit IRC (Remote host closed the connection) 2011-11-18T04:09:28 *** smiley1983 has joined #aichallenge 2011-11-18T04:13:15 *** andar has joined #aichallenge 2011-11-18T04:13:30 *** epicmonkey has quit IRC (Ping timeout: 276 seconds) 2011-11-18T04:15:49 no shit 2011-11-18T04:15:56 dimkadimon did his undergrad at utas 2011-11-18T04:16:32 *** hkraal has joined #aichallenge 2011-11-18T04:26:13 *** smiley1983 has quit IRC (Ping timeout: 240 seconds) 2011-11-18T04:26:24 *** smiley1983 has joined #aichallenge 2011-11-18T04:31:28 erm... http://pastebin.com/LSpfeHUE 2011-11-18T04:31:34 what's wrong with that first line? :\ 2011-11-18T04:33:17 antimatroid: nothing after '|' 2011-11-18T04:33:33 antimatroid: and or whitespace after \ 2011-11-18T04:34:06 antimatroid: '\' at the end of a line basically means: treat this as a single line although it contains newlines 2011-11-18T04:34:07 digger3: http://pastebin.com/QCB3mFGK that's the actual file contents :\ 2011-11-18T04:34:24 i'm retarded with these things, but i would have thought that'd be fine 2011-11-18T04:34:48 http://pastebin.com/a4ttep5Y 2011-11-18T04:34:49 antimatroid: and what is the error you're getting? 2011-11-18T04:34:50 but then i get that 2011-11-18T04:35:12 seems to run fine when just using the sample python bots :\ 2011-11-18T04:35:23 *** smiley1983 has quit IRC (Ping timeout: 252 seconds) 2011-11-18T04:35:32 *** retybok has joined #aichallenge 2011-11-18T04:36:20 janzert: from your comment about stderr is the official engine passing a pipe to the subprocess but never reading it? 2011-11-18T04:36:38 antimatroid: looks fine to me, although I have 0 experience with compiled-language-bots 2011-11-18T04:37:03 BenJackson: hmm? 2011-11-18T04:37:12 antimatroid: you could try a the full path to your bot 2011-11-18T04:37:12 shouldn't be 2011-11-18T04:38:04 digger3: that seems to have done the jobs, strange 2011-11-18T04:38:11 my comment earlier was looking for someone that could tell me if they could see their game error messages in a game 2011-11-18T04:38:24 or rather from a game on the official site 2011-11-18T04:38:46 ohh 2011-11-18T04:39:19 I thought you had a suspicion that unexpected timeouts were due to the bot filling up the stderr pipe and then blocking 2011-11-18T04:39:38 nope :) 2011-11-18T04:39:38 antimatroid: apparantly the current working directory in the script isn't what you think it is ;) 2011-11-18T04:39:49 yeah probably :P 2011-11-18T04:42:19 *** Insti has quit IRC (Remote host closed the connection) 2011-11-18T04:43:09 i'm retarded with shell scripts, how do i reference a variable so i can do like "./play_tcp.sh 5" and have it pass 5 as the parameters for number of games to play? 2011-11-18T04:43:23 *** Accoun has quit IRC () 2011-11-18T04:43:49 nm, $1? 2011-11-18T04:45:10 yeah, sounds right 2011-11-18T04:47:17 Minthos: can you look at http://aichallenge.org/visualizer.php?game=27117&user=4823 and tell me if you see you're bot stderr output on the bottom 2011-11-18T04:47:50 if you're logged in at least 2011-11-18T04:50:19 *** rajanaresh has left #aichallenge 2011-11-18T04:50:23 janzert: not sure what you mean 2011-11-18T04:51:36 if you're logged into aichallenge.org and you go to the above game, do you see below the game "Minthos - timeout" then a grey box with stderr output from your bot 2011-11-18T04:51:43 I'm trying to run the latest clone of the epsilon branch, but the engine can't load my bot (a compiled executable, from c++) with the error "Failed to start ['Chameleon']" 2011-11-18T04:53:55 * Nealefelaen scratches head 2011-11-18T04:54:00 ok, it's working now.. =P 2011-11-18T04:54:22 ./ ? 2011-11-18T04:54:30 yeah 2011-11-18T04:54:35 janzert: no grey box here 2011-11-18T04:54:44 though I was trying to load it from an above directory before moving it to the local directory 2011-11-18T04:54:47 Minthos: hmm, ok 2011-11-18T04:55:11 *** Insti has joined #aichallenge 2011-11-18T04:56:54 I think timeouts are due to cpu scheduling 2011-11-18T04:57:18 I get them locally when I run 8 bots on a quad core with aggressive turn times 2011-11-18T04:57:37 but when some of the bots time out, the rest survive for the rest of the round 2011-11-18T04:57:54 so, check the runtime of your bot regularly and abort when close to the turntime -> fixed 2011-11-18T04:58:09 I do that digger 2011-11-18T04:58:16 but sometimes it's just not enough 2011-11-18T04:59:53 Minthos: I'd seriously doubt that 2011-11-18T05:00:36 aichallenge: janzert epsilon * rac8dc2c / website/visualizer.php : Use normal equality when checking user_id for error reporting - http://git.io/hWY2zQ 2011-11-18T05:01:00 Minthos: mind checking that page again 2011-11-18T05:01:28 now I see it 2011-11-18T05:01:33 Yay :) 2011-11-18T05:01:51 *** jasox has quit IRC (Quit: Leaving) 2011-11-18T05:02:22 digger3: ubuntu has a pretty slow timer interrupt 2011-11-18T05:02:55 Minthos: what kind of slack are we talking about? 10ms? 2011-11-18T05:03:25 10 ms between interrupts on server edition yes 2011-11-18T05:03:43 if a process gets unlucky it could be starved for several times that long 2011-11-18T05:04:44 Minthos: right, > 50ms could potentionally be a problem... I see 2011-11-18T05:09:15 *** HaraKiri has quit IRC () 2011-11-18T05:09:36 *** ikaros has joined #aichallenge 2011-11-18T05:14:16 unlikely to happen in the ants setup 2011-11-18T05:14:23 I don't know what hardware the workers run on or if they're virtualized, but locally I need to set turn time somewhere between 150 and 200 ms to get my bots to behave like they do on the official server 2011-11-18T05:14:27 it would require the porcess not to be running, yet require the accurate clock 2011-11-18T05:14:56 the workers are older xeons, running vritualized 2011-11-18T05:15:05 so yes, quite a bit slower than a normal machine is quite realistic 2011-11-18T05:15:38 virtualization could add some extra starvation 2011-11-18T05:17:36 *** antimatroid has quit IRC (Ping timeout: 244 seconds) 2011-11-18T05:25:05 *** krishna has joined #aichallenge 2011-11-18T05:29:06 Ugh, I undestand almost 10% of the c++ starter. 2011-11-18T05:32:16 *** smiley1983 has joined #aichallenge 2011-11-18T05:44:28 *** smiley1983 has quit IRC (Ping timeout: 244 seconds) 2011-11-18T05:51:16 *** smiley1983 has joined #aichallenge 2011-11-18T05:52:20 *** ALplus has quit IRC (Ping timeout: 244 seconds) 2011-11-18T05:55:27 *** smiley1983 has quit IRC (Ping timeout: 245 seconds) 2011-11-18T05:56:11 *** smiley1983 has joined #aichallenge 2011-11-18T05:58:30 *** Conorach has joined #aichallenge 2011-11-18T06:02:37 *** mcstar has joined #aichallenge 2011-11-18T06:05:17 *** smiley1983 has quit IRC (Ping timeout: 276 seconds) 2011-11-18T06:12:12 *** Anilm3 has joined #aichallenge 2011-11-18T06:21:52 *** smiley1983 has joined #aichallenge 2011-11-18T06:21:53 *** u_ has joined #aichallenge 2011-11-18T06:22:43 *** u__ has joined #aichallenge 2011-11-18T06:23:38 *** liberforce has joined #aichallenge 2011-11-18T06:25:34 *** antimatroid has joined #aichallenge 2011-11-18T06:25:40 yay i think i have my "own starter bot" mostly set up 2011-11-18T06:26:07 *** u_ has quit IRC (Ping timeout: 248 seconds) 2011-11-18T06:26:07 *** u__ is now known as u_ 2011-11-18T06:31:07 *** smiley1983 has quit IRC (Ping timeout: 258 seconds) 2011-11-18T06:36:33 yay 2011-11-18T06:37:09 *** smiley1983 has joined #aichallenge 2011-11-18T06:41:29 mcstar: i noticed some fairly stupid mistakes from my old one 2011-11-18T06:41:39 like calcualating vision with water blocking :O 2011-11-18T06:41:44 *** imp5imp5 has joined #aichallenge 2011-11-18T06:41:57 in the official starter bot? 2011-11-18T06:42:00 no 2011-11-18T06:42:04 when i made it into a graph 2011-11-18T06:42:17 yeah, i have 2 boards 2011-11-18T06:42:22 one with full connectivity 2011-11-18T06:42:40 now you too 2011-11-18T06:42:46 i just have all/landNeighbour vectors for each square so yeah 2011-11-18T06:43:35 *** praveen has joined #aichallenge 2011-11-18T06:44:23 i changed a kernel option ive got to reboot 2011-11-18T06:44:25 *** mcstar has quit IRC (Quit: WeeChat 0.3.6) 2011-11-18T06:49:18 *** mcstar has joined #aichallenge 2011-11-18T06:50:34 *** grwip has joined #aichallenge 2011-11-18T06:56:04 Hi people 2011-11-18T06:57:18 *** andar has quit IRC (Ping timeout: 276 seconds) 2011-11-18T06:58:31 *** traMKer has joined #aichallenge 2011-11-18T06:58:45 *** andar has joined #aichallenge 2011-11-18T06:59:27 why does this damned windows hang in virtualbox? 2011-11-18T06:59:54 in every 2minutes cpu utilization saturates and i cant do anything in windows 2011-11-18T07:02:23 mcstar: my nemesis from planet wars was also originally from tasmania it seems :P 2011-11-18T07:02:37 (nemesis in a friendly way, just one of the people i really wanted to beat) 2011-11-18T07:04:14 and you just realized that now? 2011-11-18T07:05:03 well i knew last contest he was at anu 2011-11-18T07:05:15 "i really wanted to beat" 2011-11-18T07:05:19 but his profile now has a link to a website which said he was at utas for undergrad/honours :P 2011-11-18T07:05:21 *** HaraKiri has joined #aichallenge 2011-11-18T07:05:25 at least now you know you dont have to travel really far 2011-11-18T07:05:28 he was one of the better perfmorning aussies :P 2011-11-18T07:05:28 to beat him 2011-11-18T07:05:30 up 2011-11-18T07:05:37 he doesn't live in tassie anymore 2011-11-18T07:05:48 *** Jak_o_Shadows has quit IRC (Read error: Connection reset by peer) 2011-11-18T07:05:48 was just talking to him on facebook aha, got my stalk one 2011-11-18T07:06:03 on* 2011-11-18T07:07:24 *** praveen has quit IRC (Ping timeout: 265 seconds) 2011-11-18T07:08:11 lol 2011-11-18T07:12:47 *** xathis has joined #aichallenge 2011-11-18T07:18:27 gawd, now i have a proper terminal in windoze 2011-11-18T07:32:28 *** Anilm3 has quit IRC (Quit: Page closed) 2011-11-18T07:38:33 *** traMKer has left #aichallenge ("I'm a happy Miranda IM user! Get it here: http://miranda-im.org") 2011-11-18T07:38:49 *** netantho has joined #aichallenge 2011-11-18T07:43:08 mcstar: what terminal do you use? 2011-11-18T07:44:21 *** ace1010 has joined #aichallenge 2011-11-18T07:48:44 *** ace1010 has quit IRC (Client Quit) 2011-11-18T07:49:37 retybok: mintty 2011-11-18T07:52:22 *** foRei has joined #aichallenge 2011-11-18T07:54:16 interesting 2011-11-18T07:54:29 in virtualbox, in windows when windows is frozen 2011-11-18T07:54:33 i can still use gvim 2011-11-18T07:56:25 *** TheLinker has joined #aichallenge 2011-11-18T08:04:11 Hello, World! 2011-11-18T08:04:20 slutna 2011-11-18T08:04:21 my first windows app in many years 2011-11-18T08:05:22 mcstar: thanks, will look it up. My work migrates from linux to windows in a few months :'-( 2011-11-18T08:05:41 retybok: im sure there are better ones 2011-11-18T08:05:52 Like bash 2011-11-18T08:05:54 theres a page on wiki listing terminal emulators 2011-11-18T08:06:00 pairofdice: bash is not a terminal emulator 2011-11-18T08:06:02 wtf 2011-11-18T08:06:15 mmm, no tabs... 2011-11-18T08:06:24 Umm, yea 2011-11-18T08:06:29 retybok: its a very basic one, but still better than cmd.exe 2011-11-18T08:06:33 runs vim well 2011-11-18T08:06:37 Windows really sucks for development 2011-11-18T08:06:49 it's not hard to be better than cmd.exe :) 2011-11-18T08:07:01 it suffices for me now 2011-11-18T08:07:10 i just want to cimpile my qt app and get out of there 2011-11-18T08:07:14 o 2011-11-18T08:07:16 :) 2011-11-18T08:12:08 i could compile an example 2011-11-18T08:12:19 aarossig: i managed to get qt+mingw work 2011-11-18T08:21:30 if someone knows the answer please enlighten me: if i now compile something on windows which is 64 bit, will the exe be 32 or 64 bit? 2011-11-18T08:21:56 its more like a mingw question really, are those compilers 32 bit compilers? 2011-11-18T08:22:20 my guess is that they are 2011-11-18T08:22:57 *** buq2_ has quit IRC (Ping timeout: 245 seconds) 2011-11-18T08:24:20 *** buq2 has joined #aichallenge 2011-11-18T08:24:49 *** Accoun has joined #aichallenge 2011-11-18T08:28:44 mcstar, objdump -x f.exe, maybe 2011-11-18T08:29:12 *** smiley1983 has quit IRC (Ping timeout: 245 seconds) 2011-11-18T08:29:50 architecture: i386 2011-11-18T08:29:54 thx 2011-11-18T08:30:14 lol 2011-11-18T08:30:17 it just restarted 2011-11-18T08:30:32 g0llum: your plan worked out 2011-11-18T08:30:38 your rootkit is deployed 2011-11-18T08:31:02 lol, completely forgot, that they're actually starting it ;) 2011-11-18T08:31:45 for the info, what dlls it's plugging in, i guess 2011-11-18T08:32:39 yep i saw 2011-11-18T08:33:00 mcstar, so, recent qt version worked as is ? 2011-11-18T08:33:01 *** Raimondi has quit IRC (Ping timeout: 258 seconds) 2011-11-18T08:33:06 yes 2011-11-18T08:33:14 with mingw 4.4 2011-11-18T08:33:43 ah shit no 2011-11-18T08:33:50 i mean i havent compiled mine yet 2011-11-18T08:33:56 but i tried the examples 2011-11-18T08:33:56 mcstar: Just like in hippie-ville, you can have all sorts of crosses. 2011-11-18T08:34:28 aham, not sure what that maens 2011-11-18T08:34:36 VC++ offers x86-hosted compilers targetting x86 and x86_64, and x86_64-hosted compiler targetting x86_64. 2011-11-18T08:34:50 but this is annoying windoz froze middle-boot 2011-11-18T08:34:54 mingw-w64 offers any combination of host and target bitness, some even being multilib. 2011-11-18T08:35:02 mcstar: You seem to misspell "Windows" a lot. 2011-11-18T08:35:11 i think you do 2011-11-18T08:35:19 :D 2011-11-18T08:35:30 well this one is 32 bit 2011-11-18T08:35:33 but thats ok 2011-11-18T08:35:36 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-18T08:35:43 *** kaemo has joined #aichallenge 2011-11-18T08:35:54 i was just worried that it might build a 64 bit if im on that platform 2011-11-18T08:36:03 *** g0llum has joined #aichallenge 2011-11-18T08:36:13 *** smiley1983 has joined #aichallenge 2011-11-18T08:36:38 Zao: are you offended by my mispellings? 2011-11-18T08:36:48 mcstar: Yup. 2011-11-18T08:36:53 too bad 2011-11-18T08:36:55 I find it childish and inane. 2011-11-18T08:37:21 People ought to be able to have a neutral conversation about platforms without cute misspellings intended to upset and derail things. 2011-11-18T08:37:31 If you're not mature enough to do that, sucks to be you. 2011-11-18T08:37:46 Zao: thx, you taught me a word today: inane 2011-11-18T08:37:55 I had to look that up, too. 2011-11-18T08:38:15 Zao: i dont mind not being "mature" 2011-11-18T08:38:19 by your standards 2011-11-18T08:38:32 some people just cant have fun, too bad 2011-11-18T08:38:36 *** Anilm3 has joined #aichallenge 2011-11-18T08:38:38 Now, coffee! 2011-11-18T08:38:47 Zao: but really, thx for your help 2011-11-18T08:38:49 mcstar: I don't find aggravating trolling by fanboys "fun". 2011-11-18T08:39:37 *** buq2 has quit IRC (Ping timeout: 244 seconds) 2011-11-18T08:39:42 i think i lost you 2011-11-18T08:39:45 nvm 2011-11-18T08:39:58 Hi people 2011-11-18T08:40:22 Zao, he misspelled it once 2011-11-18T08:40:29 twice at least 2011-11-18T08:40:39 *** smiley1983 has quit IRC (Ping timeout: 244 seconds) 2011-11-18T08:40:44 its a capital crime i agree 2011-11-18T08:40:48 Mmh 2011-11-18T08:40:52 but wth, im just a man 2011-11-18T08:41:19 *** buq2 has joined #aichallenge 2011-11-18T08:42:37 i very much dislike the random tie break when spawning ants at hills 2011-11-18T08:43:54 it makes life complicated 2011-11-18T08:43:58 Zao: I think that was quite unfair and judgemental 2011-11-18T08:44:17 http://whitewhine.com/ 2011-11-18T08:44:40 misspeling the name of a ridiculous OS has nothing to do with maturity 2011-11-18T08:45:22 In fact, being offended by it is quite fanboy on your part... 2011-11-18T08:45:24 The difficulty multiplier of OSX is higher 2011-11-18T08:45:52 Oh wait... 2011-11-18T08:46:55 *** jasox has joined #aichallenge 2011-11-18T08:48:22 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-18T08:52:18 *** raemde has quit IRC (Read error: Connection reset by peer) 2011-11-18T09:09:08 *** Antimony has joined #aichallenge 2011-11-18T09:10:30 *** jasox has quit IRC (Quit: Leaving) 2011-11-18T09:11:58 wow, what a strange game: http://ants.fluxid.pl/replay.4424 2011-11-18T09:12:36 I ninja A's base (the first time I've won against A without him timing out first), then I ninja someone else's base, then I timeout and win 2011-11-18T09:12:50 `seen Yexo 2011-11-18T09:12:58 I'm here 2011-11-18T09:12:58 ah, different bot 2011-11-18T09:13:01 @seen Yexo 2011-11-18T09:13:01 Fluxid: Yexo was last seen in #aichallenge 3 seconds ago: I'm here 2011-11-18T09:13:05 .. 2011-11-18T09:13:14 lol 2011-11-18T09:13:14 Yexo: did you check the logs? 2011-11-18T09:13:27 rofl 2011-11-18T09:13:33 did you find anything? 2011-11-18T09:13:36 yes. Result: game engine has different output logged from my bot than my own log 2011-11-18T09:13:45 error log doesn't say anything about it, jsut about the timeout 2011-11-18T09:13:54 could you give me the logs of 4426 too to double-check? 2011-11-18T09:14:39 ehm, 4427 2011-11-18T09:14:43 Yexo: moved logs, just repl... ffff 2011-11-18T09:14:59 *** delt0r_ has joined #aichallenge 2011-11-18T09:15:06 * Fluxid slaps Yexo 2011-11-18T09:15:21 sorry 2011-11-18T09:15:24 Yexo: moved logs, just replace digits from links with 4427 (or 4426) ;) 2011-11-18T09:15:26 and thanks for moving them 2011-11-18T09:15:47 bah, what are people doing about storing paths to collect food etc. and the random tie break with spawning ants? 2011-11-18T09:15:55 it makes things ridiculously complicated 2011-11-18T09:16:39 what?? 2011-11-18T09:17:05 he's counting his ants before they spawn? =D 2011-11-18T09:17:31 yeah i want to make use of ants when they're going to have already been used to collect other food items and store there paths etc. 2011-11-18T09:17:45 i still dont store path 2011-11-18T09:17:47 but to keep it semi optimal you need to know where ants are going to spawn 2011-11-18T09:17:49 no need to do that 2011-11-18T09:17:52 yes there is 2011-11-18T09:17:55 i want max time for battles 2011-11-18T09:18:08 my design is completely different so i don't store paths either 2011-11-18T09:18:18 i tried, but it lead to errors 2011-11-18T09:18:27 diffusion maps, I only look one step ahead 2011-11-18T09:18:28 yeah cause you can't track where ants are going to spawn 2011-11-18T09:18:30 and my simpler approach was more efficient anyway 2011-11-18T09:18:35 assuming worst of best case for a spawn leads to problems 2011-11-18T09:18:41 worst or* 2011-11-18T09:18:47 Fluxid: same result. My own output has a move (which tcpclient also printed, so it got send to the server), but your logs have no mention of it anywhere 2011-11-18T09:19:06 i really wish we were to output spawn orders 2011-11-18T09:19:07 paclet loss? 2011-11-18T09:19:15 antimatroid: you could sacrifice one ant per hive (as in it can't do anything) and then control the spawn location and time? 2011-11-18T09:19:40 Nealefelaen: not always feasible and a pita to switch between when you do and don't have it set up 2011-11-18T09:19:50 That's very inefficient when you have a few ants 2011-11-18T09:20:04 pairofdice: it's a tcp connection, that should take care of that 2011-11-18T09:20:09 it is, but it's how you'd choose 2011-11-18T09:20:34 Right, tcp... 2011-11-18T09:21:08 when spawning at hills was implemented it really should have been done as an executed order :( 2011-11-18T09:21:22 why? 2011-11-18T09:21:28 To make it more complicated? 2011-11-18T09:21:32 because it gives the user more control 2011-11-18T09:21:35 that's less complicated 2011-11-18T09:21:49 the starter bots could just pick a hive randomly and those who don't want to control spawning don't have to worry about it 2011-11-18T09:22:04 you can only give the engine a location and a movement direction 2011-11-18T09:22:17 i'm all for having the game be simple and seeing who can come up with the best strategy 2011-11-18T09:22:26 anything more makes it more complicated.. and you can build any system you want on top of that 2011-11-18T09:22:32 not make the specs so complicated to work with that you need to make everything all fiddly to do anything cool 2011-11-18T09:23:25 if your AI needs to be able to control which hill will spawn an ant, then you have to have the overhead of implementing that =) ie, an ant sitting on top of the hives 2011-11-18T09:23:52 i don't care as much about the overhead of using ants as I do about the problem of what to do when I don't have them all controlled 2011-11-18T09:23:53 pairofdice: He has done it reliably for the last few times I've seen it. 2011-11-18T09:23:55 i'm back to the same issue 2011-11-18T09:24:18 so i guess my real question is, what's less suboptimal, assuming an ant will spawn at a hive in the worst case of time or best case of time? 2011-11-18T09:25:02 if you assume the worst case then an ant might spawn there sooner than expected and you might use an ant from say another hive to collect food much slower, if you assume the best case then you might have been able to collect the food from another hive faster than an ant actually spawns 2011-11-18T09:25:06 those are just two examples 2011-11-18T09:25:47 my AI sits an ant near where food will spawn 2011-11-18T09:26:15 mcstar: ? 2011-11-18T09:26:23 yes? 2011-11-18T09:26:34 which is less suboptimal? :P 2011-11-18T09:26:46 i didnt follow 2011-11-18T09:26:52 let me read it 2011-11-18T09:27:36 i dont think this have a tangible effect 2011-11-18T09:27:45 what do you mean? 2011-11-18T09:27:51 if an ant will be in a better position just reschedule it 2011-11-18T09:27:55 you dont waste much 2011-11-18T09:27:57 there are examples where both assumptions lead to suboptimal outcoems 2011-11-18T09:28:02 so what? 2011-11-18T09:28:15 i really think this is a minor concern 2011-11-18T09:28:15 i build my bot around playing optimally for general simple examples :P 2011-11-18T09:28:26 well, optimally under some kind of metric 2011-11-18T09:28:40 and if i could control spawning it'd be more "optimal" 2011-11-18T09:28:52 metric/measure whatever :P it's late 2011-11-18T09:29:30 Fluxid, mcstar, McLeopold: thanks for all your help. I'm still not sure where exactly the problem is, but my means to debug are exhausted. My computer sends the move, it appears nowhere in the engine logs. So for now I'm going to ignore this and disable the check for it in my bot 2011-11-18T09:29:42 i think assuming the best case might be better 2011-11-18T09:29:49 *** meatroaf_ has joined #aichallenge 2011-11-18T09:30:04 at worst you move the ant from the other hive in the opposite direction the best case number of turns 2011-11-18T09:30:39 Yexo: sure, if you have the time/willingness you could write a little bot that tests the engine 2011-11-18T09:30:57 but im still convinced the engine is correct 2011-11-18T09:31:47 with this many folks watching games and stuff, i would be surprised that there is a issue that has not already been raised 2011-11-18T09:31:58 mcstar: I might try to do that sometime next week, but that'll take quite a bit of extra time 2011-11-18T09:32:22 frankly I've had enough of it for now, the bug is not in my bot and (probably, since the logs show nothing) not in the engine either 2011-11-18T09:32:41 Yexo: sure, dont lose yourself in that 2011-11-18T09:33:09 *** raemde has joined #aichallenge 2011-11-18T09:33:28 antimatroid: i think its better to assume the worst 2011-11-18T09:33:43 mcstar: reasoning? 2011-11-18T09:33:52 there's more to it then just eating the food 2011-11-18T09:34:02 while you are reaching it you explore territory too 2011-11-18T09:34:02 what do you mean? 2011-11-18T09:34:07 and that can be more important 2011-11-18T09:34:18 even if an ant will spawn closer to that particular food 2011-11-18T09:34:25 while the other ant didnt reach it 2011-11-18T09:34:29 i am planning to do all these things similarly although with different prioirities 2011-11-18T09:34:50 the problem is whether to assume an ant will spawn in the worst or best case for any situation? 2011-11-18T09:34:58 i think you should make an ant go where a food is irrelevant if you have anant hill lcoser or not 2011-11-18T09:35:20 still worst 2011-11-18T09:35:24 no reasoning 2011-11-18T09:35:29 but then i'm not going to see if an ant spawns at the ant hill 3 moves away next turn 2011-11-18T09:35:30 i dont see the general case 2011-11-18T09:35:47 ? 2011-11-18T09:35:50 not if i'm unsure of where it will spawn 2011-11-18T09:35:55 let me try to make an example 2011-11-18T09:37:38 http://paste.openttdcoop.org/show/773/ <- Can someone with some python threading knowledge take a look at that? 2011-11-18T09:37:42 is that possible? 2011-11-18T09:38:05 I should mention that "self.inp_lines" is a normal python list 2011-11-18T09:39:44 mcstar: http://pastebin.com/Akx1maG2 2011-11-18T09:39:49 not sure if that'll make sense or not 2011-11-18T09:40:16 * food A hill? 2011-11-18T09:40:18 well, we do have a worst case, it's the first 'a' collecting the first * 2011-11-18T09:40:21 yeah 2011-11-18T09:40:44 sorry, the paste is right 2011-11-18T09:40:49 that is just an extra bit of info 2011-11-18T09:41:25 *** Areks has quit IRC (Ping timeout: 240 seconds) 2011-11-18T09:42:03 i'm going to go with best case 2011-11-18T09:42:05 what happens if you block the lower A? 2011-11-18T09:42:22 will the ant appear from the upper hill? 2011-11-18T09:42:30 yes 2011-11-18T09:42:35 but i can't block it while collecting the food? 2011-11-18T09:42:50 not with the given scenario 2011-11-18T09:42:58 well, you need another ant that opens/closes the hill 2011-11-18T09:43:06 that's not always possible 2011-11-18T09:43:19 antimatroid: i cant change the rules for you 2011-11-18T09:43:20 even if i did that some of the time i'd have to make assumptiosn when i'm not blocking all relevant hills 2011-11-18T09:44:07 my bot would never do anything if i insisted it always control hills :P 2011-11-18T09:44:18 You need to teach your ants some probability 2011-11-18T09:44:19 anyway, in a real situation what you propose is not likely to happen 2011-11-18T09:44:40 *** andar has quit IRC (Ping timeout: 240 seconds) 2011-11-18T09:44:43 I would explore with the left ant 2011-11-18T09:44:43 i guess one option is if ants> x block hills, and if hills ! blocked recalculate all paths 2011-11-18T09:45:06 *** andar has joined #aichallenge 2011-11-18T09:46:24 pairofdice: that would be assuming the best case 2011-11-18T09:47:00 Well, that ant could find more food 2011-11-18T09:48:12 my problem is that once i assign an ant to food i want to enforce that unless say the ant doesn't turn up or the path becomes invalid (from more water information) 2011-11-18T09:48:39 and if i assume the worst case then i'll assign the left 'a' to the top * and that could be rather suboptimal 2011-11-18T09:49:22 whereas if i assume the best case then the worst that can happen is i move the left 'a' down and i have to move it back up in like 1 turn 2011-11-18T09:49:41 of course it can tip more in the other direction depending on distances so i'm not really sure what's best 2011-11-18T09:50:03 i think i should just go with recalculate all the paths when i don't have full information 2011-11-18T09:51:40 or all the time 2011-11-18T09:51:48 I recalculate paths for pretty much all my ants every round 2011-11-18T09:51:50 but that's so expensive when there's lots of food 2011-11-18T09:51:56 really? 2011-11-18T09:51:59 never noticed it 2011-11-18T09:52:01 not expensive at all, when done right 2011-11-18T09:52:10 what's "done right"? 2011-11-18T09:52:20 i just needed to "lambdize" my bfs 2011-11-18T09:52:26 bfs, multiple starting points and multiple destinations 2011-11-18T09:52:32 *** twymer has joined #aichallenge 2011-11-18T09:52:40 was a* really that much slower? 2011-11-18T09:52:46 it's certainly way cooler 2011-11-18T09:53:06 a* is only meant to work with 1 starting point and 1 destination 2011-11-18T09:53:10 A* is ridiculously slow compared. i used it before, even implementing that god damn jps stuff 2011-11-18T09:53:31 hdist(x) = min mdist(x, *) :) 2011-11-18T09:53:44 that's an admissible heuristic for multiple targets, and multiple sources is easy 2011-11-18T09:53:50 so why? 2011-11-18T09:53:56 why would you do that? 2011-11-18T09:54:03 would that not be faster? 2011-11-18T09:54:06 it defies its purpuse 2011-11-18T09:54:08 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-11-18T09:54:13 mcstar: what do you mean with "lambdize"ing bfs? and how can it speed things up? 2011-11-18T09:54:21 how does that defy the purpose? 2011-11-18T09:54:26 HaraKiri: speeds up on a grand level 2011-11-18T09:54:30 Yexo: your paste is correct, that might be a problem 2011-11-18T09:54:36 it gives great flexibility to my search 2011-11-18T09:54:55 *** cbad has quit IRC (Read error: Connection reset by peer) 2011-11-18T09:55:05 my search algorithm takes a callback argument 2011-11-18T09:55:07 you guys are doing iterated bfs's yes? 2011-11-18T09:55:15 to iteratively collect the food? 2011-11-18T09:55:18 and reusing ants? 2011-11-18T09:55:18 well what do you mean with it then? as in lambda(-functions)? 2011-11-18T09:55:19 I don't use it with lambdas, but the functionality is the same 2011-11-18T09:55:34 iteratively? 2011-11-18T09:56:27 HaraKiri: yes 2011-11-18T09:56:43 *** smiley1983 has joined #aichallenge 2011-11-18T09:56:43 the end-condition is very flexible this way 2011-11-18T09:56:50 i can test virtually for anything 2011-11-18T09:56:54 ie. mark all ant locations as available in 0 turns, then bfs from ants to all food, collect a food with a shortest path and update information of when the moving ant will be available next to the food, also trying to track where the new ant will spawn, the researching with your availability information to the remaining food items iteratively? 2011-11-18T09:57:38 aren't lambda functions just tiny fuctions? 2011-11-18T09:57:41 that can be quite slow, hence why i want to store the paths across turns, but not being able to track where ants will spawn presents the problems i've indicated 2011-11-18T09:57:49 eh no, just big dumb map one ant to each food 2011-11-18T09:58:06 what do you mean by one ant to each food? 2011-11-18T09:58:06 HaraKiri: they can be of any size 2011-11-18T09:58:15 Oh man, I have a lot to fix in my ant 2011-11-18T09:58:20 that would involve f bfs's where f is the number of food squares yeah? 2011-11-18T09:58:32 each food gets exactly one ant assigned to it, the closest available in a greedy manner 2011-11-18T09:58:33 HaraKiri: the point is: i dot have to previously declare any function i would want to use as an end codition for different purposes 2011-11-18T09:58:41 no, 1 bfs for all food and all ants 2011-11-18T09:58:49 yeah well i'm trying to be a little more sophisticated :P 2011-11-18T09:59:03 I don't see the point 2011-11-18T09:59:08 If by sophisticated you mean slower ;) 2011-11-18T09:59:14 you could do that with function pointers and such, but with std::function its more convenient 2011-11-18T09:59:19 what pairofdice said 2011-11-18T09:59:26 you will see :P 2011-11-18T09:59:38 if you use one bfs for all food, it can't be perfect path's afaik 2011-11-18T09:59:42 *** Antimony has joined #aichallenge 2011-11-18T09:59:51 i search from each food 2011-11-18T10:00:11 *** cbad has joined #aichallenge 2011-11-18T10:00:11 *** cbad has joined #aichallenge 2011-11-18T10:00:14 but if the ant already has a target it will only switch targets if this is closer to it than the original 2011-11-18T10:00:21 i'm going to just assume the best case and not try and control hill spawning 2011-11-18T10:00:28 that's the least fiddly i think 2011-11-18T10:00:39 it's not "perfect" in the sense that for 2 adjacent foods they will both get an ant each even if it had been more efficient with just one for both 2011-11-18T10:00:51 Minthos: really? 2011-11-18T10:01:02 mine doesnt assign multiple ants to any target 2011-11-18T10:01:04 *** NightExcessive has joined #aichallenge 2011-11-18T10:01:13 mcstar: read what I said 2011-11-18T10:01:32 k 2011-11-18T10:02:09 *** amstan has joined #aichallenge 2011-11-18T10:02:09 *** ChanServ sets mode: +o amstan 2011-11-18T10:02:39 this can be solved very easily 2011-11-18T10:02:43 Minthos: thanks for the idea 2011-11-18T10:03:04 mcstar: i still don't see how lambda functions can speed up bfs. anything fancy you can check for in lambda that you couldn't in a regular function? 2011-11-18T10:03:31 * avdg says hi 2011-11-18T10:03:31 HaraKiri: it speeds up food collection by being flexible 2011-11-18T10:03:40 lambda is just convenient syntax for passing code as arguments to functions 2011-11-18T10:03:40 not performance wise 2011-11-18T10:04:03 oh ok 2011-11-18T10:04:05 passing code as arguments to functions is what makes the search flexible 2011-11-18T10:04:34 your point? 2011-11-18T10:05:10 why does everything have to have a point, I was just explaining it to harakiri 2011-11-18T10:05:32 i thought he is familiar with that, just misunderstood what i said 2011-11-18T10:05:43 or rather i wasnt clear 2011-11-18T10:06:04 yeah so if you're looking for a point, you just gave it 2011-11-18T10:06:14 ive never used lambdas. i should probably look into it 2011-11-18T10:06:15 HaraKiri: a lambda is an anonymous function, think of it as the function equivalent of local variables (where regular functions would be constants) 2011-11-18T10:06:58 im looking at examples at the moment. right now i wouldn't know where to use them, but sounds fancy 2011-11-18T10:07:16 Minthos: a thing that regular functions cant do is capturing variables from the context of their creation 2011-11-18T10:07:28 well, in a sense they capture them,but those are global variables 2011-11-18T10:08:10 HaraKiri: its been with us almost since the creation of proper programming languages 2011-11-18T10:08:38 mcstar: true, although you could send references/copies/whatever to the function calling the function you provide, but that gets a bit icky 2011-11-18T10:08:39 (high-level) 2011-11-18T10:08:42 https://github.com/berak/ants-tcp/pull/5 <- I hope that'll fix the problems 2011-11-18T10:08:42 i feel bad now :/. i only do stupid c++ numeric stuff, nothing fancy 2011-11-18T10:09:06 mcstar: I use D, which afaik doesn't have lambdas but something called delegates, which are functions coupled with state objects 2011-11-18T10:09:24 acheiving basically the same, just with slightly more syntax 2011-11-18T10:10:13 *** Surya has joined #aichallenge 2011-11-18T10:10:29 *** raver1975 has joined #aichallenge 2011-11-18T10:12:56 *** raver1975 has quit IRC (Client Quit) 2011-11-18T10:14:32 *** NightExcessive has quit IRC (Quit: Leaving) 2011-11-18T10:23:20 this compilation business sucks 2011-11-18T10:23:45 mingw cant compile my stuff, gcc4 asks for x11 for qt on windows 2011-11-18T10:23:46 pfff 2011-11-18T10:29:05 *** goldcaddy77 has joined #aichallenge 2011-11-18T10:29:15 If you're building Qt, ensure that you're using configure.exe and not ./configure. 2011-11-18T10:30:08 *** raemde has quit IRC (Read error: Connection reset by peer) 2011-11-18T10:31:14 dejavu 2011-11-18T10:31:27 no, i preferably dont want to rebuild it 2011-11-18T10:31:42 mingw4.4 can compile with --std=c++0x 2011-11-18T10:31:45 cant* 2011-11-18T10:36:14 The version of mingw I have accepts c++0x perfectly fine 2011-11-18T10:37:08 accepts it, but says ::swprintf is undefined or something 2011-11-18T10:37:38 Antimony: can you compile proper c++11 code with it? 2011-11-18T10:37:49 for example a lambda? thats what i tried 2011-11-18T10:37:59 lambdas aren't supported till gcc4.6 2011-11-18T10:38:05 ? 2011-11-18T10:38:09 but I can use auto, decltype, perfectly fine 2011-11-18T10:38:14 noway, it in my uploaded code 2011-11-18T10:38:17 is* 2011-11-18T10:38:36 the server has 4.5 2011-11-18T10:38:47 oops it takes 4.5 2011-11-18T10:38:51 yeah 2011-11-18T10:38:58 unfortunately I only have 4.4.1 2011-11-18T10:39:04 http://gcc.gnu.org/projects/cxx0x.html 2011-11-18T10:39:20 still, even 4.4 has a lot of useful features 2011-11-18T10:39:44 Antimony: so mingw's version number correlates with gcc's? 2011-11-18T10:39:52 I think so 2011-11-18T10:39:55 ok 2011-11-18T10:41:14 i see mingw 4.5.2 in cygwin's repo 2011-11-18T10:41:54 can i somehow see the contents of a cygwin package? 2011-11-18T10:42:07 is there a cli equivalent of the cygwin gui installer? 2011-11-18T10:42:15 (or a better alternative) 2011-11-18T10:44:07 not that I know 2011-11-18T10:44:16 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-18T10:44:49 http://cygwin.com/packages/ <- there you can search the contents of cygwin packages 2011-11-18T10:49:13 Fluxid: the issue has been fixed in berak's repo. Could you please update your server? 2011-11-18T10:50:35 *** praveen has joined #aichallenge 2011-11-18T10:51:59 *** sigh has quit IRC (Remote host closed the connection) 2011-11-18T10:52:50 Yexo: merged, i'll wait one minute for games and restart 2011-11-18T10:53:01 great, thanks :) 2011-11-18T10:53:58 Ahh, cygwin has irssi package 2011-11-18T10:55:03 done 2011-11-18T10:55:19 do you need cygwin packages or mingw packages ? 2011-11-18T10:55:20 Was that the issue of leakage 2011-11-18T10:55:46 i found it 2011-11-18T10:55:53 it is prefixed with i686 2011-11-18T10:56:17 -, the universal namespace separator 2011-11-18T10:56:23 *** amstan has quit IRC (Read error: Connection reset by peer) 2011-11-18T10:56:45 *** amstan has joined #aichallenge 2011-11-18T10:56:45 *** ChanServ sets mode: +o amstan 2011-11-18T10:57:01 mcstar: looking for mingw installer ? 2011-11-18T10:57:29 no 2011-11-18T10:57:40 i found 4.5.2 mingw in cygwin's repo 2011-11-18T10:57:51 the game on aichallenge right now is classic 2011-11-18T10:58:10 you mean gcc 4.5.2 ? 2011-11-18T10:58:28 mingw and gcc are separate struff 2011-11-18T10:58:33 *stuff 2011-11-18T11:00:05 *** smiley1983 has quit IRC (Ping timeout: 260 seconds) 2011-11-18T11:00:19 *** smiley1983 has joined #aichallenge 2011-11-18T11:00:57 liberforce: no, mingw 4.5.2 2011-11-18T11:01:23 4.5.2-1 2011-11-18T11:01:34 as far as I know, mingw is no more versioned 2011-11-18T11:01:49 it's like a rolling-release distrib 2011-11-18T11:02:03 when using mingw-get of course 2011-11-18T11:04:12 *** Antimony has joined #aichallenge 2011-11-18T11:04:36 or maybe I don't get the point and should resume what I was doing :p, sorry for the annoyance 2011-11-18T11:05:27 no 2011-11-18T11:05:30 or yes 2011-11-18T11:05:36 i mean it isnt annoyance 2011-11-18T11:06:33 *** mleise has joined #aichallenge 2011-11-18T11:13:55 *** analyst74 has joined #aichallenge 2011-11-18T11:15:49 *** aerique has quit IRC (Quit: ...) 2011-11-18T11:19:46 *** TheLinker has quit IRC (Quit: Bye) 2011-11-18T11:23:52 *** a9wegf9a has joined #aichallenge 2011-11-18T11:24:02 huh 2011-11-18T11:24:32 i managed to run a qt program under wine that i compiled in a vm'd windows with mingw 2011-11-18T11:24:54 i just copied to dll's 2011-11-18T11:25:07 so i think i wont need static linking after all 2011-11-18T11:25:12 just package the dll's 2011-11-18T11:25:43 *** a9wegf9a has quit IRC (Client Quit) 2011-11-18T11:29:37 *** pairofdice has quit IRC (Quit: Just keep breathing) 2011-11-18T11:30:42 *** Antimony_ has joined #aichallenge 2011-11-18T11:31:05 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-18T11:31:07 *** Antimony_ is now known as Antimony 2011-11-18T11:31:15 *** pairofdice has joined #aichallenge 2011-11-18T11:32:38 *** JorgeB has quit IRC (Quit: Computer has gone to sleep.) 2011-11-18T11:36:25 success 2011-11-18T11:36:28 so 2011-11-18T11:36:44 i compiled my files with i686-pc-mingw32-g++ 2011-11-18T11:36:55 and then linked with ming32-g++ 2011-11-18T11:36:59 w 2011-11-18T11:37:11 ALL GLORY TO THE HYPNOTOAD 2011-11-18T11:37:21 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-11-18T11:37:26 :D 2011-11-18T11:37:30 HaraKiri: you around? 2011-11-18T11:37:39 *** Antimony has joined #aichallenge 2011-11-18T11:38:27 mintty+irssi=not bad 2011-11-18T11:38:51 if anyone wants to try it on windows please respond 2011-11-18T11:39:09 try to compile what? 2011-11-18T11:39:12 *** smiley1983 has quit IRC (Ping timeout: 245 seconds) 2011-11-18T11:39:17 *** bearoff_w has joined #aichallenge 2011-11-18T11:39:29 *** NightExcessive has joined #aichallenge 2011-11-18T11:40:16 not compile 2011-11-18T11:40:21 if you wanna try out my app 2011-11-18T11:40:52 which app is that? 2011-11-18T11:41:21 *** smiley1983 has joined #aichallenge 2011-11-18T11:42:05 *** traMKer has joined #aichallenge 2011-11-18T11:42:14 has it been tested on osx? 2011-11-18T11:42:58 sorry 2011-11-18T11:43:03 osx? 2011-11-18T11:43:05 no 2011-11-18T11:43:07 bhasker: i dont think your bot works right anymore 2011-11-18T11:43:08 i cant do that 2011-11-18T11:43:51 yea lol 2011-11-18T11:43:52 Yexo: you can draw ant with it, and instantly resolve battle situations 2011-11-18T11:43:57 ants* 2011-11-18T11:43:58 its expansion logic is broken 2011-11-18T11:44:00 like hell 2011-11-18T11:44:05 just gets stuck in starting area 2011-11-18T11:44:10 mcstar: interesting, where can I find it? 2011-11-18T11:44:13 seems to have trouble with mazes 2011-11-18T11:44:18 i can bundle it 2011-11-18T11:44:28 yea and also has no notion of defending its hill 2011-11-18T11:44:30 source is also ok 2011-11-18T11:44:37 haven't modified it except to work with the new state stuf 2011-11-18T11:44:59 it can raze other hills but not defends its own either 2011-11-18T11:45:22 mine's pretty much the same 2011-11-18T11:45:39 yours is much better i think mine needs major rework 2011-11-18T11:45:49 hate it that they changed the stupid rules right before starting the contest 2011-11-18T11:45:58 :( 2011-11-18T11:45:58 i have a rudimentary but majorly broken hill defense 2011-11-18T11:46:09 Yexo: the source is on github 2011-11-18T11:46:18 but i went to all this trouble so you wont have to compile it 2011-11-18T11:46:22 yea i will work on it when i find time, right now at least its working good enuf for now 2011-11-18T11:46:22 it isnt trivial 2011-11-18T11:46:25 it was like this for at least a mongth before the contest 2011-11-18T11:46:38 *** jasox has joined #aichallenge 2011-11-18T11:46:41 i was out of the country! :( 2011-11-18T11:47:19 anyway i didnt even have a bot before it started. haha 2011-11-18T11:47:36 hehe , though frankly i enjoyed planet wars way more 2011-11-18T11:47:42 these games are just not as much fun to watch 2011-11-18T11:47:46 and take far too long 2011-11-18T11:47:48 i like this one more 2011-11-18T11:47:58 Ants is awesome 2011-11-18T11:48:08 *** Antimony has quit IRC (Ping timeout: 258 seconds) 2011-11-18T11:48:18 but the games do take too long, youre right there 2011-11-18T11:48:19 I didn't play planet wars but ants is great, so I think ants is better :D 2011-11-18T11:49:02 *** bearoff_w has left #aichallenge 2011-11-18T11:49:35 Minthos: closed world assumption? ;p 2011-11-18T11:49:58 yea too long imo, 750 /1000 turn games take forever 2011-11-18T11:50:16 *** amstan has quit IRC (Ping timeout: 244 seconds) 2011-11-18T11:50:24 mcstar: yes 2011-11-18T11:50:36 HaraKiri: its done you can test it now if you want to 2011-11-18T11:50:49 yay sure 2011-11-18T11:50:52 https://min.us/mDHbuSV7o 2011-11-18T11:50:58 http://ants.fluxid.pl/replay.4628 <- McLeopold something to look at for the final score. Game finishes with "rank stabilized" when orange loses it's final ant. However it keeps it's score for it's hills. This is again a situation where sacrificing your all your ants before your hill might lead to a better score 2011-11-18T11:51:02 Yexo: you too 2011-11-18T11:51:03 you are the real mcstar, right? ;p 2011-11-18T11:51:06 yes 2011-11-18T11:51:13 *** jasox has quit IRC (Client Quit) 2011-11-18T11:51:31 the minus acount belong to liquid-phynix just as my github account 2011-11-18T11:51:34 s 2011-11-18T11:51:38 that program will probably steal my ant code :D 2011-11-18T11:51:44 :D 2011-11-18T11:51:58 no it wont, but be sure you only downlload from liquid-phynix 2011-11-18T11:52:06 Yexo: I'm pretty sure that is fixed on the contest servers. I'm guessing fluxid just doesn't have the latest version of the code 2011-11-18T11:52:25 works! awesome 2011-11-18T11:52:27 Yexo: in that case, orange was the only one to have been able to improve score 2011-11-18T11:52:31 *** JorgeB has joined #aichallenge 2011-11-18T11:52:31 *** Antimony has joined #aichallenge 2011-11-18T11:52:57 mcstar: tools forum ( http://forums.aichallenge.org/viewforum.php?f=25 ) would probably like it 2011-11-18T11:53:02 McLeopold: I know. But if orange had lost it's hill before it's last ant it would've been a drawn between orange and blue 2011-11-18T11:53:14 McLeopold: but he would have a score of 0 with current code right? 2011-11-18T11:53:14 ive got to eat now 2011-11-18T11:53:17 will be back 2011-11-18T11:53:35 janzert: no 2011-11-18T11:53:49 oh, I thought that was your latest change 2011-11-18T11:53:55 only lone survivor hands out bonuses 2011-11-18T11:54:09 ahh 2011-11-18T11:54:10 only timeout or crash would remove hill points 2011-11-18T11:54:20 it's not about the bonus for yellow, but about the losing the bonus for ornage in this case 2011-11-18T11:54:45 since orange lasted longer, I'm okay with it having a better score than blue 2011-11-18T11:55:10 that's a valid argument 2011-11-18T11:55:34 so, in general, it is better to use every last ant to save a hill 2011-11-18T11:55:56 and if they can't save it, they had better die before it goes down 2011-11-18T11:56:09 yep 2011-11-18T11:56:14 even if they're across the map 2011-11-18T11:56:26 but then, why would you have ants across the map when you need to defend? 2011-11-18T11:56:32 not going to code it that way :) I've had too many games where I lost my hill but still was able to raze 1 or 2 other hills 2011-11-18T11:56:40 people do it all the time McLeopold 2011-11-18T11:56:53 you can't always see an attack coming in time 2011-11-18T11:56:54 normal isn't always good 2011-11-18T11:57:00 plus, getting food is defense 2011-11-18T11:57:31 well say you're gathering food and fighting on several fronts 2011-11-18T11:57:36 overall, I don't think this is a big scoring issue 2011-11-18T11:57:40 then one of your fronts get swarmed 2011-11-18T11:57:57 you need pretty smart ai to drop everything and defend 2011-11-18T11:58:02 It's not a big issue at all. I was just curious what you thought of it after the other recent changes 2011-11-18T11:58:07 I'm fine with it either way 2011-11-18T11:58:29 wow, I haven't submitted a new bot in over a week 2011-11-18T11:58:30 a suicide is much different from an intentional crash 2011-11-18T11:58:46 it is part of the game, so it is okay if it affects outcome 2011-11-18T11:58:49 *** Antimony_ has joined #aichallenge 2011-11-18T11:59:18 BenJackson: then you have enough games for a roughly good ranking, congrats! 2011-11-18T11:59:34 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-11-18T11:59:44 *** Antimony_ is now known as Antimony 2011-11-18T11:59:44 a1k0n: better yet, I just came in last in a game you played, catapulting you up the ranks ;-) 2011-11-18T12:00:09 heh but then i was paired with xathis in the next :( 2011-11-18T12:02:14 i think A's micro is better than xathis's now 2011-11-18T12:02:20 *** Hammerok has joined #aichallenge 2011-11-18T12:02:32 i can't help but think i'm doin it wrong against him 2011-11-18T12:02:35 I haven't really tuned my micro since I got it working the first time 2011-11-18T12:02:45 I need a new plan for my macro and I've not liked any of my new solutions 2011-11-18T12:03:13 yeah i have nothing yet... i might just tune my constants for a while 2011-11-18T12:03:19 my tip re: xathis would be to put yourself in his place (pre-hills) and think about the tradeoffs then 2011-11-18T12:03:27 who's A on the official server? is he there at all? and is xathis on any tcp servers? 2011-11-18T12:03:35 he's doing great and his bot is basically optimized for the scoring based on killing ants 2011-11-18T12:03:37 Minthos: FlagCapper 2011-11-18T12:03:56 Minthos: i'm on the fluxid tcp every once a while 2011-11-18T12:04:36 I leave an old bot on fluxid all the time so people have something to play 2011-11-18T12:04:57 ah, flagcapper just submitted a new version so he's not in the top 100 anymore 2011-11-18T12:05:17 as did many of the top 100 2011-11-18T12:05:22 *** mleise has quit IRC (Quit: Leaving.) 2011-11-18T12:05:50 but it takes like 3-4 days to get back in 2011-11-18T12:06:20 need more clouds 2011-11-18T12:06:45 Google should step up 2011-11-18T12:07:07 they did actually 2011-11-18T12:07:10 yeah it doesn't look good to have sponsored by google on something with so little computing power 2011-11-18T12:07:11 but now jeff needs to show up 2011-11-18T12:07:17 ah 2011-11-18T12:08:04 HaraKiri: im glad you liked it 2011-11-18T12:08:12 if you have suggestions pls... 2011-11-18T12:08:13 bmh is lobbying for more dollars also 2011-11-18T12:08:40 im not sure if it would fit into the tools, its is a binary 2011-11-18T12:08:49 Maybe a few subtle posts on G+... 2011-11-18T12:08:50 and i think the guys would be opposed to that 2011-11-18T12:09:06 *** smiley1983 has quit IRC (Quit: leaving) 2011-11-18T12:09:29 *** BenJackson has quit IRC (Ping timeout: 265 seconds) 2011-11-18T12:11:40 Will the engine do anything silly if I write to stderr? 2011-11-18T12:11:44 mcstar: very useful tool :) 2011-11-18T12:12:07 would be nice if you could draw gridlines so I won't have to guess were each tile is 2011-11-18T12:12:09 thx 2011-11-18T12:12:12 NightExcessive: doesn't seem to for me, I write a bunch of status to stderr each turn 2011-11-18T12:12:14 there is 2011-11-18T12:12:24 Minthos: Alright. Thanks. :D 2011-11-18T12:14:06 Yexo: http://imgur.com/S7BDD 2011-11-18T12:14:11 should look like this 2011-11-18T12:14:15 what do you see? 2011-11-18T12:14:44 wait, i dont have that picture in the background and all that grid stuff 2011-11-18T12:14:49 oh 2011-11-18T12:14:51 the buttons don't have colors and I miss the background grid on the right 2011-11-18T12:14:55 wait, ill screenshot 2011-11-18T12:15:00 HaraKiri: dont 2011-11-18T12:15:05 i run it in wine 2011-11-18T12:15:10 missing image files maybe? 2011-11-18T12:15:11 it looked like you described 2011-11-18T12:15:15 there are no image files in the zip 2011-11-18T12:15:18 Yexo: it is compiled in 2011-11-18T12:15:21 ok 2011-11-18T12:15:40 ill try to fix it 2011-11-18T12:15:54 oh ok. i just drew a penis with ants :/ 2011-11-18T12:21:26 a1k0n: in what did did google "step up" in the server problem? 2011-11-18T12:21:44 there are funds committed to this contest 2011-11-18T12:21:57 to be spent on ec2 hosting 2011-11-18T12:22:05 more than before? 2011-11-18T12:22:06 I just wrote some visualizer code that lights up the water tiles like it's rainbow road from Mario Kart 2011-11-18T12:22:27 I guess it would kind of suck running out of funds before the actual contest. 2011-11-18T12:22:27 well, last i heard was they were waiting for a response from jeff to receive them 2011-11-18T12:23:08 Whoever does the website should put a link to a PayPal account on the main splash page 2011-11-18T12:23:16 i'm not involved directly though so i don't know what's going on 2011-11-18T12:23:26 that is also being discussed 2011-11-18T12:23:29 with *** ATTENTION: NOW ACCEPTING DONATIONS TO HELP OUT WITH SERVER COSTS *** 2011-11-18T12:23:36 i'd throw down 2011-11-18T12:23:46 and so would my buddy who is in the contest 2011-11-18T12:24:04 ask for like 10 or 20 bucks a head 2011-11-18T12:24:13 *** amstan has joined #aichallenge 2011-11-18T12:24:13 *** ChanServ sets mode: +o amstan 2011-11-18T12:24:15 I think with the amount of time and effort people are putting in... 2011-11-18T12:24:21 that's not a lot to ask 2011-11-18T12:24:28 goldcaddy77: +1 for rainbow ants! http://i.imgur.com/oQSLn.png 2011-11-18T12:25:00 !rainboew it's so beautiful! *sob* 2011-11-18T12:25:04 ah damn 2011-11-18T12:25:14 !rainbow it's so beautiful! *sob* 2011-11-18T12:25:18 i give up 2011-11-18T12:26:05 @rainbow it's so beautiful! *sob* 2011-11-18T12:26:05 Yexo: it's so beautiful! *sob* 2011-11-18T12:26:20 :D 2011-11-18T12:26:20 lol 2011-11-18T12:26:26 what's that? food heatmap? 2011-11-18T12:27:01 nah it's a bfs inward from fringe of visible parts of map 2011-11-18T12:27:27 Is there any way to detect the case when my ants collide? 2011-11-18T12:27:38 dont judge me what the ants are doing or their orders. it's just for testing the visualizer :D 2011-11-18T12:27:55 collision = death, right? 2011-11-18T12:28:01 yeah mine too... 2011-11-18T12:28:08 HaraKiri: RAINBOW ROAD! http://imgur.com/2GPut 2011-11-18T12:29:34 woops, should have just linked the pic: http://i.imgur.com/2GPut.png 2011-11-18T12:29:47 HaraKiri: yes 2011-11-18T12:30:11 collison with self, correct? collision with water tile is ignored? 2011-11-18T12:30:34 *** mviel_ has quit IRC (Quit: Leaving) 2011-11-18T12:30:37 goldcaddy77: yes, I mean with self 2011-11-18T12:31:52 *** Surya has quit IRC (Quit: Page closed) 2011-11-18T12:32:15 *** Raimondi has joined #aichallenge 2011-11-18T12:34:08 *** praveen has quit IRC (Ping timeout: 265 seconds) 2011-11-18T12:34:19 collision, food, water == ignored ; self, dead == will kill you. 2011-11-18T12:34:46 I know what happens but how can you detect two of your ants have collided? 2011-11-18T12:35:02 when theyre not there anymore? 2011-11-18T12:35:16 They're suddenly dead 2011-11-18T12:35:25 how can you be sure they weren't killed by an enemy? 2011-11-18T12:35:46 If there was no enemy close 2011-11-18T12:35:57 if you have two dead ants on the same tile they must have been killed by a collision 2011-11-18T12:36:09 1 dead ant on a tile means either killed by collision with enemy or killed in a fight 2011-11-18T12:36:12 Yexo: thanks! 2011-11-18T12:36:41 I don't think ants can collide with enemies? 2011-11-18T12:36:58 close to your hill or close to your opponents hill 2011-11-18T12:37:09 HaraKiri, you get my pic? 2011-11-18T12:37:27 yes. pretty pretty lady! does it solve any practucal purpose? :D 2011-11-18T12:37:34 move next to hill, in the same turn an enemy ant spawns there. Your ant and the enemy ant are now next to eachother. If you move to the same tile they collide 2011-11-18T12:40:39 aha, so a spawned ant can actually move off its hill and crash into an enemy ant before any combat resolution is done 2011-11-18T12:41:18 or it remains there and an enemy tries to move on top of that hill 2011-11-18T12:41:26 yes, but if it does that it's likely the enemy could just take the hill 2011-11-18T12:42:53 HaraKiri: yes, but I can't share that info with you :) 2011-11-18T12:43:13 you're such a tease ;p 2011-11-18T12:47:55 *** NightExcessive has quit IRC (Ping timeout: 255 seconds) 2011-11-18T12:52:23 *** traMKer has left #aichallenge ("I'm a happy Miranda IM user! Get it here: http://miranda-im.org") 2011-11-18T12:54:49 *** NightExcessive has joined #aichallenge 2011-11-18T12:56:08 *** pairofdice has quit IRC (Quit: leaving) 2011-11-18T12:57:35 *** amstan has quit IRC (Ping timeout: 248 seconds) 2011-11-18T12:58:00 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-18T12:59:36 a1k0n: you just had to play agaist Xathis huh? 2011-11-18T12:59:41 *** pairofdice has joined #aichallenge 2011-11-18T12:59:43 that guy is a pimp 2011-11-18T13:03:57 but his bot was afraid to take the last hill which surprises me 2011-11-18T13:04:06 http://aichallenge.org/visualizer.php?game=103392 2011-11-18T13:04:14 *** Blkt has quit IRC (Remote host closed the connection) 2011-11-18T13:07:25 hmm, maybe the whole gametree-analysis didnt give a favorable result, ignoring the vast advantage of total ant numbers (just blindly guessing) 2011-11-18T13:08:01 his bot is really conservative when it comes to risking losses in order to kill enemy ants 2011-11-18T13:08:06 which is probably good 2011-11-18T13:08:31 but there's a juicy hill waiting! 2011-11-18T13:09:26 *** mj41 has quit IRC (Ping timeout: 244 seconds) 2011-11-18T13:12:11 *** Harpyon has joined #aichallenge 2011-11-18T13:20:51 he should make his ants braver in some situations 2011-11-18T13:21:02 *** treeform has joined #aichallenge 2011-11-18T13:22:10 I have a berzerk mode in my bot, you can see it in action here: http://aichallenge.org/visualizer.php?game=100212&user=4823 2011-11-18T13:22:54 of course my bot sucks compared to his, but the issue is the same and the solution works 2011-11-18T13:23:06 Minthos: your bot has an amazing tendency to pile up ants so it's good you added that, haha 2011-11-18T13:24:28 a1k0n: sorry left for a few minutes 2011-11-18T13:24:32 totally agree... 2011-11-18T13:24:45 for being that good he should have known that he had you completely outnumbered 2011-11-18T13:24:49 wow, that is awesome Minthos :D 2011-11-18T13:25:02 *** GeorgeSebastian has quit IRC (Ping timeout: 245 seconds) 2011-11-18T13:26:19 *** cuademon has joined #aichallenge 2011-11-18T13:26:32 Minthos: what finally triggered your dude to move forward? 2011-11-18T13:26:55 goldcaddy77: the number of remaining rounds going below a certain limit :p 2011-11-18T13:27:24 it could be more clever, but I have other things that are more important to fix 2011-11-18T13:27:55 *** pairofdice has quit IRC (Read error: Operation timed out) 2011-11-18T13:28:28 my current version will always go berzerk from turn 835 when the turn limit is 1000 2011-11-18T13:28:55 did you hardcode that or did you do 83.5%? 2011-11-18T13:28:56 haha 2011-11-18T13:29:20 turn counter * 1.2 > max turns 2011-11-18T13:30:36 Minthos: that is very clever, you are a scholar and a champion. 2011-11-18T13:30:58 :) 2011-11-18T13:31:18 LOL 2011-11-18T13:31:22 I'm a viking 2011-11-18T13:31:37 I think you lost 74 dudes there in the end 2011-11-18T13:31:41 hard to count 2011-11-18T13:31:47 although my attack strategy blows right now 2011-11-18T13:31:51 :) 2011-11-18T13:31:54 fun to watch 2011-11-18T13:33:16 yeah I need to adjust the ants' behavior in that mode to be just a little bit more tactical 2011-11-18T13:33:33 but they need to be more tactical in general too, so that's higher priority 2011-11-18T13:34:01 If they're sufficiently tactical without it, they don't need it 2011-11-18T13:34:12 *** Antimony has joined #aichallenge 2011-11-18T13:34:30 xathis needs something like it 2011-11-18T13:34:53 his ants are very tactical but a bit cowardly 2011-11-18T13:35:15 usually that's good but not in that game alkon posted 2011-11-18T13:36:27 *** liberforce has left #aichallenge 2011-11-18T13:36:50 xathis just needs game awareness code 2011-11-18T13:36:52 like... 2011-11-18T13:36:58 I own this whole map with my ants 2011-11-18T13:37:01 and I have you cornered 2011-11-18T13:37:02 wow, spectacular #7 servers working for the challenge! 2011-11-18T13:37:18 I'm gonna rock you even if I take losses in battle res 2011-11-18T13:37:25 2.1 games per minute, nice 2011-11-18T13:37:29 his micro strategy is flawless 2011-11-18T13:37:35 *** Redgis has joined #aichallenge 2011-11-18T13:37:48 errrr... pretty close to flawless 2011-11-18T13:38:23 Well, he is number one, but I hope he won't stay there without atleast a new version ;) 2011-11-18T13:39:04 he is #1 from eternity 2011-11-18T13:39:11 its embarrasing 2011-11-18T13:39:16 Maybe he has a secret weapon ready to upload 3 days before finals 2011-11-18T13:39:18 Yeah 2011-11-18T13:39:30 http://aichallenge.org/visualizer.php?game=103392 turn 140, coords 33,78 2011-11-18T13:39:31 ummm I would hope so 2011-11-18T13:39:34 he wrote the bot before the contest started, during the beta 2011-11-18T13:39:36 his ants back into a corner 2011-11-18T13:39:39 i'd hope he's been coding since he started 2011-11-18T13:40:40 i should be coding my bot 2011-11-18T13:40:48 and leave this qt rot in hell 2011-11-18T13:41:08 Minthos: right... so at turn 147 he should have gone east with all of those guys 2011-11-18T13:41:20 but that's looking 3-4 turns ahead 2011-11-18T13:41:21 haha 2011-11-18T13:41:28 we're not programming deep blue 2011-11-18T13:41:41 :) 2011-11-18T13:42:07 my ants use threat-aware pathfinding for retreating 2011-11-18T13:42:11 *** Raimondi has quit IRC (Read error: Connection reset by peer) 2011-11-18T13:42:22 they should be able to avoid situations like that some of the time 2011-11-18T13:42:43 *** Raimondi has joined #aichallenge 2011-11-18T13:43:21 it's not about looking far ahead, it's about realizing when you're outmatched and retreating in a safe direction 2011-11-18T13:43:47 I'm hoping to have safe zones that my ants can retreat to for reinforcements 2011-11-18T13:47:00 Ahhh gotcha minthos 2011-11-18T13:50:52 *** Antimony has quit IRC (Ping timeout: 245 seconds) 2011-11-18T13:53:52 *** Nealefelaen has quit IRC (Quit: WeeChat 0.3.2) 2011-11-18T13:54:01 Nealefelaen: it's probably a good idea to position ant-sickbays near those safe zones 2011-11-18T13:55:22 why dont you provide a retirement plan for your ants too? or maybe life-insurance? 2011-11-18T13:56:51 is anyone counting numbers of killed ants in their bots? :D 2011-11-18T13:57:11 *** bhasker has quit IRC (Quit: bhasker) 2011-11-18T13:58:04 *** GeorgeSebastian has joined #aichallenge 2011-11-18T13:58:04 *** GeorgeSebastian has joined #aichallenge 2011-11-18T14:00:51 *** Jumbala has joined #aichallenge 2011-11-18T14:01:06 *** Antimony has joined #aichallenge 2011-11-18T14:02:14 HaraKiri: no, my laywer said it could be a liability later 2011-11-18T14:02:27 * digger3 is in a foul mood, sorry :) 2011-11-18T14:02:37 *** treeform has quit IRC (Remote host closed the connection) 2011-11-18T14:02:49 this is probably all an ender's game kind of scenario. in reality, we're all controlling ant robots in a war far away 2011-11-18T14:03:09 *** treeform has joined #aichallenge 2011-11-18T14:03:46 Hi, I was just wondering something about resolution.. It says that ants that are more focused on a task will kill less focused ants... Let's say I'm attacking an ant that is focused on gathering food (one on one), am I gong to win or is food gatehring totally unrelated? 2011-11-18T14:03:54 *battle resolution 2011-11-18T14:04:42 Mmm, tactical food drops 2011-11-18T14:05:41 Jumbala: that explanation is simple, intuitive, and wrong 2011-11-18T14:06:21 Jumbala: http://aichallenge.org/specification.php#Battle-Resolution 2011-11-18T14:07:29 HaraKiri: are you on XP? 2011-11-18T14:07:36 yes 2011-11-18T14:07:56 no background no grid and no button colors 2011-11-18T14:08:09 on win7 only button colors are missing 2011-11-18T14:08:25 thats platform independent widget toolkit 2011-11-18T14:08:32 *** Jumbala has quit IRC (Quit: Page closed) 2011-11-18T14:08:47 hmm, so there should be that photo in the background? it isn't :< 2011-11-18T14:08:55 i cant even compile qt under windows in any way 2011-11-18T14:09:00 i wanted a static lib 2011-11-18T14:09:09 my linux qt compilation is flawless 2011-11-18T14:09:11 but there was no picture in the zipfile, that right? 2011-11-18T14:09:18 so im gonna package a linux binary and thats all 2011-11-18T14:09:26 HaraKiri: neednt be 2011-11-18T14:09:32 ok 2011-11-18T14:09:36 it is compiled in the exe 2011-11-18T14:09:41 fancy 2011-11-18T14:09:42 its a "resource" 2011-11-18T14:09:49 and the grid is computed, not static 2011-11-18T14:09:54 and that was missing too 2011-11-18T14:13:54 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-11-18T14:16:25 *** nle85 has joined #aichallenge 2011-11-18T14:18:18 btw a clear button would come handy 2011-11-18T14:18:19 *** Antimony has joined #aichallenge 2011-11-18T14:18:38 because now i feel that compulsory urge to clear the field every time by hand :x 2011-11-18T14:20:02 well, if you draw many ants... 2011-11-18T14:20:07 yeah i should add it 2011-11-18T14:20:52 but i dont think i will be bothered by windoz again 2011-11-18T14:21:10 sure 2011-11-18T14:22:46 wtf? are moves on top of food disallowed? 2011-11-18T14:23:10 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-18T14:23:31 yes, that struck me odd too 2011-11-18T14:23:57 it's true? i'm just inferring it from odd behavior here 2011-11-18T14:24:11 i occasionally get moves disallowed and i see two of my ants just annihilated each other next to food 2011-11-18T14:24:18 which is definitely not supposed to happen 2011-11-18T14:24:20 i had that problem when a food spawned next to my hill, and i couldnt move on it 2011-11-18T14:24:38 that seems like a bug to me 2011-11-18T14:24:45 yeah, it's apparently disallowed for whatever reason 2011-11-18T14:24:56 i don't think i know the exact reasoning 2011-11-18T14:24:59 damn it, well, i better fix that then 2011-11-18T14:26:50 i can't find that in the spec 2011-11-18T14:27:16 oh there it is. blocking. 2011-11-18T14:28:51 *** Antimony has joined #aichallenge 2011-11-18T14:30:38 *** praveen has joined #aichallenge 2011-11-18T14:30:53 what is disalowd? 2011-11-18T14:31:00 moving on top of food 2011-11-18T14:31:00 ramming your ants together? 2011-11-18T14:31:13 oh how can you move on top of food? 2011-11-18T14:31:18 i had an ant that expected to move onto food and another one that expected to take its place 2011-11-18T14:31:28 oh 2011-11-18T14:31:35 instead, the first move was blocked and they collided 2011-11-18T14:32:37 but dont ants collect the food as soon as they are next to it? 2011-11-18T14:33:32 it spawned right next to me the previous turn 2011-11-18T14:34:04 and you could see it? and it was not collected? 2011-11-18T14:34:31 yes, you move, then collect food, then more food spawns 2011-11-18T14:34:43 so if it spawns next to you you have to sit still for a turn 2011-11-18T14:37:22 it 2011-11-18T14:37:30 it's only logical, captain 2011-11-18T14:37:32 *** GeorgeSebastian has quit IRC (Ping timeout: 260 seconds) 2011-11-18T14:39:07 haha don't let the first ant move onto the food 2011-11-18T14:39:15 illegal move 2011-11-18T14:39:26 then the 2nd one wouldn't have moved there because your first guy was there 2011-11-18T14:41:04 does anyone know much about constrainted optimization? 2011-11-18T14:41:43 LP? 2011-11-18T14:45:27 *** Redgis has quit IRC (Read error: Connection reset by peer) 2011-11-18T14:45:29 *** HaraKiri has quit IRC () 2011-11-18T14:46:10 *** Redgis has joined #aichallenge 2011-11-18T14:48:37 it is a bit disturbing how kernel.org was hacked 2011-11-18T14:49:51 damn, my combat logic was improved based on tcp server data against strong bots, and now it's too conservative and loses against non-combat aware bots. :( 2011-11-18T14:50:06 *** GeorgeSebastian has joined #aichallenge 2011-11-18T14:51:23 4 servers died :( 2011-11-18T14:52:14 *** chridi has joined #aichallenge 2011-11-18T14:52:29 hi 2011-11-18T14:53:16 is there a problem with the mail activation? I Signed Up 4, 5 hours ago and still no mail (yes, also not in the spam folder) 2011-11-18T14:58:37 *** Accoun has quit IRC () 2011-11-18T15:00:37 *** arscan has joined #aichallenge 2011-11-18T15:04:46 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-18T15:09:51 *** xathis has quit IRC () 2011-11-18T15:14:06 *** Accoun has joined #aichallenge 2011-11-18T15:14:43 @tcp 2011-11-18T15:14:44 Accoun: tcp could be http://ants.fluxid.pl/howto. 2011-11-18T15:15:54 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-11-18T15:16:33 *** epicmonkey has joined #aichallenge 2011-11-18T15:18:16 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 240 seconds) 2011-11-18T15:21:49 *** mj41 has joined #aichallenge 2011-11-18T15:23:28 *** djr_ has joined #aichallenge 2011-11-18T15:25:14 *** Hammerok has quit IRC (Ping timeout: 265 seconds) 2011-11-18T15:28:18 for long, all my bot does is some exploration and gather food. How can I improve food gathering? 2011-11-18T15:31:04 A* 2011-11-18T15:31:05 BFS 2011-11-18T15:32:32 *** eomar has joined #aichallenge 2011-11-18T15:33:38 *** eomar has quit IRC (Client Quit) 2011-11-18T15:33:51 *** tekking has joined #aichallenge 2011-11-18T15:35:39 BFS am already using. But my ants directly walk into enemy's region and die 2011-11-18T15:35:57 A* - am not using as am not timing out now 2011-11-18T15:41:51 *** Conorach has quit IRC (Ping timeout: 248 seconds) 2011-11-18T15:42:37 *** Conorach has joined #aichallenge 2011-11-18T15:46:16 *** djr_ has quit IRC (Quit: Leaving) 2011-11-18T15:46:56 *** ALplus has joined #aichallenge 2011-11-18T15:47:28 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-11-18T15:47:37 *** BenJackson has joined #aichallenge 2011-11-18T15:48:00 turns out you guys weren't just quiet 2011-11-18T15:48:03 this window has been dead for hours 2011-11-18T15:48:22 *** praveen has quit IRC (Quit: Page closed) 2011-11-18T15:51:39 i guess you undeaded it 2011-11-18T15:52:44 *** Conorach_ has joined #aichallenge 2011-11-18T15:53:36 *** Conorach has quit IRC (Ping timeout: 260 seconds) 2011-11-18T15:55:44 * cuademon gets his shotgun just in case 2011-11-18T15:58:16 send "z R C" to resurrect a dead zombie the turn after you see it die 2011-11-18T15:58:26 er dead ant 2011-11-18T15:59:53 *** delt0r_ has joined #aichallenge 2011-11-18T16:04:38 :) 2011-11-18T16:04:54 Yexo: are you there? 2011-11-18T16:05:08 yes 2011-11-18T16:05:15 can you try me something? 2011-11-18T16:05:17 sure 2011-11-18T16:05:20 have you still got the binary? 2011-11-18T16:05:23 yes 2011-11-18T16:05:33 try to run it with -style 2011-11-18T16:05:44 and style might be plastique 2011-11-18T16:05:50 or windows 2011-11-18T16:06:04 maybe cleanlooks 2011-11-18T16:06:44 it seems after a bit reading that the buttons are not colored because that particular windows style doesnt support it 2011-11-18T16:06:48 all those 3 options give me coloured buttons but none of them give the background image 2011-11-18T16:06:54 yay 2011-11-18T16:07:09 are you using windows xp style by default? 2011-11-18T16:07:21 no, windows 7 with default style 2011-11-18T16:07:21 that blue/green/silver thing? 2011-11-18T16:07:25 ah ok 2011-11-18T16:07:41 on win7 the button text is colored 2011-11-18T16:07:56 and you dont see the background or the grind under win7? 2011-11-18T16:08:01 only for the clicked button 2011-11-18T16:08:06 yes 2011-11-18T16:08:11 and no background image 2011-11-18T16:08:19 that becomes "disabled" and the text becomes colored 2011-11-18T16:08:26 ok, so this solves part of the problem 2011-11-18T16:08:40 for me, win7 showed the background image and the grid 2011-11-18T16:12:02 Yexo: ok, it seems i have to provide a jpeg dll 2011-11-18T16:17:35 *** meatroaf_ has quit IRC (Quit: Page closed) 2011-11-18T16:18:34 *** JorgeB has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-11-18T16:23:20 *** McLeopold has left #aichallenge 2011-11-18T16:26:40 *** Fandekasp has quit IRC (Ping timeout: 240 seconds) 2011-11-18T16:27:27 *** retybok_ has joined #aichallenge 2011-11-18T16:27:38 *** scribble has joined #aichallenge 2011-11-18T16:37:23 *** Fandekasp has joined #aichallenge 2011-11-18T16:53:19 *** ikaros has joined #aichallenge 2011-11-18T16:55:04 *** Fandekasp has quit IRC (Ping timeout: 240 seconds) 2011-11-18T16:55:35 *** okan_ has joined #aichallenge 2011-11-18T16:56:58 *** okan_ has quit IRC (Client Quit) 2011-11-18T16:58:03 *** bergmark has joined #aichallenge 2011-11-18T17:01:20 Yexo: can you try again? https://minus.com/mK97jo7mL 2011-11-18T17:01:26 the buttons should be ok 2011-11-18T17:01:33 and i hope the background too 2011-11-18T17:01:59 that link redirects me to https://minus.com/dbb8QN9FE6MVc1.zip which says "The page does not exist" 2011-11-18T17:02:47 https://min.us/mK97jo7mL 2011-11-18T17:02:59 same problem 2011-11-18T17:03:18 Yexo: i see it too 2011-11-18T17:03:39 *** traMKer has joined #aichallenge 2011-11-18T17:04:21 *** djr_ has joined #aichallenge 2011-11-18T17:05:34 *** Zaphus has joined #aichallenge 2011-11-18T17:07:02 *** Zaphus has quit IRC (Client Quit) 2011-11-18T17:07:22 *** Zaphus has joined #aichallenge 2011-11-18T17:10:04 soo I just realized that at one level my bot plans to have two ants switch positions 2011-11-18T17:10:18 but another layer tries to avoid collisions and ends up making them do something stupid instead of switching 2011-11-18T17:10:28 *** twymer has quit IRC (Quit: leaving) 2011-11-18T17:11:24 *** goffrie has quit IRC (Read error: Operation timed out) 2011-11-18T17:11:38 *** bergmark has quit IRC (Ping timeout: 252 seconds) 2011-11-18T17:13:14 *** goffrie has joined #aichallenge 2011-11-18T17:15:25 *** foRei has quit IRC (Quit: Bye) 2011-11-18T17:16:24 *** Redgis has quit IRC (Ping timeout: 244 seconds) 2011-11-18T17:20:57 *** Antimony has joined #aichallenge 2011-11-18T17:21:40 *** cuademon has quit IRC (Quit: leaving) 2011-11-18T17:28:58 Yexo: http://min.us/mbEAxfOU2 2011-11-18T17:29:36 works fine, including background image :) 2011-11-18T17:29:38 thanks a lot 2011-11-18T17:29:42 thank you 2011-11-18T17:29:51 i changed from jpeg to png 2011-11-18T17:30:10 so you see the grid too right? 2011-11-18T17:30:17 yep 2011-11-18T17:38:25 *** treeform has quit IRC (Read error: Operation timed out) 2011-11-18T17:39:55 *** ztfw has joined #aichallenge 2011-11-18T17:40:55 *** treeform has joined #aichallenge 2011-11-18T17:49:01 *** treeform has quit IRC (Ping timeout: 240 seconds) 2011-11-18T17:51:12 *** nle85 has quit IRC (Ping timeout: 265 seconds) 2011-11-18T17:52:12 *** djr_ has quit IRC (Quit: Leaving) 2011-11-18T17:54:53 *** treeform has joined #aichallenge 2011-11-18T17:58:48 *** LouisMartin has joined #aichallenge 2011-11-18T17:59:47 Gentlemen. 2011-11-18T18:00:57 *** epicmonkey has quit IRC (Ping timeout: 276 seconds) 2011-11-18T18:02:05 Guys, if you had to bet money on a bot, which one would it be? (you can't choose yours) 2011-11-18T18:02:53 and as importantly, why this bot? 2011-11-18T18:04:02 LouisMartin: i'd bet yours 2011-11-18T18:04:18 why? bc i have such a high confidence in you 2011-11-18T18:05:12 and I'd bet yours 2011-11-18T18:05:20 bc I think money is a sin 2011-11-18T18:05:25 ;) 2011-11-18T18:05:49 I really have no bot deserving such appellation, yet 2011-11-18T18:06:41 LouisMartin: i dont find your original question really meaningful 2011-11-18T18:06:46 are you on windows? 2011-11-18T18:06:49 ranking aside, which bot has impressed you the more yet mcstar? could be a bad with a cool feature. 2011-11-18T18:07:01 i dont watch replays 2011-11-18T18:07:10 except when somebody links them in 2011-11-18T18:07:15 here 2011-11-18T18:07:19 so windows? 2011-11-18T18:08:06 The question is which bot do you think as the most odd of winning this contest based on your observations/evaluations/whatever ? 2011-11-18T18:08:39 I run MS-Dos 6.11, why? 2011-11-18T18:08:47 now really 2011-11-18T18:09:34 is this really so hard to answer? 2011-11-18T18:09:46 don't judge, it's stable as hell and the Link text web browser is perfect for my literal pron needs 2011-11-18T18:09:47 you will miss something... 2011-11-18T18:09:53 Yes, Windows 7 2011-11-18T18:09:56 k 2011-11-18T18:10:06 and Ubuntu 2011-11-18T18:10:06 http://min.us/lpE8SN0vq9cGS 2011-11-18T18:10:10 could you try this? 2011-11-18T18:10:14 let me see 2011-11-18T18:10:22 i bundled an app just couple of minutes ago 2011-11-18T18:10:28 useful little tool 2011-11-18T18:11:07 in an emulator? yes, no prob 2011-11-18T18:11:07 brb 2011-11-18T18:11:24 LouisMartin: you dont need an emulator 2011-11-18T18:11:44 and there is a linux version but i want to distribute it with static libs but its not ready 2011-11-18T18:13:10 nothing personal; I just don't run exec like that on my main win7 2011-11-18T18:13:28 booting up the emu (XP is fine?) 2011-11-18T18:13:50 yes 2011-11-18T18:14:06 LouisMartin: its ok, someone stole my nick a 2 days ago 2011-11-18T18:14:17 LouisMartin: you can trust liquid-phynix however 2011-11-18T18:14:26 my github account and my minus account 2011-11-18T18:14:46 stole your nick and did what? 2011-11-18T18:14:57 just spoke here 2011-11-18T18:15:03 but amstan kicked him 2011-11-18T18:16:57 lol ok 2011-11-18T18:18:13 oh :D that's nice mcstar 2011-11-18T18:18:42 I will use it to test some ideas I had 2011-11-18T18:18:46 k 2011-11-18T18:18:51 thanks 2011-11-18T18:18:55 right button erases 2011-11-18T18:19:04 good to know 2011-11-18T18:19:11 thats all basically 2011-11-18T18:19:16 you ca drag and paint 2011-11-18T18:19:19 can* 2011-11-18T18:19:20 so you worked on your battle algo? 2011-11-18T18:19:26 a little 2011-11-18T18:19:31 i did this app yesterday 2011-11-18T18:19:38 put the battle resolution in it 2011-11-18T18:19:47 and today i was trying to build it for windows 2011-11-18T18:19:50 what a waste of time 2011-11-18T18:20:15 theres soon gonna be a linux package if you want that... 2011-11-18T18:21:02 it's never a lost of time if you enjoyed the coding 2011-11-18T18:21:12 plus it can be useful for others 2011-11-18T18:21:22 did you posted it in the forum? 2011-11-18T18:22:36 LouisMartin: i meant trying to build it on windows, that was a major waste of time 2011-11-18T18:23:01 ah, right 2011-11-18T18:23:03 no i didnt post it there, i dont use forums 2011-11-18T18:24:02 if you plan on improving something on it, I'd suggest showing the attack range arround an ant when the mouse is over it 2011-11-18T18:24:17 and a check box to enable seeing all attack range at all time 2011-11-18T18:24:22 that'd be great 2011-11-18T18:24:57 that can be done 2011-11-18T18:25:10 but i want to improve on my bot first 2011-11-18T18:25:38 how's your bot? 2011-11-18T18:26:42 sucks at hill defense 2011-11-18T18:27:14 I see 2011-11-18T18:28:04 btw, your software just showed me I wasn't grasping correctly / completely the battle resolution .. 2011-11-18T18:28:57 which is disturbing, cause I don't put your algo in question. 2011-11-18T18:29:10 i think its correct 2011-11-18T18:29:15 i use that same al in my bot 2011-11-18T18:29:18 alg. 2011-11-18T18:29:35 and ive tested it in situations that were mentioned here 2011-11-18T18:29:37 *** Antimony has quit IRC (Ping timeout: 245 seconds) 2011-11-18T18:30:09 LouisMartin: https://github.com/liquid-phynix/ants-fight/blob/master/main.hh#L68 2011-11-18T18:30:13 here it is 2011-11-18T18:30:30 ok, thanks 2011-11-18T18:42:55 *** dedly has joined #aichallenge 2011-11-18T18:43:51 help 2011-11-18T18:44:30 IM DROWNING 2011-11-18T18:44:55 i cant work out the java tut 2011-11-18T18:45:23 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-18T18:46:23 *** dedly has quit IRC (Client Quit) 2011-11-18T18:54:16 *** analyst74 has quit IRC (Quit: HydraIRC -> http://www.hydrairc.com <- It'll be on slashdot one day...) 2011-11-18T18:55:28 *** fpcfan has quit IRC (Ping timeout: 240 seconds) 2011-11-18T19:03:18 *** sigh has joined #aichallenge 2011-11-18T19:07:22 oh no 2011-11-18T19:07:27 minecraft 1.0 just came out 2011-11-18T19:07:30 *** fpcfan has joined #aichallenge 2011-11-18T19:08:22 a1k0n: i guess your bots done then :P 2011-11-18T19:08:28 see you guys :) 2011-11-18T19:10:28 antimatroid: are you willing to test something for me? 2011-11-18T19:10:44 mcstar: depends on how much effort is involved from me? :P 2011-11-18T19:10:58 almost zero 2011-11-18T19:11:04 ... 2011-11-18T19:11:18 so, do you have qt? 2011-11-18T19:11:24 :D 2011-11-18T19:11:26 don't think so :P 2011-11-18T19:11:31 ok 2011-11-18T19:11:44 https://minus.com/liquid-phynix 2011-11-18T19:11:56 pls download from here the linux bundle with qt 2011-11-18T19:12:33 do i need to have a 64bit processor? 2011-11-18T19:12:41 you dont have? 2011-11-18T19:12:41 i think my cpu is only 32 bit 2011-11-18T19:12:45 nooo 2011-11-18T19:12:51 you machine is 64 2011-11-18T19:13:07 can i check aha? 2011-11-18T19:13:08 you told me some time ago that its better than mine 2011-11-18T19:13:50 uname -p 2011-11-18T19:13:55 checks the processor 2011-11-18T19:14:13 uname -m 2011-11-18T19:14:19 checks the architecture 2011-11-18T19:14:37 *** rogue780 has joined #aichallenge 2011-11-18T19:14:56 how do i give myself admin rights or whatever in the terminal window? 2011-11-18T19:15:05 you dont need to 2011-11-18T19:15:09 sudo 2011-11-18T19:15:30 why do you want to? 2011-11-18T19:15:49 /proc/cpuinfo is meant to give me cpu info 2011-11-18T19:15:53 but it says permission denied 2011-11-18T19:15:58 i'm not very good with linux :P 2011-11-18T19:16:07 just use uname -a and uname -m 2011-11-18T19:16:16 that said unknown 2011-11-18T19:16:34 cpuinfo doesnt tell you directly what does your os/cpu supports 2011-11-18T19:16:36 wait, no it didn't 2011-11-18T19:16:59 http://pastebin.com/FC7bMJLA 2011-11-18T19:17:11 btw i could always cat /proc/cpuinfo asa regular user 2011-11-18T19:17:39 *** chridi has quit IRC (Ping timeout: 276 seconds) 2011-11-18T19:17:45 i686 2011-11-18T19:17:47 it let me do that 2011-11-18T19:17:53 so you are 32 bit 2011-11-18T19:18:00 told you :P sorry 2011-11-18T19:19:44 *** a9wegf9a has joined #aichallenge 2011-11-18T19:19:57 so what do people make of the homework? 2011-11-18T19:20:05 I'm not asking for help/answers/anything else 2011-11-18T19:20:16 a9wegf9a: wrong channel 2011-11-18T19:20:22 poo :P 2011-11-18T19:20:24 ty mcstar 2011-11-18T19:20:26 *** a9wegf9a has left #aichallenge 2011-11-18T19:20:36 aichallenge is all homework 2011-11-18T19:20:52 antimatroid: will you pls install qt? 2011-11-18T19:21:13 and you could compile it directly from github 2011-11-18T19:21:22 and i would like to have that executable 2011-11-18T19:21:32 along with a couple of qt libs 2011-11-18T19:21:43 and i would put that alongside the others 2011-11-18T19:21:55 BenJackson: are you on 64bit? 2011-11-18T19:22:10 I ran your win build on win 7 64 2011-11-18T19:22:13 mcstar: but mehffort! :( 2011-11-18T19:22:23 ah ok 2011-11-18T19:22:30 BenJackson: you are excused 2011-11-18T19:22:33 my server is a 64-bit freebsd machine 2011-11-18T19:22:42 I don't think any of my linuxes or vms are 64 2011-11-18T19:22:49 really? 2011-11-18T19:22:56 i thought its natural by now 2011-11-18T19:23:14 well my EMC2 install is on pretty old hardware 2011-11-18T19:23:26 the rest of the linuxes I ever boot are VMs 2011-11-18T19:24:00 actually the most recent linux install I did was on a PPC (old ibook) :) 2011-11-18T19:24:43 i tried ubuntu on a powerbook ~2 years ago 2011-11-18T19:25:02 was slower than 10.5 2011-11-18T19:25:14 they quit updating OSX for this some time ago 2011-11-18T19:25:19 antimatroid: is that a no? 2011-11-18T19:25:25 and I put a CF in to replace the HD when it died 2011-11-18T19:25:28 its just a couple of commands really 2011-11-18T19:25:31 and OSX couldn't resume after that 2011-11-18T19:25:33 ubuntu can! 2011-11-18T19:25:38 mcstar: how long would it take? i have other things to do that sound more fun :P 2011-11-18T19:25:39 but if youre busy i can udnerstand 2011-11-18T19:25:48 antimatroid: minutes 2011-11-18T19:26:12 hmmmm okay pm me directions if you want 2011-11-18T19:26:18 sudo apt-get install qt or qt-devel 2011-11-18T19:26:34 mcstar: what do you need to test? 2011-11-18T19:26:52 thestinger: i need a 32-bit compiled binary 2011-11-18T19:26:58 with 32bit qt 2011-11-18T19:27:22 use a chroot :P 2011-11-18T19:27:59 *** traMKer has quit IRC (Ping timeout: 248 seconds) 2011-11-18T19:30:20 *** amstan has joined #aichallenge 2011-11-18T19:30:20 *** ChanServ sets mode: +o amstan 2011-11-18T19:30:26 http://ants.fluxid.pl/replay.5036 my combat sort of works sometimes now :) 2011-11-18T19:32:03 impressive final attack 2011-11-18T19:35:57 lol 2011-11-18T19:36:02 actually i have lib32-qt 2011-11-18T19:36:25 I need to divide up the combat stuff into separate battles now 2011-11-18T19:36:52 thestinger: you and me tooo 2011-11-18T19:36:57 was about to do that 2011-11-18T19:37:04 i have a woundrous method in mind 2011-11-18T19:37:18 *** yoden-cloud has joined #aichallenge 2011-11-18T19:39:25 ugh tvtropes won't let me code :( 2011-11-18T19:39:39 no idea how I even ended up there yet again 2011-11-18T19:39:43 heh 2011-11-18T19:39:47 I think the internet wants to kill me 2011-11-18T19:39:52 nice try, I'm not looking 2011-11-18T19:40:42 thestinger: nice, sometimes it works real nice, and sometimes they just suicide 2011-11-18T19:40:53 http://tvtropes.org/pmwiki/pmwiki.php/Main/TVTropesWillRuinYourLife 2011-11-18T19:41:13 if I'm going to waste time it's playing skyrim 2011-11-18T19:41:50 after my experience with oblivion and fallout 3, I've decided to wait a year before getting it xD 2011-11-18T19:42:10 I didn't know anything about it until someone said "it's fallout 3: middle earth" 2011-11-18T19:42:15 then I had to have it 2011-11-18T19:42:49 junkies 2011-11-18T19:43:49 I think I'll just solve my combat problems with more breadth-first searches, everything else seems like it will take more than 5 lines of code 2011-11-18T19:44:24 I keep writing elegant, fast algorithms which don't quite optmize the right thing 2011-11-18T19:44:36 I really need a mind wipe so I can approach the problem anew 2011-11-18T19:45:30 I just wrote code to detect if a map is a maze or not 2011-11-18T19:45:33 do you have hard liquor in mind? 2011-11-18T19:45:51 it seems useful to adjust behavior based on that info 2011-11-18T19:46:40 now we need half-maze maps 2011-11-18T19:46:44 BenJackson: heh. i was thinking the same 2011-11-18T19:46:51 hehe 2011-11-18T19:47:17 *** TheLinker has joined #aichallenge 2011-11-18T19:47:25 hell no, we need all maze maps. I'm good with mazes: http://aichallenge.org/visualizer.php?game=104201&user=3473 2011-11-18T19:47:33 but what if my maze detection really is just an adjustable bias? then it will correctly account for half-maze maps as well 2011-11-18T19:47:54 i'm spectacularly bad at that two-player maze map 2011-11-18T19:48:13 the hills are so far apart that i basically consider them worth 0 until someone comes to attack and then i retaliate 2011-11-18T19:50:45 I can definitely see scaling issues in that game for my code, too 2011-11-18T19:50:49 the distances are much larger than typical 2011-11-18T19:51:33 a1k0n: I actually have a semi-related problem: my bot isn't good at randomly distributing itself 2011-11-18T19:51:44 but if enemies lead it around and food lures it it will get spread out 2011-11-18T19:51:58 but once it gets clumpy it's bad at de-clumping 2011-11-18T19:52:17 that's one of the reason I'm pretty good at multi-hill maze 2011-11-18T20:02:41 i love the maze maps :P 2011-11-18T20:03:33 yeah, I dislike the really open maps 2011-11-18T20:03:53 xathis is good at those, though 2011-11-18T20:03:59 he's good at expanding to fill area 2011-11-18T20:04:08 on maze maps one can't ninja kill your oponent 2011-11-18T20:04:20 antimatroid: well.. i'm at a codepary, and i'll work at more mapgens 2011-11-18T20:04:27 you can on some of the hills on the multi-hill mazes 2011-11-18T20:04:48 a couple of the randomwalk maps have really evil one-wide passes in them which is cool 2011-11-18T20:04:57 multi hills are annoying 2011-11-18T20:05:09 my bot kind of does ok on multi hills 2011-11-18T20:05:22 i guess its more defensive 2011-11-18T20:05:48 the server is so confused about my bot's skill that it matched me up against pguillory 2011-11-18T20:05:56 needless to say, it got owned 2011-11-18T20:06:06 happens. i got xathis. 2011-11-18T20:06:10 i cant break out of 62 2011-11-18T20:08:00 a1k0n: have you stored paths between turns? 2011-11-18T20:08:03 ...and now i just got suppressed by MomoBot and teapotahedron in my 7th game 2011-11-18T20:08:06 antimatroid: no 2011-11-18T20:08:16 i'm finding that really hard to do with the random tie break on spawning ants at hills 2011-11-18T20:09:25 *** datachomper has joined #aichallenge 2011-11-18T20:10:01 I broke into the top 20 but I think I'd have to do something seriously different to get into the top 10 2011-11-18T20:11:42 i think i might just store paths for existing ants and recollect food for possible future ants each turn 2011-11-18T20:12:08 my first bot (in python) stored paths because it couldn't compute them fast enough each turn 2011-11-18T20:12:28 my food collection etc. is quite complicated 2011-11-18T20:12:38 and i want almost all the time each turn dedicated to battles 2011-11-18T20:13:00 if i could execute spawn orders my life would be a million times easier :( 2011-11-18T20:13:08 hmph 2011-11-18T20:13:09 I do a BFS from a food (marking off nearby food as part of a single cluster) and then basically do what the tutorial does 2011-11-18T20:13:34 except with distances/pathfinding from the breadth-first searches 2011-11-18T20:13:51 i iteratively collect uncollected food, reusing ants that have already moved to collect food or an ant that is expected to spawn 2011-11-18T20:13:56 @seen j3camero 2011-11-18T20:13:56 amstan: j3camero was last seen in #aichallenge 3 weeks, 6 days, 5 hours, and 20 seconds ago: If you are finding that some of the testing tools suck, please consider making better ones. We generally release whatever tools at the beginning, but the contestants always end up making better ones. 2011-11-18T20:14:06 using multi source and multi target a* to find a shortest path 2011-11-18T20:15:01 but knowing where ants are going to spawn has become retardedly hard since the addition of ant hills and the random tie break rule 2011-11-18T20:17:07 *** Antimony has joined #aichallenge 2011-11-18T20:18:19 gn 2011-11-18T20:18:20 *** mcstar has quit IRC (Quit: WeeChat 0.3.6) 2011-11-18T20:24:50 *** antimatroid has quit IRC (Ping timeout: 260 seconds) 2011-11-18T20:26:38 aarossig: http://paste.aichallenge.org/6WYDM/ 2011-11-18T20:26:50 well.. they don't look like swastikas yet 2011-11-18T20:26:53 which is good 2011-11-18T20:32:07 anyone wanna work on this with me? 2011-11-18T20:32:14 implement maze maps? 2011-11-18T20:33:28 aichallenge: Alexandru Stan amstan_mapgen * r89e5678 / ants/mapgen/symmetricmap.py : added proper rotational to maps, renamed other symmetries - http://git.io/WooGvw 2011-11-18T20:39:38 amstan: talking to me? 2011-11-18T20:39:44 maybe 2011-11-18T20:39:55 amstan: python right? 2011-11-18T20:41:06 ya 2011-11-18T20:49:09 *** Anilm3 has quit IRC (Quit: Page closed) 2011-11-18T20:53:25 *** antimatroid has joined #aichallenge 2011-11-18T20:54:27 *** flag has joined #aichallenge 2011-11-18T20:54:46 baahhh, main dev computer crashed :( 2011-11-18T20:55:51 that sucks, but naturally you backup to a remote storage frequently 2011-11-18T20:57:38 Yeah, I still have my bot, but I'm running it off my laptop right not and it's pretty slow 2011-11-18T20:58:02 Which is annoying because I intended to work on it a lot this weekend :( 2011-11-18T20:59:27 *** ztfw has quit IRC (Read error: Connection reset by peer) 2011-11-18T20:59:29 *** grwip has quit IRC (Quit: Leaving) 2011-11-18T21:06:22 *** besh has joined #aichallenge 2011-11-18T21:06:39 *** Zaphus has quit IRC (Quit: Page closed) 2011-11-18T21:08:20 * antimatroid doesn't back up his bot 2011-11-18T21:08:36 i'm terrible at backing stuff up, although i did have my thesis stuff dropboxed, wasn't risking that one 2011-11-18T21:09:08 testing 2011-11-18T21:09:09 flag: are you storing paths between turns? 2011-11-18T21:09:15 or is anyone here doing that? 2011-11-18T21:10:05 antimatroid: I used to do that, but I assume something can change every turn so I don't do it anymore 2011-11-18T21:10:27 flag: if it wasn't for the random tie break with spawning ants i think i could do waht i want fairly easily 2011-11-18T21:10:29 It's still in my code though, so I can activate it 2011-11-18T21:10:36 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-18T21:10:42 but that makes tracking where ants are going to spawn impossible, you need to make assumptions 2011-11-18T21:10:46 *** amstan has quit IRC (Read error: Connection reset by peer) 2011-11-18T21:11:06 If you have mutliple hills you can keep an ant on your hills to direct ants 2011-11-18T21:11:07 *** amstan has joined #aichallenge 2011-11-18T21:11:07 *** ChanServ sets mode: +o amstan 2011-11-18T21:11:14 yeah but you can't always ensure you can do that 2011-11-18T21:11:27 so you still need to handle the case where all hills aren't controlled and a random tie break might occur 2011-11-18T21:11:50 Is your code that specific? 2011-11-18T21:11:59 i want it to be 2011-11-18T21:12:00 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-11-18T21:12:18 i want to iteratively collect uncollected food, reusing ants that have already collected some food and ants that will spawn from collected food 2011-11-18T21:12:32 and i want to store these paths between turns so i have maximal amounts of time for battle code 2011-11-18T21:12:48 Hi. Does anyone know the algorithm top bots use for maximizing view ? 2011-11-18T21:12:52 because doing that from scratch each turn is quite slow 2011-11-18T21:13:10 I find that the time I take for pathfinding to food is relatively negligable 2011-11-18T21:13:20 it wasn't for me during beta 2011-11-18T21:13:27 what are you doing for food collection? 2011-11-18T21:13:38 Heavily-optimized bfs 2011-11-18T21:13:45 yes, but what is your bfs method? 2011-11-18T21:13:50 are you iteratively collecting food? 2011-11-18T21:13:52 reusing ants? 2011-11-18T21:13:58 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2011-11-18T21:14:23 I calculate the distances from every ant to every food, chose the shortest ant-food pairs and do bfs from that 2011-11-18T21:14:30 yeah i don't like that 2011-11-18T21:14:32 :P 2011-11-18T21:14:36 :) 2011-11-18T21:14:37 it's too simply and suboptimal 2011-11-18T21:14:40 simple* 2011-11-18T21:14:45 it hardly takes any time though 2011-11-18T21:14:58 yeah, my other way would be sweet if i stored paths between turns :P 2011-11-18T21:15:01 It doesn't account for the case where food_count > ant_count 2011-11-18T21:15:01 with ~300 ants and lots of food on the map it still takes under a few ms 2011-11-18T21:15:15 Or at least, isn't optimal for it 2011-11-18T21:15:20 But it's not too bad 2011-11-18T21:15:29 flag: do you reuse a single ant to gather a few food that are near each other? 2011-11-18T21:15:32 antimatroid: When am I going to see you on tcp? 2011-11-18T21:15:45 thestinger: depends on a few things, but sometimes yes 2011-11-18T21:15:54 flag: just set up a "base bot" that has no strategy yesterday 2011-11-18T21:16:01 although now i'm stuck on what to do about tracking spawning ants 2011-11-18T21:16:03 it's so complicated 2011-11-18T21:16:10 *** Conorach_ has quit IRC (Ping timeout: 260 seconds) 2011-11-18T21:16:27 just track all ants 2011-11-18T21:16:36 and when a new ant arrives, check if it's on a hill 2011-11-18T21:16:39 *** Antimony has joined #aichallenge 2011-11-18T21:16:42 if it is, assume it spawned 2011-11-18T21:16:44 Minthos: but you can't track where ants are going to spawn 2011-11-18T21:16:50 i want to use ants i plan to have in the future :P 2011-11-18T21:16:55 no but screw that 2011-11-18T21:17:05 you're optimizing in the wrong end 2011-11-18T21:17:15 it's a requirement for what i want to do 2011-11-18T21:17:16 antimatroid wants to solve ants before he begins writing his bot :) 2011-11-18T21:17:19 otherwise it all falls to pieces 2011-11-18T21:17:47 if i don't track future ants storing paths between turns is going to lead to really suboptimal collecting 2011-11-18T21:17:59 it worked awesome when ants spawned on food 2011-11-18T21:18:20 plus i just recalculated everything each turn back then 2011-11-18T21:18:22 but like i said, slwo 2011-11-18T21:18:27 I've decided I'll just divide up combat zones based on attackradius+1 of enemies 2011-11-18T21:18:39 i'm going to use game trees for battle 2011-11-18T21:18:40 your implementation is slow then 2011-11-18T21:18:47 it's not my implementation 2011-11-18T21:18:49 http://paste.pocoo.org/raw/5b7azNt2Zv7yt5m4bX1J/ so like that... 2011-11-18T21:18:49 it's the method 2011-11-18T21:19:05 well your method, your implementation whatever it is, you're the one doing it 2011-11-18T21:19:16 my gathering is fast 2011-11-18T21:19:20 and works well 2011-11-18T21:19:22 the method is slow, but the method is also awesome 2011-11-18T21:19:31 it doesn't just work well, it works awesomely 2011-11-18T21:19:47 eh 2011-11-18T21:20:04 no big deal when the map is full of your ants anyway 2011-11-18T21:20:26 paths from ant to food will always be short 2011-11-18T21:20:56 what about when the map isn't full of ants? :P 2011-11-18T21:20:56 not much room for inefficiency as long as you find the shortest path and don't spend too many cycles doing it 2011-11-18T21:20:58 *** yoden-cloud has quit IRC (Ping timeout: 265 seconds) 2011-11-18T21:21:06 can anyone give me hints on how to place ants to maximize view? That is after map is fully explored I think. Most top bots seem to have that and i have no clue what it could be. 2011-11-18T21:21:09 then you don't see any food either, so no problem 2011-11-18T21:21:15 you need ants to find food 2011-11-18T21:21:24 you need food to get ants 2011-11-18T21:21:25 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-18T21:22:06 But to Minthos' point, when the map isn't full of ants it doesn't really matter if your algo is suboptimal 2011-11-18T21:22:13 besh: one option is to continue "collecting not recently seen squares" 2011-11-18T21:22:30 or you can iteratively move ants so they're furthest away from their closest friend 2011-11-18T21:22:48 possible ordering which ants you move by their average distance or shortest distance to a friend or something? 2011-11-18T21:23:57 i don't mark squares as "visible", i mark them as "lastSeen" 2011-11-18T21:24:08 antimatroid: Thanks for the ideas. I tried a static division of the map in to 40 x 40 and send 1 ant per corner. But it didn't work very well. 2011-11-18T21:24:19 I just send some of my idle ants to the closest point where I don't have vision 2011-11-18T21:24:27 also using many-to-many bfs 2011-11-18T21:25:01 if you're collecting "not recently seen squares", one thing i found useful during beta was to "bomb vision information" around each location collected, so you didn't send lots of ants to the same patch of non-visible squares 2011-11-18T21:25:03 it's not a complete solution, but it seems to work fairly well 2011-11-18T21:25:29 yeah a* is out for exploring 2011-11-18T21:25:46 *** datachomper has joined #aichallenge 2011-11-18T21:25:54 but if your number of uncollected food squares is not too large then i think a* is still worthwhile 2011-11-18T21:26:37 a1k0n: did you ever do anything about symmetry detection? 2011-11-18T21:26:46 or has anyone? 2011-11-18T21:27:53 *** datachomper has quit IRC (Client Quit) 2011-11-18T21:27:57 *** treeform has quit IRC (Remote host closed the connection) 2011-11-18T21:28:24 uncollected food is usually very low, but I treat enemies near my hives as food as well so it could potentially grow quite big 2011-11-18T21:28:36 antimatroid: Yes, I also replaced the isVisible() array of the java starter package to return when it was last seen. More useful. 2011-11-18T21:28:55 and besides, it's one of several bfses I do each round, I don't see any need to make it faster 2011-11-18T21:28:58 Minthos: i plan to "protect and invade" hives before collecting food 2011-11-18T21:29:06 and run battles first up 2011-11-18T21:30:10 i think i'm going to have to just recalculate every turn 2011-11-18T21:30:14 i'll see how long it takes :( 2011-11-18T21:30:58 "hey look, there's lots and lots of food over there!" "actually I was kidding it isn't actually food it's enemies. Your nation is proud of your sacrifice!" 2011-11-18T21:32:01 lol 2011-11-18T21:32:14 my bfs was much faster when I didn't take different cost for each tile into account 2011-11-18T21:32:27 then I could run one bfs for each ant 2011-11-18T21:33:08 now I don't use it for individual pathfinding, only for groups 2011-11-18T21:33:27 and in practice, that's more than fast enough 2011-11-18T21:34:24 it turns out there aren't that many use cases where you need lots of criss-crossing paths 2011-11-18T21:40:50 *** besh has quit IRC (Quit: Page closed) 2011-11-18T21:42:26 *** datachomper has joined #aichallenge 2011-11-18T21:44:40 *** datachomper has quit IRC (Client Quit) 2011-11-18T21:48:45 *** Cyndre has quit IRC (Read error: Connection reset by peer) 2011-11-18T21:50:31 *** Cyndre has joined #aichallenge 2011-11-18T21:51:00 *** jab_bott has joined #aichallenge 2011-11-18T21:53:07 latest cavemap iteration is here: http://paste.aichallenge.org/Mr5hJ/ 2011-11-18T21:54:15 ugh - I've got a pull request in github into my vis_overlay branch and I want to merget it to my local copy so I can fix some conflicts and merge in epsilon, but I cant find the smegging URL to merge - where is it? 2011-11-18T21:55:49 jab_bott: you'll have to explain better 2011-11-18T21:57:08 amstan: I need to do: git remote add pooka url_of_pull_request <= where do I find this URL in github? 2011-11-18T21:57:20 jab_bott: main page of a git repo 2011-11-18T21:57:24 also.. 2011-11-18T21:57:50 there's a merge button in the pull request 2011-11-18T21:58:19 but that's the url of the repo - doesn't the pull request have its own url that only includes those commits that were included in the pull request and not others that may have been added since? 2011-11-18T21:58:20 *** emiel_ has quit IRC (Read error: Connection reset by peer) 2011-11-18T21:58:34 *** emiel_ has joined #aichallenge 2011-11-18T21:59:17 amstan: I don't want to merge it remotely as I need to also merge in epsilon from the main repo and fix a conflict between the merges 2011-11-18T21:59:18 no 2011-11-18T21:59:22 you can't just pull like that 2011-11-18T21:59:35 you need to add their repo as a remote 2011-11-18T21:59:41 git fetch their_repo 2011-11-18T21:59:58 git checkout your_branch_where_you_want_to_merge_into 2011-11-18T22:00:01 git merge their_repo/their_branch 2011-11-18T22:00:05 ahh ok, so I just add the whole repo then cherry-pick the commits from the pull request - yeah I'll do that 2011-11-18T22:00:07 git push 2011-11-18T22:00:13 or that 2011-11-18T22:00:40 cheers - still getting to grips with git and github 2011-11-18T22:00:41 they should have had the pull changes in a separate branch if they wanted to do other magic as well 2011-11-18T22:02:42 *** Harpyon has quit IRC (Quit: Computer has gone to sleep.) 2011-11-18T22:03:20 amstan: oh, they probably didn't make any other changes - I'm just overly anal 2011-11-18T22:03:31 jab_bott: then just merge 2011-11-18T22:03:36 cherry picks look ugly in the history 2011-11-18T22:03:39 they have 2 authors 2011-11-18T22:04:16 won't the cherry-picks look exactly the same - since it's the same commits? 2011-11-18T22:06:38 *** rogue780 has quit IRC (Quit: This computer has gone to sleep) 2011-11-18T22:07:25 nope, cherry picks add your name to the log(as commiter, and his as author) and it'll change the hash 2011-11-18T22:07:28 amstan, fair enough? http://forums.aichallenge.org/viewtopic.php?f=26&t=1872 2011-11-18T22:08:13 LouisMartin: yeah, get a github pull request/issue, with the change already made(if possible) and ping janzert about it 2011-11-18T22:09:04 amstan: cool 2011-11-18T22:11:37 aichallenge: Alexandru Stan amstan_mapgen * re9a989c / (5 files): added default terrain for terrain.py, added hill functions in map.py, added space around hills, cavemap carver now adds land instead of water - http://git.io/oZsvLA 2011-11-18T22:16:29 lol, I've caused some weird movement bugs while implementing combat >.< 2011-11-18T22:17:44 http://paste.aichallenge.org/BSKrh/ my ants are now willing to take some losses which is an improvement (1 dead in this scenario near the end) 2011-11-18T22:25:44 *** Antimony has joined #aichallenge 2011-11-18T22:29:45 thestinger: did you worked on the pathfinding algo we talked about? 2011-11-18T22:31:24 are you sure it was me you talked to? I don't remember that xD 2011-11-18T22:32:50 the stellation algo? 2011-11-18T22:33:02 that was someone else :P 2011-11-18T22:33:14 ahh lol 2011-11-18T22:33:18 alright 2011-11-18T22:33:40 I'm just using BFS to make distance maps, with a larger increment than 1 when going from a tile with an ant to another tile 2011-11-18T22:33:57 btw, what is that map and the "lazybot" standard in the package? 2011-11-18T22:34:09 I just took hunterbot and deleted all the logic 2011-11-18T22:34:29 the map is the tutorial map with some ants pre-added (you can use --scenario to start with ants/food on the map) 2011-11-18T22:35:14 okay, I see. I'll check that. thanks 2011-11-18T22:35:41 http://paste.pocoo.org/show/510064/ that's how you do it basically (that's the map file) 2011-11-18T22:36:09 I guess you must know "focus battle resolution" well by now. can you take a look at this? http://i.imgur.com/8aXEi.png 2011-11-18T22:36:55 I basically just took the python code from that page and wrote a simulate() function in C++ 2011-11-18T22:37:04 the top-left would be equal to the top-right in terms of who dies? (the result at the bottom for the top-right one) 2011-11-18T22:38:04 cool. I'm stealing it :) 2011-11-18T22:39:33 aarossig: http://paste.aichallenge.org/qxXVx/ 2011-11-18T22:40:14 amstan: looksgoodman.jpg 2011-11-18T22:40:15 *** mj41 has quit IRC (Ping timeout: 244 seconds) 2011-11-18T22:40:18 LouisMartin: I think all 3 ants die in that top left one 2011-11-18T22:40:21 I could be wrong 2011-11-18T22:40:48 well, all 3 on both sides 2011-11-18T22:41:05 yes, makes sense to me 2011-11-18T22:41:32 How do people calculate manhattan distances on the map? 2011-11-18T22:41:39 because you can't just take the diff between points 2011-11-18T22:41:45 I don't anymore 2011-11-18T22:41:50 but the top-right scenario would have the same 6 ants to die even if blue is overpowering the 2 top red ants that much? 2011-11-18T22:41:57 thestinger: how do you pathfind then? gradients? 2011-11-18T22:42:08 BFS out from a target, marking distances in a 2D array 2011-11-18T22:42:23 can do a multi-source BFS for multiple targets, start them off at something other than 0, etc. 2011-11-18T22:42:38 wel, i need to pathfind between bases to check if they're ocnnected 2011-11-18T22:42:39 that way I get "real" (assuming unexplored areas are land) distances, and pathfinding 2011-11-18T22:43:14 i was hoping to do a* so it's fast 2011-11-18T22:43:14 ((rowa - rowb + map.rows) % map.rows) + same thing for cols 2011-11-18T22:43:48 Minthos: that gives you lowest diff? 2011-11-18T22:44:19 if my math is correct yes, but it's 4.44 am here so it may not be 2011-11-18T22:44:45 amstan: http://paste.pocoo.org/show/tDVWZJ2uKmcfM6dZ7LoQ/ I was doing that when I still used manhattan distance 2011-11-18T22:44:49 i don't think that'll work 2011-11-18T22:44:56 what if i have a point at 1 and mapsize-1 2011-11-18T22:45:03 i should get something like 2 for diff 2011-11-18T22:45:12 yeah 2011-11-18T22:45:12 but i'll get the long way 2011-11-18T22:45:15 no 2011-11-18T22:45:23 % is modulo 2011-11-18T22:45:35 rowa=1, rowb=9, rows=10 2011-11-18T22:45:36 well, % is remainder is C/C++ :P 2011-11-18T22:45:40 in* 2011-11-18T22:45:46 1-9+10=2 2011-11-18T22:45:47 fine 2011-11-18T22:45:50 what about the other way? 2011-11-18T22:46:01 9-1+10=18%10=8 2011-11-18T22:46:09 not good 2011-11-18T22:46:16 remainder, modulo same difference when you make sure it's always positive by adding the map dimension before moduloing 2011-11-18T22:46:30 thestinger: min you say! good idea 2011-11-18T22:46:40 y = Y2 - Y1 % Map.Height; if (y < 0) y = y + Map.Height; and same for X 2011-11-18T22:47:42 LouisMartin: i'm pretty sure my python % is always positive 2011-11-18T22:47:47 LouisMartin: but that's not good, i need min 2011-11-18T22:50:54 yeah, python's % is mod 2011-11-18T22:50:55 *** NightExcessive has quit IRC (Ping timeout: 255 seconds) 2011-11-18T22:51:01 % in C is remainder 2011-11-18T22:51:14 Prelude> 123 `mod` (-5) 2011-11-18T22:51:16 -2 2011-11-18T22:51:18 Prelude> 123 `rem` (-5) 2011-11-18T22:51:20 3 2011-11-18T22:51:32 yay for haskell being sane :P 2011-11-18T22:55:45 amstan: http://i.imgur.com/8aXEi.png can you confirm or infirm that the right scenario has the same 6 ants killed than the left one? (bottom = result for the right scenario) 2011-11-18T22:56:14 LouisMartin: no idea, try play_game.py --scenario 2011-11-18T22:56:31 ok 2011-11-18T22:56:49 play_game is the ultimate test, unless it has a bug 2011-11-18T22:56:58 but play_game is used on the servers, so.. 2011-11-18T22:57:07 *** TwistedLogic has joined #aichallenge 2011-11-18T22:59:57 *** TwistedLogic has quit IRC (Client Quit) 2011-11-18T23:00:33 *** AntDroid has joined #aichallenge 2011-11-18T23:00:47 *** GeorgeSebastian has quit IRC (Ping timeout: 248 seconds) 2011-11-18T23:02:58 *** rogue780 has joined #aichallenge 2011-11-18T23:06:45 LouisMartin: I think in Windows you can press alt+printscrn to take an image of the current active window to the clipboard (but then I haven't used windows for years so it could have changed) 2011-11-18T23:07:22 amstan: I fixed my problem with canvas invalidation thanks to mleise messages he left me :) 2011-11-18T23:07:38 jab_bott: good to know, thanks 2011-11-18T23:08:05 amstan: is it @tell msg to leave msgs? 2011-11-18T23:08:14 later tell 2011-11-18T23:09:15 @later tell mleise Thanks - your messages helped. I've fixed the problem now by adding a cached value of aistatePlayer in the canvas and checking it in checkState 2011-11-18T23:09:16 jab_bott: Aye, aye, sir 2011-11-18T23:09:25 amstan: thanks :) 2011-11-18T23:12:06 [18:17] if i don't track future ants storing paths between turns is going to lead to really suboptimal collecting 2011-11-18T23:12:17 unless your code that selects the moves is stable 2011-11-18T23:12:36 that is that any ant that moves toward a goal is even more likely to be picked for that goal next time 2011-11-18T23:12:43 that's why people have had success with CD 2011-11-18T23:13:41 what's CD? 2011-11-18T23:14:07 collaborative diffusion 2011-11-18T23:15:37 *** JorgeB has joined #aichallenge 2011-11-18T23:17:27 you mean influence maps 2011-11-18T23:17:33 at least call it by the right name 2011-11-18T23:18:11 seen BenJackson 2011-11-18T23:18:17 Antimony: Diffusion would be an algorithm running on Influence Map 2011-11-18T23:18:19 the forum people started calling it CD 2011-11-18T23:18:27 *** Ionic_Groove has quit IRC (Read error: No route to host) 2011-11-18T23:18:31 at least if I recall my AI book well 2011-11-18T23:18:36 that's because it was all started by someone linking to a papaer that called it that 2011-11-18T23:18:47 because the paper was written by idiots who had never heard of influence maps 2011-11-18T23:18:55 and thought they discovered something new 2011-11-18T23:19:14 I invented something similar for this challenge before I saw that post too 2011-11-18T23:19:23 I just don't have a reason to publish papers ;-) 2011-11-18T23:19:37 I'm having a lot of trouble with my cosntraint solver 2011-11-18T23:19:46 I rewrote it to use UCS/A* but it's too slow 2011-11-18T23:19:57 it has to expand 19k nodes to prove it found the optimal solution 2011-11-18T23:19:58 Antimony: remember that outside gaming, you often *have* to invent new inane names so you can avoid linking too much to "AI" 2011-11-18T23:20:19 I don't think I understand how my bot works anymore >.< 2011-11-18T23:20:26 thestinger: but does it work? 2011-11-18T23:20:30 sometimes 2011-11-18T23:20:50 http://ants.fluxid.pl/player/strcat_combatzones this new branch seems to work quite well, but it doesn't do what I thought I coded 2011-11-18T23:21:09 I like how you are implementing all of your ideas and trying them out 2011-11-18T23:21:19 I am too lazy so I don't bother as soon as I think of a fatal flaw 2011-11-18T23:21:35 but I might make more progress if I implemented the half-baked stuff 2011-11-18T23:22:08 I'm in the fatal flaw stage right now 2011-11-18T23:22:22 unless I can figure out a way to propogate soft cosntraints, it will never be fast enough 2011-11-18T23:22:42 the branch I started for combat has turned into a complete rewrite 2011-11-18T23:22:59 yeah, most of my work has been in coming up with clever data structures that let me search the solution space quickly 2011-11-18T23:23:00 *** TheLinker has quit IRC (Quit: Bye) 2011-11-18T23:23:24 http://paste.pocoo.org/show/510067/ diff to master branch from my combat branch o_o 2011-11-18T23:23:50 the problem is that my current heuristic is too optimistic and so I need to expand 19k nodes 2011-11-18T23:23:51 *** BenJackson_ has joined #aichallenge 2011-11-18T23:23:57 wtf webchat 2011-11-18T23:24:04 well I missed anything you just said 2011-11-18T23:24:05 just to solve 13 ants 2011-11-18T23:24:18 and that's without even considering combat 2011-11-18T23:24:45 BenJackson_: http://paste.pocoo.org/show/sSoINQoN0CuIxCgI3DMC/ <- what you missed 2011-11-18T23:25:07 well, I dunno how much you actually missed but it's in there :P 2011-11-18T23:25:27 *** Scryer has joined #aichallenge 2011-11-18T23:25:32 thestinger: are you finding scenario maps worthwhile? 2011-11-18T23:25:40 *** LouisMartin has quit IRC (Ping timeout: 265 seconds) 2011-11-18T23:25:40 *** arscan has quit IRC (Ping timeout: 265 seconds) 2011-11-18T23:25:52 BenJackson_: yeah, definitely 2011-11-18T23:26:09 *** jab_bott has quit IRC (Ping timeout: 265 seconds) 2011-11-18T23:26:09 *** goldcaddy77 has quit IRC (Ping timeout: 265 seconds) 2011-11-18T23:26:12 I also look at a lot of tcp replays 2011-11-18T23:26:37 but I can't just switch around code and redo the situations in the tcp replay 2011-11-18T23:26:38 *** imp5imp5 has quit IRC (Ping timeout: 265 seconds) 2011-11-18T23:26:38 *** praveeen has quit IRC (Ping timeout: 265 seconds) 2011-11-18T23:27:07 *** AntDroid has quit IRC (Ping timeout: 265 seconds) 2011-11-18T23:27:12 debugging with scenarios is a lot easier 2011-11-18T23:27:36 *** BenJackson has quit IRC (Ping timeout: 265 seconds) 2011-11-18T23:27:36 *** jab_bott has joined #aichallenge 2011-11-18T23:28:03 Scenario maps? Is that where you can set up a specific situation to test against? 2011-11-18T23:28:09 yeah 2011-11-18T23:28:37 I like my old code vs my new code with a preset situation 2011-11-18T23:28:41 Sounds very useful. 2011-11-18T23:28:42 I like testing* 2011-11-18T23:29:04 *** BenJackson_ is now known as BenJacksno 2011-11-18T23:29:07 *** BenJacksno is now known as BenJackson 2011-11-18T23:29:19 when I'm looking for something specific running a bunch of games and looking through them to find something like that is a waste 2011-11-18T23:29:55 I get situations like "OK, why did I forget that enemy hill was there and go into my far-back backup random mode?" Not sure how to set that up with a specific map. 2011-11-18T23:30:29 well, you can make a scenario with ants/water/hills placed in the same locations as when you had the problem 2011-11-18T23:30:31 thestinger: I see you have 0.6 more skill than me on aichallenge now 2011-11-18T23:30:47 :), might just be a lucky streak 2011-11-18T23:30:55 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-18T23:31:07 http://aichallenge.org/profile.php?user=2255 my pre-combat version is somehow rank 15 2011-11-18T23:31:11 I've been resisting "tweaking" mine because I think I need to get out of a rut 2011-11-18T23:31:12 seriously, no awareness of enemy ants o_o 2011-11-18T23:31:23 probably closer to the end I'll resign myself to trying to notch a few more positions 2011-11-18T23:31:32 really? wow 2011-11-18T23:31:41 actually the version on the site might pathfind a bit different based on enemy ants, I forget exactly how old it is 2011-11-18T23:31:47 definitely no combat though 2011-11-18T23:32:10 I really can't tell if it's just getting lucky or doing something right 2011-11-18T23:32:54 hey I hadn't watched this one: 2011-11-18T23:32:55 http://aichallenge.org/visualizer.php?game=103264&user=2255 2011-11-18T23:32:57 gang's all here 2011-11-18T23:33:06 in that one I'd say you were lucky! 2011-11-18T23:33:17 I've been making a git tag for each version I upload so I can actually check and see which revision that is 2011-11-18T23:33:18 seeing as how you got your points after losing your hill :) 2011-11-18T23:34:05 yeah, that was before combat and proper foraging xD 2011-11-18T23:35:03 I'm surprised at how well my v7 is doing considering some major bugs I found in the combat code 2011-11-18T23:36:03 *** Antimony has joined #aichallenge 2011-11-18T23:36:06 the version on the site does have some primitive avoidance of enemies, but it actually made my bot worse than it was based on tcp testing (and all that old code is deleted now) 2011-11-18T23:36:29 So what are scenario maps and how do you get them? 2011-11-18T23:36:30 yeah one of my early hacks to just avoid suiciding into things I didn't *mean* to suicide into helped 2011-11-18T23:36:34 *** rogue780 has quit IRC (Quit: Leaving) 2011-11-18T23:36:46 Antimony: you just edit a map file (or make one manually) 2011-11-18T23:36:47 but real combat superceded that 2011-11-18T23:37:24 Antimony: if you run play_game with --scenario, it uses pre-placed food/ants 2011-11-18T23:37:43 example: http://pastie.org/2886770 2011-11-18T23:37:48 oh 2011-11-18T23:37:55 that would be very helpful 2011-11-18T23:37:57 how do I run the standalone visualizer? 2011-11-18T23:40:52 that's quite a challenge for player 0 2011-11-18T23:41:15 if your map is asymmetric like that one you need --food none I think 2011-11-18T23:44:04 yeah, I use --food none with that one 2011-11-18T23:44:16 ok, this will be an interesting challenge 2011-11-18T23:44:21 http://paste.aichallenge.org/BSKrh/ 2011-11-18T23:44:33 I'll see if I can get an updated bot of mine to beat a prior version as blue 2011-11-18T23:44:41 I have some weird movement code bugs I need to fix 2011-11-18T23:45:22 one I just found was that my low-level code prevented ants trading positions 2011-11-18T23:45:26 but the combat code would do that sometimes 2011-11-18T23:45:36 so the low-level code would fuck up the combat 2011-11-18T23:45:49 if blue is actually a sane bot that will be pretty impossible xD 2011-11-18T23:46:04 well just to prove I've fixed my bugs 2011-11-18T23:47:53 *** delt0r_ has quit IRC (Ping timeout: 252 seconds) 2011-11-18T23:48:02 *** jab_bott has quit IRC (Quit: Page closed) 2011-11-18T23:50:45 my ants will trade positions, but they prefer going around the other ant if it doesn't make them go farther from the target 2011-11-18T23:50:51 I'm not really sure what's going on with this bug 2011-11-18T23:51:02 I caused it by deleting old combat heuristics 2011-11-18T23:51:49 *** AntDroid has joined #aichallenge 2011-11-18T23:55:06 *** datachomper has joined #aichallenge 2011-11-18T23:59:01 *** AntDroid has quit IRC (Ping timeout: 265 seconds)