2010-09-21T00:00:08 Zarathust: That python engine is available in source form from the Google Code project. 2010-09-21T00:00:50 ai-contest source: http://code.google.com/p/ai-contest 2010-09-21T00:00:54 *** rogue780 has quit IRC (Quit: Leaving) 2010-09-21T00:04:31 bss03: for enemy planets, isn't that approximately the current number of ships on the planet? 2010-09-21T00:05:04 sigh: Right. 2010-09-21T00:05:30 sigh: I'm not currently considering, reinforments that are not yet in the air. 2010-09-21T00:05:38 then it doesn't really care about distance 2010-09-21T00:05:52 sigh: I do consider reinforcing flees that are knows. 2010-09-21T00:06:01 sigh: I also consider the growth of the planet. 2010-09-21T00:06:04 does your bot often attack enemies far away in preference to closer targets? 2010-09-21T00:07:12 No, it likes closer targets. It will "just" continue a drawn-out contest for a hostile planet when it would be better off doing more expansion onto neutral planets. 2010-09-21T00:07:36 thanks sigh 2010-09-21T00:07:38 hmm... ok, I'm probably misunderstanding something then 2010-09-21T00:07:43 carry on :) 2010-09-21T00:09:11 sigh: I build a Map from [0..n] -> GameState, then run a fitness function against ALL the planets, with the "time" being part of the input. 2010-09-21T00:09:40 *** naglijo_ has quit IRC (Ping timeout: 240 seconds) 2010-09-21T00:09:50 Then I filter based on ship availablity, ship need, and sort. 2010-09-21T00:10:27 Right now "n" is maxTransitTime -- the maximum distance between any two planets. 2010-09-21T00:10:46 yeah, I get that, I'm just a bit confused by your evaluation function... but I don't have time to think about that right now 2010-09-21T00:11:53 sigh: Why not? You are in the top 5, right? 2010-09-21T00:11:54 @rank 2010-09-21T00:11:54 bss03: Error: The command "rank" is available in the Alias, ChannelStats, and Factoids plugins. Please specify the plugin whose command you wish to call by using its name as a command before "rank". 2010-09-21T00:11:55 *** Olathe has quit IRC (Ping timeout: 240 seconds) 2010-09-21T00:11:59 @alias rank 2010-09-21T00:12:00 bss03: Top 10 players: McLeopold(3512), kcm1700(3506), dmj111(3434), sighbot(3372), EBraun(3372), adumlah(3369), petesimard(3363), montanalow(3354), paulus(3335), Tracer(3290) 2010-09-21T00:12:10 bss03: I'm at work :P 2010-09-21T00:12:23 ok i've got the svn checkout 2010-09-21T00:12:25 sigh: Ah, carry on, then. :) 2010-09-21T00:12:38 how should i start engine? 2010-09-21T00:13:09 Zarathust: I don't know. Maybe there are some docs? I'm just using PlayGame.jar here. 2010-09-21T00:13:30 Ok 2010-09-21T00:13:33 Zarathust: I wrote enough of an engine to do my predictive analysis based on the spec. in Haskell. 2010-09-21T00:13:46 nice 2010-09-21T00:13:56 Zarathust: So, I don't really know my way around the Java or Phyton source, sorry. :( 2010-09-21T00:14:27 I'm in C# 2010-09-21T00:14:49 i crash after a few dozens of turns with very minimal code 2010-09-21T00:14:55 I wished to see the source to debug better 2010-09-21T00:15:30 Zarathust: I think the suggestion for getting extra debug information is to write stuff to a file. 2010-09-21T00:15:57 Zarathust: PlayGame.jar seems to swallow stdout and stderr from your program, but opening a file and logging to that should work. 2010-09-21T00:16:01 Yes i already do this 2010-09-21T00:16:14 But stepping through the engine could be better 2010-09-21T00:16:30 *** naglijo has joined #aichallenge 2010-09-21T00:16:47 with java you can backtrack some executions 2010-09-21T00:17:11 this competition is really starting to take its toll on me 2010-09-21T00:17:20 i'm so damn tired today, and i even made it to bed by 1am last night 2010-09-21T00:17:57 antimatroid1: Walk away from it for a few days. You'll probably some back with better ideas anyway. >:) 2010-09-21T00:18:25 antimatroid1: Let it soak into your subconscious. 2010-09-21T00:19:59 i have ideas to implement currently but have to take a few days off 2010-09-21T00:20:09 i have quite a few uni assignments to do this week and a test tomorrow 2010-09-21T00:28:42 @tcp-server 2010-09-21T00:28:43 bss03: tcp-server = http://www.benzedrine.cx/planetwars/ 2010-09-21T00:28:50 there's something about coding for cache locality that just feels warm and fuzzy 2010-09-21T00:29:02 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-21T00:31:17 jmcarthur: my code make CPU warm and fuzzy... 2010-09-21T00:31:22 heh 2010-09-21T00:31:23 *makes 2010-09-21T00:31:33 *** kozlovsky has joined #aichallenge 2010-09-21T00:31:42 my code is god awful 2010-09-21T00:31:48 but it's much cleaner than last time 2010-09-21T00:33:03 my guy still seems to suck on some maps 2010-09-21T00:33:15 i really need to address that problem, some use being best on some 2010-09-21T00:33:23 no use* 2010-09-21T00:35:57 *** Waco has quit IRC (Quit: ( www.nnscript.com :: NoNameScript 4.22 :: www.esnation.com )) 2010-09-21T00:37:44 *** otherAntimatroid has joined #aichallenge 2010-09-21T00:37:56 *** jesionaj has joined #aichallenge 2010-09-21T00:37:59 *** jesionaj has left #aichallenge 2010-09-21T00:38:05 *** jesionaj has joined #aichallenge 2010-09-21T00:40:29 *** Utkarsh has quit IRC (Ping timeout: 276 seconds) 2010-09-21T00:41:20 *** John__ has quit IRC (Ping timeout: 252 seconds) 2010-09-21T00:43:06 *** pjreddie has joined #aichallenge 2010-09-21T00:45:02 *** john__ has joined #aichallenge 2010-09-21T00:45:17 is something going on with the server? i submitted at like 10 this morning, no emails, no games, nothing... 2010-09-21T00:48:15 *** kozlovsky has quit IRC (Quit: Page closed) 2010-09-21T00:50:15 *** kozlovsky has joined #aichallenge 2010-09-21T00:51:07 server is not very good 2010-09-21T00:56:27 Top 10 players: McLeopold(3497), kcm1700(3489), dmj111(3422), sighbot(3373), EBraun(3361), adumlah(3353), petesimard(3351), montanalow(3341), paulus(3311), Tracer(3271) 2010-09-21T00:59:51 *** mightybyte has left #aichallenge 2010-09-21T01:10:40 *** hellman has joined #aichallenge 2010-09-21T01:12:44 *** Snowplan has left #aichallenge 2010-09-21T01:19:21 drewrobb: lost that top spot aye ;) 2010-09-21T01:21:16 arggg 2010-09-21T01:21:23 these matches are so brutal 2010-09-21T01:21:58 i really need to fix my bot up to win when it can 2010-09-21T01:22:18 *** icky has joined #aichallenge 2010-09-21T01:22:46 its not like your bot loses when it should have won 2010-09-21T01:24:04 no, i think i have that part going alright 2010-09-21T01:24:16 i just need to fix up a few combat things and work on my start 2010-09-21T01:24:29 then it's time to try algorithms to generate better moves 2010-09-21T01:25:13 awesome, i didn't realize that the arrow keys controlled the canvas visualizer 2010-09-21T01:25:28 jmcarthur: thanks 2010-09-21T01:25:33 time to try search for me 2010-09-21T01:25:40 oh main, does it fast forward too 2010-09-21T01:25:41 best 2010-09-21T01:25:44 yeah 2010-09-21T01:25:52 it goes fast if you hold it down 2010-09-21T01:26:05 hmm not working for me 2010-09-21T01:26:13 jmcarthur: I changed mine so that I can skip to a move by clicking on the chart 2010-09-21T01:26:14 it's just right and left 2010-09-21T01:26:17 really handy 2010-09-21T01:26:21 sigh: nice 2010-09-21T01:26:28 drewrobb: make sure you click on the canvas? 2010-09-21T01:26:37 i don't have to click on it 2010-09-21T01:26:53 my friend is trying to fix up the c++ one a big for testing with 2010-09-21T01:26:58 bit* 2010-09-21T01:27:21 yeah, i'm clicking, holding down doesn't work 2010-09-21T01:27:29 antimatroid1: yeah, I'm essentially doing the same with the canvas one... displaying planet ids and such 2010-09-21T01:27:32 i have a really weird system so that is probably why 2010-09-21T01:28:09 hope to turn it into a display with a heap of useful info 2010-09-21T01:28:31 *** hornairs has quit IRC (Quit: hornairs) 2010-09-21T01:34:28 antimatroid1: weird how you don't have much trouble with deepblue 2010-09-21T01:36:57 deepblue is easy :P 2010-09-21T01:37:35 rwa's top ranking is a sham 2010-09-21T01:38:16 if he doesn't play, he can't lose! 2010-09-21T01:38:57 i should have done that when my elo was 800+ :P 2010-09-21T01:42:55 its a bit silly to get excited about doing well this early 2010-09-21T01:43:13 but it is the only thing that keeps me programming i suppose 2010-09-21T01:45:40 *** Palmik has joined #aichallenge 2010-09-21T01:45:42 yeah, i was starting to lose interest when people were owning me last week 2010-09-21T01:45:56 doing well keeps one motivated 2010-09-21T01:50:00 oh man, this is just sad 2010-09-21T01:50:03 http://72.44.46.68/canvas?game_id=71843 2010-09-21T01:50:48 note how locutus takes the 99 0 growth center on turn 0 2010-09-21T01:51:01 ouch! 2010-09-21T01:51:46 he then proceeds to waste ships on neutrals in your area 2010-09-21T01:51:53 you seem to defend too much? 2010-09-21T01:52:06 he is exploiting a weird bug in my code 2010-09-21T01:52:38 my bot ignores 0 growth planets, and only attacks the nearest enemy under certain conditions 2010-09-21T01:52:59 sneaky sneaky 2010-09-21T01:53:32 ah, so because the 0 growth was the closest planet? 2010-09-21T01:53:36 i have had troubles with predicting the closest neighbour and algorithms based of that information for similar reasons 2010-09-21T01:53:44 closest enemy* 2010-09-21T01:53:57 algorithms == heuristics 2010-09-21T01:54:47 this is just a hilarious corner case 2010-09-21T01:56:27 Top 10 players: McLeopold(3495), kcm1700(3486), dmj111(3395), sighbot(3368), EBraun(3363), adumlah(3356), montanalow(3332), paulus(3311), petesimard(3284), Tracer(3263) 2010-09-21T02:05:26 *** Tempest has joined #aichallenge 2010-09-21T02:08:06 *** hellman has quit IRC (Remote host closed the connection) 2010-09-21T02:10:39 *** phreeza has joined #aichallenge 2010-09-21T02:10:53 drewrobb: current versions seem pretty well matched 2010-09-21T02:11:40 *** amstan has quit IRC (Ping timeout: 240 seconds) 2010-09-21T02:13:40 Does anyone know how a fleet's ID is assigned? 2010-09-21T02:14:28 indeed 2010-09-21T02:14:36 maybe time to release 1.4 2010-09-21T02:14:42 *** Appleman1234 has joined #aichallenge 2010-09-21T02:17:09 Tempest: unless you are talking about a starter pack, fleets don't inherently have IDs 2010-09-21T02:17:24 i'm not familiar enough with the starters to answer if that's that you are talking about 2010-09-21T02:17:34 *if that's what 2010-09-21T02:21:24 Tempest: Fleets don't have IDs, per se. Their index in the Fleets() array may change as fleets arrive. 2010-09-21T02:21:40 does change... 2010-09-21T02:22:10 since the order of arrival is not the same as the order of dispatch 2010-09-21T02:23:28 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-09-21T02:23:35 *** Appleman1234 has quit IRC (Ping timeout: 265 seconds) 2010-09-21T02:29:14 *** phreeza has joined #aichallenge 2010-09-21T02:30:01 Thank you; the variable 'fleet_id' in PlanetWars.GetFleet() seems to be misnamed 2010-09-21T02:30:54 in the starter pack 2010-09-21T02:31:52 now that my forum login is fixed, I can't log in anymore to the main site. 2010-09-21T02:34:19 *** aerique has joined #aichallenge 2010-09-21T02:34:28 *** boegel has joined #aichallenge 2010-09-21T02:34:45 *** boegel has joined #aichallenge 2010-09-21T02:34:59 hiya y'all 2010-09-21T02:35:18 I just made a new submission, and received two email concerning the compilation 2010-09-21T02:35:40 the first one said it failed ("collect2: ld terminated with signal 11 [Segmentation fault]"), the second said it succeeded 2010-09-21T02:35:59 is that expected? does the system try to compile the submission again when it fails? 2010-09-21T02:36:10 that happeend to me once and the submission seemed to work fine 2010-09-21T02:36:15 i'd leave it a little bit and see what happens 2010-09-21T02:36:47 *** Baphomet has quit IRC (Ping timeout: 276 seconds) 2010-09-21T02:37:08 hmm, first game played seems like it's playing the old version 2010-09-21T02:37:28 I can easily tell, because my old bot didn't work at all, this is my first somewhat decent submission 2010-09-21T02:38:08 resubmitted, just to be sure :) 2010-09-21T02:38:44 how long does it take for the system to pick up the new version? 2010-09-21T02:39:12 should be pretty much straight away, but it's hardly giving out games atm 2010-09-21T02:39:17 is it possible that it was playing with the older version during the submission, and that it'll take those odl results into account? 2010-09-21T02:39:18 http://www.benzedrine.cx/planetwars/ try that 2010-09-21T02:39:32 it shouldn't play your old versions ever 2010-09-21T02:39:39 and your elo is pretty much wiped everytime you resubmit 2010-09-21T02:40:00 antimatroid1: hmm, cool 2010-09-21T02:40:09 antimatroid1: hmmk, I see 2010-09-21T02:40:25 *** Baphomet has joined #aichallenge 2010-09-21T02:41:20 in this game, the system is definitely using my old bot: http://ai-contest.com/visualizer.php?game_id=4615883 2010-09-21T02:41:29 and it was played after my most recent submission 2010-09-21T02:41:43 antimatroid1: is it possible a game was in play *during* my submission? 2010-09-21T02:42:04 haha, that would be a terrible race 2010-09-21T02:42:05 because the game result came in just 12s after my submission 2010-09-21T02:42:17 Zannick: yeah, my old bot was utter crap, full of bugs 2010-09-21T02:42:19 i wonder if it does... 2010-09-21T02:42:26 no, i mean race condition 2010-09-21T02:42:42 boegel: it would wipe any old games when you resubmit (or at least that's waht it's done for me) 2010-09-21T02:43:01 Zannick: ah, ok 2010-09-21T02:43:31 well, the version it's playing doesn't match with what I'm seeing locally, so my best guess is that it's playind the old version 2010-09-21T02:44:33 well, the 12s thing is also a good indicator it might be 2010-09-21T02:45:52 Zannick: I thought so too 2010-09-21T02:46:16 antimatroid1: the benzedrine version allows you to play as many games as you want in a short period of time? 2010-09-21T02:46:28 antimatroid1: and thus get a quick estimate of the quality of your bot? 2010-09-21T02:47:31 antimatroid1: how does it know if you're playing a new version? the ranking is cumulative, I assume, no matter which version you're using? 2010-09-21T02:47:48 *** aerique has quit IRC (Quit: ...) 2010-09-21T02:49:57 *** bss03 is now known as bss03-AFK 2010-09-21T02:54:09 boegel: you specify a username when running your bot, so if you want a new score, pick a new name 2010-09-21T02:54:19 hence why you see people using different version numbers on tehre 2010-09-21T02:55:40 *** perror has joined #aichallenge 2010-09-21T02:55:45 boegel: E.g. I use git to develop my bot. I use git tag each time I submit to the official server. I generate my tcp-server username using git describe. 2010-09-21T02:56:28 Top 10 players: McLeopold(3507), kcm1700(3496), dmj111(3390), sighbot(3366), EBraun(3363), adumlah(3358), montanalow(3319), paulus(3300), petesimard(3282), Tracer(3262) 2010-09-21T02:56:41 My tags are DaTwinkDaddy-s#. So, my username on the TCP server is DaTwinkDaddy-s#--. 2010-09-21T03:06:29 bss03-AFK: thanks, that's good 2010-09-21T03:06:49 do you also check that everything is committed? 2010-09-21T03:07:19 *** sinan has joined #aichallenge 2010-09-21T03:18:12 antimatroid1: I see 2010-09-21T03:19:03 antimatroid1: you can just pick a password, and it'll be registred the first time you use a name or something? 2010-09-21T03:22:23 *** Tempest has quit IRC (Quit: Page closed) 2010-09-21T03:26:43 *** kozlovsky has quit IRC (Ping timeout: 252 seconds) 2010-09-21T03:32:04 *** aerique has joined #aichallenge 2010-09-21T03:35:17 *** sigh has quit IRC (Remote host closed the connection) 2010-09-21T03:39:58 *** pc has joined #aichallenge 2010-09-21T03:40:20 *** pc is now known as xvinyl 2010-09-21T03:40:21 *** mega1 has quit IRC (Read error: Connection reset by peer) 2010-09-21T03:40:22 *** mega1_ has joined #aichallenge 2010-09-21T03:40:42 *** dufflebunk has joined #aichallenge 2010-09-21T03:41:16 I'm getting a failure to compile my (Haskell) bot on the server, but the email just lists some compile warnings. 2010-09-21T03:41:27 tmp/ghc18909_0/ghc18909_0.s: Assembler messages: 2010-09-21T03:41:28 /tmp/ghc18909_0/ghc18909_0.s:15882:0: Warning: 0000000757b12c00 shortened to 0000000057b12c00 2010-09-21T03:41:40 HAs anyone has anything like that before? 2010-09-21T03:43:13 *** delt0r_ has joined #aichallenge 2010-09-21T03:43:38 dufflebunk: using Haskell too? brave, like me :P 2010-09-21T03:43:58 More interesting than C or Java. 2010-09-21T03:44:47 *** delt0r___ has quit IRC (Ping timeout: 265 seconds) 2010-09-21T03:48:25 *** [1]adesst <[1]adesst!~adesst@180.214.233.19> has joined #aichallenge 2010-09-21T03:50:17 <[1]adesst> rrr... i've modified my bot, and still lossing the game.. :( 2010-09-21T03:51:14 *** pc has joined #aichallenge 2010-09-21T03:51:53 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T03:52:19 <[1]adesst> is there anyway that i could use ShowGame.jar to read an txt file, lets say the var data="blabla" of the website result 2010-09-21T03:52:44 <[1]adesst> i am using python package 2010-09-21T03:52:47 <[1]adesst> tq 2010-09-21T03:56:25 *** john__ has quit IRC (Ping timeout: 252 seconds) 2010-09-21T03:56:29 Top 10 players: McLeopold(3510), kcm1700(3495), dmj111(3450), EBraun(3368), sighbot(3354), adumlah(3344), montanalow(3315), petesimard(3307), Tracer(3305), paulus(3300) 2010-09-21T03:57:30 *** mega1_ has quit IRC (Quit: No Ping reply in 180 seconds.) 2010-09-21T03:57:34 *** mega1 has joined #aichallenge 2010-09-21T03:58:25 *** pc is now known as xvinyl 2010-09-21T03:58:50 *** Utkarsh has joined #aichallenge 2010-09-21T04:01:09 *** naglijo has quit IRC (Ping timeout: 252 seconds) 2010-09-21T04:03:25 *** [2]adesst <[2]adesst!~adesst@180.214.233.7> has joined #aichallenge 2010-09-21T04:03:57 *** [1]adesst <[1]adesst!~adesst@180.214.233.19> has quit IRC (Ping timeout: 276 seconds) 2010-09-21T04:04:13 *** jaspervdj has quit IRC (Quit: leaving) 2010-09-21T04:04:31 *** jaspervdj has joined #aichallenge 2010-09-21T04:04:55 Should a compile warning cause the compilation to fail? 2010-09-21T04:09:54 *** mog_ has joined #aichallenge 2010-09-21T04:11:12 *** mega1 has quit IRC (Quit: No Ping reply in 180 seconds.) 2010-09-21T04:12:38 *** mega1 has joined #aichallenge 2010-09-21T04:14:00 *** Naktibalda has quit IRC (Remote host closed the connection) 2010-09-21T04:15:49 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T04:16:04 *** xvinyl has joined #aichallenge 2010-09-21T04:35:45 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-09-21T04:36:48 *** [2]adesst <[2]adesst!~adesst@180.214.233.7> has quit IRC (Ping timeout: 264 seconds) 2010-09-21T04:37:28 *** Naith has quit IRC (Ping timeout: 265 seconds) 2010-09-21T04:38:24 *** sigh has joined #aichallenge 2010-09-21T04:40:12 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T04:40:39 *** xvinyl has joined #aichallenge 2010-09-21T04:41:38 *** mega1 has quit IRC (Write error: Broken pipe) 2010-09-21T04:41:45 *** mega1 has joined #aichallenge 2010-09-21T04:42:10 *** Utkarsh has joined #aichallenge 2010-09-21T04:44:51 *** z-247 has quit IRC (Quit: leaving) 2010-09-21T04:45:11 *** mog_ has quit IRC (Ping timeout: 252 seconds) 2010-09-21T04:45:30 *** bss03-AFK has quit IRC (Read error: Connection reset by peer) 2010-09-21T04:46:01 *** bss03-AFK has joined #aichallenge 2010-09-21T04:47:35 aerique: I have a submission that would print some useful info about the server, but my login stopped working. 2010-09-21T04:47:37 it's my starter package with a small mod in MyBot.lisp 2010-09-21T04:47:38 if you would be so kind as to submit it ... 2010-09-21T04:47:50 *** Naktibalda has joined #aichallenge 2010-09-21T04:48:51 *** mega1 has quit IRC (Quit: No Ping reply in 180 seconds.) 2010-09-21T04:48:55 *** mega1 has joined #aichallenge 2010-09-21T04:49:52 or if any admin sees this and they fix my login that would be even better 2010-09-21T04:52:40 never mind, I'm in 2010-09-21T04:53:02 the password firefox remembered became somehow out of sync with reality 2010-09-21T04:54:06 *** Gorovkhaia2 has joined #aichallenge 2010-09-21T04:56:17 *** sinan has quit IRC (Ping timeout: 240 seconds) 2010-09-21T04:56:29 Top 10 players: McLeopold(3508), kcm1700(3494), dmj111(3445), EBraun(3380), sighbot(3351), adumlah(3343), montanalow(3310), paulus(3302), Tracer(3301), petesimard(3280) 2010-09-21T04:58:09 *** Appleman1234 has joined #aichallenge 2010-09-21T05:07:14 mega1: i never get mails about compilation errors :-| 2010-09-21T05:07:33 aerique: I do 2010-09-21T05:07:36 and it works :-) 2010-09-21T05:07:44 will post the updated starter package 2010-09-21T05:07:53 mega1: what was the problem? 2010-09-21T05:08:21 I think the problem was what you found. 2010-09-21T05:08:52 in my pack another problem was that compiler output was redirected to stderr 2010-09-21T05:09:03 to which the compile daemon didn't take kindly. 2010-09-21T05:10:04 aerique: ah, ofcourse that's probably a problem in my starter pack as well then 2010-09-21T05:10:31 * aerique wonders why he's talking to himself 2010-09-21T05:10:51 it's a good way to maintain focus 2010-09-21T05:16:27 *** lemonedo has joined #aichallenge 2010-09-21T05:18:56 mega1: so your starter pack is playing games now? 2010-09-21T05:19:01 yes 2010-09-21T05:19:06 http://ai-contest.com/profile.php?user_id=8565 2010-09-21T05:19:13 it's your dummy bot 2010-09-21T05:19:25 *** MrCeeJ has joined #aichallenge 2010-09-21T05:19:28 as you could tell, surely 2010-09-21T05:21:05 mega1: great, finally :) 2010-09-21T05:21:20 yeah, took ages 2010-09-21T05:22:04 mega1: shall we stay with 1.0.11 then or push our luck and ask them to install 1.0.42? 2010-09-21T05:22:16 ask them to install 2010-09-21T05:22:33 mega1: oh wait, they're considering allowing binaries so that would make the point moot 2010-09-21T05:22:35 I don't have the time now to prepare a binary for them 2010-09-21T05:22:57 that would make sense 2010-09-21T05:23:54 *** smellyhippy has quit IRC (Ping timeout: 276 seconds) 2010-09-21T05:27:14 *** lemonedo has quit IRC (Remote host closed the connection) 2010-09-21T05:32:41 *** Naith has joined #aichallenge 2010-09-21T05:46:20 oh aerique is here \o/ 2010-09-21T05:47:04 What would you recommend for getting started with lisp? Your package? Anwyn's? Are there o thers? 2010-09-21T05:53:03 *** sinan has joined #aichallenge 2010-09-21T05:54:31 cYmen: at the moment I would use mega1's, he just got his package working 2010-09-21T05:55:02 I'm not sure how that follows. :) 2010-09-21T05:55:12 Here is something that barely works. Use that. 2010-09-21T05:55:25 mega1: no offense.. :) 2010-09-21T05:55:33 well, use mine then which doesn't work ;-) 2010-09-21T05:55:36 the other packages do not work at all 2010-09-21T05:55:52 ah I see :) 2010-09-21T05:56:03 other than that mine is a what I consider to be a cleanup of aerique's 2010-09-21T05:56:03 *** sinan has quit IRC (Client Quit) 2010-09-21T05:56:27 cYmen: he also placed 11th in the Tron compo while I was in the 100's so you might want to follow his example 8-D 2010-09-21T05:56:30 Top 10 players: McLeopold(3490), kcm1700(3480), dmj111(3448), EBraun(3363), sighbot(3341), petesimard(3330), adumlah(3326), Tracer(3284), paulus(3279), montanalow(3270) 2010-09-21T05:57:49 Just got mine working as well: http://ai-contest.com/visualizer.php?game_id=4617780http://ai-contest.com/visualizer.php?game_id=4617780 2010-09-21T05:58:05 ah, broken mouse button 2010-09-21T05:58:44 oh you're that mega1 :) 2010-09-21T05:58:46 mega1: i muffled compiler notes and style warnings 2010-09-21T05:59:32 cYmen: which "that"? From #lisp? yes 2010-09-21T05:59:58 aerique: do they go to stderr? 2010-09-21T06:01:02 mega1: apparently 2010-09-21T06:01:21 bah 2010-09-21T06:01:30 mega1: (declaim (sb-ext:muffle-conditions sb-ext:compiler-note style-warning)) 2010-09-21T06:01:53 at the top of MyBot.lisp 2010-09-21T06:02:09 okay 2010-09-21T06:04:08 aerique: the exit code of sbcl is 1 on error (when --disable-debugger) 2010-09-21T06:04:28 mega1: that's good no? 2010-09-21T06:04:34 so stderr can be simply redirected to stdout to see all the notes 2010-09-21T06:04:49 (which I'd get in the email from the server) 2010-09-21T06:05:13 I have to check if the server actually checks the exit code. 2010-09-21T06:05:22 *** hakand has joined #aichallenge 2010-09-21T06:05:41 mega1: my impressions is it simply checks for output on stderr but i'd have to look at the code 2010-09-21T06:06:04 perhaps Janzert knows :) 2010-09-21T06:07:28 def check_path(path): 2010-09-21T06:07:28 if not os.path.exists(path): 2010-09-21T06:07:28 return "Failure: output file " + str(path) + " was not created\n" 2010-09-21T06:07:44 and it's called with "MyBot" 2010-09-21T06:07:59 so no, it doesn't check the exit code but we are still good 2010-09-21T06:26:00 *** boegel has quit IRC (Quit: Leaving) 2010-09-21T06:30:14 *** kozlovsky has joined #aichallenge 2010-09-21T06:31:16 *** Macuyiko has joined #aichallenge 2010-09-21T06:32:32 *** hellman has joined #aichallenge 2010-09-21T06:36:08 *** icky has quit IRC (Ping timeout: 252 seconds) 2010-09-21T06:40:45 *** Naith has quit IRC (Ping timeout: 255 seconds) 2010-09-21T06:45:04 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-09-21T06:49:13 *** Naith has joined #aichallenge 2010-09-21T06:51:44 *** Utkarsh has joined #aichallenge 2010-09-21T06:53:44 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-21T06:56:30 Top 10 players: kcm1700(3495), McLeopold(3472), dmj111(3434), EBraun(3395), petesimard(3328), adumlah(3326), sighbot(3325), paulus(3312), montanalow(3265), Tracer(3253) 2010-09-21T06:58:28 *** tobiassjosten has joined #aichallenge 2010-09-21T07:09:41 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-09-21T07:10:25 *** dhartmei has joined #aichallenge 2010-09-21T07:15:14 *** Utkarsh has joined #aichallenge 2010-09-21T07:20:31 *** kozlovsky has quit IRC (Quit: Page closed) 2010-09-21T07:30:55 *** Sean_McEligot has joined #aichallenge 2010-09-21T07:42:06 *** hakand has quit IRC (Remote host closed the connection) 2010-09-21T07:43:18 *** hakand has joined #aichallenge 2010-09-21T07:44:10 getting addicted to terminal (: 2010-09-21T07:54:46 *** delt0r___ has joined #aichallenge 2010-09-21T07:56:15 *** delt0r_ has quit IRC (Ping timeout: 265 seconds) 2010-09-21T07:56:31 Top 10 players: kcm1700(3497), McLeopold(3481), dmj111(3442), EBraun(3415), adumlah(3352), sighbot(3330), paulus(3313), petesimard(3306), Tracer(3277), Loks1982(3262) 2010-09-21T07:58:14 *** hakand has quit IRC (Quit: leaving) 2010-09-21T08:01:59 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-09-21T08:03:50 *** hakand has joined #aichallenge 2010-09-21T08:04:29 you people are awfully quiet tonight 2010-09-21T08:05:22 working on their secret strategies (: 2010-09-21T08:05:50 i don't have time for that tonight ): 2010-09-21T08:06:36 and its 3pm in here, working :/ 2010-09-21T08:06:51 it's 10pm here :P 2010-09-21T08:08:20 *** Utkarsh has joined #aichallenge 2010-09-21T08:08:22 1pm :) 2010-09-21T08:11:57 mouse=ultime weapon of productivity destruction 2010-09-21T08:12:46 *** hakand is now known as HakanD 2010-09-21T08:13:16 *** ermau has joined #aichallenge 2010-09-21T08:17:36 *** ermau is now known as ermau|wor 2010-09-21T08:17:40 *** ermau|wor is now known as ermau|work 2010-09-21T08:17:54 *** HakanD has quit IRC (Quit: leaving) 2010-09-21T08:18:19 *** HakanD has joined #aichallenge 2010-09-21T08:18:32 *** sinan has joined #aichallenge 2010-09-21T08:23:19 *** otherAntimatroid has left #aichallenge 2010-09-21T08:23:30 *** amstan has joined #aichallenge 2010-09-21T08:23:30 *** ChanServ sets mode: +o amstan 2010-09-21T08:26:00 mega1: seems that declaim of mine doesn't work on 1.0.11 2010-09-21T08:27:16 *** artyom_shegeda has joined #aichallenge 2010-09-21T08:28:48 aerique: I ended up with this: 2010-09-21T08:28:50 http://quotenil.com/git/?p=planet-wars.git;a=blobdiff;f=MyBot.lisp;h=3322ddf5e02d837c5fef2b0c35de75819ad739f3;hp=5657cba30408aaa77b02f71fa76d72018bc09cda;hb=b51c33399126d4ba83f611a45ac00b3abf88586d;hpb=ad96df470ac84cc021010a5b806bc070b5a3ed82 2010-09-21T08:29:31 bah, pasting the whole file would be shorter than that url 2010-09-21T08:30:37 mega1: no needed, already grabbed it from your git earlier :) 2010-09-21T08:30:54 cool 2010-09-21T08:31:44 *** HakanD has quit IRC (Quit: leaving) 2010-09-21T08:32:37 *** HakanD has joined #aichallenge 2010-09-21T08:35:29 hmmm, with so much variation of bots on different maps, would it work better if each time bots were paired, they had a match off on a few maps to decide which one was the winner/loser out of them? 2010-09-21T08:36:16 *** ratatata has joined #aichallenge 2010-09-21T08:36:52 implement suitable strategies for different kind of maps? (: 2010-09-21T08:38:29 i have categorized maps based on distance of neutrals, enemy home etc, and choosing different strategies for each 2010-09-21T08:39:12 i am, i'm just pondering ways of better ranking bots 2010-09-21T08:40:21 yep, sth needs to be done about ranking 2010-09-21T08:41:15 at least for final round 2010-09-21T08:41:38 problem is you may always end up with some cyclic kind of preferences 2010-09-21T08:41:47 which makes it hard to give a perfect ranking 2010-09-21T08:42:01 but i still think it could be improved upon what it's doing currently 2010-09-21T08:46:20 haha, twitter fail again (: 2010-09-21T08:47:48 some serious hacking going on 2010-09-21T08:47:54 these guys will never learn about security 2010-09-21T08:48:47 *** artyom_shegeda has quit IRC (Quit: Page closed) 2010-09-21T08:48:55 how so? 2010-09-21T08:49:50 via XSS, with their new css feature 2010-09-21T08:51:10 http://mashable.com/2010/09/21/twitter-mouseover-bug/ 2010-09-21T08:52:28 how unfortunate 2010-09-21T08:52:33 they need to sanitize more 2010-09-21T08:52:54 Poor TwitterWhale... a hard work today :P https://twitter.com/twltterwhale 2010-09-21T08:53:18 they have only 1 input area, with no protection for xss 2010-09-21T08:54:13 great way of developing world-class app (: 2010-09-21T08:55:11 bah, dont even use twitter.com for now 2010-09-21T08:55:34 even for trying, as this is a big vulnerability 2010-09-21T08:56:32 Top 10 players: kcm1700(3493), McLeopold(3472), dmj111(3450), EBraun(3374), adumlah(3340), paulus(3303), sighbot(3298), petesimard(3298), montanalow(3272), Tracer(3255) 2010-09-21T08:56:39 well, must admit it is funny (: 2010-09-21T08:56:49 hmz 2010-09-21T08:56:57 is someone claiming to be me on benzedrine ? 2010-09-21T08:57:08 hahaha, that's hilarious 2010-09-21T08:57:13 haha 2010-09-21T08:57:19 what's the name 2010-09-21T08:57:29 bartwe.11 2010-09-21T08:58:03 that's awesome 2010-09-21T08:58:03 you ai expanded and took control :P 2010-09-21T08:58:13 har 2010-09-21T08:58:19 i don't like this 2010-09-21T08:58:23 why? 2010-09-21T08:58:33 it's not like it really matters 2010-09-21T08:58:39 matters to me 2010-09-21T08:59:03 don't think you can really prevent it 2010-09-21T08:59:14 true 2010-09-21T08:59:23 unless you literally go through and claim every possible bartwe username /version number 2010-09-21T08:59:24 the password feature doesn't support the versioning 2010-09-21T08:59:47 i guess i could ask dhartmei about that 2010-09-21T09:00:04 probably more effort than it's worth tbh imo :P 2010-09-21T09:00:26 *** Naith has quit IRC (Quit: Naith) 2010-09-21T09:00:56 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T09:01:10 *** xvinyl has joined #aichallenge 2010-09-21T09:01:25 guess i should feel flattered 2010-09-21T09:02:00 yeah, no ones taking the time to imitate me 2010-09-21T09:02:31 *** sigh is now known as antimatroid2 2010-09-21T09:02:40 don't worry 2010-09-21T09:02:45 (: 2010-09-21T09:03:34 :) 2010-09-21T09:03:40 btw, why does your name have a 1 at the end? 2010-09-21T09:03:45 where is antimatroid? 2010-09-21T09:03:49 ... 2010-09-21T09:03:57 ~ . ~ 2010-09-21T09:03:57 hi amstan! 2010-09-21T09:04:20 @sigh 2010-09-21T09:04:20 i know... 2010-09-21T09:04:22 i probably registered antimatroid and it kicked me out for not logging in with it, or it timed out, came back with this and stayed as that, either way meh 2010-09-21T09:04:23 *** Naith has joined #aichallenge 2010-09-21T09:04:29 contestbot: rainbow F A B U L O U S 2010-09-21T09:04:29 pgpaskar_: F A B U L O U S 2010-09-21T09:04:36 lol that will always be hilarious 2010-09-21T09:04:40 :) 2010-09-21T09:04:52 contestbot: rainbow R A I N B O W 2010-09-21T09:04:52 antimatroid1: R A I N B O W 2010-09-21T09:04:59 ha, i no longer hate contestbot 2010-09-21T09:05:04 btw, amstan, I have an idea for contest bot 2010-09-21T09:05:11 *** Naith has quit IRC (Client Quit) 2010-09-21T09:05:22 give it a soldering iron and have ti build robots 2010-09-21T09:05:26 *** Accoun has quit IRC (Read error: Connection reset by peer) 2010-09-21T09:05:31 could you make it post the game info (like the @game command) automatically when someone posts link to a game? 2010-09-21T09:05:32 antimatroid2: cool 2010-09-21T09:05:39 amstan: could we get the hourly rank updates in rainbow form? 2010-09-21T09:05:42 @game 111111 2010-09-21T09:05:43 amstan: An error has occurred and has been logged. Please contact this bot's administrator for more information. 2010-09-21T09:05:51 antimatroid1: it needs a valid gameid 2010-09-21T09:05:52 @rankings 2010-09-21T09:05:53 pgpaskar_: Top 10 players: kcm1700(3491), McLeopold(3470), dmj111(3449), EBraun(3371), adumlah(3338), paulus(3299), petesimard(3298), sighbot(3297), montanalow(3270), Tracer(3252) 2010-09-21T09:05:56 oh wow 2010-09-21T09:06:00 yeah, I know 2010-09-21T09:06:01 @rainbow [rankings] 2010-09-21T09:06:02 amstan: Top 0 players: kcm1(9), McLeopold(), dmj(), EBraun(), adumlah(), paulus(), petesimard(2), (1 more message) 2010-09-21T09:06:02 *** antimatroid2 is now known as sigh 2010-09-21T09:06:05 mcleopold fell off the #1 spot 2010-09-21T09:06:08 dun dun dunnnnnnn! 2010-09-21T09:06:09 haha 2010-09-21T09:06:17 auch 2010-09-21T09:06:23 yay for c++ 2010-09-21T09:06:24 @rainbow OwNeD!!! 2010-09-21T09:06:24 pgpaskar_: OwNeD!!! 2010-09-21T09:06:28 :P 2010-09-21T09:06:40 like is someone pastes: http://www.ai-contest.com/visualizer.php?game_id=4619923 2010-09-21T09:06:49 @game 4619923 2010-09-21T09:06:50 amstan: Game played at 2010-09-21 07:30:30, on map 685, kcm1700 beat Tracer. http://ai-contest.com/visualizer_canvas.php?game_id=4619923 2010-09-21T09:06:53 contestbot will post the players and results 2010-09-21T09:06:54 sigh: I'm sorry Dave, I cannot will. 2010-09-21T09:07:08 he cannot will, lol! 2010-09-21T09:07:10 * pgpaskar_ slaps contestbot with a telephone 2010-09-21T09:07:20 :D 2010-09-21T09:07:26 contest do that. 2010-09-21T09:07:29 contestbot do that. 2010-09-21T09:07:30 bartwe: You have no gotten any error messages recently, so here's a random one just to let you know that we care. 2010-09-21T09:07:32 *** askhader has joined #aichallenge 2010-09-21T09:07:41 Hm, it appears your bot has achieve sentience... 2010-09-21T09:07:44 ehmz 2010-09-21T09:07:48 haha 2010-09-21T09:07:56 We simply must do something about this. 2010-09-21T09:08:06 contestbot: @success 2010-09-21T09:08:07 amstan: You have no gotten any error messages recently, so here's a random one just to let you know that we care. 2010-09-21T09:08:09 this cannot stand 2010-09-21T09:08:10 contestbot: success 2010-09-21T09:08:10 amstan: Ready to serve, my lord. 2010-09-21T09:08:20 contestbot: allyourspace 2010-09-21T09:08:20 are belong to us 2010-09-21T09:08:25 even though he's sentient, he obeys us 2010-09-21T09:08:34 the three rules of robotics 2010-09-21T09:08:39 contestbot what are the three laws ? 2010-09-21T09:08:39 bartwe: Error: No factoid matches that key. 2010-09-21T09:08:51 1) do not talk about robots, 2) do not talk about robots, 3) robots have no rules 2010-09-21T09:08:52 oh dear, he seems to have lost that part 2010-09-21T09:08:56 @help what 2010-09-21T09:08:56 sigh: (what ) -- Alias for "whatis $*". 2010-09-21T09:09:10 contestbot: whatis live? 2010-09-21T09:09:10 pgpaskar_: Error: No factoid matches that key. 2010-09-21T09:09:16 contestbot: whatis love? 2010-09-21T09:09:16 pgpaskar_: Error: No factoid matches that key. 2010-09-21T09:09:19 * pgpaskar_ slaps contestbot 2010-09-21T09:09:36 contestbot: whatis bot 2010-09-21T09:09:36 HakanD: I do not know about 'bot', but I do know about these similar topics: 'botrepo', 'bottopic', 'botdoc' 2010-09-21T09:10:09 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-09-21T09:10:10 amstan: so is it possible to automate that @game command when someone posts a link? or too hard? 2010-09-21T09:10:37 sigh: i was trying too, but i can't make the snarfer api work 2010-09-21T09:10:39 shouldnt be that hard to do gameid=[0-9]+ or something 2010-09-21T09:10:44 then have it look up the gameid 2010-09-21T09:11:00 amstan: ok, no worries 2010-09-21T09:12:18 *** bartwe has left #aichallenge 2010-09-21T09:12:50 *** bartwe has joined #aichallenge 2010-09-21T09:12:59 *** Olathe has joined #aichallenge 2010-09-21T09:16:09 *** Utkarsh has joined #aichallenge 2010-09-21T09:16:25 *** Sean_McEligot has quit IRC (Ping timeout: 264 seconds) 2010-09-21T09:19:16 who's fatbot? 2010-09-21T09:19:23 they just stormed tcp 2010-09-21T09:19:45 i demand a match against them aha 2010-09-21T09:20:57 haha... funny game: http://72.44.46.68/canvas?game_id=78147 2010-09-21T09:22:08 3 of his games against drewrobb are like this 2010-09-21T09:22:28 yeah, my bot has similar problems with not taking it's win efficiently 2010-09-21T09:22:34 its**** 2010-09-21T09:22:39 *** shul has joined #aichallenge 2010-09-21T09:23:55 http://72.44.46.68/canvas?game_id=78305 2010-09-21T09:23:59 not sure how i won that 2010-09-21T09:26:34 drewrobb doesn't redistribute his ships 2010-09-21T09:27:29 *** Sean_McEligot has joined #aichallenge 2010-09-21T09:27:36 yeah, he will cane once he does 2010-09-21T09:27:59 i should stop assuming every competitor is male, but i'm yet to come across a counterexample 2010-09-21T09:28:09 antimatroid1: are you waiting in the start on purpose ? 2010-09-21T09:28:22 *** wvdschel has joined #aichallenge 2010-09-21T09:28:25 the enemy of slimbot ? 2010-09-21T09:28:35 waiting where? i made a move on the first go there 2010-09-21T09:28:47 i have certain things in place to try and eliminate stupid moves yes 2010-09-21T09:28:57 but it's a very bad hack that doesn't work very well 2010-09-21T09:29:00 antimatroid1: gg :) 2010-09-21T09:29:04 hmmmm, I've written a bot, and it seems to crash whenever it kills all enemy planets, which kind of sucks 2010-09-21T09:29:11 or should I say ggl (good game logic) 2010-09-21T09:29:35 but I've added debug prints to everything, and it doesn't get any input after it's last orders, so I'm not sure what I'm doing wrong 2010-09-21T09:29:37 wvdschel: does it happen when the game ends? and you won ? 2010-09-21T09:29:39 antimatroid1: not sure if i disagree with fatbots strategyagainst drewrobb 2010-09-21T09:29:45 shul, yes 2010-09-21T09:30:05 wvdschel: so why does it suck? you won :) ... but you should check what happened 2010-09-21T09:30:11 with the Jars I got from a starter pack 2010-09-21T09:30:25 shul, because the PlayGame.jar output says the other bot won! 2010-09-21T09:30:49 wvdschel: that's what I meant 2010-09-21T09:30:55 are you doing something like say finding nearest enemy planet at the end and it then not existing? 2010-09-21T09:30:59 and bot crashing or something 2010-09-21T09:31:12 antimatroid1: that's not good 2010-09-21T09:31:26 it was a question not a statement 2010-09-21T09:31:38 although my bot seems to chew up lots of memory right at the end of a game 2010-09-21T09:31:44 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-09-21T09:31:45 and yet to work out what's causing it 2010-09-21T09:31:54 antimatroid1, that's what I thought at first, but I added a few more debug prints, and it doesn't even call my make_move function 2010-09-21T09:32:17 so there is no calculating going on 2010-09-21T09:33:09 an extract from my source: 2010-09-21T09:33:11 case 'go': 2010-09-21T09:33:11 this.make_move(); 2010-09-21T09:33:25 (from the input parser) 2010-09-21T09:33:36 this.make_move = function() { 2010-09-21T09:33:36 sys.debug("Start"); 2010-09-21T09:33:42 awww man, it's so boring when people play like 12 games to get their bot up near the top then stop running it 2010-09-21T09:33:51 and yet no Start appears on my debug output 2010-09-21T09:34:20 antimatroid1, with the TCP client? 2010-09-21T09:34:27 *** Utkarsh has joined #aichallenge 2010-09-21T09:34:59 shul, well, it's no good if my brilliant tactics destroy every bit of the enemy, while after I do so they get the victory on a technicality 2010-09-21T09:35:23 got to go. will be back 2010-09-21T09:35:25 *** wvdschel has quit IRC (Quit: ragequit) 2010-09-21T09:36:36 got to play fatbot 2010-09-21T09:36:44 got to beat* fatbot 2010-09-21T09:36:50 oh it was a timeout, no fun 2010-09-21T09:37:25 *** sinan has quit IRC (Ping timeout: 264 seconds) 2010-09-21T09:44:08 *** kozlovsky has joined #aichallenge 2010-09-21T09:45:01 *** Appleman1234 has quit IRC (Ping timeout: 265 seconds) 2010-09-21T09:45:43 *** Utkarsh has quit IRC (Ping timeout: 255 seconds) 2010-09-21T09:47:14 *** s_denchev has joined #aichallenge 2010-09-21T09:49:14 arg... my future state prediction is completely and utterly broken 2010-09-21T09:49:20 *** s_denchev is now known as s_denchev__at_wo 2010-09-21T09:49:44 *** s_denchev__at_wo is now known as s_denchev|work 2010-09-21T09:49:59 how my bot is functioning is beyond my knowledge 2010-09-21T09:50:32 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T09:50:39 haha, i'm pretty sure mine works, but there's lots of things i need to fix up 2010-09-21T09:50:45 *** xvinyl has joined #aichallenge 2010-09-21T09:50:46 if only there was moar time 2010-09-21T09:50:54 http://72.44.46.68/canvas?game_id=78633 2010-09-21T09:50:59 intresting game i think 2010-09-21T09:51:06 meager start 2010-09-21T09:51:11 strange stagnation in the middle 2010-09-21T09:51:14 stupid end 2010-09-21T09:51:15 I have a question about 'fair play'. it is probably better on the forums but I thought i'd ask here anyway. When moving ships from a->b to reinforce, moving 50 ships in 50 1 ship fleets is obviously going to cause more pain for other bots than one single fleet, but is clearly not making your bot any stronger tactically, just exploiting a possible weakness of your opponent (i.e. bad sportsmanship)... is it against the rules / spirit of the competition? ( 2010-09-21T09:52:05 MrCeeJ: have you checked if the game server doesn't combine those attacks ? 2010-09-21T09:52:19 i'm pretty sure they fixed this 2010-09-21T09:52:24 *** Utkarsh has joined #aichallenge 2010-09-21T09:52:30 i pointed it out to them ages ago 2010-09-21T09:52:41 same as you can't send fleets of size 0 anymore 2010-09-21T09:52:43 No i didn't want to implement it. I read they removed size 0 fleets, but there was a comment to the effect of 'size 1 and greater fleets are fleets' 2010-09-21T09:52:58 yeah, but i think the game engine merges them 2010-09-21T09:53:08 that's what I've heard also 2010-09-21T09:53:18 thanks 2010-09-21T09:53:19 I think it was mentioned in that same thread on teh form 2010-09-21T09:53:21 *forum 2010-09-21T09:53:43 by all means feel free to spam ships to every planet from yours, but it's not going to cripple my so i don't care 2010-09-21T09:53:55 me* 2010-09-21T09:54:47 *** amstan has quit IRC (Ping timeout: 272 seconds) 2010-09-21T09:54:54 * sigh is off to build a spam bot 2010-09-21T09:54:55 It was more that by simply overriding the sendShips method to loop and send them individually you might get and advantage for (a) bad programming (b) bad sportsmanship and (c) no effort 2010-09-21T09:55:28 it is fixed 2010-09-21T09:55:31 game engine merges them now 2010-09-21T09:55:45 ok 2010-09-21T09:56:32 Top 10 players: kcm1700(3495), McLeopold(3474), dmj111(3450), Arekku(3390), EBraun(3376), adumlah(3340), paulus(3304), sighbot(3297), petesimard(3281), montanalow(3275) 2010-09-21T09:56:46 *** Blkt has joined #aichallenge 2010-09-21T09:56:51 2nd related question, having your n planets send n-1 fleets size one to each other each turn might also hurt the other bot, but is actually consuming your own resources. So would this use of ships be prevented under fair play? 2010-09-21T09:57:17 that's fine, but a rather stupid tactic imo 2010-09-21T09:57:25 MrCeeJ: you would lose with such a tactics 2010-09-21T09:57:37 also alot of bots come pretty close to doing that allready 2010-09-21T09:57:38 if the bot is stupid enough to be crippled by it, chances are you would have beaten them anyway 2010-09-21T09:57:48 if it isn't, then it's going to own you 2010-09-21T09:57:53 MrCeeJ, You really can't send enough to time out a properly coded bot 2010-09-21T09:58:08 *** ermau|work has quit IRC (Quit: ermau|work) 2010-09-21T09:58:43 yeah, above a certain level the games are legitimately full of hundreds of fleets 2010-09-21T09:59:03 if a bot can't deal with that, it is screwed 2010-09-21T09:59:04 it's really not hard to handle 2010-09-21T09:59:06 *** Sean_McEligot has quit IRC (Ping timeout: 252 seconds) 2010-09-21T09:59:21 you only care about a fleets destination, arrival time and size 2010-09-21T09:59:25 yup 2010-09-21T09:59:28 use that as you wish 2010-09-21T09:59:33 and fleet owner 2010-09-21T09:59:52 well yeah, but that i have combined into fleets with both players having 0+ ships 2010-09-21T09:59:53 I can parse and sort 400000 fleets in under 0.1 secs here on my machine... 2010-09-21T10:00:23 So as long as there is no silly n^3 or something lurking --its going to handle the fleets 2010-09-21T10:00:35 :) 2010-09-21T10:01:28 I can't think of any situation that would involve require looking at all pairs of fleets, let alone any higher order 2010-09-21T10:01:40 doubtless there are some, I just can't think of any 2010-09-21T10:02:12 delt0r: question is, what are the specs of the contest machine 2010-09-21T10:02:28 I was more concerned about the unknown processing power of the servers and the 1sec rule. 21 planets with 20 fleets generated per turn for the last 10 turns is 4000, if you then parse that list for each of you 20 planets you have 80,000 iterations of your loop to get through 2010-09-21T10:02:29 *** sinan has joined #aichallenge 2010-09-21T10:02:37 true... but the point remains. fleet counts are unlikely to be an issue 2010-09-21T10:03:20 MrCeeJ: you shouldn't need to process all fleets for each planet 2010-09-21T10:03:52 the real point i am getting at is are actions designed soley to make life harder for the other player to process against the spirit: 2010-09-21T10:03:55 "Entries which are deemed to violate the spirit of fair and sportsmanlike competition will be disqualified without any opportunity for appeal" 2010-09-21T10:04:15 *** retybok has quit IRC (Read error: No route to host) 2010-09-21T10:04:33 MrCeeJ: in this case the obvious trick is splitting up your fleets is mitigated by the server merging them 2010-09-21T10:04:44 any other use of fleets is fair game 2010-09-21T10:05:40 There is a legit reason to have fleets sent on the same turn from the same planet source and destination 2010-09-21T10:05:48 *** Zeiris has quit IRC (Ping timeout: 265 seconds) 2010-09-21T10:06:10 yep, i often do it now 2010-09-21T10:06:36 that is the bot may decide that with 0.5 sec left it wants to send 10 --but later ups that to 20 2010-09-21T10:07:19 delt0r: I agree, I was just talking in the context of attempting to cause mischief :P 2010-09-21T10:07:43 my code does it too 2010-09-21T10:08:20 mine does not at this stage... you can't "uncommit" troops --so i manage that myself 2010-09-21T10:08:27 so essentially all that is banned is actual cheating as described by "memory scanning, intentionally losing games, and behavior conditional on the opponent's identity" rather than attempting to trap / trick / crash the opposing bots 2010-09-21T10:08:49 MrCeeJ, and also having lots of accounts 2010-09-21T10:08:55 yeah 2010-09-21T10:09:11 shame as it takes so long to get a stable rank :/ 2010-09-21T10:09:19 yeah :( 2010-09-21T10:09:28 loose agaist yourself --easiest ladder hack --hardest to detect and stop 2010-09-21T10:09:31 you might want to play on the tcp server for quicker feedback 2010-09-21T10:10:10 *** kcm1700 has joined #aichallenge 2010-09-21T10:10:10 yeah I havn't looked into setting that up yet 2010-09-21T10:10:25 but I have some powerful hardware at home, and didn't want to write code that fell over on the server 2010-09-21T10:10:30 MrCeeJ: it takes all of 2 minutes :) 2010-09-21T10:10:39 MrCeeJ: yes, of course 2010-09-21T10:11:59 I shall probably see some of you on tcp later on this evening. Thanks for the responses :) 2010-09-21T10:12:16 *** hornairs has joined #aichallenge 2010-09-21T10:12:38 good lunk! 2010-09-21T10:12:40 *luck 2010-09-21T10:13:45 hi 2010-09-21T10:13:52 i need some suggestions for my bot 2010-09-21T10:14:00 the .10 series is not really beating the .8 series 2010-09-21T10:14:00 haha 2010-09-21T10:14:09 and i'm losing momentum .... 2010-09-21T10:14:09 .11 seems pretty good :P 2010-09-21T10:14:24 antimatroid1: .11 is worse then .8 and also, not mine 2010-09-21T10:14:40 that last part was the point ;) 2010-09-21T10:15:16 geh, 2010-09-21T10:15:24 do you guys model coordinated strikes ? 2010-09-21T10:15:35 bartwe: I do 2010-09-21T10:15:49 how do you model somethign like that ? 2010-09-21T10:16:04 even picking the fleet seems very combinatorial 2010-09-21T10:16:14 I do loop over each destination planet and decide the best way of attacking it 2010-09-21T10:16:18 *** pc has joined #aichallenge 2010-09-21T10:16:25 where that attack can involve multiple sources 2010-09-21T10:17:02 I do each destination independently and score them based on the cost of attacking them, distance and other factors 2010-09-21T10:17:45 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T10:18:30 previous to that I looped over all sources and picked the best destination to send ships to, that was much simpler but didn't allow as sophisticated attacks 2010-09-21T10:20:06 sigh: i do much the same, but it feels way to sub optimal to do well later 2010-09-21T10:20:33 i'm trying to work out ways to solve the problem better, but it gets complicated with fleets being 0 or above x 2010-09-21T10:21:17 antimatroid1: sure, but I'm no where near reaching the potential of this current approach 2010-09-21T10:21:39 yeah neither, my point is the above is what i see as that key :P 2010-09-21T10:22:11 one issue with coordinated strikes is the combination of timing and reevaluation of not yet departed fleet bits of a strike 2010-09-21T10:22:19 i see no good way to model that 2010-09-21T10:22:45 you can compute a surprising amount :P 2010-09-21T10:22:51 :) 2010-09-21T10:23:26 bartwe: example of my coordinated attacks, see the end of this game: http://www.ai-contest.com/visualizer.php?game_id=4615907 2010-09-21T10:23:51 last 15 moves or so when I take the last planets 2010-09-21T10:23:56 the timing seems off 2010-09-21T10:24:22 a bit, enemy does stuff in the mean time to screw up timing, and I have bugs too 2010-09-21T10:24:58 the neutral 44 you take on turn 29/30, you hit first with 24 then with 27 a turn later 2010-09-21T10:25:09 possibly not rounding up the distance ? 2010-09-21T10:26:13 MrCeeJ: nah, it wont be a rounding error... it's either a bug, or I'm being over defensive 2010-09-21T10:26:34 probably bug :) 2010-09-21T10:26:54 point I wanted to illustrate was the coordinated attack, not the timing :P 2010-09-21T10:27:05 so stop picking on me :D 2010-09-21T10:27:08 :p 2010-09-21T10:27:08 hehe 2010-09-21T10:27:13 looks great 2010-09-21T10:27:18 how do you pick the sources ? 2010-09-21T10:27:29 secret atm :P 2010-09-21T10:27:32 I was under the impression that the coordination (timing) was the point :p 2010-09-21T10:28:04 MrCeeJ: timing is more of a detail... the hard part is getting different source planets to act in unison 2010-09-21T10:28:22 *** justin_pdx has joined #aichallenge 2010-09-21T10:28:37 *** kunaljain has joined #aichallenge 2010-09-21T10:29:03 if you know what sources you want to use, then it is easy to brute force possible timings 2010-09-21T10:29:12 it is only a matter of perspective, you can look at all the sources and pick a destination, or you can look at all the destinations and pick a (combination of) sources 2010-09-21T10:29:23 *easy being relative here 2010-09-21T10:29:23 if you are out of turn, then which player wins ? The player with more number of planets or more number of Army ? 2010-09-21T10:29:33 army size 2010-09-21T10:29:47 MrCeeJ: thanks 2010-09-21T10:29:52 those are sort of dual problems 2010-09-21T10:29:57 kunaljain: very hard to have superior army size at move 200 without a equal or better planet position 2010-09-21T10:30:08 in that they're really the same, and you don't necessarily have to lose that detail doing the former 2010-09-21T10:30:25 sigh: does coordinated strikes reduce or increase the complexity of moving units forward ? 2010-09-21T10:30:40 moving units forward? 2010-09-21T10:30:49 like redistributing them? 2010-09-21T10:30:59 yes, into a position of influence 2010-09-21T10:31:13 for me those two part of the code are completely independant 2010-09-21T10:31:28 *completely is an approximation 2010-09-21T10:31:43 so it doesn't really impact me, but that's just my approach 2010-09-21T10:32:30 *** phreeza has joined #aichallenge 2010-09-21T10:33:17 yeah, i've split my makemove function up into different components based on priority 2010-09-21T10:33:27 so do you reevaluate the sending of units in a coordinated strike for which the initial units are inflight ? 2010-09-21T10:34:12 bartwe: at the moment my bot has no notion of state... it doesn't remember that the previous move it wanted to do a coordinated strike :P 2010-09-21T10:34:57 It will be a long time before I change that, because making my bot stateful will make testing and debugging much harder 2010-09-21T10:35:17 *** s_denchev|work has quit IRC (Quit: Page closed) 2010-09-21T10:35:48 i want to know if we're guaranteed distances of at least 2 in the final tournament of not 2010-09-21T10:35:59 it would make things a LOT easier if it was 2010-09-21T10:36:15 I have seen some planets that overlap and cause issues for bots 2010-09-21T10:36:18 dhartmei said yes... but looking at the generating script, I can't see it 2010-09-21T10:36:24 sigh: genius 2010-09-21T10:36:24 but it is possible tehy were 2 away and just large 2010-09-21T10:36:26 *** sorki has joined #aichallenge 2010-09-21T10:36:34 bartwe: what is? 2010-09-21T10:36:45 sigh: replotting the coordinated strike in every timestep 2010-09-21T10:36:49 how do I translate "if (pw.Production(1) >= pw.Production(2)) {" to C++ ? 2010-09-21T10:36:54 it may make you vunerable to overcommit 2010-09-21T10:36:55 unless they are at exactly the same point, they are distanced at least 1 apart 2010-09-21T10:36:56 but very cool 2010-09-21T10:37:04 :) 2010-09-21T10:37:09 *** amstan has joined #aichallenge 2010-09-21T10:37:09 *** ChanServ sets mode: +o amstan 2010-09-21T10:37:11 but you never see a fleet from the enemy if they send it to a planet distanced 1 from its soure 2010-09-21T10:37:15 it was just eisier that way 2010-09-21T10:37:22 *easier 2010-09-21T10:37:33 shul: that's already fine? 2010-09-21T10:38:07 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-21T10:38:21 antimatroid1: what ? 2010-09-21T10:40:51 *** xvinyl has joined #aichallenge 2010-09-21T10:40:55 *** pc has quit IRC (Remote host closed the connection) 2010-09-21T10:41:22 *** mightybyte has joined #aichallenge 2010-09-21T10:42:58 *** phreeza has quit IRC (Ping timeout: 240 seconds) 2010-09-21T10:43:01 sigh: still don't see how your cost function accepts the coordinated attck :) 2010-09-21T10:43:06 but very cool stuff 2010-09-21T10:44:07 *** jorge_ has joined #aichallenge 2010-09-21T10:44:09 antimatroid1: I remember that C++ starter pack hasn't Production method :) 2010-09-21T10:44:14 hi 2010-09-21T10:44:21 bartwe: yeah, cost function is something I think about all the time... my current one makes no sense 2010-09-21T10:44:25 but it does the job 2010-09-21T10:44:42 shul: that comparison is useless, you can delete it now 2010-09-21T10:44:48 is it possible to have a visualization engine locally like the one in the web? I mean, smooth :) 2010-09-21T10:45:06 pacotabaco: hi :) 2010-09-21T10:45:15 *** phreeza has joined #aichallenge 2010-09-21T10:45:21 jorge_: code is here: http://github.com/DanielVF/Planet-Wars-Canvas-Visualizer 2010-09-21T10:45:35 *** iouri_ has joined #aichallenge 2010-09-21T10:45:43 use visualize_localy.py as your visualiser 2010-09-21T10:45:43 sigh: thanks! 2010-09-21T10:46:04 oh, and download up to date version of visualizer.js 2010-09-21T10:46:23 oh? that's not up to date? 2010-09-21T10:47:12 it doesn't display ship count/growth rate 2010-09-21T10:47:48 ah 2010-09-21T10:48:21 http://planetwars.zerd.net/ 2010-09-21T10:48:22 *** Accoun has joined #aichallenge 2010-09-21T10:48:46 http://planetwars.zerd.net/visualizer.js was updated 2 days ago 2010-09-21T10:48:58 aha 2010-09-21T10:48:59 thanks 2010-09-21T10:49:37 it's used in tcp site 2010-09-21T10:49:50 yeah, I've noticed that 2010-09-21T10:50:38 it's a very handy feature :) 2010-09-21T10:50:46 *** bartwe has left #aichallenge 2010-09-21T10:50:52 *** bartwe has joined #aichallenge 2010-09-21T10:52:20 hi jorge_ ;) 2010-09-21T10:56:32 Top 10 players: kcm1700(3524), McLeopold(3483), dmj111(3475), Arekku(3405), EBraun(3382), adumlah(3374), petesimard(3317), montanalow(3300), sighbot(3285), paulus(3280) 2010-09-21T10:56:58 everytime contestbot comes around I drop a rank :( 2010-09-21T10:57:20 it's a conspiracy ;) 2010-09-21T10:57:23 :P 2010-09-21T10:57:51 oh well, it's no suprise... nearly everyone in the top 10 has beaten me 2010-09-21T10:58:08 *** justin_pdx has joined #aichallenge 2010-09-21T10:58:09 my bot is flopping miserably on there 2010-09-21T10:58:20 is it the one on tcp? 2010-09-21T10:58:24 yep 2010-09-21T10:58:28 o_O 2010-09-21T10:58:59 my one is the same as sigh-1A that I had on tcp the last two days 2010-09-21T10:59:21 yeah, i think mine is only good on some maps 2010-09-21T10:59:36 yeah, that can be a problem 2010-09-21T10:59:40 then fails on others, which doesn't do well on the main site 2010-09-21T10:59:58 but that's alright, just need to fix up move selection 2010-09-21T11:00:10 :) 2010-09-21T11:00:11 i'm too agressive sometimes, to the point that i can't back up my claims 2010-09-21T11:00:51 sigh-1 reached the top of tcp on sunday BECAUSE it was too aggressive 2010-09-21T11:01:05 but then everyone caught up and it started being demolished 2010-09-21T11:01:08 yeah, works well often, but not always 2010-09-21T11:01:57 oh well, guess you had the same effect 2010-09-21T11:01:58 *** justin_pdx has quit IRC (Client Quit) 2010-09-21T11:02:03 reached the top with aggression :P 2010-09-21T11:02:34 shouldn't be that surprising, with the quality of rage bot and all :P 2010-09-21T11:03:04 :D 2010-09-21T11:03:29 i have a love hate relationship with ragebot aha 2010-09-21T11:03:39 mostly love these days 2010-09-21T11:04:23 ok, time to sleep 2010-09-21T11:04:31 night 2010-09-21T11:04:37 see you all later 2010-09-21T11:04:43 *** sigh has quit IRC (Quit: Leaving...) 2010-09-21T11:09:14 *** ajhager has quit IRC (Ping timeout: 245 seconds) 2010-09-21T11:09:33 argh, stupid moves :(((( 2010-09-21T11:12:44 *** elijahbal has quit IRC (Quit: ZNC - http://znc.sourceforge.net) 2010-09-21T11:13:18 *** tobiassjosten has quit IRC (Ping timeout: 240 seconds) 2010-09-21T11:14:22 *** elijahbal has joined #aichallenge 2010-09-21T11:17:42 *** hornairs has quit IRC (Quit: hornairs) 2010-09-21T11:29:11 *** Tempest has joined #aichallenge 2010-09-21T11:29:17 hi 2010-09-21T11:32:11 may i assume the growth rate of any planet is always less than or equal to 5? 2010-09-21T11:33:32 hmm 2010-09-21T11:33:51 *** mightybyte has quit IRC (Quit: leaving) 2010-09-21T11:33:59 why do you need this assumption ?? 2010-09-21T11:34:55 *** danielvf has joined #aichallenge 2010-09-21T11:35:31 *** ajhager has joined #aichallenge 2010-09-21T11:36:10 *** kunaljain has quit IRC (Quit: Page closed) 2010-09-21T11:36:47 It looks like its always between 1 and 5 based on map_generator.py, but I know I've seen a 0 growth on tcp 2010-09-21T11:37:05 No idea if that's a custom or what 2010-09-21T11:37:25 jesionaj: There are 0's on the server too 2010-09-21T11:38:06 oh, there's two different ranges in map_generator 2010-09-21T11:38:35 center can be 0, others are 1-5 from the looks of it 2010-09-21T11:42:39 *** tires has joined #aichallenge 2010-09-21T11:44:21 *** Kastra has joined #aichallenge 2010-09-21T11:48:01 dufflebunk: those are weird warnings 2010-09-21T11:48:25 woah i was way scrolled up 2010-09-21T11:53:52 *** tires has quit IRC (Quit: Page closed) 2010-09-21T11:54:55 *** Sean_McEligot has joined #aichallenge 2010-09-21T11:55:11 *** sinan has quit IRC (Read error: Operation timed out) 2010-09-21T11:56:16 *** amstan has quit IRC (Quit: Konversation terminated!) 2010-09-21T11:56:31 *** amstan has joined #aichallenge 2010-09-21T11:56:31 *** ChanServ sets mode: +o amstan 2010-09-21T11:56:34 Top 10 players: kcm1700(3523), McLeopold(3485), dmj111(3478), Arekku(3403), EBraun(3381), adumlah(3374), petesimard(3318), montanalow(3299), sighbot(3285), paulus(3280) 2010-09-21T12:01:17 *** ajhager has quit IRC (Ping timeout: 272 seconds) 2010-09-21T12:08:08 *** aerique has quit IRC (Quit: ...) 2010-09-21T12:11:33 where is the link to tcp.zip for windows ? 2010-09-21T12:12:59 *** ratatata has quit IRC (Quit: Page closed) 2010-09-21T12:14:57 *** mightybyte has joined #aichallenge 2010-09-21T12:15:49 *** Kastra has quit IRC (Ping timeout: 265 seconds) 2010-09-21T12:20:48 *** boegel has joined #aichallenge 2010-09-21T12:27:36 *** Blkt has quit IRC (Ping timeout: 252 seconds) 2010-09-21T12:28:52 http://www.benzedrine.cx/planetwars/tcp.zip 2010-09-21T12:31:27 jesionaj: thanks, but it says the server was not found, does it work for you (maybe its a dns problem on my end) ? 2010-09-21T12:32:19 $ host www.benzedrine.cx 2010-09-21T12:32:19 www.benzedrine.cx is an alias for insomnia.benzedrine.cx. 2010-09-21T12:32:19 insomnia.benzedrine.cx has address 213.3.30.106 2010-09-21T12:32:45 *** Baphomet has quit IRC (Ping timeout: 240 seconds) 2010-09-21T12:33:18 Works fine for me, maybe try: 213.3.30.106/planetwars/tcp.zip 2010-09-21T12:35:40 *** Sean_McEligot has left #aichallenge 2010-09-21T12:37:11 *** Baphomet has joined #aichallenge 2010-09-21T12:38:07 my bot is going to be trivially parallelizable. i'm debating whether i should use more than one core for the tcp server 2010-09-21T12:38:18 maybe i could just have my bot under a different name for that 2010-09-21T12:39:36 jmcarthur: the real contest only allows 1 thread though, right? 2010-09-21T12:39:51 or was it 1 process 2010-09-21T12:39:58 both 2010-09-21T12:40:46 yeah the real one only allows a single thread 2010-09-21T12:41:13 although i'm not entirely sure whether that's enforced technically, i will not be exploiting that 2010-09-21T12:41:34 wish i could though. it would give me a significant advantage 2010-09-21T12:41:43 *i think 2010-09-21T12:43:46 youre using haskell? 2010-09-21T12:43:52 *** Alex_____ has joined #aichallenge 2010-09-21T12:44:00 can you even controll that with ghc? 2010-09-21T12:44:27 oh yeah 2010-09-21T12:44:35 *** __Duh__ <__Duh__!534dff82@gateway/web/freenode/ip.83.77.255.130> has joined #aichallenge 2010-09-21T12:45:34 for me it's just a matter of building with -threaded and running with +RTS -N (or i could wrap it with a C main and have those flags passed to the haskell runtime automatically) 2010-09-21T12:45:42 this of course depends on getting to upload a binary 2010-09-21T12:46:48 wish i had more than two cores at my disposal for the tcp server 2010-09-21T12:47:32 i have a nettop with a dual core Atom that has hyperthreading, making it 4 virtual cpus, but i doubt it would really be much faster if at all 2010-09-21T12:47:35 *** ntroPi has joined #aichallenge 2010-09-21T12:47:50 hi 2010-09-21T12:48:04 hyello 2010-09-21T12:50:03 *** Alex_____ has quit IRC (Quit: Page closed) 2010-09-21T12:56:19 *** justin_pdx has joined #aichallenge 2010-09-21T12:56:33 Top 10 players: kcm1700(3574), dmj111(3555), McLeopold(3508), adumlah(3395), EBraun(3393), sighbot(3341), montanalow(3336), paulus(3305), Tracer(3290), Arekku(3257) 2010-09-21T12:56:57 *** amstan has quit IRC (Quit: Konversation terminated!) 2010-09-21T12:57:07 out of curiosity, why haskell? 2010-09-21T12:57:09 *** amstan has joined #aichallenge 2010-09-21T12:57:09 *** ChanServ sets mode: +o amstan 2010-09-21T12:57:38 *** __Duh__ <__Duh__!534dff82@gateway/web/freenode/ip.83.77.255.130> has quit IRC (Quit: Page closed) 2010-09-21T12:58:29 woo, new personal record 2010-09-21T12:58:38 9489 units at round 200 2010-09-21T12:59:51 bartwe: link? 2010-09-21T13:00:16 http://72.44.46.68/canvas?game_id=81457 2010-09-21T13:00:53 pjreddie: i could spout the evangelism, but then i would just appear to be a fanboy 2010-09-21T13:01:35 i've done it enough here as it is ;) 2010-09-21T13:02:13 lol so many mirrors 2010-09-21T13:02:25 bartwe: lol 2010-09-21T13:02:45 *** joanpujol has joined #aichallenge 2010-09-21T13:02:55 Hi 2010-09-21T13:03:48 Anyone from the organitzation? 2010-09-21T13:03:55 bartwe: lol 2010-09-21T13:04:23 Or that can look at compile error messages? I've problem submiting my updated java code 2010-09-21T13:04:43 I've tried to ask help in the forums but not solved my problem 2010-09-21T13:04:50 bartwe: Meanie. heh. 2010-09-21T13:05:15 *** hornairs has joined #aichallenge 2010-09-21T13:05:26 *** amstan has quit IRC (Ping timeout: 276 seconds) 2010-09-21T13:05:52 jesionaj: thanks, it worked ! 2010-09-21T13:05:57 *** Kastra has joined #aichallenge 2010-09-21T13:06:02 np 2010-09-21T13:06:05 now the question is whether my firewall stops 995 2010-09-21T13:06:31 *** jorge_ has quit IRC (Remote host closed the connection) 2010-09-21T13:07:07 *** sinan has joined #aichallenge 2010-09-21T13:07:25 None had java compile errors? 2010-09-21T13:07:33 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-21T13:07:56 *** danielvf has joined #aichallenge 2010-09-21T13:09:32 whoa... http://ai-contest.com/visualizer.php?game_id=4623894 2010-09-21T13:10:40 jmcarthur: Yes they are. I can't find anything online resembling them. 2010-09-21T13:10:57 small difference at the beginning makes subtle difference at the end of the game. 2010-09-21T13:11:23 kcm1700: nice feedforward system 2010-09-21T13:11:35 dufflebunk: i'm tempted to just blame all the old software on the server 2010-09-21T13:11:38 dmj111 seems to use coordinated attack 2010-09-21T13:11:56 should give you a larger window of attack possily 2010-09-21T13:12:40 he does make you pay for the neutrals 2010-09-21T13:13:03 no clue why he doesn't take the neutral in his area 2010-09-21T13:13:54 ademar: p0wned :) http://72.44.46.68/canvas?game_id=81747 2010-09-21T13:15:07 nice ragebot 2010-09-21T13:15:20 kcm1700: ןמאקרקדאןמע 2010-09-21T13:15:27 kcm1700: interesting 2010-09-21T13:16:02 thanks a lot.... what were those characters? ןמאקרקדאןמע ??? 2010-09-21T13:17:05 kcm1700: I'll let you guess :-), hint: my nick is related to that 2010-09-21T13:17:19 do you need an ai-contest login to use the TCP server ? 2010-09-21T13:18:00 *** JCS^ has joined #aichallenge 2010-09-21T13:18:01 nope 2010-09-21T13:18:12 just pick a name and optionally password 2010-09-21T13:18:27 jmcarthur: so am I, but that doesn't help me with getting my bot to work. 2010-09-21T13:18:34 *** ajhager has joined #aichallenge 2010-09-21T13:18:47 dufflebunk: i'm really close to convincing the relevant people to allow binary submissions 2010-09-21T13:18:58 which would solve all that :D 2010-09-21T13:19:30 jmcarthur: I'm guessing it has something to do with with the use of System.Timeout.timeout but I'm not sure. It would suck if it is. 2010-09-21T13:19:42 contestbot seen j3camero 2010-09-21T13:19:42 Janzert: j3camero was last seen in #aichallenge 5 days, 14 hours, 29 minutes, and 32 seconds ago: Cyndre: are you interested in building it? That would be soooo bitchin. I have seen some simple online IDEs before, and I gotta say it's pretty cool. We could build something unprecedented! 2010-09-21T13:19:47 bah 2010-09-21T13:20:00 dufflebunk: System.Timeout.timeout unfortunately doesn't do quite what would want it to 2010-09-21T13:20:17 dufflebunk: in the tron contest i ended up writing an FFI binding so i could use SIGALRM 2010-09-21T13:20:25 might end up using that again 2010-09-21T13:20:49 *** drewrobb has quit IRC (Read error: Operation timed out) 2010-09-21T13:20:50 *what i would want it to 2010-09-21T13:21:04 It does what I want it to do, stops the evaluation after 900 ms. I just had to be sure to useControl.Exception.evaluate with it 2010-09-21T13:21:11 *** sinan has quit IRC (Ping timeout: 276 seconds) 2010-09-21T13:21:32 dufflebunk: my problem with it was that it can't give you partial information if it has to stop early 2010-09-21T13:21:34 hi my bot suddenly starts to send "invalid input" (see: http://72.44.46.68/canvas?game_id=81524) 2010-09-21T13:21:48 dufflebunk: and my tron bot (and current bot) had partial information to give 2010-09-21T13:22:06 i log all output data and seems to be correctlly formatted 2010-09-21T13:22:06 jmcarthur: Ah, yeah, like best move found so far. Mine's not that advanced yet. 2010-09-21T13:22:43 no 0 ships fleets and no fleets to self planet, there are more things to check? 2010-09-21T13:23:02 pacotabaco: negative fleets? 2010-09-21T13:23:03 sending more ships than you have on a planet 2010-09-21T13:23:13 no 2010-09-21T13:23:48 "12 8 12\n4 19 17\n3 5 9\n9 2 26\ngo" 2010-09-21T13:23:57 \n after go too :P 2010-09-21T13:23:58 pacotabaco: not owning the source planet? 2010-09-21T13:24:20 mmm i'll check 2010-09-21T13:26:34 usually it'll give a reason in log.txt if you run the java engine 2010-09-21T13:26:39 (PlayGame.jar) 2010-09-21T13:27:06 my first silly try : http://72.44.46.68/canvas?game_id=81946 2010-09-21T13:27:20 how do u add . and all sorts of stuff after it ? 2010-09-21T13:28:07 that's just the username 2010-09-21T13:28:26 *** sinan has joined #aichallenge 2010-09-21T13:28:28 jmcarthur: couldn't you use an MVar and timeout? Put the partial info or final solution into the MVar, then the main thread will have access to it if the worker is stopped 2010-09-21T13:28:44 * dufflebunk hasn't done much multi-threaded haskell 2010-09-21T13:29:03 dufflebunk: i did that but ran into some issue that i don't remember 2010-09-21T13:29:09 dufflebunk: multithreading is against the rules 2010-09-21T13:29:15 shul: you can use green threads 2010-09-21T13:29:57 dufflebunk: what i ended up doing was sharing a couples MVars with the SIGALRM signal handler, last time 2010-09-21T13:29:58 jmcarthur: you are right... sening from an enemy planet :P (shit happens!) 2010-09-21T13:30:08 hehe, Civ5 is out :) 2010-09-21T13:30:34 if only i had a machine fast enough to play it 2010-09-21T13:30:35 dufflebunk: when the signal is received it waits for the worker to finish its current iteration, grabs the current result, then pushes it to another MVar that the IO thread is waiting on 2010-09-21T13:30:52 then the IO thread pushes back to the worker thread when it's done 2010-09-21T13:31:06 if only i have time enough to play it :) 2010-09-21T13:31:09 in fact, i had the IO thread push back to the worker between sending orders and receiving the next game state 2010-09-21T13:31:14 enough time- 2010-09-21T13:31:16 shul: yeah, no OS threads so it should be fine 2010-09-21T13:31:22 and then grab it back again when the next state update is received 2010-09-21T13:31:38 so it was really working all the time except during actual IO and during signal handling 2010-09-21T13:31:51 HakanD: civ5 is the biggest enemy of ai-contest :) 2010-09-21T13:32:10 civ5 is my friend because i never got into civ 2010-09-21T13:32:11 and minecraft, it turns out 2010-09-21T13:32:18 therefore is gives me an advantage in the contest 2010-09-21T13:32:20 *it 2010-09-21T13:32:28 *** perror has quit IRC (Quit: Bye all !) 2010-09-21T13:32:46 Starcraft2 also 2010-09-21T13:32:54 powned by mcleopold's :-) 2010-09-21T13:32:59 jmcarthur: interesting 2010-09-21T13:33:59 sid meyer is jmcarthur friend, it was all pre-orchestrated years ago 2010-09-21T13:38:00 minecraft is out ? oh... 2010-09-21T13:38:30 finally.... 2010-09-21T13:38:52 10.13.b is only somewhat worse then .8.2 that cost days, horrible 2010-09-21T13:39:19 *** boegel has quit IRC (Ping timeout: 240 seconds) 2010-09-21T13:39:28 *** ajhager has quit IRC (Ping timeout: 272 seconds) 2010-09-21T13:40:03 *** delt0r has quit IRC (Remote host closed the connection) 2010-09-21T13:40:38 *** ermau has joined #aichallenge 2010-09-21T13:40:42 *** boegel has joined #aichallenge 2010-09-21T13:42:26 *** ermau is now known as ermau|work 2010-09-21T13:43:22 *** mark_jn has joined #aichallenge 2010-09-21T13:43:48 *** sinan has quit IRC (Ping timeout: 265 seconds) 2010-09-21T13:44:36 *** smellyhippy has joined #aichallenge 2010-09-21T13:44:36 *** smellyhippy has joined #aichallenge 2010-09-21T13:45:56 *** HakanD has quit IRC (Quit: brb) 2010-09-21T13:46:12 it's interesting how everybody's attempts to improve their bots is resulting in disaster 2010-09-21T13:46:17 *are 2010-09-21T13:47:07 yea ... one little change and i lose agains ragebot on a map where i usually win in 60 turns. 2010-09-21T13:48:10 i wonder why this is 2010-09-21T13:49:23 because noone _really_ understands their heuristics for choosing planets ;) 2010-09-21T13:49:37 thats the case with my current bot anyways 2010-09-21T13:51:21 just feels right and works somehow :-D 2010-09-21T13:53:41 *** hornairs has quit IRC (Quit: hornairs) 2010-09-21T13:54:27 because heristics work best when they are simple, and when multiple heristics fits together in transparent and predictable ways 2010-09-21T13:54:51 hey, where'd my "u" go? 2010-09-21T13:54:56 is there a way in which you can always win? 2010-09-21T13:55:15 I'm not asking to explain how, I just wonder if there is 2010-09-21T13:55:21 boegel: yes. solve the big equation 2010-09-21T13:55:35 dwchandler: :) 2010-09-21T13:55:46 dwchandler: but not in 1s, I presume 2010-09-21T13:56:08 *** Ceridan has joined #aichallenge 2010-09-21T13:56:14 no. there's always pruning of some kind needed 2010-09-21T13:56:34 Top 10 players: kcm1700(3584), McLeopold(3551), dmj111(3522), adumlah(3412), EBraun(3408), montanalow(3355), Tracer(3335), sighbot(3332), paulus(3317), Loks1982(3259) 2010-09-21T13:57:55 it almost seems like there should be a closed-form solution to this game, but i don't know that anybody will find it 2010-09-21T13:59:25 I just profiled my javascript bot - 55% of CPU time in the garbage collector 2010-09-21T13:59:30 Maybe I need to fix that. 2010-09-21T14:00:09 danielvf: :) 2010-09-21T14:00:28 danielvf: as long as you don't run into time-outs, who cares? :) 2010-09-21T14:00:37 *** kay_ is now known as Guest95353 2010-09-21T14:01:21 boegel: yeah, but I do on the server. :D 2010-09-21T14:02:03 danielvf: ah :) 2010-09-21T14:02:09 danielvf: then you should fix it 2010-09-21T14:02:38 Or I just need to find out how to disable garbage collection. Surely I can't make too much trash in 200 turns 2010-09-21T14:03:24 I wouldn't be too sure about that... 2010-09-21T14:04:09 err 2010-09-21T14:04:14 yeah i wouldn't try that 2010-09-21T14:04:29 lol 2010-09-21T14:04:44 *** Gorovkhaia2 has joined #aichallenge 2010-09-21T14:05:04 i love this game http://ai-contest.com/visualizer.php?game_id=4623664 2010-09-21T14:05:11 sadly, It's not an option 2010-09-21T14:06:34 Gorovkhaia2: haha, I can see why :) 2010-09-21T14:06:38 Gorovkhaia2: finish fail 2010-09-21T14:08:31 *** Bobng has joined #aichallenge 2010-09-21T14:09:26 danielvf: are you bumping against 1 second? if not, don't worry about profiling yet 2010-09-21T14:09:45 dwchandler: yeah, I'm getting 3 turns of play on the server 2010-09-21T14:09:54 ouch. 2010-09-21T14:10:20 dwchandler: This wasn't coded for speed, so I'm not too worried. 2010-09-21T14:10:25 *** HakanD has joined #aichallenge 2010-09-21T14:11:00 *** shul has quit IRC (Ping timeout: 252 seconds) 2010-09-21T14:13:16 not much javascript is coded for speed anyway 2010-09-21T14:13:30 jmcarthur: any idea if there's a way to control the compile options? It may only happen with -O2 2010-09-21T14:14:01 dufflebunk: you may be able to try {-# OPTIONS -O0 #-}, but i think command line flags override pragmas 2010-09-21T14:14:10 Tried it, no luck 2010-09-21T14:14:31 no way that i know of then, aside from waiting for binaries to be supported 2010-09-21T14:14:33 actually, it was GHC_OPTIONS I tried 2010-09-21T14:14:47 no big difference 2010-09-21T14:15:32 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-09-21T14:15:48 the only thing i can suggest if you don't want to wait for binary support is set up a VM with ubuntu server 8.04 64 bit, install the distro's ghc, and try building with it 2010-09-21T14:16:19 I might have to do that to track it down. 2010-09-21T14:18:04 *** RainCT has joined #aichallenge 2010-09-21T14:21:03 *** pc has joined #aichallenge 2010-09-21T14:21:09 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T14:27:20 I guess I'm the victim of unknown timeouts or invalid commands 2010-09-21T14:27:29 *** shades has quit IRC (Ping timeout: 276 seconds) 2010-09-21T14:27:42 Since I can run a while local game in 5 seconds 2010-09-21T14:28:11 probably timeouts. i *always* had issues with those in the last contest 2010-09-21T14:28:22 *** shades has joined #aichallenge 2010-09-21T14:28:43 i'm still convinced that the server occasionally giving us less than 0.1s sometimes 2010-09-21T14:28:46 *gave 2010-09-21T14:29:05 never was able to verify since we couldn't do stderr output 2010-09-21T14:30:45 but even setting my timeout to 0.1s would still give me timeouts sometimes, even though local testing showed that my bot would always complete in a timely manner on my machine 2010-09-21T14:31:20 *** mb has joined #aichallenge 2010-09-21T14:31:46 *** mb is now known as Guest64701 2010-09-21T14:32:06 my latest bot is doing surprisingly well, given the fact that I can make it a lot more aggressive and targetted, w00! 2010-09-21T14:33:10 *** SophieB has joined #aichallenge 2010-09-21T14:33:30 *** smhanov has quit IRC (Ping timeout: 240 seconds) 2010-09-21T14:34:27 great, ubuntu network manager decided to play hide and seek 2010-09-21T14:34:42 this is the first time i've used this repa library. purely functional regular arrays are taking some time to get used to, but so far it's pretty cool 2010-09-21T14:35:09 i'm not used to indexing being so abstracted, is the main mental stumbling block i'm having 2010-09-21T14:37:12 for example, replication along any dimension is just an indexing transformation rather than an actual copy, so my function for precomputing planet distances basically replicates the array of coordinates twice, along two different dimensions, then applies the distance computation element-wise to "both" arrays (although there is actually only one, just with different indexing transformations under 2010-09-21T14:37:14 two different names) 2010-09-21T14:41:22 *** ysulaiman has joined #aichallenge 2010-09-21T14:41:53 *** McLeopold has joined #aichallenge 2010-09-21T14:42:22 http://ai-contest.com/visualizer.php?game_id=4624224 2010-09-21T14:42:27 *** Naith has joined #aichallenge 2010-09-21T14:43:17 weird 2010-09-21T14:44:35 I though that matched looked good for me. Yet I can keep the #1 stop even when I bet him? 2010-09-21T14:44:42 beat 2010-09-21T14:45:14 McLeopold: Thats Bayesian for you... 2010-09-21T14:45:33 yeah, too many losses on maps I'm weak on 2010-09-21T14:45:46 McLeopold: it looks like you've lost to more bots that are worse than kcm 2010-09-21T14:45:50 change the prior change the result. Personally there should be a bunch of extra probablity mass right over me ;) 2010-09-21T14:45:57 *** xvinyl has joined #aichallenge 2010-09-21T14:46:13 *** smellyhippy has quit IRC (Ping timeout: 252 seconds) 2010-09-21T14:46:29 McLeopold: i wish the confidence interval was shown 2010-09-21T14:46:29 uh, everyone's worse that kcm...and me 2010-09-21T14:46:35 that would probably put you at ease about it 2010-09-21T14:46:42 that's strange... his bot is better than mine.. 2010-09-21T14:46:56 *** pc has quit IRC (Remote host closed the connection) 2010-09-21T14:46:59 Maybe he has tighter confidence? 2010-09-21T14:47:19 I wonder if the averages are switched? 2010-09-21T14:47:21 McLeopold: no, i mean you probably have overlapping confidence intervals 2010-09-21T14:47:37 Is it possible to look? 2010-09-21T14:47:47 McLeopold: tighter confidence wouldn't be reflected in the rank, which is only the mean 2010-09-21T14:47:48 Or is that secret info? 2010-09-21T14:48:01 i doubt it's secret, but i don't know of a way to see it without having an admin show you 2010-09-21T14:48:06 I thought it was the mean - confidence? 2010-09-21T14:48:12 no, it's just mean 2010-09-21T14:48:20 i have been pushing for the lower confidence bound 2010-09-21T14:48:24 but to no avail 2010-09-21T14:48:27 That puts me at less ease. :) 2010-09-21T14:48:49 It makes more sense when you begin. Start at the bottom. 2010-09-21T14:49:07 After you are established, it can be tricky. 2010-09-21T14:49:16 McLeopold: it could be that you have a very tight confidence interval, which would mean that games affect your rank less 2010-09-21T14:50:19 i agree. i don't care if it takes a while to get to the top. in fact, i think that is better 2010-09-21T14:50:36 I see. So I could be 3530-3534, and kcm could be 3520-3602 or something. 2010-09-21T14:50:49 yeah 2010-09-21T14:50:57 My lower bound could be higher, but not the mean. 2010-09-21T14:51:15 exactly 2010-09-21T14:51:20 *** smellyhippy has joined #aichallenge 2010-09-21T14:51:20 *** smellyhippy has joined #aichallenge 2010-09-21T14:52:19 Well, when I lose to dmj, it won't matter. :) 2010-09-21T14:52:25 heh 2010-09-21T14:55:30 It would be cool to see a graph of the top bots with their confidence intervals. 2010-09-21T14:56:02 i thought about that too 2010-09-21T14:56:06 And, a list of up and coming undefeated new bots. 2010-09-21T14:56:19 Like a rookie list or something. 2010-09-21T14:56:33 Top 10 players: kcm1700(3561), McLeopold(3532), dmj111(3503), adumlah(3400), EBraun(3389), montanalow(3331), Tracer(3317), sighbot(3286), paulus(3256), Loks1982(3239) 2010-09-21T14:56:34 Maybe winning streak. 2010-09-21T14:56:40 Up and down arrows. 2010-09-21T14:57:15 the bots with the most played games that still have confidence intervals overlapping the top bots' intervals 2010-09-21T14:58:01 Yeah, that would get you a list of potential new winners. 2010-09-21T14:58:40 what i think would be really awesome is if there was an ajaxy page showing current games 2010-09-21T14:58:49 is the 'size' of the maps known? i.e. how many turns would it take to go from one side to the other? 2010-09-21T14:58:52 so people would stop complaining that there were no games being played :P 2010-09-21T14:59:00 boegel: not (yet) specified 2010-09-21T14:59:06 Yeah, or at least a number of current games. 2010-09-21T14:59:11 jmcarthur: hmmk, thx 2010-09-21T14:59:17 boegel: The most I've seen is about 30 2010-09-21T14:59:23 *** Ari2 has joined #aichallenge 2010-09-21T14:59:34 McLeopold: a live readout would be awesome though, especially showing winners of recent games and stuff 2010-09-21T15:00:59 I have made a graph of rank changes of only very few people. http://kcm1700.snucse.org/new/rankgraph.png 2010-09-21T15:01:18 kcm1700: that's pretty cool! 2010-09-21T15:01:58 *** ysulaiman has quit IRC (Quit: Page closed) 2010-09-21T15:02:44 kcm1700: you didn't happen to record elo's as well, did you? 2010-09-21T15:03:14 I didn't.. but probably i'm gonna start work on elo score. 2010-09-21T15:03:45 That's awesome. Does it get updated regularly? 2010-09-21T15:04:15 probably... hope so. i'm using crontab to get it updated. 2010-09-21T15:04:19 danielvf: thank you for investigating the timeout issue! nobody did that last year, to my knowledge, despite my (and others') loud complaints 2010-09-21T15:04:30 s/last year/last contest/ 2010-09-21T15:04:34 jmcarthur: NP 2010-09-21T15:04:42 Just getting around to it. 2010-09-21T15:06:39 kcm1700: http://ai-contest.com/visualizer.php?game_id=4623894 2010-09-21T15:06:50 Does that game look like it could go either way? 2010-09-21T15:08:26 I think giving more turns could give a different result. 2010-09-21T15:09:07 yeah, or at least make you victory more clear. It's hard to tell with all the fleets flying around 2010-09-21T15:09:08 to me it looks like the potential was there for one of the bots to screw up, but i think the win was deserved if there would have been no screwups in a longer game 2010-09-21T15:09:22 *abnormal screwups, i mean 2010-09-21T15:09:24 *** ysulaiman has joined #aichallenge 2010-09-21T15:09:55 If dmj's logic would have gone for the neutral, who knows. 2010-09-21T15:10:05 i'd give it a 25% chance of being a different outcome for more turns 2010-09-21T15:10:31 The turn limit is too short. 2010-09-21T15:10:35 yeah, small but still there. 2010-09-21T15:10:40 * jmcarthur goes to make up more BS numbers in the kitchen 2010-09-21T15:10:46 *** joanpujol has quit IRC (Ping timeout: 252 seconds) 2010-09-21T15:10:50 i agree, we should have more turns 2010-09-21T15:10:57 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T15:11:01 *** pc has joined #aichallenge 2010-09-21T15:11:05 at least 300-400 2010-09-21T15:11:35 and I'll need to change my bot not to crash after 200 turns. 2010-09-21T15:11:40 me too 2010-09-21T15:11:45 it crashes on 201st turn :-) 2010-09-21T15:11:51 mine too ;) 2010-09-21T15:12:11 well, i'm making it take an optional argument for turns, but default to 200 for the server 2010-09-21T15:12:19 *** Naktibalda has joined #aichallenge 2010-09-21T15:13:59 divide by zero? 2010-09-21T15:14:23 and out or index too. 2010-09-21T15:16:06 *** icky has joined #aichallenge 2010-09-21T15:16:14 *** ntroPi has quit IRC (Quit: cya) 2010-09-21T15:17:06 *** ysulaiman has quit IRC (Quit: Page closed) 2010-09-21T15:17:31 you make an array of 200 states ? 2010-09-21T15:17:41 and use that to simulate ? 2010-09-21T15:22:04 *** mark_jn has quit IRC (Quit: leaving) 2010-09-21T15:34:55 *** ermau|work has quit IRC (Quit: ermau|work) 2010-09-21T15:37:01 *** Accoun has quit IRC () 2010-09-21T15:40:23 *** dstufft has quit IRC (Quit: Reconnecting) 2010-09-21T15:40:26 Lafi: :) 2010-09-21T15:40:30 *** dstufft has joined #aichallenge 2010-09-21T15:41:06 not necessarily that 2010-09-21T15:41:18 but it's nice to be able to simulate a game to completion 2010-09-21T15:41:30 since out-of-turns is an endgame condition, it must be in any correct simulation 2010-09-21T15:46:08 *** mega1 has quit IRC (Ping timeout: 272 seconds) 2010-09-21T15:47:12 way too late, but it would be awesome if number of turns were included in game state 2010-09-21T15:47:15 moin 2010-09-21T15:47:35 http://code.google.com/p/ai-contest/issues/detail?id=74 2010-09-21T15:48:00 dwchandler: not really... well i don't think so 2010-09-21T15:48:09 but i an't nobody 2010-09-21T15:48:13 *** Accoun has joined #aichallenge 2010-09-21T15:48:15 j3camero has given provisional approval 2010-09-21T15:49:03 although that was on friday and he seems to have disappeared afterward 2010-09-21T15:49:22 # : key value 2010-09-21T15:49:23 bartwe: I have no idea what you mean. 2010-09-21T15:49:25 makes sense 2010-09-21T15:49:35 # rainbow Hello world 2010-09-21T15:49:35 bartwe: Hello world 2010-09-21T15:50:01 heh 2010-09-21T15:50:21 cool bot-- mine can't do that 2010-09-21T15:50:25 *** nullkuhl has quit IRC (Read error: Connection reset by peer) 2010-09-21T15:50:50 *** nullkuhl has joined #aichallenge 2010-09-21T15:50:55 # rainbow Bot Pride 2010-09-21T15:50:55 dwchandler: Bot Pride 2010-09-21T15:51:34 Janzert: thanks for the link. I hope it gets in 2010-09-21T15:54:09 i do too 2010-09-21T15:55:14 *** delt0r_ has joined #aichallenge 2010-09-21T15:56:20 *** delt0r___ has quit IRC (Ping timeout: 240 seconds) 2010-09-21T15:56:33 Top 10 players: McLeopold(3522), kcm1700(3518), dmj111(3513), adumlah(3388), EBraun(3362), montanalow(3331), Tracer(3312), sighbot(3280), Loks1982(3254), Arekku(3233) 2010-09-21T16:01:10 *** pc has quit IRC (Remote host closed the connection) 2010-09-21T16:01:12 *** xvinyl has joined #aichallenge 2010-09-21T16:05:44 # rainbow ABCDEFG 2010-09-21T16:05:44 kcm1700: ABCDEFG 2010-09-21T16:06:13 is the starting budget of 100 ships a given? 2010-09-21T16:06:21 no 2010-09-21T16:06:28 at least not yet 2010-09-21T16:06:35 jmcarthur: hmmk, thx :) 2010-09-21T16:08:35 how can I win a game with invalid input, see http://72.44.46.68/getplayer?player=boegel.20100921.9e1fca9872d9913bbd9725ff0c9601d5 2010-09-21T16:08:48 if I look at the game, it doesn't seem a win at all either 2010-09-21T16:09:01 is that a bug in the evaluation I'm seeing then? 2010-09-21T16:09:37 boegel: your opponent probably bugged out 2010-09-21T16:09:44 pretty sure it means allchemist sent an illegal order 2010-09-21T16:12:10 jmcarthur: ah, it doens't mean it was my bot that sent the wrong order, ok, my mistake 2010-09-21T16:12:22 *** seon has joined #aichallenge 2010-09-21T16:14:35 *** hellman has quit IRC (Remote host closed the connection) 2010-09-21T16:16:04 *** drewrobb has joined #aichallenge 2010-09-21T16:17:26 *** seon has quit IRC (Ping timeout: 240 seconds) 2010-09-21T16:21:56 *** retybok_ has joined #aichallenge 2010-09-21T16:22:48 Has the protocol changed? I keep getting these "lost invalid" games on dhartmei's tcp server, but I don't have them on my machine when I play against the example bots... 2010-09-21T16:25:24 ouch, mysql error on the rankings page. 2010-09-21T16:25:57 too many connects. you'd think that would be cached? 2010-09-21T16:26:42 contestbot: rankings 10 2010-09-21T16:26:53 McLeopold: Error: There was a problem accessing the interface to ai-contest.com 2010-09-21T16:27:35 long live dataases 2010-09-21T16:28:04 was that typo on purpose? I'm trying to pronounce it... 2010-09-21T16:28:52 daaases 2010-09-21T16:29:01 Looks like it's down for the count. 2010-09-21T16:29:07 McLeopold: don't worry you're still ranked first :-D 2010-09-21T16:29:10 who needs pesky consonants 2010-09-21T16:29:15 my '' key is roken at times :) 2010-09-21T16:29:23 (i can still access it) 2010-09-21T16:29:31 bartwe: lol 2010-09-21T16:29:35 I missed the crossover. I want to know how I bumped up again? 2010-09-21T16:30:29 my bot made another loss. 2010-09-21T16:30:33 got back in, 2010-09-21T16:30:37 yea, I'm watching 2010-09-21T16:30:49 was too greedy 2010-09-21T16:30:55 hmz, damn close to using full cpu without bounded search 2010-09-21T16:31:02 time for the profiler 2010-09-21T16:31:24 I don't think you were greedy enough? Unless your refering to the top center planet. 2010-09-21T16:32:10 *** Guest64701 has quit IRC (Ping timeout: 252 seconds) 2010-09-21T16:32:41 kcm1700: I think it is your misdirected traffic issue again. 2010-09-21T16:41:21 woot 2010-09-21T16:42:32 what would you guys suggest for the red side to keep this combat less chaotic ? http://72.44.46.68/canvas?game_id=85055 2010-09-21T16:42:50 *** smhanov has joined #aichallenge 2010-09-21T16:43:42 *** customKoncept has joined #aichallenge 2010-09-21T16:43:53 PlayGame.jar crashes and in the log file there's only one word, "initializing"... 2010-09-21T16:44:12 looked to me as though the log file isn't actually used 2010-09-21T16:44:42 use a different log file 2010-09-21T16:44:58 log.txt is rumored to be used by the visualizer, too 2010-09-21T16:45:37 Zannick: wow thank you it works! 2010-09-21T16:48:55 *** makg has joined #aichallenge 2010-09-21T16:50:51 *** Mathnerd314 has quit IRC (Ping timeout: 265 seconds) 2010-09-21T16:54:05 *** smhanov has quit IRC (Ping timeout: 272 seconds) 2010-09-21T16:56:03 bartwe: sending a fleet of 70 at turn 41 lost you that side of the board temporarily. 2010-09-21T16:56:14 yeah, its weird 2010-09-21T16:56:21 *** Kastra has quit IRC (Ping timeout: 276 seconds) 2010-09-21T16:56:34 Top 10 players: McLeopold(3521), kcm1700(3519), dmj111(3512), adumlah(3389), EBraun(3362), montanalow(3329), Tracer(3316), sighbot(3282), Loks1982(3254), _Shurf(3235) 2010-09-21T16:56:44 That's the type of problem I'm working on now. 2010-09-21T16:57:50 looks like a stupendously stupid move to expend that much to a neutral . 2010-09-21T16:58:14 especially when you can already count the incoming fleets to your growth 5 planet just below 2010-09-21T16:58:15 i think the problem is in the worstcase predictor, it fails to take the arrivals into account 2010-09-21T16:58:52 it fails to see inflight enemy fleets as a source of incomming attacks 2010-09-21T16:59:09 +potential 2010-09-21T16:59:35 *** Mathnerd314 has joined #aichallenge 2010-09-21T17:02:13 *** Palmik has quit IRC (Remote host closed the connection) 2010-09-21T17:04:13 *** ajhager has joined #aichallenge 2010-09-21T17:04:17 *** choas has joined #aichallenge 2010-09-21T17:06:28 *** customKoncept has quit IRC (Quit: Page closed) 2010-09-21T17:10:32 ah ha! 0x757b12c00 is the number of milliseconds in a year 2010-09-21T17:12:19 jmcarthur: that's probably too big for an Int32 2010-09-21T17:13:08 it's 9 bytes 2010-09-21T17:13:48 lastlog 757b12 2010-09-21T17:14:22 yeah, the thing pasted earlier was an int64 2010-09-21T17:14:43 er, not 9 bytes 2010-09-21T17:14:54 it's 5 bytes 2010-09-21T17:15:02 07 57 b1 2c 00 2010-09-21T17:16:29 right, Int64, maybe too big for that too. I'll just skip that section. It'll just cause at most one incorrect calculation per year. 2010-09-21T17:17:05 why does your bot need such a number? 2010-09-21T17:17:56 I'm making sure the (green) thread doesn't go over 900 ms, and the time library gives me differences in a silly structure 2010-09-21T17:18:09 ah 2010-09-21T17:18:13 where it gives be the change in years, months, days, ... and you have to add it all up 2010-09-21T17:19:11 Yep, that was it. No more warnings or errors from the compile script. 2010-09-21T17:20:34 *** FlagCapper has joined #aichallenge 2010-09-21T17:20:44 *** dhartmei has quit IRC (Quit: good night) 2010-09-21T17:21:57 cool 2010-09-21T17:22:13 with several days of work, i'm now at parity with my old bot :) 2010-09-21T17:22:32 only this one has a sensible cost function 2010-09-21T17:22:36 *** smhanov has joined #aichallenge 2010-09-21T17:26:15 bartwe: I know what you mean. 2010-09-21T17:26:22 I can't beat myself at all! 2010-09-21T17:26:31 geh 2010-09-21T17:27:07 i guess i could build somethign to automaticly walk the statespace of som of the magical constants 2010-09-21T17:27:19 that could keep benzedrine busy for a few days 2010-09-21T17:27:35 magical constants? 2010-09-21T17:28:17 Why don't you just do that locally? 2010-09-21T17:28:29 against what and whom ? 2010-09-21T17:28:42 what magical constants you are referring to ? 2010-09-21T17:29:11 meh, how to scale fudge factors in the cost function regarding cost of movement and such 2010-09-21T17:34:27 *** smhanov has quit IRC (Ping timeout: 264 seconds) 2010-09-21T17:35:06 *** smhanov has joined #aichallenge 2010-09-21T17:35:55 yes, each improvement i make in intelligence makes it play worse 2010-09-21T17:37:20 I think that my latest version is too cautious 2010-09-21T17:39:58 *** linux-m00n has joined #aichallenge 2010-09-21T17:46:17 * Naktibalda is thinking about unit tests for bot 2010-09-21T17:49:59 *** seon has joined #aichallenge 2010-09-21T17:52:41 *** MatthewR has joined #aichallenge 2010-09-21T17:53:03 *** tapwater has joined #aichallenge 2010-09-21T17:53:25 there must be a lot of bugs in my code 2010-09-21T17:53:43 *** kozlovsky has quit IRC (Quit: Page closed) 2010-09-21T17:56:28 Naktibalda> cause you lose 2010-09-21T17:56:29 ? 2010-09-21T17:56:35 Top 10 players: McLeopold(3515), kcm1700(3507), dmj111(3503), adumlah(3374), montanalow(3357), EBraun(3339), Tracer(3302), sighbot(3274), Loks1982(3231), Arekku(3227) 2010-09-21T17:56:37 *** smhanov has quit IRC (Ping timeout: 272 seconds) 2010-09-21T17:56:44 no, because I see them in action :) 2010-09-21T17:57:03 me too 2010-09-21T17:57:05 *** icky has quit IRC (Ping timeout: 252 seconds) 2010-09-21T17:57:15 I've found that fixing bugs improve performance a lot more than implementing new features :) 2010-09-21T17:57:21 but it's not as much fun :) 2010-09-21T17:57:40 hmm, that's good, looks like I have a lot of them 2010-09-21T17:58:33 what are the bugs? 2010-09-21T18:00:07 for example: sending more ships then necesarry 2010-09-21T18:00:11 *** RainCT has quit IRC (Remote host closed the connection) 2010-09-21T18:00:58 yesterday I discovered, that I clean targets list at the wrong level of nested loops, so it sends multiple fleets of the same size if there are enough 2010-09-21T18:01:15 *** Ceridan has quit IRC () 2010-09-21T18:01:45 Naktibalda> do you prefer to send more ships? 2010-09-21T18:01:46 now I see that my planet size calculator after capturing works wrong, so it sends not enough ships to get a cheap capture 2010-09-21T18:01:55 *** iouri_ has quit IRC (Quit: Page closed) 2010-09-21T18:02:16 no, I prefer to attack more targets at the same time :) 2010-09-21T18:03:17 another bug that I see, is that if one planet is under attack, another planet sends multiple fleets at the same turn. 2010-09-21T18:03:39 *** jaspervdj has quit IRC (Quit: leaving) 2010-09-21T18:04:46 how many ships do you use before retunr? 2010-09-21T18:05:27 *** icky has joined #aichallenge 2010-09-21T18:07:44 *** sorki has quit IRC (Ping timeout: 240 seconds) 2010-09-21T18:08:03 *** delt0r_ has quit IRC (Quit: Leaving) 2010-09-21T18:13:39 1 ship matters 2010-09-21T18:14:12 *** MatthewR has quit IRC (Quit: Page closed) 2010-09-21T18:14:42 I just reduced a size of reinforcement fleet by one, and my bot lost in 200 turns by score instead of being dominated in 178 2010-09-21T18:18:53 *** makg has quit IRC (Ping timeout: 252 seconds) 2010-09-21T18:21:12 *** linux-m00n has quit IRC (Quit: Leaving) 2010-09-21T18:21:19 *** seon has quit IRC (Quit: Leaving.) 2010-09-21T18:22:36 one more fix and game is almost close 2010-09-21T18:23:05 *** mightybyte has quit IRC (Quit: leaving) 2010-09-21T18:24:50 yep, old version had only 2 planets more at the end 2010-09-21T18:25:08 old aggresive version of my bot is my main opponent :) 2010-09-21T18:33:18 no use, it's loosing 79 games out of 100 allready :) 2010-09-21T18:33:51 *** Appleman1234 has joined #aichallenge 2010-09-21T18:35:21 *** Mathnerd314 has quit IRC (Ping timeout: 240 seconds) 2010-09-21T18:36:51 *** Zeiris has joined #aichallenge 2010-09-21T18:42:49 *** retybok_ has quit IRC (Quit: leaving) 2010-09-21T18:47:55 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-09-21T18:48:47 *** icky has quit IRC (Ping timeout: 252 seconds) 2010-09-21T18:50:42 *** boegel has quit IRC (Quit: Leaving) 2010-09-21T18:51:43 *** smhanov has joined #aichallenge 2010-09-21T18:52:30 *** ysulaiman has joined #aichallenge 2010-09-21T18:54:51 *** ysulaiman has quit IRC (Client Quit) 2010-09-21T18:56:02 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-21T18:56:34 Top 10 players: McLeopold(3512), kcm1700(3507), dmj111(3496), adumlah(3379), montanalow(3349), EBraun(3339), Arekku(3313), Tracer(3306), sighbot(3254), Loks1982(3215) 2010-09-21T18:58:51 *** sigh has joined #aichallenge 2010-09-21T19:04:51 *** jimmerton has joined #aichallenge 2010-09-21T19:05:38 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T19:05:52 *** xvinyl has joined #aichallenge 2010-09-21T19:08:30 *** choas has quit IRC (Quit: leaving) 2010-09-21T19:20:25 *** amstan has joined #aichallenge 2010-09-21T19:20:27 *** ChanServ sets mode: +o amstan 2010-09-21T19:21:45 *** smhanov has quit IRC (Read error: Operation timed out) 2010-09-21T19:25:13 *** Bobng has quit IRC (Quit: Leaving) 2010-09-21T19:25:47 hi, just a question... the number of turns of a game affects the ELO? 2010-09-21T19:27:07 for ex: if some one win in 50 turns get more ELO points than if he do it in 200? 2010-09-21T19:27:26 nope 2010-09-21T19:27:46 all wins count the same 2010-09-21T19:27:59 ok, it's just win/lose/draw? 2010-09-21T19:28:04 yup 2010-09-21T19:28:10 ok, tnks :) 2010-09-21T19:29:37 *** Mathnerd314 has joined #aichallenge 2010-09-21T19:32:00 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T19:36:46 *** jimmerton_ has joined #aichallenge 2010-09-21T19:36:54 *** smhanov has joined #aichallenge 2010-09-21T19:37:59 *** yuwei has quit IRC (Quit: leaving) 2010-09-21T19:39:18 *** jimmerton has quit IRC (Ping timeout: 255 seconds) 2010-09-21T19:39:18 *** jimmerton_ is now known as jimmerton 2010-09-21T19:40:13 *** McLeopold has joined #aichallenge 2010-09-21T19:40:30 contestbot: rankings 7 2010-09-21T19:40:31 McLeopold: Top 7 players: McLeopold(3512), kcm1700(3508), dmj111(3496), adumlah(3379), montanalow(3349), EBraun(3340), Arekku(3314) 2010-09-21T19:40:49 *** SophieB is now known as zz_SophieB 2010-09-21T19:40:53 Arekku is going straight to #1 2010-09-21T19:42:45 http://ai-contest.com/visualizer.php?game_id=4620680 2010-09-21T19:42:56 That is what most games will start to look like. 2010-09-21T19:44:31 wow, nice defense 2010-09-21T19:45:37 heh, nice cross-galaxy spread http://ai-contest.com/visualizer.php?game_id=4624633 2010-09-21T19:46:38 nice 2010-09-21T19:47:07 *** zz_SophieB is now known as SophieB 2010-09-21T19:51:25 interesting route of petesimard ships 2010-09-21T19:51:47 I only hope he plays kcm1700 first. :) 2010-09-21T19:55:09 *** SophieB is now known as zz_SophieB 2010-09-21T19:55:10 Is ... is Arekku mimicking? 2010-09-21T19:55:37 No 2010-09-21T19:55:54 He is building very fast. And defending well. 2010-09-21T19:56:04 Arekku impressive move... 2010-09-21T19:56:34 Top 10 players: McLeopold(3501), kcm1700(3499), dmj111(3487), adumlah(3360), montanalow(3336), EBraun(3330), Tracer(3283), Arekku(3250), sighbot(3243), Willjarn(3235) 2010-09-21T20:00:21 *** zz_SophieB is now known as SophieB 2010-09-21T20:03:00 *** jimmerton_ has joined #aichallenge 2010-09-21T20:04:22 *** xvinyl has joined #aichallenge 2010-09-21T20:05:10 *** jimmerton has quit IRC (Ping timeout: 265 seconds) 2010-09-21T20:05:11 *** jimmerton_ is now known as jimmerton 2010-09-21T20:07:00 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-21T20:07:21 *** jimmerton_ has joined #aichallenge 2010-09-21T20:07:59 *** hamstah has joined #aichallenge 2010-09-21T20:08:01 hey 2010-09-21T20:08:48 why does playGame.jar make 1 turn more than given in parameter? 2010-09-21T20:09:07 compute the last moves to see the winner I guess 2010-09-21T20:09:35 last turn just updates 2010-09-21T20:09:36 *** jimmerton has quit IRC (Read error: Connection reset by peer) 2010-09-21T20:09:37 *** jimmerton_ is now known as jimmerton 2010-09-21T20:10:23 I made scripts to run bots/game with netcat. kind of the same as tcpserver, but between 2 persons. is there a place I should add a link to them? 2010-09-21T20:12:07 *** SophieB is now known as zz_SophieB 2010-09-21T20:12:45 *** Macuyiko has quit IRC (Ping timeout: 265 seconds) 2010-09-21T20:13:18 *** ulope has quit IRC (Ping timeout: 276 seconds) 2010-09-21T20:13:48 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-21T20:15:03 *** fahi has quit IRC (Ping timeout: 276 seconds) 2010-09-21T20:15:07 Can add it to the forums 2010-09-21T20:15:23 2 persons or 2 bots? 2010-09-21T20:15:31 *** fahi has joined #aichallenge 2010-09-21T20:15:33 *** ulope has joined #aichallenge 2010-09-21T20:15:35 ah, it's p2p 2010-09-21T20:15:50 Naith, k 2010-09-21T20:16:10 its convenient if you want to play about someone in particular without giving out your code 2010-09-21T20:16:12 *** cYmen has quit IRC (Ping timeout: 255 seconds) 2010-09-21T20:16:29 *** cYmen has joined #aichallenge 2010-09-21T20:16:44 Sounds cool thanks I'll check it out later. 2010-09-21T20:17:09 does it work without external IP? 2010-09-21T20:17:19 works on localhost as well 2010-09-21T20:17:26 anything netcat can connect to is fine 2010-09-21T20:17:39 I tested it with a few people 2010-09-21T20:17:55 *** zz_SophieB is now known as SophieB 2010-09-21T20:25:25 *** jimmerton_ has joined #aichallenge 2010-09-21T20:26:07 *** SophieB is now known as zz_SophieB 2010-09-21T20:26:32 *** iFire has joined #aichallenge 2010-09-21T20:27:23 *** cYmen has quit IRC (Read error: Operation timed out) 2010-09-21T20:27:57 *** cYmen has joined #aichallenge 2010-09-21T20:28:00 *** jimmerton has quit IRC (Ping timeout: 272 seconds) 2010-09-21T20:28:00 *** jimmerton_ is now known as jimmerton 2010-09-21T20:33:26 *** zz_SophieB is now known as SophieB 2010-09-21T20:37:47 *** jimmerton_ has joined #aichallenge 2010-09-21T20:39:31 *** fahi has quit IRC (Ping timeout: 240 seconds) 2010-09-21T20:39:42 *** jimmerton has quit IRC (Read error: Connection reset by peer) 2010-09-21T20:39:43 *** jimmerton_ is now known as jimmerton 2010-09-21T20:41:24 *** cYmen has quit IRC (Ping timeout: 255 seconds) 2010-09-21T20:41:29 *** cYmen has joined #aichallenge 2010-09-21T20:42:27 ok posted into the forums. 2010-09-21T20:42:33 *** fahi has joined #aichallenge 2010-09-21T20:42:37 bbl, good night 2010-09-21T20:45:15 *** rogue780 has joined #aichallenge 2010-09-21T20:45:44 *** SophieB is now known as zz_SophieB 2010-09-21T20:48:10 *** Appleman1234 has quit IRC (Quit: Leaving) 2010-09-21T20:51:59 *** zz_SophieB is now known as SophieB 2010-09-21T20:55:47 *** jimmerton_ has joined #aichallenge 2010-09-21T20:56:35 Top 10 players: McLeopold(3499), kcm1700(3497), dmj111(3483), adumlah(3357), montanalow(3333), EBraun(3328), Tracer(3280), Arekku(3247), sighbot(3241), Willjarn(3230) 2010-09-21T20:56:44 *** smf68 has quit IRC (Remote host closed the connection) 2010-09-21T20:56:49 *** SophieB is now known as zz_SophieB 2010-09-21T20:58:02 *** jimmerton has quit IRC (Ping timeout: 252 seconds) 2010-09-21T20:58:03 *** jimmerton_ is now known as jimmerton 2010-09-21T21:00:09 *** jimmerton_ has joined #aichallenge 2010-09-21T21:02:04 *** JCS^ has quit IRC (Ping timeout: 265 seconds) 2010-09-21T21:02:22 *** zz_SophieB is now known as SophieB 2010-09-21T21:02:43 *** jimmerton has quit IRC (Ping timeout: 276 seconds) 2010-09-21T21:02:43 *** jimmerton_ is now known as jimmerton 2010-09-21T21:04:38 *** jimmerton_ has joined #aichallenge 2010-09-21T21:05:25 *** Cyndre_ has joined #aichallenge 2010-09-21T21:05:49 *** Cyndre has quit IRC (Ping timeout: 276 seconds) 2010-09-21T21:06:42 *** jimmerton has quit IRC (Read error: Connection reset by peer) 2010-09-21T21:06:43 *** jimmerton_ is now known as jimmerton 2010-09-21T21:09:14 *** SophieB is now known as zz_SophieB 2010-09-21T21:09:15 *** jimmerton_ has joined #aichallenge 2010-09-21T21:11:52 *** jimmerton has quit IRC (Ping timeout: 272 seconds) 2010-09-21T21:11:52 *** jimmerton_ is now known as jimmerton 2010-09-21T21:14:11 *** MarcAndrePC has joined #aichallenge 2010-09-21T21:14:26 Hello - Allo 2010-09-21T21:16:01 *** zz_SophieB is now known as SophieB 2010-09-21T21:18:40 *** Mathnerd314 has quit IRC (Quit: ChatZilla 0.9.86-rdmsoft [XULRunner 1.9.2.8/20100722155716]) 2010-09-21T21:22:08 *** SophieB is now known as zz_SophieB 2010-09-21T21:24:51 *** Snowplan has joined #aichallenge 2010-09-21T21:25:19 is there a prob with the server Im getting connection refused 2010-09-21T21:25:43 prob just noobing it up 2010-09-21T21:27:02 *** zz_SophieB is now known as SophieB 2010-09-21T21:29:03 IP of tcp server was changed 2010-09-21T21:29:20 ok whats the new IP 2010-09-21T21:29:34 I didn't read that much into it 2010-09-21T21:29:43 read http://www.benzedrine.cx/planetwars/ 2010-09-21T21:29:47 thanks 2010-09-21T21:30:03 rankings 2010-09-21T21:30:09 (don't forget to change the port as well) 2010-09-21T21:30:11 @rankinigs 2010-09-21T21:30:12 sigh: I have no idea what you mean. 2010-09-21T21:30:17 @ranking 2010-09-21T21:30:19 sigh: Top 10 players: McLeopold(3506), kcm1700(3497), dmj111(3487), adumlah(3369), montanalow(3339), EBraun(3332), Tracer(3284), Arekku(3251), sighbot(3245), Willjarn(3221) 2010-09-21T21:30:27 Tempest: there you go :P 2010-09-21T21:30:51 thx 2010-09-21T21:32:18 lol forgot to change the port 2010-09-21T21:32:32 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T21:32:55 Snowplan: what did I say :P 2010-09-21T21:33:31 should change my name to noobplan 2010-09-21T21:34:12 nah, the only reason I mentioned that is because others made the same mistake 2010-09-21T21:36:46 I'm running off the java client is there a way to visualize 2010-09-21T21:37:45 damn thought my bot would do a bit better 2010-09-21T21:37:57 what's your rank? 2010-09-21T21:38:59 what to visualize? you can watch games in server 2010-09-21T21:39:41 945 now, no real AI work in my bot. Thought a rough version of my bot would at least be around 500 2010-09-21T21:40:38 alls I'm looking at on TCP is statements and you lose 2010-09-21T21:40:39 lol 2010-09-21T21:41:04 can I pipe the showgame 2010-09-21T21:41:49 http://72.44.46.68/getratings find yourself 2010-09-21T21:43:17 *** SophieB is now known as zz_SophieB 2010-09-21T21:43:18 Thanks 2010-09-21T21:44:52 you don't need to hunt through the ratings list, you can go to: http://72.44.46.68/getplayer?player= 2010-09-21T21:47:17 *** jimmerton_ has joined #aichallenge 2010-09-21T21:49:46 *** zz_SophieB is now known as SophieB 2010-09-21T21:50:10 *** jimmerton has quit IRC (Ping timeout: 276 seconds) 2010-09-21T21:50:11 *** jimmerton_ is now known as jimmerton 2010-09-21T21:50:18 once again thanks sigh. 2010-09-21T21:51:10 why is the tcp server so much better then the waterloo one 2010-09-21T21:51:57 the waterloo one has 1400 bots to play against each other 2010-09-21T21:52:02 it's a completely different architecture with completely different goals 2010-09-21T21:52:34 dufflebunk: 1884 bots, actually 2010-09-21T21:53:37 a single round robin would require over 3.5 million games 2010-09-21T21:53:45 and this number is still climbing 2010-09-21T21:53:56 s/1400/more bots than I can count on my fingers and toes/ 2010-09-21T21:53:56 oh i did that math very wrong though 2010-09-21T21:54:16 my bad, "only" 1.8 million games 2010-09-21T21:54:20 lol 2010-09-21T21:54:34 any idea on the prize yet 2010-09-21T21:54:47 * jmcarthur shrugs 2010-09-21T21:54:55 the prize could end up just being fame and glory 2010-09-21T21:55:04 and a million recruiters breathing down your neck :) 2010-09-21T21:55:20 great point 2010-09-21T21:55:38 Snowplan: a1k0n had that "problem" after the last contest 2010-09-21T21:56:27 It's always good to sweeten the pot. 2010-09-21T21:56:35 Top 10 players: McLeopold(3509), kcm1700(3504), dmj111(3492), adumlah(3376), EBraun(3350), montanalow(3335), Tracer(3283), sighbot(3249), Arekku(3242), Loks1982(3225) 2010-09-21T21:57:58 *** SophieB is now known as zz_SophieB 2010-09-21T22:01:35 *** jimmerton_ has joined #aichallenge 2010-09-21T22:02:20 Again I do thank all the people who made this contest. This is fun. 2010-09-21T22:03:08 *** icky has joined #aichallenge 2010-09-21T22:03:30 *** jimmerton has quit IRC (Ping timeout: 252 seconds) 2010-09-21T22:03:30 *** jimmerton_ is now known as jimmerton 2010-09-21T22:04:48 *** zz_SophieB is now known as SophieB 2010-09-21T22:05:42 *** FlagCapper has quit IRC (Quit: Page closed) 2010-09-21T22:10:07 lol I made a stupid bot and rank 228, but the work to make it better is so much bigger O_o 2010-09-21T22:11:55 I want to make a bot that can be stable in 300. Def pat yourself on the back 2010-09-21T22:13:01 everyone is making his bot better 2010-09-21T22:14:36 what is AI actually? 2010-09-21T22:14:39 is there a way to make the simulator quicker? 2010-09-21T22:14:45 it takes a couple of seconds per game on my machine 2010-09-21T22:14:48 why? 2010-09-21T22:14:52 seems very slow 2010-09-21T22:14:59 to run through a few hundred games 2010-09-21T22:15:15 jimmerton: make your bot faster :) 2010-09-21T22:15:33 hah, you think it's my bot that's slow? 2010-09-21T22:15:37 how many turns your games last? 2010-09-21T22:15:38 I assumed it was the game engine 2010-09-21T22:15:43 100-odd 2010-09-21T22:16:14 are you talking about the engine or the visualizer? 2010-09-21T22:16:18 the engine 2010-09-21T22:16:32 maybe try one of the c++ engines from the forums? 2010-09-21T22:16:33 java -jar tools/PlayGame 2010-09-21T22:16:37 ah I didn't see those 2010-09-21T22:16:41 I'll try one 2010-09-21T22:16:41 couple of seconds for a game is very fast, you are allowed to use 1 second per turn 2010-09-21T22:16:42 there's at least one 2010-09-21T22:16:51 not sure if the Qt one is also c++ (probably is) 2010-09-21T22:16:52 mine sucks 2010-09-21T22:17:00 my bot is pretty stupid Naktibalda 2010-09-21T22:17:06 so it should be quick 2010-09-21T22:17:11 what language is it? 2010-09-21T22:17:22 python 2010-09-21T22:17:33 my bot is going to run full game simulations many times per turn, so i'm working hard to make it pretty fast 2010-09-21T22:18:18 how many turns into the future are you expecting to simulate? 2010-09-21T22:18:29 why do you want to make it faster? It takes me a few minutes to analyze 1 game 2010-09-21T22:18:47 for bulk testing 2010-09-21T22:18:53 I want my bot to eventually realize this game is trivial and pursue global dominations 2010-09-21T22:19:03 sometimes I run a new version against all the bots on all the maps 2010-09-21T22:19:14 it takes about 10 minutes 2010-09-21T22:19:27 I'd like it to be quicker 2010-09-21T22:19:44 jimmerton: i'm simulating all the way to the end of the match, just not all possible matches 2010-09-21T22:20:26 as in, there is no branching factor 2010-09-21T22:20:42 how is there any value in that? 2010-09-21T22:20:49 i shan't tell 2010-09-21T22:20:53 heh, ok 2010-09-21T22:21:03 are you assuming the enemy will do what your bot would do? 2010-09-21T22:21:09 not at all 2010-09-21T22:21:26 i am making no assumptions about the enemy, actually 2010-09-21T22:21:37 good to know 2010-09-21T22:21:57 I can't see much value in evaluating branches which are unlikely to occur 2010-09-21T22:22:13 i assure you there is. i'm just not going to say what 2010-09-21T22:23:21 do you have a bot in the competition? 2010-09-21T22:23:26 I think a real AI thinks on its own thats why I'm running through cases to implement an indefinate strategy 2010-09-21T22:23:47 not yet 2010-09-21T22:24:07 I'm just hand coding sensible strategy for now 2010-09-21T22:24:15 I'll look at proper ai methods later 2010-09-21T22:24:21 when I have better understanding of the game myself 2010-09-21T22:24:24 heh, "proper" 2010-09-21T22:24:27 no such thing 2010-09-21T22:24:40 what I mean is 2010-09-21T22:24:44 machine learning algorithms 2010-09-21T22:24:48 ah 2010-09-21T22:24:52 that's an unusual approach 2010-09-21T22:24:56 *** Naith has quit IRC (Quit: Naith) 2010-09-21T22:25:03 do you think so? 2010-09-21T22:25:06 anyone use linux 2010-09-21T22:25:14 <-- uses linux 2010-09-21T22:25:25 I'm planning on hand coding most of the strategy and tweaking the value functions using machine learning 2010-09-21T22:25:30 linux and os x, what's up? 2010-09-21T22:25:33 It's actially awsome 2010-09-21T22:25:33 jimmerton: yeah. the last contest was won by search algorithms 2010-09-21T22:25:41 actually 2010-09-21T22:25:41 No one, ever, uses linux. why? 2010-09-21T22:26:03 i don't think search applies as strongly this time 2010-09-21T22:26:11 because the search space is so huge 2010-09-21T22:26:15 Defintely 2010-09-21T22:26:21 Tron was all about search 2010-09-21T22:26:23 I think the value functions will be important 2010-09-21T22:26:28 nonetheless, i am working with a search algorithm of my own 2010-09-21T22:26:38 the value of a planet based on the planets around it 2010-09-21T22:26:47 etc 2010-09-21T22:26:49 You could definitely turn this into a search problem... 2010-09-21T22:27:10 yeah you just have to have insane pruning, or don't generate moves that you have to prune in the first place... 2010-09-21T22:27:48 it also looks a lot like some sort of linear programming problem to me, but i haven't determined how to go about that approach yet 2010-09-21T22:28:11 the rules are just ever so slightly more complicated than i think straight up linear programming can tackle 2010-09-21T22:28:20 but maybe some generalization... 2010-09-21T22:28:35 might be too computationally expensive if it even exists 2010-09-21T22:29:11 and finding it would probably be thesis-worthy 2010-09-21T22:30:01 I've been trying to nut out some solid maths for this, but failing other than in very simple situations 2010-09-21T22:30:26 my attempts to micro-optimize my simulation code has proven to be surprisingly fruitful in this area 2010-09-21T22:30:33 of understanding the game on a more mathematical level, i mean 2010-09-21T22:30:57 hmmm... any hints? ;) 2010-09-21T22:31:16 that *was* a hint :P 2010-09-21T22:31:23 haha 2010-09-21T22:31:58 *** antiUni has joined #aichallenge 2010-09-21T22:32:24 optimization has the nice property that it's most effective when you can weed out redundant information and distill the problem to its core 2010-09-21T22:32:41 redundant and unnecessary information, that is 2010-09-21T22:32:54 yeah 2010-09-21T22:32:58 i love going back through and clearing out crap i originally thought i needed 2010-09-21T22:33:23 bot has finally made it a bit higher on the official leaderboard agian 2010-09-21T22:34:48 less then 4 planets my bot sucks 2010-09-21T22:34:53 *** hornairs has joined #aichallenge 2010-09-21T22:35:21 *** Appleman1234 has joined #aichallenge 2010-09-21T22:35:28 using this purely functional multidimensional array library has really made me think about the problem differently. i didn't quite expect that 2010-09-21T22:36:08 thinking about anything in a purely functional way makes you think about the problem differently 2010-09-21T22:36:35 well yeah, but i have been using haskell as my primary language for 3+ years now 2010-09-21T22:36:43 :) 2010-09-21T22:36:51 but this library is a really fresh look at multidimensional arrays for me, regardless 2010-09-21T22:37:08 *** SophieB is now known as zz_SophieB 2010-09-21T22:37:08 I keep meaning to get into haskell 2010-09-21T22:37:27 *** zz_SophieB is now known as SophieB 2010-09-21T22:38:20 i like how it emphasizes cache coherency by offering functions that pretty much exclusively operate on the "inner" dimensions of the arrays 2010-09-21T22:38:56 hmmm... hadn't thought of it that way 2010-09-21T22:39:05 interesting 2010-09-21T22:39:13 and to work on out dimensions you have to transpose the array, which is actually a free operation because it only changes the indexing function, but having to do that explicitly forces you to consider the cache effects you are about to inflict 2010-09-21T22:39:17 *on outer 2010-09-21T22:40:17 and then finally you can balance it all out with a force function that rewrites the array into a cache coherent form for the current index transformation 2010-09-21T22:40:22 antiUni: what rank is your bot? 2010-09-21T22:40:29 15 2010-09-21T22:40:33 "youwillfail" 2010-09-21T22:40:34 wow, nice 2010-09-21T22:40:47 haha... it's antimatroid1 2010-09-21T22:41:00 what sort of strategy are you using at this point? 2010-09-21T22:41:07 have you done any machine learning stuff? 2010-09-21T22:41:22 *** Naktibalda has quit IRC (Remote host closed the connection) 2010-09-21T22:44:20 no machine learning stuff, still stricly heuristic based off of information 2010-09-21T22:44:52 word 2010-09-21T22:45:29 planning to move beyond that soon though 2010-09-21T22:45:44 i too need to jump in and play around with haskell sometime, people seem to think it's wonderful 2010-09-21T22:46:22 is it possible to enter a haskell bot? 2010-09-21T22:46:29 what language are you using currently? 2010-09-21T22:47:14 i'm using c++, you can use haskell though 2010-09-21T22:47:18 i believe jmcarthur is 2010-09-21T22:48:03 *** icky has quit IRC (Ping timeout: 252 seconds) 2010-09-21T22:49:23 *** dmj111_ has joined #aichallenge 2010-09-21T22:49:33 *** dmj111_ is now known as dmj111 2010-09-21T22:49:40 yeah there are a few using haskell 2010-09-21T22:49:48 the server currently has a super crappy version of ghc though 2010-09-21T22:50:11 jmcarthur: how goes your campaign for binaries? 2010-09-21T22:50:37 sigh: no objections on the mailing list. i think it will go through. i just have to convince an admin to actually take the steps to allow it 2010-09-21T22:50:45 cool 2010-09-21T22:50:56 sigh: i had written a patch for it ages ago. just needs to be merged in and an entry added to the db 2010-09-21T22:51:57 jmcarthur: what is your role in this comp, you seem quite involved, but you aren't an admin? 2010-09-21T22:52:11 no i'm just a contributor 2010-09-21T22:52:18 the project being open source means it's easy to help out :) 2010-09-21T22:52:23 ah, i see... keep up the good work :) 2010-09-21T22:53:04 pretty much the only people who are disqualified from entering are the maintainers (those who have access to the machines running the bots and stuff) 2010-09-21T22:53:17 yeah, that's fair 2010-09-21T22:53:38 some of the maintainers are actually people who were originally going to compete but sacrificed that to assist with the technicals 2010-09-21T22:53:48 i do not envy them, but i am very grateful 2010-09-21T22:53:53 :) 2010-09-21T22:54:24 I think it's a great competition 2010-09-21T22:54:34 me too 2010-09-21T22:54:41 I'm trying to get more of my friends involved 2010-09-21T22:55:22 *** SophieB is now known as zz_SophieB 2010-09-21T22:55:44 I've got friends who hassel me when my rank starts dropping too low, but no one I know is participating 2010-09-21T22:55:50 *hassle 2010-09-21T22:56:10 McLeopold beat Arekku 2010-09-21T22:56:37 Top 10 players: McLeopold(3507), dmj111(3494), kcm1700(3493), adumlah(3366), montanalow(3331), EBraun(3321), Tracer(3266), sighbot(3247), Arekku(3225), Loks1982(3215) 2010-09-21T22:56:55 McLeopold beats everyone :P 2010-09-21T22:57:02 he didn't think he would beat Arekku 2010-09-21T22:57:21 why Arekku specifically? 2010-09-21T22:57:55 watch some of Arekku's games 2010-09-21T22:58:05 they a methodical murders 2010-09-21T22:58:08 *they are 2010-09-21T22:58:49 k, I'll study them tonight 2010-09-21T23:01:44 *** zz_SophieB is now known as SophieB 2010-09-21T23:06:39 *** SophieB is now known as zz_SophieB 2010-09-21T23:10:17 *** xvinyl has joined #aichallenge 2010-09-21T23:10:48 *** lemonedo has joined #aichallenge 2010-09-21T23:10:57 *** lemonedo has left #aichallenge 2010-09-21T23:11:10 *** zz_SophieB is now known as SophieB 2010-09-21T23:11:13 hey, McLeopold gave Arekku the first loss http://ai-contest.com/visualizer.php?game_id=4628636 2010-09-21T23:16:02 *** MarcAndrePC has quit IRC (Quit: Page closed) 2010-09-21T23:21:20 *** SophieB is now known as zz_SophieB 2010-09-21T23:24:43 *** jimmerton has quit IRC (Quit: jimmerton) 2010-09-21T23:27:20 *** zz_SophieB is now known as SophieB 2010-09-21T23:31:29 *** pc has joined #aichallenge 2010-09-21T23:32:06 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-21T23:32:11 *** SophieB is now known as zz_SophieB 2010-09-21T23:34:27 *** duallain has joined #aichallenge 2010-09-21T23:35:03 *** Alystair has joined #aichallenge 2010-09-21T23:38:19 *** zz_SophieB is now known as SophieB 2010-09-21T23:42:51 *** Appleman1234 has quit IRC (Quit: Leaving) 2010-09-21T23:43:23 *** Appleman1234 has joined #aichallenge 2010-09-21T23:44:49 *** SophieB is now known as zz_SophieB 2010-09-21T23:46:00 *** McLeopold has joined #aichallenge 2010-09-21T23:48:36 *** zz_SophieB is now known as SophieB 2010-09-21T23:49:06 *** duallain has quit IRC (Quit: Page closed) 2010-09-21T23:52:17 *** antiUni has quit IRC (Quit: Page closed) 2010-09-21T23:55:50 *** jmcarthur has quit IRC (Quit: WeeChat 0.3.0) 2010-09-21T23:56:36 Top 10 players: McLeopold(3499), dmj111(3489), kcm1700(3487), adumlah(3341), EBraun(3312), Arekku(3309), montanalow(3298), Tracer(3261), Loks1982(3249), paulus(3222)