2010-10-13T00:00:53 Top 10 players: rsergio(4088), Hazard(3941), novastar(3898), dmj111(3890), felixcoto(3871), davidjliu(3787), asavis(3786), cfaftw(3772), MariusCG(3763), savchenko(3750) 2010-10-13T00:03:07 *** yasith has joined #aichallenge 2010-10-13T00:08:02 *** yasith has quit IRC (Ping timeout: 252 seconds) 2010-10-13T00:12:07 *** rebelxt_ has quit IRC (Quit: Page closed) 2010-10-13T00:15:10 *** delt0r_ has joined #aichallenge 2010-10-13T00:16:32 *** delt0r___ has quit IRC (Ping timeout: 250 seconds) 2010-10-13T00:24:36 *** janzert_ has quit IRC (Quit: Ex-Chat) 2010-10-13T00:46:57 amstan: did you get a chance to make that schema dump? 2010-10-13T00:54:58 *** javagamer has quit IRC (Ping timeout: 245 seconds) 2010-10-13T01:00:54 Top 10 players: rsergio(4087), Hazard(3950), novastar(3896), dmj111(3887), felixcoto(3871), davidjliu(3785), asavis(3782), MariusCG(3767), cfaftw(3757), savchenko(3752) 2010-10-13T01:01:10 *** davidjliu has quit IRC (Quit: Page closed) 2010-10-13T01:06:11 janzert: i'm looking at it now 2010-10-13T01:06:36 schema or the patch tournament_manager? 2010-10-13T01:06:51 the patch 2010-10-13T01:07:18 ok, it's still untested cause I keep running into schema changes and finally gave up on guessing what it's really suppose to be 2010-10-13T01:07:27 *** iwantorganizatio has quit IRC (Quit: Page closed) 2010-10-13T01:08:22 *** JCS^ has quit IRC (Quit: \x00) 2010-10-13T01:09:43 janzert: hey 2010-10-13T01:09:46 sorry, i was afk 2010-10-13T01:09:50 what's the command? 2010-10-13T01:09:53 np 2010-10-13T01:10:07 *** javagamer has joined #aichallenge 2010-10-13T01:10:12 mysqldump -u username -p'password' --no-data contest 2010-10-13T01:10:24 mysqldump -u username -p'password' --no-data contest > new_schema 2010-10-13T01:10:36 janzert: it'll end up in home.hypertriangle.com/~alex/ 2010-10-13T01:10:40 thanks 2010-10-13T01:10:55 that's a very weak password 2010-10-13T01:11:16 janzert: i know man, get a better password in the future 2010-10-13T01:11:18 :) most common password out there in a lot of common password lists 2010-10-13T01:11:26 might as well join the crowd 2010-10-13T01:11:27 so, just going off the code 2010-10-13T01:11:44 everything looks good 2010-10-13T01:12:15 janzert: btw.. just use -p next time 2010-10-13T01:12:22 was a little worried about making the temp folder in the submissions dir, but then i read mkdtemp and it has prefix="tmp" as default 2010-10-13T01:12:27 having the password in bash history is not cool 2010-10-13T01:12:46 ahh, yeah that is true. 2010-10-13T01:13:09 yep 2010-10-13T01:13:15 janzert: refresh 2010-10-13T01:13:21 it makes names like "tmp123456" or so 2010-10-13T01:13:55 thanks much, amstan 2010-10-13T01:14:13 yeah, and that doesn't conflict since the non-tmp directories are numbers only 2010-10-13T01:14:22 just one thing i want you to do 2010-10-13T01:14:24 really minor 2010-10-13T01:14:29 get rid of the semicolon :) 2010-10-13T01:14:37 *** hellman has joined #aichallenge 2010-10-13T01:15:02 oops, how did that slip in there :P 2010-10-13T01:15:22 it was in the original 2010-10-13T01:15:36 ahh, yes 2010-10-13T01:15:40 not your fault :) and it does nothing anyway 2010-10-13T01:15:47 yep 2010-10-13T01:16:20 bah, I'm going to have to let my tower break down on its own now 2010-10-13T01:17:42 so yeah, looks good to me, but i haven't done any testing either 2010-10-13T01:18:43 I'll definitely test it and check it in on a branch for someone to do one more look over it 2010-10-13T01:19:46 I'll follow your lead though and just branch the backend directory, a little nicer that way for this 2010-10-13T01:20:55 yeah 2010-10-13T01:21:12 branch the backend directory of another branch, hahaha 2010-10-13T01:21:21 yep :) 2010-10-13T01:25:41 whee, 2 new tables and 3 new columns added to existing tables plus misc. other modifications to existing columns, various added indexes, etc. :) 2010-10-13T01:36:46 *** tapwater has quit IRC (Read error: Operation timed out) 2010-10-13T01:37:20 hmmm I made a bot that doesn't attack at all. I wonder if it will be good :) 2010-10-13T01:37:21 *** tapwater has joined #aichallenge 2010-10-13T01:38:04 it'll tie against copybot :) 2010-10-13T01:38:27 copybot? lol 2010-10-13T01:39:22 actually it's strange because my old bot was ranked 90 and this bot beat my old one most of the time... 2010-10-13T01:40:12 *** tapwater_ has joined #aichallenge 2010-10-13T01:41:17 *** hornairs has quit IRC (Remote host closed the connection) 2010-10-13T01:43:45 *** tapwater has quit IRC (Ping timeout: 260 seconds) 2010-10-13T01:43:46 *** tapwater_ is now known as tapwater 2010-10-13T01:45:16 what is the site to view the replay of the tcp server? 2010-10-13T01:48:17 *** tapwater_ has joined #aichallenge 2010-10-13T01:48:33 !tcp server 2010-10-13T01:48:34 antimatroid: I do not know about 'tcp server', but I do know about these similar topics: 'tcp-server', 'tcpserver' 2010-10-13T01:48:39 !tcpserver 2010-10-13T01:48:40 antimatroid: tcpserver = http://www.benzedrine.cx/planetwars/ 2010-10-13T01:50:29 *** tapwater has quit IRC (Ping timeout: 240 seconds) 2010-10-13T01:50:30 *** tapwater_ is now known as tapwater 2010-10-13T01:51:17 oh, elo scores have broken 4000 2010-10-13T01:51:20 didn't notice 2010-10-13T01:51:51 *** Palmik has joined #aichallenge 2010-10-13T01:53:31 does it mean something? 2010-10-13T01:54:26 can one planet send more then one fleet per turn? 2010-10-13T01:54:33 decipher: yes. 2010-10-13T01:54:38 ty 2010-10-13T01:55:11 Gudradain: in theory, if you are 500 points above someone you should win 95% of the time 2010-10-13T01:55:25 ah really 2010-10-13T01:55:59 so larger spread of elo scores implies larger range of skill 2010-10-13T01:56:17 contest started at 600 2010-10-13T01:56:38 nvm, 54 2010-10-13T01:57:00 Zannick: is bswolf still copybot (with bug fix)? 2010-10-13T01:57:07 yeah 2010-10-13T01:57:19 it's rank 2104! 2010-10-13T01:57:31 pretty far down :P 2010-10-13T01:57:45 you're above over 1000 people 2010-10-13T01:57:59 that's ridiculous 2010-10-13T02:00:53 *** hellman has quit IRC (Remote host closed the connection) 2010-10-13T02:00:54 Top 10 players: rsergio(4078), Hazard(3941), novastar(3890), dmj111(3878), felixcoto(3862), davidjliu(3776), asavis(3772), MariusCG(3750), savchenko(3748), cfaftw(3743) 2010-10-13T02:02:44 *** hellman has joined #aichallenge 2010-10-13T02:04:22 *** tapwater_ has joined #aichallenge 2010-10-13T02:06:38 *** tapwater has quit IRC (Ping timeout: 250 seconds) 2010-10-13T02:06:38 *** tapwater_ is now known as tapwater 2010-10-13T02:20:24 lulz "There's a long-standing bug in the x86 architecture that allows you to boot Windows." 2010-10-13T02:23:38 *** Gudradain has quit IRC (Quit: Page closed) 2010-10-13T02:26:00 *** decipher has quit IRC (Ping timeout: 265 seconds) 2010-10-13T02:43:20 *** P_Hansson has joined #aichallenge 2010-10-13T02:43:46 *** Frontier has joined #aichallenge 2010-10-13T02:43:50 *** sigh has quit IRC (Ping timeout: 264 seconds) 2010-10-13T02:49:08 how does server schedule games to be played? are recently uploaded bots played more often, or something like that? 2010-10-13T02:54:31 *** tty2 has joined #aichallenge 2010-10-13T02:55:38 the bot with the least number of games is chosen as player 1 and player 2 is chosen randomly weighted towards another player with a close rank 2010-10-13T02:55:50 *** tty2 has quit IRC (Changing host) 2010-10-13T02:55:50 *** tty2 has joined #aichallenge 2010-10-13T02:57:35 *** tty1 has quit IRC (Ping timeout: 252 seconds) 2010-10-13T03:00:26 and yes, top bots gets less games than avg bots 2010-10-13T03:00:38 plus new bot gets instantly one game 2010-10-13T03:00:54 Top 10 players: rsergio(4057), Hazard(3921), novastar(3872), felixcoto(3854), dmj111(3842), davidjliu(3766), asavis(3760), cfaftw(3744), MariusCG(3743), savchenko(3732) 2010-10-13T03:02:51 *** stevenheidel has joined #aichallenge 2010-10-13T03:06:41 You should add a "remember me" type feature to the logins on the website 2010-10-13T03:07:08 *** jukkerknott has joined #aichallenge 2010-10-13T03:11:47 *** stevenheidel has quit IRC (Remote host closed the connection) 2010-10-13T03:26:39 *** jaspervdj has joined #aichallenge 2010-10-13T03:27:55 *** jukkerknott has quit IRC (Quit: Page closed) 2010-10-13T03:36:33 *** boegel has joined #aichallenge 2010-10-13T03:42:23 *** P_Hansson has quit IRC (Ping timeout: 265 seconds) 2010-10-13T03:43:52 Zannick: in case you're still around the review is up for the tournament_manager changes 2010-10-13T03:44:39 http://code.google.com/p/ai-contest/issues/detail?id=200&show=review 2010-10-13T03:47:54 the game map extractor bookmarklet (http://ai-contest.com/forum/viewtopic.php?f=18&t=790 2010-10-13T03:48:05 ) doesn't work for me on the tcp server. 2010-10-13T03:48:12 how to beat dmj in 8 moves 2010-10-13T03:48:16 http://72.44.46.68/canvas?game_id=462673 2010-10-13T03:54:07 *** hellman has quit IRC (Remote host closed the connection) 2010-10-13T04:00:55 Top 10 players: rsergio(4066), Hazard(3927), novastar(3877), felixcoto(3872), dmj111(3851), davidjliu(3787), asavis(3766), MariusCG(3753), cfaftw(3744), sequoh(3737) 2010-10-13T04:09:16 *** wvdschel has joined #aichallenge 2010-10-13T04:17:23 *** Florian95 has joined #aichallenge 2010-10-13T04:24:39 *** Frontier has quit IRC (Ping timeout: 276 seconds) 2010-10-13T04:25:25 *** Doddzy has joined #aichallenge 2010-10-13T04:25:50 *** Frontier has joined #aichallenge 2010-10-13T04:46:14 *** sigh has joined #aichallenge 2010-10-13T04:56:45 *** Naktibalda has joined #aichallenge 2010-10-13T05:00:56 Top 10 players: rsergio(4069), Hazard(3932), felixcoto(3876), dmj111(3852), davidjliu(3790), novastar(3769), asavis(3769), cfaftw(3756), MariusCG(3749), sequoh(3739) 2010-10-13T05:03:46 *** Doddzy has quit IRC (Ping timeout: 255 seconds) 2010-10-13T05:03:52 *** tty1 has joined #aichallenge 2010-10-13T05:05:49 *** tty2 has quit IRC (Ping timeout: 245 seconds) 2010-10-13T05:16:55 *** Florian95 has quit IRC (Quit: Leaving...) 2010-10-13T05:20:35 *** Florian95 has joined #aichallenge 2010-10-13T05:29:38 *** byteroot has joined #aichallenge 2010-10-13T05:30:20 Hi, did fights sopped from yesturday just for me, or for everyone ? 2010-10-13T05:31:08 i get one game if i upload and that's it 2010-10-13T05:31:19 try tcp server if you want to test your bot 2010-10-13T05:32:37 Thanks. 2010-10-13T05:34:22 what time is it now in server? 2010-10-13T05:34:34 I got too slow pc here for tcp server =) 2010-10-13T05:35:02 my last games: Oct 12 12:52:52 PM, Oct 12 04:17:50 PM, Oct 12 07:51:30 PM, Oct 13 02:07:04 AM, Oct 13 04:06:40 AM 2010-10-13T05:35:08 Forum says It is currently Wed Oct 13, 2010 8:33 am. 2010-10-13T05:35:09 very long breaks 2010-10-13T05:35:21 Yep, seems we all have that break 2010-10-13T05:35:46 *** mega1 has quit IRC (Ping timeout: 265 seconds) 2010-10-13T05:35:50 Naktibalda, top php bot ? 2010-10-13T05:35:57 no 2010-10-13T05:36:09 *** vizier has joined #aichallenge 2010-10-13T05:36:33 err, I missed a word php :) 2010-10-13T05:36:39 =)) 2010-10-13T05:37:32 no competition here 2010-10-13T05:38:18 GenericBob had a higher ranking for a few times, but he didn't submit since 28th of September 2010-10-13T05:38:27 yep 2010-10-13T05:39:40 Did you ever been timeouted, or server speed is enough ? 2010-10-13T05:40:02 Don't think php is fastest choice =) 2010-10-13T05:40:51 I'm not doing enough calculations to have such issue 2010-10-13T05:42:04 I had, at my virtual pc, but it is really slow and low mem. 2010-10-13T05:46:07 where is that server status page? 2010-10-13T05:47:11 don't think there any 2010-10-13T05:48:15 there is a page, that displays an average number of games per minute (+ average of last hour, 24hours) on each servers 2010-10-13T05:48:24 server 2010-10-13T05:49:04 the main server can generate ~6 games per minute 2010-10-13T05:49:32 hm 2010-10-13T05:49:36 2 other servers are some kind of virtual Amazon servers that works only when computing time is cheap 2010-10-13T05:51:10 please post link here if you find it 2010-10-13T05:52:41 http://www.ai-contest.com/game_stats.php 2010-10-13T05:53:21 nice, thanks a lot 2010-10-13T05:53:55 27 games last 24 hours =( 2010-10-13T05:54:02 seems something is broken 2010-10-13T06:00:56 Top 10 players: rsergio(4066), Hazard(3929), felixcoto(3873), dmj111(3854), davidjliu(3785), cfaftw(3752), asavis(3748), MariusCG(3746), sequoh(3742), savchenko(3733) 2010-10-13T06:04:51 *** tapwater has quit IRC (Quit: tapwater) 2010-10-13T06:25:27 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-10-13T06:28:20 *** iFire has joined #aichallenge 2010-10-13T06:30:55 I think they migh tonly run games when the ec2 slots are a certain price 2010-10-13T06:31:16 *** Meatkat has joined #aichallenge 2010-10-13T06:32:00 wow, 2 games in 12 hours... a new record! /wine 2010-10-13T06:32:13 ou bien /whine? 2010-10-13T06:33:05 I really wish I could play on the tcp server... :( /whine 2010-10-13T06:33:20 <-- needs to stop 2010-10-13T06:34:38 *** JensT1 has quit IRC (Quit: Leaving.) 2010-10-13T06:35:47 *** hellman has joined #aichallenge 2010-10-13T06:40:22 *** hellman has quit IRC (Remote host closed the connection) 2010-10-13T06:45:18 *** jesionaj has quit IRC (Ping timeout: 240 seconds) 2010-10-13T06:49:10 =) 2010-10-13T06:54:40 *** Olathe has joined #aichallenge 2010-10-13T07:00:57 Top 10 players: rsergio(4070), Hazard(3933), felixcoto(3877), dmj111(3858), davidjliu(3789), cfaftw(3756), asavis(3752), MariusCG(3750), sequoh(3746), savchenko(3737) 2010-10-13T07:08:12 Meatkat: why can't you? 2010-10-13T07:09:17 *** Flavorador has joined #aichallenge 2010-10-13T07:11:14 brx: my connection is terrible, almost every game I time out 2010-10-13T07:11:31 brx: I'm in Burkina Faso 2010-10-13T07:27:58 oh, hrm. 2010-10-13T07:28:17 *** BtbN has joined #aichallenge 2010-10-13T07:32:21 *** luy has joined #aichallenge 2010-10-13T07:34:44 *** Flavorador has quit IRC (Quit: jmIrc-m v0.35e by Archangel (http://jmirc-m.net.ru/)) 2010-10-13T07:38:04 Meatkat: why you can't play on tcp? 2010-10-13T07:38:21 oh 2010-10-13T07:38:24 just asked :D 2010-10-13T07:38:50 so you need somekinda proxy tcp :) 2010-10-13T07:39:22 send your sourcecode to that which compiles it and runs with tcp and gives results 2010-10-13T07:40:49 use a shell host (free or payed) 2010-10-13T07:41:11 nice idea 2010-10-13T07:44:02 I had network issues too before they increased timeout 2010-10-13T07:44:52 I used shellmix.com for 2 weeks. It's the only freeshell server that has an easy registration 2010-10-13T07:46:02 I couldn't find another when shellmix.com was down 2010-10-13T07:46:29 *** Olathe has quit IRC (Ping timeout: 240 seconds) 2010-10-13T07:47:37 *** Olathe has joined #aichallenge 2010-10-13T07:52:15 *** mega1 has joined #aichallenge 2010-10-13T07:54:15 *** Olathe has quit IRC (Ping timeout: 240 seconds) 2010-10-13T08:00:57 Top 10 players: rsergio(4075), Hazard(3938), felixcoto(3882), dmj111(3863), davidjliu(3794), cfaftw(3762), asavis(3757), MariusCG(3756), sequoh(3752), savchenko(3742) 2010-10-13T08:09:28 *** mega1_ has joined #aichallenge 2010-10-13T08:09:29 *** mega1 has quit IRC (Read error: Connection reset by peer) 2010-10-13T08:10:46 Hey, I just noticed it's possible to send fleets of size 0, is this supposed to happen? 2010-10-13T08:12:03 *** krokokrusa has joined #aichallenge 2010-10-13T08:12:10 javagamer: no... i think it will be or is a loss to do that. 2010-10-13T08:12:59 dmj111, alright, I'll fix my bot up so it stops 2010-10-13T08:13:22 it's ignored by server, but a valid move 2010-10-13T08:13:30 i think that too 2010-10-13T08:13:37 Naktibalda: thanks... i wasn't sure. 2010-10-13T08:13:54 but, if you send 0 size fleet and then on the same turn 1 size fleet to same planet 2010-10-13T08:13:57 then it's invalid move 2010-10-13T08:14:01 i think.. 2010-10-13T08:14:02 :) 2010-10-13T08:14:55 why? 2010-10-13T08:15:03 server merges fleets 2010-10-13T08:15:20 you can send more than 1 fleet from a to b 2010-10-13T08:15:21 Naktibalda, does the server merge fleets now? 2010-10-13T08:15:39 *** delt0r___ has joined #aichallenge 2010-10-13T08:16:23 Naktibalda: Thanks for the idea, I'll check out shellmix 2010-10-13T08:17:16 *** delt0r_ has quit IRC (Ping timeout: 255 seconds) 2010-10-13T08:17:22 javagamer: I think that I read it somewhere. I'm not 100% sure, because it isn't important to me :) 2010-10-13T08:19:29 *** perror has joined #aichallenge 2010-10-13T08:23:26 *** Frontier has left #aichallenge 2010-10-13T08:33:47 *** jorge has joined #aichallenge 2010-10-13T08:36:46 *** luy has quit IRC (Quit: Page closed) 2010-10-13T08:39:11 *** Acount has quit IRC (Ping timeout: 240 seconds) 2010-10-13T08:41:35 *** acount has joined #aichallenge 2010-10-13T08:48:00 Hmmm... it seems the TCP server is doing something weird, it's sending me two turn 0s 2010-10-13T08:48:23 Wait, nm! 2010-10-13T08:56:52 Yes! Just won my first TCP game. 2010-10-13T08:57:00 Oh, he timed out :/ 2010-10-13T08:57:58 *** aerique has joined #aichallenge 2010-10-13T09:00:07 Well, at least I got a legitimate win against RageBot 2010-10-13T09:00:58 Top 10 players: rsergio(4068), Hazard(3932), felixcoto(3876), dmj111(3857), davidjliu(3788), cfaftw(3756), asavis(3751), MariusCG(3749), sequoh(3745), savchenko(3736) 2010-10-13T09:03:04 *** HakanD has joined #aichallenge 2010-10-13T09:10:30 *** delt0r___ has quit IRC (Ping timeout: 240 seconds) 2010-10-13T09:10:47 *** delt0r_ has joined #aichallenge 2010-10-13T09:11:24 *** delt0r_ has joined #aichallenge 2010-10-13T09:11:58 *** Appleman1234 has quit IRC (Quit: Leaving) 2010-10-13T09:16:26 *** n has joined #aichallenge 2010-10-13T09:16:52 *** n is now known as Guest83817 2010-10-13T09:19:34 *** HakanD has quit IRC (Ping timeout: 240 seconds) 2010-10-13T09:26:58 *** Guest83817 has quit IRC (Quit: Page closed) 2010-10-13T09:35:00 *** Azrathud has quit IRC (Ping timeout: 245 seconds) 2010-10-13T09:49:49 *** phreeza has joined #aichallenge 2010-10-13T09:50:05 hi all 2010-10-13T09:50:07 been a while 2010-10-13T09:58:07 any groundbreaking new ideas come up in the last month or are the bots still just bags of assorted heuristics? :) 2010-10-13T09:58:56 nothing new 2010-10-13T09:59:10 *** hellman has joined #aichallenge 2010-10-13T09:59:54 well, maybe somebody manage to sort heuristics :) 2010-10-13T10:00:58 Top 10 players: rsergio(4079), Hazard(3944), felixcoto(3887), dmj111(3868), davidjliu(3799), cfaftw(3768), asavis(3762), MariusCG(3761), sequoh(3757), savchenko(3747) 2010-10-13T10:01:08 *** rebelxt_ has joined #aichallenge 2010-10-13T10:02:38 hey all, any update on server slowdown? 2010-10-13T10:06:44 not here 2010-10-13T10:07:50 :( cloud servers keep falling down for some reason 2010-10-13T10:07:53 *** wh1teside has quit IRC (Ping timeout: 276 seconds) 2010-10-13T10:11:14 does the contest run on amazon or rackspace or something now? 2010-10-13T10:11:32 or still on UW servers? 2010-10-13T10:12:20 they use Amazon EC2 when it's cheap 2010-10-13T10:12:28 amazon EC2 2010-10-13T10:12:35 probably it didnt hapeen for more than 20 hours 2010-10-13T10:12:44 ahh, so do they use spot instances? 2010-10-13T10:13:34 "Last 24 hours" average is dropping fast, so it's getting close to 24hours without EC2 2010-10-13T10:15:21 current spot price on small instances is 0.03$ per hour which is normal, so i think there is another problem - not pricing/availability of servers 2010-10-13T10:17:36 *** sigh has quit IRC (Remote host closed the connection) 2010-10-13T10:19:35 contestbot: later tell amstan have not seen any cloud servers online for 20+ hours. are you guys aware of the issue? thanks 2010-10-13T10:19:35 rebelxt_: Ready to serve, my lord. 2010-10-13T10:19:42 thats pretty cool, didn't know you could do that... buy by the hour depending on proce 2010-10-13T10:19:45 price 2010-10-13T10:21:06 phreeza: yeah it's a great model. really cheap too. but there is always a risk your instance can get shut down if market price exceeds price threshold you set 2010-10-13T10:22:13 rebelxt_: can you also specify a budget? if so it would be very neat to coule it with a paypal donation button... 2010-10-13T10:22:27 want to have more plays? hit the donate button :) 2010-10-13T10:23:17 hahaha 2010-10-13T10:23:42 we fortunetly have tcp server with unlimited free play ;-) 2010-10-13T10:23:53 *** JamesMG has joined #aichallenge 2010-10-13T10:24:38 rebelxt_ true 2010-10-13T10:25:42 *** wh1teside has joined #aichallenge 2010-10-13T10:28:10 wow the tcp server has really grown since i was last working on this... 2010-10-13T10:36:15 *** cfaftw has joined #aichallenge 2010-10-13T10:38:18 *** hellman has quit IRC (Remote host closed the connection) 2010-10-13T10:38:24 *** aerique has quit IRC (Quit: ...) 2010-10-13T10:41:32 *** Frontier has joined #aichallenge 2010-10-13T10:55:10 *** JensTi has joined #aichallenge 2010-10-13T10:56:40 Naktibalda: I'm on shellmix but it isn't letting me execute binaries. Is there a way around that? 2010-10-13T10:56:54 oh really? 2010-10-13T10:57:25 =>Wed 13 16:47<= meatkat(at)shellmix ~/PlanetWars/PlanetWars> ./tcp 2010-10-13T10:57:34 works for me 2010-10-13T10:57:36 ELF binary type "0" not known. 2010-10-13T10:57:43 compile it in shellmix 2010-10-13T10:57:44 ./tcp: Exec format error. Binary file not executable. 2010-10-13T10:59:00 javac TCP.java 2010-10-13T10:59:02 Error occurred during initialization of VM 2010-10-13T10:59:02 Could not reserve enough space for object heap 2010-10-13T10:59:02 javac in free(): warning: junk pointer, too high to make sense 2010-10-13T10:59:02 Could not create the Java virtual machine. 2010-10-13T10:59:45 I can give you my tcp binary file (not java) 2010-10-13T10:59:48 I've got the other file, what is it? gcc -c tcp.c? 2010-10-13T10:59:53 the c version 2010-10-13T11:00:07 gcc -o tcp tcp.c 2010-10-13T11:00:59 Top 10 players: rsergio(4092), Hazard(3935), felixcoto(3891), dmj111(3867), davidjliu(3795), MariusCG(3765), asavis(3761), sequoh(3753), cfaftw(3750), savchenko(3742) 2010-10-13T11:01:32 acount: are you Accoun on the forums? 2010-10-13T11:01:37 *** boegel has quit IRC (Ping timeout: 252 seconds) 2010-10-13T11:01:44 delt0r_: yes 2010-10-13T11:02:07 mmmm 2010-10-13T11:02:11 :) 2010-10-13T11:02:12 *** amstan sets mode: -q *!*@213.88.2.78 2010-10-13T11:03:15 hey i was gonna be the good guy! 2010-10-13T11:03:22 sweet, it runs now! I tried putting it in a run.sh file and it says it doesn't want to run that 2010-10-13T11:03:34 ./tcp_play.sh & 2010-10-13T11:03:35 [1] 29611 2010-10-13T11:03:35 ./tcp_play.sh: Command not found. 2010-10-13T11:03:38 any idea why? 2010-10-13T11:03:54 have you such file? 2010-10-13T11:04:00 indeed I do 2010-10-13T11:04:06 run.sh != tcp_play.sh 2010-10-13T11:04:09 amstan: he/she is well known? 2010-10-13T11:04:36 right, sorry I meant tcp_play.sh 2010-10-13T11:05:05 chmod it 2010-10-13T11:05:08 *** JensTi has quit IRC (Quit: Leaving.) 2010-10-13T11:06:35 delt0r_: he's known like this 2010-10-13T11:06:40 contestbot: quote accoun 2010-10-13T11:06:40 amstan: all book about atomic nuclear bomp prodiction writeln in russian )))) 2010-10-13T11:06:46 Naktibalda: already done, i tried bashing it and it started to run, but game me an error with my python file. I'm going to recompile it and see if that works. Any idea why bash would work but ./run.sh wouldn't? 2010-10-13T11:06:56 lol 2010-10-13T11:07:12 contestbot: quote delt0r 2010-10-13T11:07:12 delt0r_: Error: I couldn't find a matching quotegrab for delt0r. 2010-10-13T11:07:15 delt0r_: also like this http://pastebin.com/7GnS6cmw 2010-10-13T11:07:47 nuf said 2010-10-13T11:08:01 what i was suspecting..... 2010-10-13T11:12:49 *** deepblue has joined #aichallenge 2010-10-13T11:16:40 1 2010-10-13T11:16:58 *** acount is now known as Accoun 2010-10-13T11:17:02 пидпрасы 2010-10-13T11:18:32 Accoun: is your bot running on the TCP server right now? 2010-10-13T11:19:09 *** rebelxt_ has quit IRC (Ping timeout: 265 seconds) 2010-10-13T11:19:32 JamesMG> Y 2010-10-13T11:20:15 *** wvdschel has quit IRC (Quit: Leaving) 2010-10-13T11:21:33 Naktibalda: so the server seems very against running my .pyc or .pyo files, even when I've chmod +x to them is there a way around that? 2010-10-13T11:21:55 Meatkat: have you a proper #! line? 2010-10-13T11:22:11 err, I don't know python, can't help 2010-10-13T11:22:24 Naktibalda: would that have to be in my MyBot.py file? 2010-10-13T11:23:42 yeah, it's probably related to the #! line, right now mine is set at /user/bin/env python, which I imagine I don't have access to on the server 2010-10-13T11:23:58 /usr/ 2010-10-13T11:26:17 *** hellman has joined #aichallenge 2010-10-13T11:26:56 *** Gudradain has joined #aichallenge 2010-10-13T11:35:48 Naktibalda: /usr/bin/env python - doesn't work I'll google around and see if I can find anything, rawr this is annoying 2010-10-13T11:38:38 *** jorge_ has joined #aichallenge 2010-10-13T11:38:41 game are slow right now, only 3 game in 11 hours... 2010-10-13T11:39:09 amstan: did you svn update compile_anything.py after the lisp bugfix yesterday? 2010-10-13T11:39:26 nobody told me 2010-10-13T11:39:37 amstan: could you do that now? :) 2010-10-13T11:40:54 rev 432 2010-10-13T11:42:05 Gudradain: me too, really slow :( 2010-10-13T11:42:10 *** byteroot has quit IRC (Quit: Page closed) 2010-10-13T11:42:25 yup I wonder what the problem is 2010-10-13T11:42:59 amstan: can you explain what happened to servers 2010-10-13T11:43:04 i have no idea 2010-10-13T11:43:13 who has? 2010-10-13T11:43:22 jeff maybe 2010-10-13T11:43:25 *** Frontier has left #aichallenge 2010-10-13T11:43:34 how do you send a message to someone like -> name: message? 2010-10-13T11:43:35 !see j3camero 2010-10-13T11:43:36 Naktibalda: I'm sorry Dave, I cannot see. 2010-10-13T11:43:40 !seen j3camero 2010-10-13T11:43:40 Naktibalda: j3camero was last seen in #aichallenge 1 week, 1 day, 13 hours, 23 minutes, and 54 seconds ago: amstan: yes, but no names, haha. 2010-10-13T11:44:49 *** lavalamp has joined #aichallenge 2010-10-13T11:45:16 Man, contestbot is the ultimate IRC stalker 2010-10-13T11:45:29 JamesMG: of course, it logs everything 2010-10-13T11:45:46 *** rebelxt has joined #aichallenge 2010-10-13T11:46:11 amstan: thanks 2010-10-13T11:47:43 !exists test 2010-10-13T11:47:43 deepblue: User error -- Replace user. 2010-10-13T11:47:55 lol 2010-10-13T11:48:12 I hoped for a 'I'm sorry Dave, I cannot exist.' 2010-10-13T11:48:12 ;-) 2010-10-13T11:48:14 reply 2010-10-13T11:49:31 @error 2010-10-13T11:49:31 amstan: Run as fast as you can and don't look back. 2010-10-13T11:50:23 :D 2010-10-13T11:51:58 anyone is playing on tcp right now? 2010-10-13T11:52:13 @error 2010-10-13T11:52:14 Gudradain: User error -- Replace user. 2010-10-13T11:52:22 meh 2010-10-13T11:52:58 *** JensTi has joined #aichallenge 2010-10-13T11:53:44 Gudradain: yes me 2010-10-13T11:53:59 what's your name? 2010-10-13T11:54:07 phreeza :) 2010-10-13T11:54:15 ok MarcAndrePC on tcp 2010-10-13T11:55:08 ok and? 2010-10-13T11:55:08 oh hey, I just played you 2010-10-13T11:55:22 agains me? 2010-10-13T11:55:40 *** jukkerknott has joined #aichallenge 2010-10-13T11:57:33 Gudradain: interesting, you try to keep your front line planets at around 100? against JamesMG that led to a kind of stalemate 2010-10-13T11:58:33 *** JensTi1 has joined #aichallenge 2010-10-13T11:58:58 that game is making me think I broke my expansion code 2010-10-13T11:59:05 phreeza: but I was winning 2010-10-13T11:59:15 Gudradain: yep 2010-10-13T11:59:22 I should've picked up that 90 ship planet on the bottom right to bring it even at least 2010-10-13T11:59:53 I just remove all my attacking code and just put defense and expansion, than I just send the overstock (over 100) to the nearest planet 2010-10-13T12:00:09 really dumb but it beat my old bot that was rank 90... 2010-10-13T12:00:19 I wonder why 2010-10-13T12:00:26 my expansion code also screwed up on turn 21 2010-10-13T12:00:35 handed you a planet for basically free, hehe 2010-10-13T12:00:59 Top 10 players: rsergio(4101), Hazard(3936), felixcoto(3901), dmj111(3879), davidjliu(3805), MariusCG(3791), sequoh(3773), asavis(3771), ruilov(3762), savchenko(3756) 2010-10-13T12:01:06 yea ty 2010-10-13T12:01:38 *** JensTi has quit IRC (Ping timeout: 276 seconds) 2010-10-13T12:02:23 4101 elo, it's ridiculous 2010-10-13T12:02:55 *** Azrathud has joined #aichallenge 2010-10-13T12:03:29 yeaj he 2010-10-13T12:03:31 it's high, but not clear if rsergio can hold it. servers are not playing any games at the moment :( 2010-10-13T12:03:33 he's quite good 2010-10-13T12:06:09 It's strange because Hazard and rsergio are playing completely differently but are both on top 2010-10-13T12:06:29 Hazard is all on attack, while rsergio is a bit more defensive 2010-10-13T12:07:26 it's nice to see different strategies work well 2010-10-13T12:08:34 next it will be galaxy wars 2010-10-13T12:09:35 I want to see the map size increased 2010-10-13T12:09:49 :D 2010-10-13T12:09:50 and amounts of planets 2010-10-13T12:09:54 like 1000 planets or so 2010-10-13T12:09:58 that would be interesting to see 2010-10-13T12:10:06 yea 2010-10-13T12:10:08 deepblue: map_size *= 1000000 ;-) 2010-10-13T12:10:13 would be nice but it would mess up my strat 2010-10-13T12:10:14 but how long to play a game 2010-10-13T12:10:16 would see a lot of bots timeout at first I'm sure 2010-10-13T12:10:20 deepblue: my bot would time out on 2nd turn! 2010-10-13T12:10:28 and how long before the server even got through a round 2010-10-13T12:11:28 i know 2010-10-13T12:11:32 but I want it! 2010-10-13T12:11:38 I would need to rewrite my bot too :) 2010-10-13T12:11:39 build your own server 2010-10-13T12:11:47 deepblue, start your own server 2010-10-13T12:11:54 and only puts maps that have 100+ planets 2010-10-13T12:11:56 good idea 2010-10-13T12:12:08 hehe 2010-10-13T12:12:08 and 2500 turns 2010-10-13T12:12:08 set max turns to 1000 or so 2010-10-13T12:12:13 http://72.44.46.68/canvas?game_id=468316 2010-10-13T12:12:17 * Naktibalda was typing the same thing 2010-10-13T12:12:24 lol so tigh 2010-10-13T12:13:52 Interesting, looks like your direct attacks instead of redirected were more efficient? 2010-10-13T12:14:33 he doesn't send all his fleet like me 2010-10-13T12:26:54 *** Blkt has joined #aichallenge 2010-10-13T12:28:24 *** Bobng has joined #aichallenge 2010-10-13T12:29:16 meh there is aikon 0.4.1 and aikon 0.2 2010-10-13T12:30:21 hehe 2010-10-13T12:33:31 is there a way to cancel a fight on the tcp if you dont want to fight agains someone? 2010-10-13T12:33:41 control C 2010-10-13T12:33:44 but you'll loose 2010-10-13T12:34:22 if I see many of my game end before turn 1 does it mean the fight was cancel or just some connection problem? 2010-10-13T12:34:49 the list of your games will tell you if it was disconnection or not 2010-10-13T12:35:05 where do I get this list? 2010-10-13T12:35:16 http://72.44.46.68/getplayer?player=deepblue 2010-10-13T12:35:37 http://72.44.46.68/getratings 2010-10-13T12:37:09 ty 2010-10-13T12:38:09 I guess it was timeout :( 2010-10-13T12:39:57 *** jukkerknott has quit IRC (Quit: Page closed) 2010-10-13T12:41:30 *** Azrathud has quit IRC (Ping timeout: 240 seconds) 2010-10-13T12:47:19 *** jorge___ has joined #aichallenge 2010-10-13T12:48:05 *** jorge_ has quit IRC (Ping timeout: 264 seconds) 2010-10-13T12:49:22 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-10-13T12:49:34 *** JensTi1 has quit IRC (Quit: Leaving.) 2010-10-13T12:55:28 *** amstan_ has joined #aichallenge 2010-10-13T12:55:28 *** ChanServ sets mode: +o amstan_ 2010-10-13T13:00:59 Top 10 players: rsergio(4093), Hazard(3928), felixcoto(3893), dmj111(3871), davidjliu(3797), MariusCG(3783), sequoh(3765), asavis(3763), ruilov(3750), savchenko(3748) 2010-10-13T13:01:40 *** amstan_ has quit IRC (Ping timeout: 240 seconds) 2010-10-13T13:03:14 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-10-13T13:04:24 *** hellman has quit IRC (Remote host closed the connection) 2010-10-13T13:06:11 *** hellman has joined #aichallenge 2010-10-13T13:12:52 *** jorge___ has quit IRC (Remote host closed the connection) 2010-10-13T13:13:03 *** tapwater has joined #aichallenge 2010-10-13T13:17:29 *** Florian95 has quit IRC (Remote host closed the connection) 2010-10-13T13:22:04 *** bhasker has joined #aichallenge 2010-10-13T13:22:07 morning 2010-10-13T13:23:10 yo 2010-10-13T13:23:22 where are you that it's morning? west coast? 2010-10-13T13:26:24 not really morning there, it's like 11 I think 2010-10-13T13:26:30 you on tcp atm cfaftw? 2010-10-13T13:26:37 meatkat, no 2010-10-13T13:27:02 I finally got mine working! I'm filled with joy 2010-10-13T13:27:11 and my bot fell off the top 10 :( 2010-10-13T13:27:20 but... that bot was defective anyway 2010-10-13T13:27:21 My current bot on there isn't that good, but just having him up there makes me happy 2010-10-13T13:27:23 oh noes! 2010-10-13T13:27:24 it's more of a status thing, haha 2010-10-13T13:27:31 indeed 2010-10-13T13:27:45 and now with games being played so slowly... it'll take years to get back up there 2010-10-13T13:28:12 glad you got it working 2010-10-13T13:28:31 thanks, here he is -> http://72.44.46.68/getplayer?player=Meatkat_15.3 2010-10-13T13:29:08 yeah i already started watching some of your games :) 2010-10-13T13:29:58 haha, the one where I beat CCCF-ex9 is interesting 2010-10-13T13:30:16 you are about to get another win 2010-10-13T13:30:23 I disabled my "don't send too many from one planet" calculations for this version, I need to put that back in so I stop looking like a noob 2010-10-13T13:31:05 JamesMG: why do you say that? 2010-10-13T13:31:11 I know my bot :) 2010-10-13T13:32:23 your the honolulu series? 2010-10-13T13:32:43 no 2010-10-13T13:32:48 Meatkat, win against CCCF was good one 2010-10-13T13:33:41 yup west coast here 2010-10-13T13:34:00 rebelxt: thanks 2010-10-13T13:34:17 my bot has gotten a little better but still have bugs to fix :-\ 2010-10-13T13:34:30 *** vizier has quit IRC (Ping timeout: 265 seconds) 2010-10-13T13:34:34 Meatkat: bug you need to fix http://72.44.46.68/canvas?game_id=469340 2010-10-13T13:36:12 bhasker, yeah he said he took out that code for this version 2010-10-13T13:36:15 bhasker: yeah... hence my previous comment about disabling my "don't send too many away from one planet" and "looking like a noob" lol 2010-10-13T13:37:25 aah :D 2010-10-13T13:38:04 actually some of the top bots seem to lack any control over sending too many ships from one planet 2010-10-13T13:38:23 meaning, on turn 1 they send all their ships 2010-10-13T13:38:34 but they are only like 4 turns from the enemy starting planet 2010-10-13T13:38:40 instant death 2010-10-13T13:38:57 cfaftw: profit! 2010-10-13T13:39:21 $_$ 2010-10-13T13:39:23 otherwise your bot kicked my bots ass mostly:-\ 2010-10-13T13:39:44 also i guess the bots on tcp have gotten much better 2010-10-13T13:39:49 elo rating climb is slow and painful 2010-10-13T13:39:55 *** lavalamp has quit IRC (Quit: Page closed) 2010-10-13T13:39:56 bots are constantly getting better 2010-10-13T13:40:21 yea my bot has been climbing like 5 elo points every 15-20 minutes or so 2010-10-13T13:40:36 cfaftw: not true, my bot is getting worse! ;-) 2010-10-13T13:41:03 http://72.44.46.68/canvas?game_id=469644 2010-10-13T13:41:11 lol 2010-10-13T13:41:13 ^ 2010-10-13T13:41:15 well... as a general rule :P 2010-10-13T13:41:38 cfaftw: where is your jetbot? i wanna try my new suicidal bot against it 2010-10-13T13:41:54 ah jetbot is your cfaftw interesting bot 2010-10-13T13:42:00 what was the reason for the name 2010-10-13T13:42:03 give me a minute... i'll put it up on the TCP server 2010-10-13T13:42:31 bhasker, a long time ago, some my friends nicknamed me "The Jet" 2010-10-13T13:42:44 so i just named one of my versions after that 2010-10-13T13:42:45 for no reason 2010-10-13T13:43:44 *** lavalamp has joined #aichallenge 2010-10-13T13:44:16 who is lovebot? 2010-10-13T13:44:17 any idea whose bot is "lovebot"? he seems to be running 24x7 so lots of games against him 2010-10-13T13:44:22 Meatkat: haha 2010-10-13T13:44:37 *** Naktibalda has joined #aichallenge 2010-10-13T13:44:47 rebelxt: jinx! 2010-10-13T13:45:11 cfaftw: have you seen this video about Benny "The Jet" Urquidez? :) http://www.youtube.com/watch?v=ZT-Lkqz_dnk 2010-10-13T13:45:16 Meatkat: you are still running your bot, right? i started mine a few minutes ago, hope we get a game or two 2010-10-13T13:45:29 no, but i will 2010-10-13T13:45:52 rebelxt: yup, I think I put it to run 1000 games 2010-10-13T13:45:53 lovebot is quite good 2010-10-13T13:45:58 so... it'll be there for a while 2010-10-13T13:45:58 keeps kicking my bots ass :( 2010-10-13T13:46:15 i've gotta admit i'm scared of jetbot showing up in a few minutes and stomping all over my buggy bot ;-) 2010-10-13T13:46:30 eh... mine's not that great anymore 2010-10-13T13:46:38 but we'll see 2010-10-13T13:46:41 Meatkat: cool i'm in for ~100 games so we'll probably get matched soon 2010-10-13T13:46:44 it's mostly just that i haven't worked on it 2010-10-13T13:46:52 which are you bhasker? 2010-10-13T13:46:58 malazan-0.1 2010-10-13T13:47:00 *** Titankiller has joined #aichallenge 2010-10-13T13:47:05 was running them under my name earlier 2010-10-13T13:47:08 i've been working on a new (similar) bot, that is more neatly organized and maintainable 2010-10-13T13:47:10 switched the name around 2010-10-13T13:47:17 after i rewrote the core 2010-10-13T13:47:33 oh, malazan is you. That thing likes to kill my bot 2010-10-13T13:47:38 hehe:D 2010-10-13T13:47:59 JamesMG which is your bot? 2010-10-13T13:48:04 JamesMG.dev 2010-10-13T13:48:17 my bot is kinda stupid though sometimes it will beat like a top 10 bot 2010-10-13T13:48:25 and then on some maps it will go lose to -300 elo rated bot 2010-10-13T13:48:26 :( 2010-10-13T13:48:38 was trying a different expansion method recently that hasn't worked particularly well 2010-10-13T13:48:59 bhasker: probably an issue with ship distribution during first 1-2 turns 2010-10-13T13:49:07 yea my second turn needs fixing 2010-10-13T13:49:09 first is fine 2010-10-13T13:49:18 still thinking on a better second turn strat 2010-10-13T13:49:23 it picks bad planets sometimes 2010-10-13T13:49:33 and ends up overcommitting to acquiring a new neutral 2010-10-13T13:49:46 when it should be trying to defend the one it sent ships to in turn 1 2010-10-13T13:49:55 i end up with a very bad ship differential 2010-10-13T13:50:25 what i found is that scoring function needs to be different during first 1-2 turns.. well, the same function, but a few parameters are different.. dunno if others had similar experience 2010-10-13T13:50:42 yup 2010-10-13T13:50:49 i use a different algorithm altogether on turn 1 2010-10-13T13:50:59 but i am thinking maybe i need to do it differently for turn 2 2010-10-13T13:51:04 as well 2010-10-13T13:51:13 it is different, all cheap targets are taken on turn 1 2010-10-13T13:51:32 and then devise 198 different algorithms for the remaining turns.. 2010-10-13T13:52:14 well turn 2-200 is the same strat for my bot ^ 2010-10-13T13:52:20 brb 2010-10-13T13:52:20 but i am thinking of adding some kamikaze moves 2010-10-13T13:52:24 *** JCS^ has joined #aichallenge 2010-10-13T13:52:26 *** Gudradain has quit IRC (Quit: Page closed) 2010-10-13T13:52:28 like if i am losing 2010-10-13T13:52:39 krokokrusa: lol 2010-10-13T13:52:43 do like a shock attack of launching a large number of ships 2010-10-13T13:52:49 or something 2010-10-13T13:52:54 maybe the other bot does something stupid 2010-10-13T13:52:56 kamikaze sounds fun 2010-10-13T13:53:01 but last time i tried that 2010-10-13T13:53:05 i lost all my planets 2010-10-13T13:53:07 *** Palmik has quit IRC (Remote host closed the connection) 2010-10-13T13:53:09 and didnt do squat 2010-10-13T13:53:13 especially close to the end, when it's clear what you're losing 2010-10-13T13:53:17 yup 2010-10-13T13:53:45 rebelxt what is your bots name? 2010-10-13T13:53:47 i suspect some bots strategies don't work when the number of attacking ships is huge 2010-10-13T13:53:52 mine for sure :) 2010-10-13T13:54:11 not huge, but bigger than usual i'd say 2010-10-13T13:54:14 yea 2010-10-13T13:54:25 one thing that i have seen screws over my bot is 2010-10-13T13:54:34 if i lose a core planet to a large force 2010-10-13T13:54:39 its game over 2010-10-13T13:54:40 Meatkat: rebelxt100 - it's a new buggy bot though ;-) 2010-10-13T13:54:44 they rapidly take over all the planets 2010-10-13T13:54:47 since most of my ships are upfront 2010-10-13T13:55:06 err rebelxt110 2010-10-13T13:55:09 and the way my scoring function works i end up trying to defend my planets instead of trying to take his 2010-10-13T13:55:10 bhasker: true 2010-10-13T13:55:21 so my front planet starts sending ships back 2010-10-13T13:55:27 I hate it when huge fleet starts traveling through my planets :) 2010-10-13T13:55:30 instead of ahed into enemy territory:( 2010-10-13T13:55:49 this usually happens when enemy sends a huge fleet a a neutral near me 2010-10-13T13:55:52 and i just ignore it 2010-10-13T13:55:57 thinking it needs too many ships 2010-10-13T13:55:58 to acquire 2010-10-13T13:56:08 and then at X i have a huge fleet sitting outside my base 2010-10-13T13:56:19 so have to add some lookahead for neutrals with huge incoming fleets 2010-10-13T13:56:21 and counter them 2010-10-13T13:56:55 http://72.44.46.68/canvas?game_id=469847 2010-10-13T13:57:01 why o why do people do stuff like that 2010-10-13T13:57:44 though i won that because of a bug in my code ^ 2010-10-13T13:58:02 well not a bug, basically my bot decided that none of the planets nearby were worth acquiring 2010-10-13T13:58:09 so it did nothing on turn 1 2010-10-13T13:58:17 er...he sent 33 ships to a planet with 59. 2010-10-13T13:58:26 some bots do that 2010-10-13T13:58:30 partial ships 2010-10-13T13:58:46 but then ignored it 2010-10-13T13:58:50 as you conquered him 2010-10-13T13:58:57 yup 2010-10-13T14:00:48 *** krokokrusa has left #aichallenge ("going to sleep") 2010-10-13T14:01:00 ok 2010-10-13T14:01:00 Top 10 players: rsergio(4077), Hazard(3922), felixcoto(3877), dmj111(3853), davidjliu(3779), MariusCG(3761), sequoh(3747), asavis(3746), ruilov(3734), savchenko(3733) 2010-10-13T14:01:03 jetbot is going 2010-10-13T14:01:24 on the topic of huge fleet traveling through your planets - i don't think it's worth hard-coding anything to prevent this. this can probably be addressed by a better scoring function which takes neighbor planets into account. it's tricky though :( 2010-10-13T14:01:39 bah loss to CCCF-ex9 2010-10-13T14:01:41 hi, jetbot! ;-) 2010-10-13T14:01:45 by 1 ship 2010-10-13T14:01:46 :( 2010-10-13T14:02:11 who is CCF-ex9? it's a very pretty bot whoever it is 2010-10-13T14:02:16 CCCF* 2010-10-13T14:02:24 no idea who it is 2010-10-13T14:02:34 bhasker: wow he attacked 14-neutral one turn ahead of you and won 2010-10-13T14:02:47 yea 2010-10-13T14:02:52 cause i ignore 1 growth planets on turn 1 2010-10-13T14:02:58 CCCF is felixcoto (rumor not a fact) 2010-10-13T14:03:10 i decided after watching a bunch of games 2010-10-13T14:03:31 even if it's 1-distance 1-ship 1-growth neutral? 2010-10-13T14:03:31 oh sh*t, here comes jetbot! 2010-10-13T14:03:38 yea for now yes 2010-10-13T14:03:41 but i think i will change that 2010-10-13T14:04:02 mostly i wnat to ignore any 1 growth planets that require more than 4-5 ships to take 2010-10-13T14:04:11 i found that its generally not useful to burn more ships on a 1 growth 2010-10-13T14:04:15 on turn 1 2010-10-13T14:04:31 lol 2010-10-13T14:04:39 already lost to dmj's bot with -200 elo 2010-10-13T14:05:32 cfaftw: that's not gonna matter in a few games. that's dmj's new bot that he started a few minutes earlier it appears. it's gonna climb way up 2010-10-13T14:06:14 *** Titankiller has quit IRC (Quit: Nettalk6 - www.ntalk.de) 2010-10-13T14:06:22 oh 2010-10-13T14:07:04 climbing rapidly 2010-10-13T14:07:06 already at 137 elo 2010-10-13T14:07:28 ok, got it running automatically now 2010-10-13T14:08:00 There are so many games to look at! It's a sensory overload. I'm used to having to wait for the 1/hour server games. Analyze every move several times, then do it again for fun. They take a nap and voila! a new game! 2010-10-13T14:08:21 But now.... all the games I could ever want 2010-10-13T14:08:22 that's why the TCP server was such a good idea 2010-10-13T14:08:33 agree++ 2010-10-13T14:09:10 what's the oppo. column in the tcp server stats? 2010-10-13T14:09:12 official server is actually doing fewer games per minute than tcp server at the moment. /sigh 2010-10-13T14:09:22 Meatkat: avg opponent's elo rating 2010-10-13T14:09:33 *** Mathnerd314 has quit IRC (Ping timeout: 276 seconds) 2010-10-13T14:09:51 oic 2010-10-13T14:10:21 my bot on the official server has got 2 games total today 2010-10-13T14:10:23 :-\ 2010-10-13T14:10:33 yeah 2010-10-13T14:10:37 i should upload a new version 2010-10-13T14:10:52 bhasker: don't upload until they fix cloud servers 2010-10-13T14:10:55 yea 2010-10-13T14:11:00 don't want to lose the rank 2010-10-13T14:11:01 lol 2010-10-13T14:11:11 and it's not worth waiting :) 2010-10-13T14:11:14 will take you a week to climb back up at this rate 2010-10-13T14:11:30 probably longer 2010-10-13T14:11:31 *** Mathnerd314 has joined #aichallenge 2010-10-13T14:12:09 yup 2010-10-13T14:12:14 not that my rank is very high 2010-10-13T14:12:17 its at 123 :-\ 2010-10-13T14:12:20 *** phreeza has joined #aichallenge 2010-10-13T14:12:23 been hovering around 100 2010-10-13T14:14:19 *** Bobng has quit IRC (Quit: Leaving) 2010-10-13T14:17:58 rebelxt, what's the name of your new bot? 2010-10-13T14:18:08 i mean, whatever bot you're running on TCP 2010-10-13T14:18:22 rebelxt109 hasn't been played in 13 hours 2010-10-13T14:18:31 rebelxt110 2010-10-13T14:19:52 new dmj bot went to first after 8 games! 2010-10-13T14:19:52 ... 2010-10-13T14:21:43 just lost to rebelxt110 2010-10-13T14:22:56 *** jmcarthur has quit IRC (Read error: Operation timed out) 2010-10-13T14:22:57 *** P_Hansson has joined #aichallenge 2010-10-13T14:23:25 any ETA if more than 1 server will be back up? 2010-10-13T14:25:39 haha merry go round http://72.44.46.68/canvas?game_id=470175 2010-10-13T14:27:11 *** amstan_ has joined #aichallenge 2010-10-13T14:27:12 *** ChanServ sets mode: +o amstan_ 2010-10-13T14:31:41 nice 2010-10-13T14:32:06 cfaftw: running rebelxt110 at the moment 2010-10-13T14:32:23 *** jmcarthur has joined #aichallenge 2010-10-13T14:33:05 ok 2010-10-13T14:34:22 cfaftw: good game. current bot is buggy and this explains my win pretty much 2010-10-13T14:34:46 rebelxt, no, your bot is good 2010-10-13T14:35:15 rebelxt, but i took down JetBot.2 and put up JetBot.1, which is my official submission bot with one bug fix that i know worked 2010-10-13T14:35:28 JetBot.2 had a bunch of fixes that i thought would help, but apparently don't 2010-10-13T14:35:51 *** tty1 has quit IRC (Ping timeout: 245 seconds) 2010-10-13T14:35:59 regardless... i'm turning my back on that bot 2010-10-13T14:36:12 :( Poor guy! 2010-10-13T14:36:18 IT'S DEAD TO ME 2010-10-13T14:36:29 jetbot.1 just beat me:( 2010-10-13T14:36:36 http://72.44.46.68/canvas?game_id=470352 2010-10-13T14:36:42 oh that's yours? 2010-10-13T14:36:46 cfaftw: lol 2010-10-13T14:36:48 poor planet selection at turn 2 2010-10-13T14:36:49 :( 2010-10-13T14:36:54 i took a 1 growth planet 2010-10-13T14:36:57 :-\ 2010-10-13T14:37:22 bhasker, LOL 2010-10-13T14:37:26 that's hilarious 2010-10-13T14:37:29 yea:( 2010-10-13T14:37:36 as i said my heuristic is fucked up 2010-10-13T14:37:39 does stupid stuff like that 2010-10-13T14:37:42 need to fix that 2010-10-13T14:37:49 more importantly its greedy 2010-10-13T14:37:59 *** jesionaj has joined #aichallenge 2010-10-13T14:38:04 it will rather take the 1 growth this turn 2010-10-13T14:38:07 than a 5 growth next turn 2010-10-13T14:38:12 as long as it has enough ships 2010-10-13T14:38:26 yeah you have to fix that 2010-10-13T14:38:28 yup 2010-10-13T14:38:53 we just played rebelxt: http://72.44.46.68/canvas?game_id=470309 2010-10-13T14:39:00 *** Luy has joined #aichallenge 2010-10-13T14:39:49 Exact same until I open up the middle planet by taking the 80, need to fix that 2010-10-13T14:41:16 *** rebelxt has quit IRC (Ping timeout: 245 seconds) 2010-10-13T14:41:25 here we go meatkat 2010-10-13T14:41:48 btw who is honolulu? 2010-10-13T14:41:54 oh snaps! 2010-10-13T14:42:23 *** Luy has quit IRC (Client Quit) 2010-10-13T14:42:52 meatkat, you're the one that is connecting from africa, right? 2010-10-13T14:43:13 yeah, got an account on a server in the states now though 2010-10-13T14:43:14 ftw 2010-10-13T14:43:52 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-10-13T14:44:57 cfaftw gg 2010-10-13T14:45:01 *** tty1 has joined #aichallenge 2010-10-13T14:45:01 *** tty1 has joined #aichallenge 2010-10-13T14:45:21 that was a gg 2010-10-13T14:46:10 *** Bobng has joined #aichallenge 2010-10-13T14:46:35 yeah, i need to get a lot better at neutral denying, not just taking 2010-10-13T14:47:27 neither of us had solid control of those two middle planets on turn 1, yours took the chance and mine made it worth it by sending off to some secondary targets 2010-10-13T14:47:35 amstan do you have a good way to ping j3camero about the cloud servers being down? 2010-10-13T14:47:47 or amstan_ 2010-10-13T14:47:58 janzert: that was mostly danielvf's work, i think 2010-10-13T14:47:58 *** rebelxt has joined #aichallenge 2010-10-13T14:48:04 jeff@ai-contest.... 2010-10-13T14:48:12 this is an incredible win by lovebot 2010-10-13T14:48:13 http://72.44.46.68/canvas?game_id=470477 2010-10-13T14:48:14 yeah, j3camero should be able to bring them up though 2010-10-13T14:48:34 doh lost internet connection for 10 minutes :( 2010-10-13T14:48:38 amstan_: ok, wasn't sure if you had a better way 2010-10-13T14:48:40 my bot was winning, then he launched an attack... by turn 200 he was up by 40 ships 2010-10-13T14:48:56 janzert: you can ping him, but he's at work till like ~18:00 2010-10-13T14:49:12 if it had played out longer i think my bot would have won 2010-10-13T14:49:36 ahh 2010-10-13T14:50:15 cfaftw: yeah you would have, couldn't you have defended against that last attack? 2010-10-13T14:50:16 *** Frontier has joined #aichallenge 2010-10-13T14:50:36 huge tactical error, there 2010-10-13T14:50:38 probably could have :) but it has bugs 2010-10-13T14:50:49 nice game :D 2010-10-13T14:50:55 yeah, didn't defend, and attacked two neutrals 2010-10-13T14:51:16 that explains the loss of the lead, you went -170 ships from that 2010-10-13T14:51:42 lol 2010-10-13T14:51:45 attacking the neutrals 2010-10-13T14:51:56 i didn't notice that 2010-10-13T14:51:57 silly bot 2010-10-13T14:53:12 nice comeback by me if I may say so: http://72.44.46.68/canvas?game_id=470532 2010-10-13T14:53:17 *** jaspervdj has quit IRC (Quit: NEVER GONNA GIVE YOU UP NEVER GONNA LET YOU DOWN) 2010-10-13T14:53:54 *** jaspervdj has joined #aichallenge 2010-10-13T14:54:36 wow nice 2010-10-13T14:54:58 CCCF needed to do more with his big stack, but it can only be at one place at a time 2010-10-13T14:55:51 my OMFG RUN! strat actually worked, lol 2010-10-13T14:55:53 http://ai-contest.com/forum/viewtopic.php?f=18&t=917 2010-10-13T14:56:19 ^need it 2010-10-13T14:56:32 Accoun: dude.. you know you can do it yourself 2010-10-13T14:56:39 +33% to bot testvng speed 2010-10-13T14:56:59 that was a nice game meatkat 2010-10-13T14:56:59 Accoun: you can just download it and compile yourself 2010-10-13T14:57:07 easier to use the c++ version? 2010-10-13T14:57:07 Accoun: it's really not high priority 2010-10-13T14:57:11 amstan> отстань от меня 2010-10-13T14:57:16 and not bother with the java version at all 2010-10-13T14:57:16 CCCF (felixcoto) bot seems to favor the center planets 2010-10-13T14:57:16 ?? 2010-10-13T14:57:35 amstan> i not need u words 2010-10-13T14:57:58 thanks bhasker 2010-10-13T14:58:06 someone java coders please do it 2010-10-13T14:58:07 Accoun: what encoding do you use for irc? 2010-10-13T14:58:08 alright, I'm going to go get some programming done 2010-10-13T14:58:12 be back in a bit 2010-10-13T14:58:21 laterz 2010-10-13T14:58:24 *** Meatkat has quit IRC (Quit: Leaving) 2010-10-13T14:58:53 Accoun: I can write Java; how much are you paying? 2010-10-13T14:59:07 talk to him ^ 2010-10-13T14:59:08 Accoun: this again? 2010-10-13T14:59:12 Accoun: I charge $30 / hour, with a minimum of 8 hours billes. 2010-10-13T14:59:20 *** bss03-AFK is now known as bss03 2010-10-13T14:59:21 good price 2010-10-13T14:59:39 Oo what's going on guys 2010-10-13T14:59:51 deepblue: accoun's crying again 2010-10-13T15:00:24 i'm to lazy to recompile my java version with that change myself too 2010-10-13T15:00:41 but mostly because I hate java :D 2010-10-13T15:00:56 why do you need it anyway, it'll only increase speed during testing 2010-10-13T15:01:01 Top 10 players: rsergio(4067), Hazard(3899), felixcoto(3869), dmj111(3843), davidjliu(3782), sequoh(3751), MariusCG(3741), ruilov(3741), asavis(3739), savchenko(3710) 2010-10-13T15:01:14 I should sell my current version of my bot 2010-10-13T15:01:16 ;) 2010-10-13T15:01:22 heh 2010-10-13T15:01:24 deepblue: what rank? 2010-10-13T15:01:35 for free 2010-10-13T15:01:40 http://ai-contest.com/profile.php?user_id=5898 2010-10-13T15:01:43 16 2010-10-13T15:01:50 15 2010-10-13T15:01:51 :) 2010-10-13T15:02:01 Accoun: you should give him some uranium for it instead 2010-10-13T15:02:02 ok... now my bot is just playing lovebot over and over again 2010-10-13T15:02:08 i need to slow down my game time 2010-10-13T15:02:20 uranium 2010-10-13T15:02:32 nice 2010-10-13T15:02:32 Accoun: I only do free code under the LGPL or GPL; I don't think either are entirely compatible with the existing PlayGame.jar license. 2010-10-13T15:02:39 amstan> leave me alone- i dont need u words 2010-10-13T15:03:40 amstan> it compatible trast me 2010-10-13T15:03:41 Accoun: for all the work you have gone through to get somebody else to do it for you you could have already installed svn and done it yourself 2010-10-13T15:03:49 bss03> it compatible trast me 2010-10-13T15:03:54 i know, eh? 2010-10-13T15:04:07 Accoun: Sorry, you aren't my legal coucil. 2010-10-13T15:04:34 Accoun: I'd check with him, but that might cost money. I'd have to have some hours already lined up to made sure I could cover it. 2010-10-13T15:04:46 * Accoun ignored amstan_ 2010-10-13T15:04:52 * Accoun ignored jmcarthur 2010-10-13T15:05:11 Accoun: you're awesome 2010-10-13T15:05:21 my favorite person on irc 2010-10-13T15:05:24 contestbot: echo Accoun so like.. we're ops, you know we can just mute you forever 2010-10-13T15:05:24 Accoun so like.. we're ops, you know we can just mute you forever 2010-10-13T15:05:25 * Accoun ignored bss03 2010-10-13T15:05:36 * Accoun ignored contestbot 2010-10-13T15:05:39 lol 2010-10-13T15:05:47 why not just ignore the whole channel? 2010-10-13T15:05:47 Accoun: You do relize the amstan_ and jmcarthur are getting along much better with the contest orgs. than you are right now. 2010-10-13T15:05:50 we can help 2010-10-13T15:05:56 bss03: he ignored you 2010-10-13T15:05:59 yeah.. let's 2010-10-13T15:06:48 *** amstan_ sets mode: +q *!*@213.88.2.78 2010-10-13T15:06:54 is it a ban? 2010-10-13T15:06:56 Bah, and I was really close to doing the svn checkout. 2010-10-13T15:07:01 not a ban 2010-10-13T15:07:07 why does it say ban then? 2010-10-13T15:07:11 i know.. it's q, not b 2010-10-13T15:07:19 it's a ban on speaking 2010-10-13T15:07:22 I have the svn URL in my paste buffer. 2010-10-13T15:07:26 btw.. b looks like q rotated 180 degs 2010-10-13T15:07:54 Well, I think I'll just ignore him; assuming he ever gets "-q". 2010-10-13T15:08:35 *** rebelxt has left #aichallenge 2010-10-13T15:13:05 *** tj has joined #aichallenge 2010-10-13T15:13:32 *** tj is now known as Guest91744 2010-10-13T15:14:15 *** Guest91744 has quit IRC (Client Quit) 2010-10-13T15:19:41 hello everyone 2010-10-13T15:19:47 sup 2010-10-13T15:19:51 hi 2010-10-13T15:20:09 I'm trying to figure out if there's a way I can speed up getting the server to accept golang submissions.... 2010-10-13T15:20:40 I'm anxious to see how my bot (http://72.44.46.68/getplayer?player=pwwax.12) would do there... 2010-10-13T15:20:57 Any ideas? 2010-10-13T15:22:19 Got an issue open? 2010-10-13T15:22:48 If not, open one. If so: Got a patch written? 2010-10-13T15:23:12 the code for compiling Go has been there from the start...i think all that's left is having the go compiler on the servers and the entry in the database 2010-10-13T15:23:18 http://code.google.com/p/ai-contest/issues/detail?id=44 2010-10-13T15:23:19 If not, write one. If so: Attached / Linked from issue? 2010-10-13T15:23:33 I wrote a starter package 2010-10-13T15:23:50 no patch necessary, other than starter package (which is not required for support) 2010-10-13T15:24:02 Hrm, the compile logic appears to be in place, according to the ticket. 2010-10-13T15:24:05 Zannick: btw, I got the tournament_manager tested locally and put up a branch for review 2010-10-13T15:24:11 Which is linked there and on the forum (http://ai-contest.com/forum/viewtopic.php?f=18&t=883) 2010-10-13T15:24:13 janzert: i saw 2010-10-13T15:24:19 Have you tried submitting a Go-lang bot? 2010-10-13T15:24:29 yes, it doesn't compile 2010-10-13T15:24:38 At least it didn't yesterday 2010-10-13T15:24:47 i'll give it more attention tonight 2010-10-13T15:25:02 lavalamp: with what error? just a standard couldn't find your main code file? 2010-10-13T15:25:07 The error message says it can't figure out what language 2010-10-13T15:25:10 yeah 2010-10-13T15:25:12 like Zannick said, while support is in the code, I don't think golang has been installed on the server 2010-10-13T15:25:18 i suspect that the entry isn't in the database 2010-10-13T15:25:32 and that might be because the server doesn't have it 2010-10-13T15:25:41 I don't think that's something I can submit a patch for :) 2010-10-13T15:25:44 actually I think amstan said the same thing on the forum and had actually checked the server 2010-10-13T15:25:46 it isn't :) 2010-10-13T15:26:20 I emailed Jeff, as amstan suggested but no word back 2010-10-13T15:26:53 !seen j3camero 2010-10-13T15:26:53 Zannick: j3camero was last seen in #aichallenge 1 week, 1 day, 17 hours, 7 minutes, and 8 seconds ago: amstan: yes, but no names, haha. 2010-10-13T15:27:29 Someone should update the issue, incidating that compile_anything has support, but that Go-lang isn't installed on the server. 2010-10-13T15:27:46 Should I do that? 2010-10-13T15:29:26 go for it 2010-10-13T15:29:33 bss03: give me instructions and i'll install 2010-10-13T15:29:39 but i have no idea what to do for the cloud 2010-10-13T15:29:40 ok 2010-10-13T15:29:57 *** mceier has joined #aichallenge 2010-10-13T15:30:19 I thought there was an Ubuntu package for Go by now. But... 2010-10-13T15:30:24 I can't find one. :( 2010-10-13T15:30:50 No package, you need to use hg: http://golang.org/doc/install.html 2010-10-13T15:30:55 http://72.44.46.68/canvas?game_id=471028 2010-10-13T15:31:00 rebelxt: 2010-10-13T15:31:10 I think they do that as it's still evolving 2010-10-13T15:31:59 meatkat http://72.44.46.68/canvas?game_id=470918 2010-10-13T15:33:33 oh shoot, I just saw that my branch creation last night actually checked in unrelated fixes to the cloud server build process. 2010-10-13T15:35:00 it's actually a minor(?) security fix, so I'm not sure if I should back it out or not 2010-10-13T15:35:09 *** RainCT has joined #aichallenge 2010-10-13T15:37:28 I added a comment to the issue. Hopefully someone who can do something about it will notice... 2010-10-13T15:38:12 Zannick: have an opinion on if I should back out the accidently checked in patch from this issue 2010-10-13T15:38:14 http://code.google.com/p/ai-contest/issues/detail?id=196#makechanges 2010-10-13T15:38:25 actual check in is here http://code.google.com/p/ai-contest/source/detail?r=430 2010-10-13T15:39:09 *** chl is now known as lovebot 2010-10-13T15:39:18 *** Accoun has quit IRC () 2010-10-13T15:39:59 http://72.44.46.68/canvas?game_id=471208 2010-10-13T15:40:02 i have no idea what these changes do 2010-10-13T15:40:22 saves the jailuser iptables rules accross reboots 2010-10-13T15:40:43 and blocks them from any network communication instead of just tcp protocol 2010-10-13T15:41:18 *** Olathe has joined #aichallenge 2010-10-13T15:41:36 I think I'll go ahead and back it out since it went in completely by accident. 2010-10-13T15:41:39 i don't want to say just that it's okay... 2010-10-13T15:41:41 *** McLeopold has joined #aichallenge 2010-10-13T15:41:43 yeah 2010-10-13T15:41:58 even though I have tested it here 2010-10-13T15:43:40 *** ChanServ sets mode: +o McLeopold 2010-10-13T15:48:35 backed out in r433 2010-10-13T15:49:55 *** amstan_ has quit IRC (Quit: Konversation terminated!) 2010-10-13T15:50:00 *** amstan_ has joined #aichallenge 2010-10-13T15:50:01 *** ChanServ sets mode: +o amstan_ 2010-10-13T15:54:52 *** amstan_ has quit IRC (Remote host closed the connection) 2010-10-13T15:57:57 *** Accoun has joined #aichallenge 2010-10-13T16:01:02 Top 10 players: rsergio(4044), Hazard(3872), felixcoto(3843), dmj111(3821), davidjliu(3757), MariusCG(3729), ruilov(3724), asavis(3715), savchenko(3711), sequoh(3711) 2010-10-13T16:03:25 mcleopold, where's your new bot? 2010-10-13T16:05:52 so we can use udp logging now to debug our bots remotely. great! ;) 2010-10-13T16:11:36 *** Bobng has quit IRC (Remote host closed the connection) 2010-10-13T16:14:07 *** boegel has joined #aichallenge 2010-10-13T16:14:16 except the cloud servers aren't running anyway :P 2010-10-13T16:16:54 cfaftw: it was too slow. major re-write. 2010-10-13T16:17:55 mcleopold, i'm currently rewriting the majority of mine... it got too messy and buggy 2010-10-13T16:18:30 yeah: i know what you mean. It's been 2 days and I haven't gotten it back to where it was yet. But it is much faster to process. 2010-10-13T16:18:38 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-10-13T16:18:54 seems like everybody is rewriting heavily 2010-10-13T16:18:55 :) 2010-10-13T16:19:12 creating those extra bots helped alot. 2010-10-13T16:19:21 My framework is becoming very solid. 2010-10-13T16:19:47 i have a great new idea for mine but it too requires major rewriting 2010-10-13T16:19:51 maybe this weekend 2010-10-13T16:19:51 *** iFire has joined #aichallenge 2010-10-13T16:20:00 makes me glad there's so much time 2010-10-13T16:20:14 *** rebelxt has joined #aichallenge 2010-10-13T16:20:17 I've been surprised by how much time I can sink into this 2010-10-13T16:20:25 due to rewriting/refactoring 2010-10-13T16:20:35 Sometime tonight or tomorrow, AntBot will make it's appearance on tcp. 2010-10-13T16:20:41 My bot took over my life... :/ 2010-10-13T16:20:46 McLeopold: looking forward to it 2010-10-13T16:20:56 woho 2010-10-13T16:21:05 I hope I catch a game or two against mine ;) 2010-10-13T16:21:21 yea, me too 2010-10-13T16:21:30 antbot... i love how your bot names are meaningful 2010-10-13T16:21:51 will there be life after Nov 27th? only if there is another contest.. ;-) 2010-10-13T16:21:53 I try and name the predicted behavior. 2010-10-13T16:22:05 i might as well name my bot "DoesSomethingMaybeBot" 2010-10-13T16:22:20 SlothBot? 2010-10-13T16:22:49 Incorrigibot 2010-10-13T16:23:07 WhenInDoubtAttackBot 2010-10-13T16:23:32 SPISPOPDBot 2010-10-13T16:23:48 WinBot? 2010-10-13T16:23:49 (good luck with that reference...) 2010-10-13T16:23:51 TimeOutBot 2010-10-13T16:24:05 NerdRageBot 2010-10-13T16:24:23 FailBot#2454646 2010-10-13T16:24:25 *** Appleman1234 has joined #aichallenge 2010-10-13T16:24:35 too many options 2010-10-13T16:24:41 *** McLeopold has quit IRC (Quit: Leaving) 2010-10-13T16:25:58 i wonder if anyone has a test case or a few for his/her bot?? it feels like test cases might be a great idea due to regression bugs most of us are being hit with 2010-10-13T16:26:30 i.e. automated unit tests 2010-10-13T16:27:17 *** Bobng has joined #aichallenge 2010-10-13T16:28:45 I've just been running mine against all the sample bots in mirrored matches on specific maps 2010-10-13T16:28:54 and then dumping it on TCP and waiting for weird issues to appear 2010-10-13T16:29:33 i thought about that yesterday @rebelxt 2010-10-13T16:29:39 i don't have any 2010-10-13T16:29:55 but it would be nice to create something to give the bot certain situations and see how it reacts 2010-10-13T16:31:09 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-10-13T16:32:27 I was considering the same thing this morning, the problem would be that you can't really just through a bot in a situation b/c afaik most bots keep information from turn to turn 2010-10-13T16:33:28 *** iFire has joined #aichallenge 2010-10-13T16:34:14 *** cfaftw has quit IRC (Quit: Leaving) 2010-10-13T16:34:17 *** Appleman1234 has quit IRC (Ping timeout: 264 seconds) 2010-10-13T16:35:14 you'd just have to write a small extension that would fill everything and jump to where you need to be 2010-10-13T16:39:59 *** McLeopold has joined #aichallenge 2010-10-13T16:40:03 *** ChanServ sets mode: +o McLeopold 2010-10-13T16:42:38 javagamer: Hrm, my bot still doesn't keep any state from turn to turn. 2010-10-13T16:43:02 javagamer: It's just easier (for me) that way in Haskell. 2010-10-13T16:43:22 I do it for performance reasons. 2010-10-13T16:43:42 Plus, I can check if my state matches the server state and look for errors in my predictions. 2010-10-13T16:43:45 The only thing my bot keeps between turns is the table of planet distances... 2010-10-13T16:43:54 bss03: why is it easier in haskell? 2010-10-13T16:44:02 McLeopold: Right now, the only thing like that I would hold is planet distances. 2010-10-13T16:44:38 bss03: do you track fleets or create fleet objects? 2010-10-13T16:44:43 i'm keeping state in mine and it's at least as easy as it would be in any other language 2010-10-13T16:44:57 bss03: that was the first thing I got rid of from the starter pack 2010-10-13T16:45:04 *** boegel has quit IRC (Quit: *poof!*) 2010-10-13T16:45:29 McLeopold: Not directly. My engine/simulation does that. 2010-10-13T16:45:33 I haven't even bothered to look, was the Distance method eating a lot of time? 2010-10-13T16:45:35 The first thing I did was spend a bunch of time organizing data I could hold from turn to turn 2010-10-13T16:45:47 (back) 2010-10-13T16:45:53 Only now am I actually working on how my bot actually acts 2010-10-13T16:46:04 McLeopold: I expose a function that just does the "deployment" step to my main function, so it creates fleets there. 2010-10-13T16:46:05 *** sigh has joined #aichallenge 2010-10-13T16:46:07 the first thing i did was create a PersistentState subclass of PlanetWars in c++ 2010-10-13T16:46:14 i don't keep fleets either. i just keep how many ships are expected to arrive at each planet and when 2010-10-13T16:46:23 the second thing i did was implement copybot using that. 2010-10-13T16:46:32 the third thing i did was start over in c. :) 2010-10-13T16:46:37 heh 2010-10-13T16:46:37 deepblue: on test cases - agreed, i think it's time for me to write some.. getting too many stupid regressions 2010-10-13T16:46:54 McLeopold: Some a "fake" fleets that are used for further calculations -- I send the order to my internal engine, but I don't actually write them out on stdout. 2010-10-13T16:46:55 I'll share my test cases in another week or so. 2010-10-13T16:47:10 i have two separate data types for the game state. one is for data that is the same for the entire game, and the other is for data that changes on every turn 2010-10-13T16:47:16 *** ratatata has joined #aichallenge 2010-10-13T16:47:30 jmcarthur: are you playing on tcp yet? 2010-10-13T16:47:35 McLeopold: nice 2010-10-13T16:47:47 still no 2010-10-13T16:47:54 :( 2010-10-13T16:47:59 the last couple days have been writing tests and fixing bugs 2010-10-13T16:48:00 jmcarthur: My doTurn is GameState -> [Order], using jaspervdj's bot fuction to run it. 2010-10-13T16:48:03 well worth it 2010-10-13T16:48:22 bss03: ah, i am just writing my own loop 2010-10-13T16:48:46 i'm not bothering a whole lot with nice abstractions this time around 2010-10-13T16:48:49 My test cases are pretty simple. Kinda like chess board that say "Mate in 3" 2010-10-13T16:48:54 jmcarthur: I haven't moved to using GameState -> IO [Order] and an IORef to or similar, yet. 2010-10-13T16:49:13 McLeopold: i'm generating test data and checking properties on it on the fly 2010-10-13T16:49:14 *** Appleman1234 has joined #aichallenge 2010-10-13T16:49:18 McLeopold: found lots of bugs that way 2010-10-13T16:49:19 jmcarthur: There's nothing I've found other than planet distances (and turns remaining) that I really want to keep track of. 2010-10-13T16:49:51 bss03: i've not really used any IO aside from stdin/stdout, either 2010-10-13T16:50:08 still, a reader monad is called far, IMO 2010-10-13T16:50:12 for the static state, at least 2010-10-13T16:50:14 McLeopold: test case ideas - sniping, anti-sniping, precise defense.. trying to think of more intelligent tests 2010-10-13T16:50:16 *** sigh_ has joined #aichallenge 2010-10-13T16:50:33 rebelxt: keep generating some, they will help you out 2010-10-13T16:50:51 jmcarthur: Contrary to my assertion eariler that I wouldn't re-write Control.Monad.State, I do have a number... 2010-10-13T16:51:04 Once I get my own tcp server, I'll have a port for playing the test cases with other bots. 2010-10-13T16:51:12 positional advantage test cases would be nice 2010-10-13T16:51:17 jmcarthur: of function with type GameState -> ([Order], GameState) that I manually chain together. 2010-10-13T16:51:17 bss03: i'm still depending on being able to upload binaries :\ 2010-10-13T16:51:48 They don't compile haskell? 2010-10-13T16:51:51 bss03: yeah, i hate it when i see that and don't want to pull in a monad lib 2010-10-13T16:52:00 McLeopold: they do, but with a crappy old compiler 2010-10-13T16:52:05 McLeopold: and i mean *crappy* 2010-10-13T16:52:19 :), yeah, python is 2.5. I wish it were py3k 2010-10-13T16:52:33 jmcarthur: I don't mind that it is GHC 6.8.2, just that they don't have the MTL installed. 2010-10-13T16:52:41 McLeopold: even if it wasn't so bad, i would still absolutely need ghc 7, which is still only in release candidate stages, to get good performance from what i'm doing 2010-10-13T16:52:56 jmcarthur: Even in 6.12, MTL is still a separate package in Ubuntu / Debian-land, IIRC. 2010-10-13T16:53:16 i also just want to be able to use libraries like a sane haskeller would 2010-10-13T16:53:21 jmcarthur: so, if you had bad performance, where are you at for per/turn processing time? 2010-10-13T16:53:33 2 days ago I was at 3-5 seconds. 2010-10-13T16:53:41 McLeopold: i'm just using the whole second, so it would be the same anyway 2010-10-13T16:53:49 ah 2010-10-13T16:53:50 i just get to run more iterations if it's faster 2010-10-13T16:53:54 Until my last set of changes I was at around 300ms or less per turn. 2010-10-13T16:54:05 *** BtbN has quit IRC (Quit: Verlassend) 2010-10-13T16:54:21 McLeopold: the difference between ghc 6.12 and ghc 7 for my code should be around 10x though. it's that important in this case 2010-10-13T16:54:24 Now, I average probably 500ms per turn, but there are some sideways cases where I take 3-5s on a turn and timeout locally. 2010-10-13T16:54:42 I ran on the TCP server for 2 days with the same code and never timed out there, though. :P 2010-10-13T16:54:42 and between 6.8 and 6.12 is also a consideration... 2010-10-13T16:55:02 I should get into haskell some time. 2010-10-13T16:55:31 McLeopold: Two pretty good books about it, both available to read for free on the web. 2010-10-13T16:55:34 I'd like to see how you do given multiple threads and 5 seconds a turn. :) 2010-10-13T16:55:42 heh 2010-10-13T16:55:54 McLeopold: Learn You A Haskell (LYAH for short) and Real World Haskell (RWH for short). 2010-10-13T16:55:56 probably pretty well by comparison :) 2010-10-13T16:56:11 although right now i have stuck some code in that is not parallelism friendly 2010-10-13T16:56:16 i could fix it pretty easily though i think 2010-10-13T16:56:21 bss03: tcp server has 10-sec time-out (unlike dhartmei adjusted it), so be careful 2010-10-13T16:56:25 You should play your bot against itself with one given shorter time. That would be interesting. 2010-10-13T16:57:08 rebelxt: That might be why I didn't see it then, I thought the time-out was still like 1500ms on the TCP server. 2010-10-13T16:57:26 bss03: no, it's 10 seconds - i just checked the source code 2010-10-13T16:57:39 10... seconds...? 2010-10-13T16:58:09 it really is just for testing i guess :) 2010-10-13T16:58:10 rebelxt: I'm pretty sure I can get the time down quite a bit though. I'm fairly sure I'm got a lambda that GHC 6.12 isn't floating... 2010-10-13T16:58:13 jmcarthur: actually, sorry it's 5 seconds! was looking at a wrong line 2010-10-13T16:58:18 ah 2010-10-13T16:58:20 that's more reasonable 2010-10-13T16:58:55 rebelxt: that I float manually fairly easily, and drop some redundant calculations on the way. 2010-10-13T16:59:36 I just haven't gotten around to it. Got a lot of ideas bouncing around that I haven't sat down to write in code. 2010-10-13T16:59:44 i personally think such optimizations should just be written by hand anyway. it's much more clear how your performance will be that way and reduces redundancies in your code 2010-10-13T17:00:07 I just realized my defense code would be nearly identical to my sniping code, I can just tweak it slightly and use the same code for both 2010-10-13T17:00:10 many such optimizations anyway 2010-10-13T17:00:12 not all 2010-10-13T17:00:39 javagamer: you know you should copy and paste 2010-10-13T17:00:44 javagamer: it doubles your productivity 2010-10-13T17:00:48 I'm trying to let them congeal; I'm pretty sure implementing one of them correctly will make the others much easier. 2010-10-13T17:00:49 heh 2010-10-13T17:00:54 so much more code written in so much less time! 2010-10-13T17:01:02 Top 10 players: rsergio(4026), Hazard(3845), felixcoto(3823), dmj111(3804), davidjliu(3739), MariusCG(3714), ruilov(3711), asavis(3699), savchenko(3692), sequoh(3681) 2010-10-13T17:01:31 javagamer: My defence, sniping, expansion, and attack code is all the same function. :P 2010-10-13T17:02:14 javagamer: I have separate functions for moving ships to the front line and fleeing planets I'm going to lose, but most of it is in one large function. 2010-10-13T17:02:15 bss03, I may end up with something like the eventually, but for now I'm pretty far from that 2010-10-13T17:02:32 bss03, you flee planets you'll lose? 2010-10-13T17:02:46 *** sigh has quit IRC (Remote host closed the connection) 2010-10-13T17:03:28 javagamer: I think I'm going to move away from it (all one function) in some respects. It's started to grow some special cases, so I need to refactor some stuff into smaller functions and re-build on that. 2010-10-13T17:03:32 I have a function that decides how many ships I need where and when, and another one that figures out where to get them from 2010-10-13T17:03:58 Well, the first one is multiple functions. 2010-10-13T17:04:13 javagamer: If I'm behind in ship count, and the planet is going to fall this turn, and there's a planet I think I will hold by the time I can get to it, I flee. 2010-10-13T17:04:44 javagamer: If I have 25 ships and they have 50 ships, it is very important I don't trade 20 ships 1-for-1. 2010-10-13T17:04:56 bss03, true, didn't consider that 2010-10-13T17:05:24 My code is primarily written for when I'm winning, it's not very good at turning around games yet 2010-10-13T17:05:27 bss03: I think do that, too, but it's an unintended consequence... 2010-10-13T17:05:43 javagamer: However, I do need to tweak that code to flee *less* that that even. 2010-10-13T17:05:57 if my defense goal fails, it frees up all those ships to take something else... 2010-10-13T17:06:25 javagamer: Cause the rest of my code doesn't take into account the fact that I flee, so it doesn't expect the large hostile troops there and doesn't "win back" the planet the way it should. 2010-10-13T17:07:05 So, I'm thinking only flee if I don't have fleets headed toward that planet. 2010-10-13T17:08:05 If you're going to lose it anyway, you might as well flee. You can't get the ships you sent towards the doomed planet back. 2010-10-13T17:09:19 My problem is that my fleeing ships like to attack neutral planets... >.< 2010-10-13T17:09:21 lavalamp: Well, depends. Say they are about to hit my growth 3 planet with 24 ships, and I have 20 there right now, but 13 arriving in a couple of turns. 2010-10-13T17:09:51 bss03: ah, good point, if you can retake it you shouldn't flee, true 2010-10-13T17:10:00 lavalamp: If I keep my ships there, I'll retake the planet. If I flee, I'll lose those 13, plus I'll lose some growth. 2010-10-13T17:10:14 true 2010-10-13T17:10:41 *** delt0r___ has joined #aichallenge 2010-10-13T17:10:59 bss03: I just realized I need to trigger my attack code on planets of my own that I'm going to lose, thanks 2010-10-13T17:11:40 lavalamp: No problem. My bot is LGPL anyway, so it's not like I'm accidentally giving away strategy. :P 2010-10-13T17:11:44 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2010-10-13T17:15:24 bss03: Does that mean your source code is available? 2010-10-13T17:15:37 lavalamp: mine kind of works that way too 2010-10-13T17:15:59 lavalamp: figure out where and when i need ships, then figure out where to get them from, that is 2010-10-13T17:16:12 lavalamp: http://github.com/stephen-smith/ai-contest-2010 2010-10-13T17:16:30 lavalamp: i'm having more trouble with the latter than the former... 2010-10-13T17:16:50 *** jukkerknott has joined #aichallenge 2010-10-13T17:17:00 * jmcarthur looks at bss03's code 2010-10-13T17:17:32 jmcarthur: It's not pretty. Or well-documented. 2010-10-13T17:17:52 jmcarthur: Yeah, more specifically I list a bunch of goals. I turn the goals into "sinks" (x ships should be at planet y at turn z). Then I fulfill as many sinks as I can, starting with the highest priority goal. 2010-10-13T17:18:09 jmcarthur: But, it is hanging on to 120th on the official server, last I checked. 2010-10-13T17:18:15 lavalamp: ah, mine is different from that 2010-10-13T17:18:40 lavalamp: That's sort of similar to mine. 2010-10-13T17:18:53 lavalamp: i determine *exactly* how many ships i want at each planet, but not exactly how soon, then i try to figure out the quickest way to make it so (fewest number of turns) 2010-10-13T17:19:53 lavalamp: or, rather, i determine how many ships i want *heading toward* each planet, not even really how many i want there on a particular turn 2010-10-13T17:19:56 That might be better, I had a bug where attacks were failing because the goal setting code was calculating the distance from the wrong planet 2010-10-13T17:19:59 bss03: did you open source your bot? 2010-10-13T17:20:35 bss03: I'll take a look, maybe I'll learn something about Haskell 2010-10-13T17:21:00 lavalamp: I calculate for each time t (0 <= t <= 28 or so), for each planet, how many ships to send to take it, reward for taking it. Sort by reward, send "available" ships, to each one in order, if I have enough available ships. 2010-10-13T17:21:19 rebelxt: Yes, it is LGPL. Has been since day -1. 2010-10-13T17:21:29 rebelxt: (2010-09-09). 2010-10-13T17:21:39 bss03: wow, your doTurn is massive! 2010-10-13T17:21:45 bss03: got it. do look at the discussion at http://ai-contest.com/forum/viewtopic.php?f=17&t=886 2010-10-13T17:21:51 but your line count is much smaller than mine 2010-10-13T17:21:55 Hm, maybe my solution is overly complicated 2010-10-13T17:22:00 *** P_Hansson has quit IRC (Quit: Page closed) 2010-10-13T17:22:44 jmcarthur: Most of doTurn is attackOrders. 2010-10-13T17:23:04 bss03: in a nutshell, if your bot is good (top100), people will be pissed about it being open source... 2010-10-13T17:23:05 bss03: my code is at 800 lines and still growing, but that's counting all the "starter" code and test code 2010-10-13T17:23:12 jmcarthur: My *Orders function are all local so they can use the "distances" map. 2010-10-13T17:23:19 but that's also *not* counting the numerous libraries i'm using too 2010-10-13T17:23:29 ah 2010-10-13T17:23:37 bss03: oh so *that's* how you're avoiding passing state :P 2010-10-13T17:24:05 rebelxt: I read it, but I don't think it'll be quite an issue with my bot. Mainly because I'm writing in Haskell. 2010-10-13T17:24:19 bss03: hey, you could use implicit parameters for your distances if you don't want to do a reader monad 2010-10-13T17:25:17 jmcarthur: Well, the "distances" map is still recalculated each time doTurn is executed, so I don't actually save the state between turns. 2010-10-13T17:25:18 doesn't clear up your type signatures, but it keeps your implementation code simpler at least 2010-10-13T17:25:23 ah 2010-10-13T17:25:57 i don't think i've ever seen so many where clauses in a single function :P 2010-10-13T17:26:32 What do your bots end up rated on the TCP server? (if you've run them there, that is) 2010-10-13T17:27:00 i have not yet 2010-10-13T17:27:35 lavalamp: Not good right now. I might still be listed. DaTwinkDaddy-s11. 2010-10-13T17:27:39 oh ok haha 2010-10-13T17:27:42 bss03: cool, if you read it, you know my opinion on it.. i don't personally care much, but expect to see unhappy haskell people when they find out ;) 2010-10-13T17:27:51 *** amstan_ has joined #aichallenge 2010-10-13T17:27:52 *** ChanServ sets mode: +o amstan_ 2010-10-13T17:27:55 lavalamp: Elo of 150 (tcp server) or so, I guess. 2010-10-13T17:27:55 bss03 DaTwinkDaddy is your bot 2010-10-13T17:28:02 your bot and mine have duked it out a lot 2010-10-13T17:28:10 my last version that is 2010-10-13T17:28:12 Oh that's you-- my latest bot was beating you I think (pwwax.*) 2010-10-13T17:28:16 i'd only be unhappy if it's a top 50 bot, and even then.... there aren't many people willing to learn haskell just to get a top 50 bot to start with 2010-10-13T17:28:17 *** amstan_ has quit IRC (Remote host closed the connection) 2010-10-13T17:28:34 jmcathur: Especially my ugly Haskell. 2010-10-13T17:28:37 lol 2010-10-13T17:28:48 but you were trouncing me for quite a while 2010-10-13T17:28:53 *** bartwe has left #aichallenge 2010-10-13T17:28:56 *** jukkerknott has quit IRC (Ping timeout: 265 seconds) 2010-10-13T17:29:10 I was up to the Top 60 at one point. 2010-10-13T17:29:37 bss03: Nice. 2010-10-13T17:29:40 But, other bots have continued to get better and I haven't gotten much better. 2010-10-13T17:30:01 My bot would be running right now but apparently my wireless died at home again... :( 2010-10-13T17:30:16 rebelxt: Re: open-sourced bot; I do actually request in the LICENCE file that people not submit the bot unmodified. 2010-10-13T17:31:44 *** justin_pdx has joined #aichallenge 2010-10-13T17:32:14 Yeah, I'm DaTwinkDaddy on both the official and TCP servers. On the TCP server I use DaTwinkDaddy-s#-gHASH as generated by "git describe"; I have (some of) my submission tagged in the git true as DaTwinkDadd-s$submission_number. 2010-10-13T17:32:40 So, Right now, DaTwinkDaddy-s11 on the TCP server is my bot on the official server. 2010-10-13T17:32:52 bss03: Ah, I was thinking those were a random seed or something... 2010-10-13T17:32:58 *** amstan_ has joined #aichallenge 2010-10-13T17:32:58 *** ChanServ sets mode: +o amstan_ 2010-10-13T17:33:46 *** Azrathud has joined #aichallenge 2010-10-13T17:34:01 bss03: How hard is Haskell to set up? My bot could use a local sparring partner... 2010-10-13T17:34:40 lavalamp: apt-get install haskell-platform 2010-10-13T17:35:00 bss03: I think I can manage that, hehe 2010-10-13T17:35:01 lavalamp: ghc --make -O2 WhatEver.hs 2010-10-13T17:35:09 *** elijahbal has quit IRC (Ping timeout: 272 seconds) 2010-10-13T17:35:42 lavalamp: I don't think Lenny has haskell-platform, might only be Squeeze and unstable. 2010-10-13T17:35:58 bss03: Thanks, I may try that later 2010-10-13T17:36:14 *** justin_pdx has quit IRC (Client Quit) 2010-10-13T17:36:14 lavalamp: Not completely sure about the availability on Ubuntu; ISTR 10.04 not having it, but maybe 10.10 has it. 2010-10-13T17:36:45 *** elijahbal has joined #aichallenge 2010-10-13T17:36:50 I forget what I'm on. 2010-10-13T17:38:11 *** justin_pdx has joined #aichallenge 2010-10-13T17:39:36 lavalamp: I do recommdn the haskell-platform for real experiementation, but just ghc6 should be enough to compile and run my bot. 2010-10-13T17:40:20 lavalamp: And ghc6 has been around for a while, so should be readily available. I think "ghc6" is the only Haskell package on the official server. 2010-10-13T17:40:38 bss03: Oh, ok-- even easier, great 2010-10-13T17:40:42 Which is why I can't use all of nifty libraries. :( 2010-10-13T17:41:23 bss03: does your bot draw against itself? 2010-10-13T17:42:41 Mine does. 2010-10-13T17:42:52 McLeopold: I haven't checked; probably. 2010-10-13T17:43:25 McLeopold: Except for the iterative-refinement branch, I don't try and guess my opponents moves. 2010-10-13T17:43:29 If, on the official server, you ever see a mirrored game, you'll know someone submitted your code. 2010-10-13T17:44:01 McLeopold: The iterative-refinement branch is still not enough improvement to be worth it. 2010-10-13T17:44:20 Assuming there's an optimal strategy, shouldn't all games end in draws? 2010-10-13T17:44:32 McLeopold: So, it hasn't be merged with master. 2010-10-13T17:44:38 lavalamp: we are a ways out from testing that, but I think yes 2010-10-13T17:45:08 lavalamp: I don't think we will know for a long time. 2010-10-13T17:45:18 lavalamp: That would be my assumption. But, optimal strategy is beyond what we have to play with, I think. 2010-10-13T17:46:11 bss03: btw, I'm in favor of not having open-source bots until after the competition, regardless of rank 2010-10-13T17:46:21 and language 2010-10-13T17:46:32 lavalamp: I know my bot surprised me on at least one map. By coming up with something that was better than what I thought was "optimal strategy". 2010-10-13T17:46:33 Some maps are obviously draws (home worlds too close, neutrals too expensive or far) 2010-10-13T17:46:53 I'd like to work with dmj111 on better ways people can test bots against good ones without giving up source code. 2010-10-13T17:46:54 bss03: that's pretty cool 2010-10-13T17:47:16 McLeopold: If you guys need hosting for that when it's ready let me know 2010-10-13T17:47:32 I'd be interested in a binary exchange, maybe, definitely if there were some easy way to run them in a sandbox 2010-10-13T17:47:32 JamesMG: great, I'll keep you in mind. 2010-10-13T17:47:36 *shrug* I believe in development in the open. 2010-10-13T17:47:42 JamesMG: what kinda of server can you offer? 2010-10-13T17:47:55 I was thinking of just paying for an ec2 instance 2010-10-13T17:48:01 A rising tide lifts all boats is fine with me. 2010-10-13T17:48:29 unless it has to run all the binaries itself it should work 2010-10-13T17:48:31 It would be nice if people could submit a bot to a server, and then others could request to play it. 2010-10-13T17:48:33 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-10-13T17:48:52 *** bhasker_ has joined #aichallenge 2010-10-13T17:48:53 But, if it's gonna be an issue -- like driving down participation, I'd be cool with pulling it down and only releasing after the event. 2010-10-13T17:50:17 bss03: my worry is a bunch of submissions that people didn't write. Pseudo code is fine with me, just not something that doesn't have to be understood to use. 2010-10-13T17:51:04 *** bhasker has quit IRC (Ping timeout: 255 seconds) 2010-10-13T17:51:19 I do request in the LICENSE that they not submit the code unmodified. 2010-10-13T17:51:32 Yeah, but I not worried about honest people. :) 2010-10-13T17:51:34 bss03, appreciate your readyness to take it down. it's a good idea not to open source full bot code until Nov 27th 2010-10-13T17:51:58 And, my code is rather... ugly... so they'd have to spend some time getting familiar with it before they could usefullly modify it. 2010-10-13T17:52:05 My bot can't consistently beat yours still bss03, hah 2010-10-13T17:52:15 it gets lucky on some maps 2010-10-13T17:53:15 *** jorge has joined #aichallenge 2010-10-13T17:54:23 bss03: i was glad when ziloq8 took down his source code. open sourcing bots is kind of against the nature of the competition no matter what language (before Nov 27th) 2010-10-13T17:54:58 I don't really see how it is against the nature of the competition. 2010-10-13T17:55:26 *** Utkarsh has joined #aichallenge 2010-10-13T17:56:17 How about if I just chainge the LICENSE so that it is only LGPL after Nov. 27, and non-redistributable until then? 2010-10-13T17:57:11 *** Accoun has quit IRC () 2010-10-13T17:57:19 bss03: If you can get the organizers to double check the entries to make sure people comply, then that would be good. 2010-10-13T17:57:40 *** jorge has quit IRC (Remote host closed the connection) 2010-10-13T17:58:11 bss03, http://ai-contest.com/forum/viewtopic.php?f=17&t=886 . problem is that other people will get pissed when someone re-submits your code having changed one line. then someone who is in 200th spot feels like they have to fight copies of your bot 2010-10-13T17:58:16 amstan / amstan_ : What's your opinion on my open-source bot? Should I (try to) take down the code until Nov. 27? 2010-10-13T17:59:17 RebelXT: I read the thread. "A rasing tide make all ships float" isn't a problem IMO. 2010-10-13T17:59:19 Entries which are deemed to violate the spirit of fair and sportsmanlike competition will be disqualified 2010-10-13T17:59:35 bss03: imagine your bot is top10 in a week, and still open source... 2010-10-13T17:59:45 You may be enabling unsportsmanlike competition, but it's a very gray area. 2010-10-13T18:00:04 McLeopold: I don't see how releasing my code is at all unfair or unsportsmanlike. 2010-10-13T18:00:18 "Entries which are deemed to violate the spirit of fair and sportsmanlike competition will be disqualified" -- this is not going to be enforced. we just have to trust people don't cheat 2010-10-13T18:00:25 McLeopold: It's fair because everyone gets the same rights to it. 2010-10-13T18:00:29 Not you being unsportsman, but helping other to be 2010-10-13T18:01:03 Top 10 players: rsergio(4012), Hazard(3832), felixcoto(3811), dmj111(3791), davidjliu(3727), MariusCG(3711), ruilov(3697), asavis(3687), savchenko(3673), sequoh(3668) 2010-10-13T18:01:47 I think sharing your ideas as pseudo code is great, just not the actual code. And again, it's a grey area. 2010-10-13T18:02:14 *** mogron has joined #aichallenge 2010-10-13T18:02:23 I can't find it now, but one of my games on the tcp server ended as a draw for an interesting reason: My bot lost on the board at the exact same turn the other bot timed out... 2010-10-13T18:02:36 hahaha 2010-10-13T18:03:03 It's kinda like giving away money to people while playing monopoly. Sure, you can be fair about it, but it's just ruins the game. 2010-10-13T18:03:23 completely agreed with McLeopold. sharing utilities code is probably fine too. just not the "core" algorithm 2010-10-13T18:04:06 bss03: and I'd like amstan's opinion as well... 2010-10-13T18:04:10 bss03, i share your passion for open source too btw! it really drives innovation 2010-10-13T18:04:31 I'm totally going open-source after Nov. 2010-10-13T18:04:53 I'm going to like reading some of these after the 27th 2010-10-13T18:06:02 JamesMG: can't wait too. hope everyone comments their code ;-) 2010-10-13T18:06:09 bss03: also, the competition is more than just who wins the whole thing. There may be local or group competitions that you might affect. 2010-10-13T18:06:27 Like, who's the best bot at a university or something. 2010-10-13T18:07:13 rebelxt: Just think of my code as one big library then. :P 2010-10-13T18:08:04 rebelxt: 'Cause really, it's the same thing. I can be arbitrily close to a complete submission with a "library". 2010-10-13T18:08:15 *** sigh_ is now known as sigh 2010-10-13T18:09:22 *** jukkerknott has joined #aichallenge 2010-10-13T18:09:52 Also, I am still hoping I can get others to submit patches to me. 2010-10-13T18:10:02 :) 2010-10-13T18:10:10 i wonder how hard it would be to solve a particular map statically 2010-10-13T18:10:11 Or, that I'll be able to pull commits from git trees based on my bot. 2010-10-13T18:10:12 http://github.com/apinkin/planetwars-python-kit - this is my compilation of python kit (based on work of ulope and many others). my bot is in private repository though. Nov 27th i flip the switch 2010-10-13T18:10:23 *** Accoun has joined #aichallenge 2010-10-13T18:10:25 by how hard i mean how long 2010-10-13T18:10:30 could even be a very simple map 2010-10-13T18:10:40 just complex enough to be interesting for testing, like four planets or something 2010-10-13T18:10:50 turns or cpu time? 2010-10-13T18:11:15 *** mceier has quit IRC (Quit: leaving) 2010-10-13T18:11:27 i just mean given any amount of time, can a simple but non-trivial map be solved quickly enough to perhaps generate a game plan to test your real bot against 2010-10-13T18:11:32 bss03: then you should change your account to an "open source team" and compete in a different division ;-) 2010-10-13T18:11:44 jmcarthur: you might want to ask antimatroid about that, he's really into game theory and is/wants to investigate that sort of thing 2010-10-13T18:11:48 like have a bot to play against, just for testing your real bot, that has "beaten" the map already 2010-10-13T18:12:01 RebelXT: That's the beauty of open-source, *everyone* has the same base. 2010-10-13T18:12:20 jmcarthur: yes, most of my test maps are designed that way 2010-10-13T18:12:26 bss03: and that's precisely the downside in this case 2010-10-13T18:12:27 But, by hand. 2010-10-13T18:13:29 jmcarthur: I don't see it as a downside. 2010-10-13T18:13:58 bss03, yeah like i said i love open source. it's the main innovation driver in software 2010-10-13T18:13:59 bss03: you don't see it as a downside that everybody could instantly be at place 40 with no work and then have to beat that just to get above average? 2010-10-13T18:14:23 jmcarthur: Everyone has access to the code of Gnome, KDE, and XFCE, but there's still dozens of different window managers with their own user base. 2010-10-13T18:14:44 Just because there's an open-source bot, doesn't mean all or even many of the entires must be basaed on it. 2010-10-13T18:14:44 bss03: but gnome, kde, and xfce are not competing in an ai contest 2010-10-13T18:14:46 XWindows is not a competition with a deadline. 2010-10-13T18:15:20 jmcarthur: Since when has raising the bar for competition been bad? 2010-10-13T18:15:55 raising the bar should mean writing a better bot, not giving away source code to everyone 2010-10-13T18:15:55 If you can't beat code I'm willing to give to anyone, so you really be placed above average in the competition? 2010-10-13T18:15:58 bss03: let's clarify with organizers and have them make decision on this and update the rules section on the web site 2010-10-13T18:16:08 Yeah, I agree. 2010-10-13T18:16:13 bss03: raising the bar is fine. giving everybody a trampoline isn't, because then everybody has to learn how to beat a trampoline 2010-10-13T18:16:23 *** awr41 has joined #aichallenge 2010-10-13T18:16:24 Still, I'm gonna push up a commit to remove access to my bot via github in the meantime. 2010-10-13T18:16:24 i don't think we should change the rules in the middle of the contest 2010-10-13T18:16:51 but for a later contest i would vote against allowing already-open-source bots to enter 2010-10-13T18:16:54 If a T.O. gives a nod to open-sourcing the bot, I can always but it back up then. 2010-10-13T18:16:54 jmcarthur: it wouldn't be a rule change, just a broader definition of unsportsman like 2010-10-13T18:17:04 If they think it is a grey area, I'll leave it down. 2010-10-13T18:17:17 bss03: thanks for listening. I think waiting for T.O. is great. 2010-10-13T18:17:49 jmcarthur: I'd recommend a FAQ entry. 2010-10-13T18:17:57 i kind of see an open source bot like i would see a poker player showing his hands while he plays and talking about why he is making particular decisions openly for all spectators to hear 2010-10-13T18:18:02 "Should I open-source my bot while the tournament is ongoing?" etc. 2010-10-13T18:18:19 jmcarthur: I don't think other poker players would have much of a problem with that. 2010-10-13T18:18:20 in a poker tournament 2010-10-13T18:18:34 contestbot later tell amstan Many would like an official word on the fairness of open-sourcing good bots. 2010-10-13T18:18:34 McLeopold: I come to serve. 2010-10-13T18:18:43 jmcarthur: It makes him pretty easy to beat; just one real improvement on his strategy and BAM! 2010-10-13T18:18:44 not a problem, per se, but it does mean that all the other poker players have to be at least that good now 2010-10-13T18:18:48 well...i would say more that he's doing that during the tournament, but not actually while playing 2010-10-13T18:18:54 yeah, that's not the aspect i was trying to focus on 2010-10-13T18:18:56 *** RainCT has quit IRC (Remote host closed the connection) 2010-10-13T18:19:07 Zannick: yeah 2010-10-13T18:19:20 bss03: exactly, i was actually considering downloading your bot and using it as a benchmark like the starter bots 2010-10-13T18:19:24 like standing over a table and talking about the game while the players are trying to come up with their own tactics 2010-10-13T18:20:05 Or like that scene from "The Office" where Jim and Creed play chess... 2010-10-13T18:20:17 never saw that one 2010-10-13T18:20:35 but i can guess 2010-10-13T18:21:08 *** jukkerknott has quit IRC (Ping timeout: 265 seconds) 2010-10-13T18:22:59 bss03: do appreciate you listening too. 2010-10-13T18:24:08 as do it. hard to do that while disagreeing 2010-10-13T18:24:11 *as do i 2010-10-13T18:24:46 there's also gray area. for example, we don't complain about the example bots 2010-10-13T18:24:52 not that that says much, but still 2010-10-13T18:25:06 surely even slightly more advanced bots would be permissible to be open source. where do we draw the line? 2010-10-13T18:25:14 RebelXT: Does GitHub provide private repos for free accounts? 2010-10-13T18:25:40 bitbucket.org does 2010-10-13T18:25:43 bss03: no. 2010-10-13T18:25:47 but it uses hg, not git 2010-10-13T18:25:48 bss03: unfortunetly it's 5$ per month to have private repos :( bitbucket does i think 2010-10-13T18:26:10 bitbucket used to be 1 private repo per free account, but has opened to more :D 2010-10-13T18:26:17 cause they got bought 2010-10-13T18:27:04 *shrug* I guess I'll just reconfigure things to push/fetch from my VPS. 2010-10-13T18:27:24 It was nice using GitHub as staging between my desktop and laptop. 2010-10-13T18:28:10 bss03: some people use git and dropbox for that - works fine if there is only one developer sharing code between diff machines 2010-10-13T18:28:47 I pretty much always have SSH access to my VPS, so that'll work for me. 2010-10-13T18:29:02 i.e. http://stackoverflow.com/questions/1960799/using-gitdropbox-together-effectively 2010-10-13T18:29:14 ahh cool 2010-10-13T18:31:49 alephnullplex still has references to some of my commits. 2010-10-13T18:32:28 Looks like basically submission 9 and it's big bugfix. 2010-10-13T18:35:30 *** Bobng has quit IRC (Ping timeout: 264 seconds) 2010-10-13T18:36:38 http://github.com/stephen-smith/ai-contest-2010 -- should be just the starter kit plus a note now. 2010-10-13T18:36:43 *** JamesMG has quit IRC (Quit: Leaving.) 2010-10-13T18:39:05 contestbot later tell amstan How to the organizers feel about me developing my bot in the open during the contest? I had it on GitHub, but have removed it temporarily until I can get some official feedback on that. 2010-10-13T18:39:05 bss03: I come to serve. 2010-10-13T18:47:07 *** mager has joined #aichallenge 2010-10-13T18:47:30 hello 2010-10-13T18:49:20 I'm trying to get started with the C++ starter kit, but I have no idea how to use the visualizer with it. 2010-10-13T18:51:33 in the same way as with java 2010-10-13T18:51:54 just put a name of your file instead of java -jar .... 2010-10-13T18:52:12 java -jar example_bots/RageBot.jar 2010-10-13T18:52:17 bss03: where? 2010-10-13T18:52:31 what do i compile it as? exe or dll? 2010-10-13T18:53:16 are you able to compile it to exe? 2010-10-13T18:53:18 jmcarthur, McLeopold, bss03: i'm here, so what? 2010-10-13T18:53:19 *** bhasker has joined #aichallenge 2010-10-13T18:53:42 what should i compile it as? 2010-10-13T18:53:55 because im using the default mybot.cc right now and its giving errors 2010-10-13T18:54:07 well, most people will use g++ which will compile into an executable with no extension 2010-10-13T18:54:20 what compiler? 2010-10-13T18:54:37 the starter pack came with a makefile, no? 2010-10-13T18:54:42 *** bhasker_ has quit IRC (Ping timeout: 240 seconds) 2010-10-13T18:55:03 the makefile isn't g++ plus "make" probably doesn't work on windows 2010-10-13T18:55:03 :P 2010-10-13T18:55:19 learn 2 use make 2010-10-13T18:55:25 i use Visual C++ 2008 2010-10-13T18:55:33 ok, sorry 2010-10-13T18:55:48 amstan_: can we get an official word on the fairness of open-sourced bots? 2010-10-13T18:55:56 mager: what's the error? 2010-10-13T18:55:58 WARNING: player 2 crashed 2010-10-13T18:56:05 *** davidjliu has joined #aichallenge 2010-10-13T18:56:09 To me, it seems to violate good sportsmanship. 2010-10-13T18:56:14 i'm not sure, this should be on a case by case basis 2010-10-13T18:56:14 oh, so you did get it to compile? 2010-10-13T18:56:36 Exception in thread "main" java.lang.NullPointerException at Engine.main 2010-10-13T18:56:38 by sportmanship we mostly meant don't deprive other bots of cpu power 2010-10-13T18:56:56 using other people's code is fine 2010-10-13T18:57:02 it was in the RPS one 2010-10-13T18:57:08 RPS? 2010-10-13T18:57:16 rock paper scissors 2010-10-13T18:57:38 Exception in thread "Thread-2" java.lang.ArrayIndexOutOfBoundsException : 1 at ViewerPanel.run at java.lang.Thread.run 2010-10-13T18:57:49 So, if I made it to the #1 spot, then released by code so everyone could use it, you'd be okay with that? 2010-10-13T18:58:11 Released in such a form that everyone could submit it as is. 2010-10-13T18:58:13 well 2010-10-13T18:58:22 you'd only doing harm to yourself 2010-10-13T18:58:35 true facts 2010-10-13T18:58:41 I would argue I'd harm the whole competition. 2010-10-13T18:58:44 it does deprive noobs of the learning opportunity, but this is a contest 2010-10-13T18:58:46 not a classroom 2010-10-13T18:59:00 McLeopold: not really, how? 2010-10-13T18:59:01 so if i made it to the #1 spot, then released older code that looks right but actually doesn't quite work... 2010-10-13T18:59:02 amstan_: no, this breaks the whole competition imho 2010-10-13T18:59:02 a contest where you cheat and win? 2010-10-13T18:59:27 What about the rule that you can't have control of more than one account, but multiple people must join as a team? 2010-10-13T18:59:28 could someone kindly tell me w hat those runtime errors mean :) 2010-10-13T18:59:36 personally i think more than winning the competition its the learning experience that matters for me 2010-10-13T18:59:41 If there is an open-source effort, how do you police that? 2010-10-13T18:59:50 McLeopold: that's for protecting ranks, so you can't artifically boost them 2010-10-13T18:59:57 yeah, next competition let's have a game where the bots have to cheat to win 2010-10-13T19:00:11 mager: so, "player 2 crashed" means there was a problem running the bot 2010-10-13T19:00:23 what happens when you run MyBot.exe from the command line? 2010-10-13T19:00:38 yes - but its the default bot that came with the starter package 2010-10-13T19:00:40 So, your only concern is for bots to play fairly, not people? 2010-10-13T19:00:46 it just sits at a blank black cmd prompt 2010-10-13T19:01:00 okay, type "go" and hit enter 2010-10-13T19:01:04 Top 10 players: rsergio(3992), Hazard(3807), felixcoto(3791), dmj111(3772), davidjliu(3705), MariusCG(3683), ruilov(3679), asavis(3666), savchenko(3663), sequoh(3648) 2010-10-13T19:01:06 McLeopold: it's just impossible to police such things 2010-10-13T19:01:17 if it says "go" back to you, it's working 2010-10-13T19:01:25 it repeated 'go' to me 2010-10-13T19:01:27 Yes, but you could have an official position on the matter even if you don't enforce it. 2010-10-13T19:01:30 so instead of giving reasons for them to try to avoid them and potentially break even more important rules, we should just allow them 2010-10-13T19:01:31 amstan_ : I had my bot on GitHub, LGPL license. I had a few requests to take it down. A few people thought having my bot available was unfair or was encouraging unsportsmanlike conduct in others. 2010-10-13T19:01:33 okay, good 2010-10-13T19:01:46 so if it works...why doesn't it work lol :P 2010-10-13T19:01:52 bss03: let them complain them 2010-10-13T19:01:54 so the problem is probably with the command you gave the PlayGame.jar 2010-10-13T19:02:00 i don't think we can ban people for giving their code away 2010-10-13T19:02:25 No, but you could discourage it. But if official word is it's okay... 2010-10-13T19:02:33 ...Ouch. 2010-10-13T19:02:44 i don't see the point of such things though 2010-10-13T19:02:54 hah finally 2010-10-13T19:02:55 http://72.44.46.68/canvas?game_id=474056 2010-10-13T19:02:57 *** deepblue has quit IRC (Quit: Page closed) 2010-10-13T19:03:28 don't make it a rule, but strongly frown upon? idk 2010-10-13T19:03:28 amstan_: right, no need to ban or police anyone. an official statement on code sharing would be helpful though 2010-10-13T19:03:28 i copied the command from the 5 minute quick tutorial 2010-10-13T19:03:31 Would you be willing to make that more clear in the rules section, that open-source and code sharing is allowed? 2010-10-13T19:03:39 amstan_: I just wanted some official input as to whether it was okay; I don't want to be skirting the rules. 2010-10-13T19:03:51 mcleopold check that game out:D 2010-10-13T19:03:55 so... if there's such concerns i guess it would be nice to give other users the learning opportunity of making the bot from scratch 2010-10-13T19:04:02 but we cannot make it a rule, it's impossible to enforce 2010-10-13T19:04:03 i finally managed to beat CCCF:D 2010-10-13T19:04:19 bss03: you are not against the rules.. afaik 2010-10-13T19:04:27 amstan_: I have a request (not a requirement) in the LICENSE that they not submit the bot unmodified, but it really was there so anyone (that honors the LGPL) could build upon it. 2010-10-13T19:04:48 sure 2010-10-13T19:04:57 so... it's the same as reading a book 2010-10-13T19:05:18 you try to include spoiler alerts everywhere 2010-10-13T19:05:24 amstan_: I'd like to see the competition be, "write the best bot", not "tweak the best bot", but that is my opinion. 2010-10-13T19:05:51 it's write the best bot with whatever resources you can find 2010-10-13T19:05:54 amstan_: and you're right, it's unenforable. 2010-10-13T19:06:00 Well, if they modfied it, they'd have to use the GPL or LGPL for their bot so I could use their code as well. ;) 2010-10-13T19:06:14 bss03: not really 2010-10-13T19:06:21 amstan_, can you guys discuss this issue among organizers before you make a decision? we had a lengthy discussion here, and also on the forums - http://ai-contest.com/forum/viewtopic.php?f=17&t=886 on this topic. 2010-10-13T19:06:24 bss03: for that you need to make it gpl affero 2010-10-13T19:06:33 bss03: they're not obligated to send you the source, nor the binary 2010-10-13T19:06:43 amstan_: Yes really; uploading it to the competition is distribution. 2010-10-13T19:07:03 *** mogron has quit IRC (Ping timeout: 265 seconds) 2010-10-13T19:07:18 bss03: you only have to give the source to the people that also have the binary 2010-10-13T19:07:32 bss03: since nobody has the binary, nobody gets the source either 2010-10-13T19:07:33 amstan_: True 'nuf. 2010-10-13T19:07:35 amstan_, i'm strongly with McLeopold on this - let people share binaries and libraries (utility code, starter packs) but draw a line somewhere to disallow mindless bot tweaking. just my $0.02 of course 2010-10-13T19:08:00 rebelxt: so.. idk, i actually did this in RPS, that's how i won it 2010-10-13T19:08:06 you might boo me all you want 2010-10-13T19:08:25 i just took a previously made up bot with some extra 50 lines of code. 2010-10-13T19:08:38 for detecting if others were also this epic bot 2010-10-13T19:08:50 i ran 2 instances of it 2010-10-13T19:09:41 bhasker: why is that game interesting? 2010-10-13T19:09:55 :D cause i managed to beat CCCF never done that before ! 2010-10-13T19:10:15 Well, I'm gonna leave the code down for now and mull it over. I like developing in the open, and if more people send up being able to read/write some Haskell, I'll count that as a win... 2010-10-13T19:10:30 ...even if I lose to a slightly tweaked version of my bot. 2010-10-13T19:11:17 amstan_, no worries - i'm not here to boo anyone. i'm here to have fun.. i was just expressing my personal opinion.. i have 5 co-workers in the same office with me, all of us actively participating in the contest. all 5 think open sourcing top bots until Nov 27th is a a move which will hurt/break the competition 2010-10-13T19:12:08 well.. who do you think will be negatively affected if you do this? 2010-10-13T19:12:23 I wouldn't call my top a "Top Bot" anymore; is top has been getting tougher and my bot hasn't really gotten much better is a few weeks. 2010-10-13T19:12:27 *** Olathe has quit IRC (Remote host closed the connection) 2010-10-13T19:13:31 bss03: what language is it? 2010-10-13T19:13:34 amstan_: everyone bss03 can beat now? 2010-10-13T19:13:48 amstan_: My bot? Haskell. 2010-10-13T19:13:58 McLeopold: so.. you're worried about person B, that's at the same rank as person A 2010-10-13T19:14:07 but person A copied the bot from bss03? 2010-10-13T19:14:25 yes 2010-10-13T19:14:51 but person B has the same opportunity to copy bss03 2010-10-13T19:15:42 true, but right now, some people think that is unfair. So an official word from the officials would help clear it up. 2010-10-13T19:16:08 That way everyone has the same expectations of what is considered fair. 2010-10-13T19:16:19 amstan_: well, i'm in 20 place overall, and there are 712 other python submissions. so i flip a switch and open source my bot. 200 "cheaters" take my code and submit it as theirs. 200 cheaters fly on the leaderboard and crush the competition.. people affected are those who want to keep working on their bot using their own ideas and code 2010-10-13T19:17:10 amstan_, that's the point, tons of people don't feel like having to copy someone else's code or algorithm just to stay competetive 2010-10-13T19:17:13 The sportsmanship rule does not make it clear that it applies to bots and not people. 2010-10-13T19:17:41 *** davidjliu has quit IRC (Ping timeout: 265 seconds) 2010-10-13T19:17:50 Well... it does say entries, but still, I think some clarification is in order. 2010-10-13T19:17:52 *** Appleman1234 has quit IRC (Ping timeout: 264 seconds) 2010-10-13T19:17:58 and some more intrepid "cheaters" tweak their code to beat yours 2010-10-13T19:19:00 "people affected are those who want to keep working on their bot using their own ideas and code" << this is also an argument against sharing strategies 2010-10-13T19:19:47 i will share my strategies only when i have determined they are insufficient to put me in the top n. 2010-10-13T19:19:52 for small n 2010-10-13T19:20:04 Also, the rule about only one account should have some clarification as well. You have a collaborative open-source effort, you a effectively putting code in someone else's submition. 2010-10-13T19:21:05 sigh: well, kind of, but lots of room for idea sharing - i pretty much listed most of my heuristics on the forums at a high level.. http://ai-contest.com/forum/viewtopic.php?f=17&t=856 2010-10-13T19:21:45 amstan_ must be rewriting bss03's entry and submitting it right now... 2010-10-13T19:22:46 *** Meatkat has joined #aichallenge 2010-10-13T19:23:11 Hey all, who's medrimonia_15.2 on the tcp? 2010-10-13T19:24:29 *** jukkerknott has joined #aichallenge 2010-10-13T19:24:46 I'm just saying a lot of arguments about this are not black and white... it seems ultimately that people just find it intrinsically unfair that someone would get a higher ranking with little/no work. This is not a bad sentiment, jut that they try to rationalise with (I find) arguments that are encompass a lot more than they intend 2010-10-13T19:27:23 It appears the biggest impact of releasing a top bot will be people' perceptions of the tournament, and people's motivations. This in 2010-10-13T19:28:04 (Ignore the "This in") 2010-10-13T19:28:33 Did someone else post source code to a top bot? or are we talking about the one that was already posted and removed? 2010-10-13T19:28:45 hypotheticals, now 2010-10-13T19:29:01 sigh: exactly right 2010-10-13T19:30:39 sigh: agree+=1 2010-10-13T19:30:46 "but person B has the same opportunity to copy bss03" - i really hope this does not become official word on open sourcing bots 2010-10-13T19:31:13 *** jukkerknott has quit IRC (Ping timeout: 265 seconds) 2010-10-13T19:32:47 Meatkat: yes, there was another open source bot which was top50 at the time.. he took it down after forum discussion and described his logic in pseudo code: http://ai-contest.com/forum/viewtopic.php?f=17&t=886 2010-10-13T19:33:13 rebelxt: yeah, that's the one I was talking about 2010-10-13T19:38:45 btw, was this an issue in the tron contest? 2010-10-13T19:38:58 how was is handled there? 2010-10-13T19:39:01 http://ai-contest.com/forum/viewtopic.php?f=18&t=888&view=unread#p5933 that post is very much what I would expect to see happen if the compile race bug was actually occurring on the cloud server 2010-10-13T19:39:40 *** amstan_ has left #aichallenge ("Konversation terminated!") 2010-10-13T19:40:01 *** McLeopold has quit IRC (Quit: Leaving) 2010-10-13T19:40:21 *** lavalamp has quit IRC (Quit: Page closed) 2010-10-13T19:40:51 janzert: the two compilations for one submission here: http://code.google.com/p/ai-contest/issues/detail?id=194 is what i'd expect to see if the compile race was occurring 2010-10-13T19:41:16 no, the cloud compilations aren't recorded or reported back to the user 2010-10-13T19:41:22 *** rebelxt has left #aichallenge 2010-10-13T19:41:25 er, oh right 2010-10-13T19:41:40 its been about a minute and the visualizer still isn't popping up : ( 2010-10-13T19:41:58 the worker=0 comments are probably due to it being the only one most of the time 2010-10-13T19:42:05 i assume it takes longer when its 648 turns long of a battle. 2010-10-13T19:42:12 oh, yeah. 2010-10-13T19:42:19 : ( 2010-10-13T19:42:34 well worker=0 is the main server and I think is using different code in many places than the other servers 2010-10-13T19:42:37 change the second 1000 to 200. 2010-10-13T19:42:46 so not too surprising it has different bugs 2010-10-13T19:43:52 for instance I'm pretty sure the main server is not recording errors still 2010-10-13T19:45:02 that was what i was thinking wrt those errors 2010-10-13T19:45:40 * janzert nods 2010-10-13T19:47:23 the two emails thing, I have to wonder if he tried uploading and when it was taking to long refreshed the page and caused it to upload a second time 2010-10-13T19:47:41 so even though he thought it was only one submission it ended up being two 2010-10-13T19:47:56 doesn't explain why the first one failed though of course 2010-10-13T19:48:06 it would explain that, maybe 2010-10-13T19:48:33 if it tried to start compiling a source file, failed quietly because the other thread was compiling it... 2010-10-13T19:48:38 be nice to know what profile it was so you could see if it was two submission ids 2010-10-13T19:48:47 yeah, that doesn't make enough sense since each file is a separate comman 2010-10-13T19:48:49 command 2010-10-13T19:49:52 http://ai-contest.com/profile.php?user_id=9336 2010-10-13T19:50:43 oct 13: 9:43, 9:44 2010-10-13T19:50:58 er, no, that's two different languages 2010-10-13T19:51:27 yeah, that's kinda weird in itself 2010-10-13T19:51:36 maybe accidental uploaded the wrong file 2010-10-13T19:52:02 then uploaded the correct one really fast, causing two threads to try to compile his latest submission 2010-10-13T19:52:13 there is 9:44 9:45 but it says both compile fine 2010-10-13T19:52:58 * Zannick shrugs 2010-10-13T19:53:05 I guess it doesn't really help :/ 2010-10-13T19:53:48 definitely need to guard against races better 2010-10-13T19:53:55 well gotta run for a few hours 2010-10-13T19:54:11 separate uploads should be getting different submission ids so no race 2010-10-13T19:54:22 well, the compile fix solved the problem with C not compiling :D http://ai-contest.com/language_profile.php?lang=C 2010-10-13T19:54:22 * janzert shrugs 2010-10-13T19:54:35 yay 2010-10-13T19:54:44 only if they're using submission ids, and not user ids to retrieve the latest id 2010-10-13T19:56:32 IIRC compile_daemon doesn't look at user ids at all does it? 2010-10-13T19:56:43 been a while since I was digging in it though 2010-10-13T19:57:00 i'll look later 2010-10-13T19:57:07 * Zannick work for 1 more hour 2010-10-13T19:57:31 yeah, actually I need to go finish off some work as well 2010-10-13T19:57:36 cya later 2010-10-13T19:57:39 see ya 2010-10-13T20:00:28 *** bhasker_ has joined #aichallenge 2010-10-13T20:01:04 Top 10 players: rsergio(3921), Hazard(3790), felixcoto(3783), dmj111(3758), davidjliu(3705), MariusCG(3675), ruilov(3674), savchenko(3660), asavis(3656), sequoh(3626) 2010-10-13T20:01:26 hmm....my AI gets owned by the DualBot 2010-10-13T20:01:46 not very intelligent :) 2010-10-13T20:03:57 Artificial ~intelligence 2010-10-13T20:04:00 *** bhasker has quit IRC (Ping timeout: 272 seconds) 2010-10-13T20:04:13 artificial innelegance :) 2010-10-13T20:04:26 Heh. 2010-10-13T20:04:32 lol :P 2010-10-13T20:05:18 There really needs to be a programming contest tracker. 2010-10-13T20:06:39 *** jaspervdj has quit IRC (Quit: NEVER GONNA GIVE YOU UP NEVER GONNA LET YOU DOWN) 2010-10-13T20:07:01 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-10-13T20:15:25 *** jukkerknott has joined #aichallenge 2010-10-13T20:27:42 *** jukkerknott has quit IRC (Quit: Page closed) 2010-10-13T20:31:48 *** rwa has quit IRC (Quit: rwa) 2010-10-13T20:34:32 my AI is really failing... 2010-10-13T20:34:47 improve it 2010-10-13T20:34:51 trying 2010-10-13T20:39:41 *** tapwater has quit IRC (Remote host closed the connection) 2010-10-13T20:39:54 *** bhasker_ has quit IRC (Quit: bhasker_) 2010-10-13T20:39:54 *** tapwater has joined #aichallenge 2010-10-13T20:41:01 *** jukkerknott has joined #aichallenge 2010-10-13T20:42:37 *** bhasker has joined #aichallenge 2010-10-13T20:43:55 *** bhasker_ has joined #aichallenge 2010-10-13T20:45:50 *** Azrathud has quit IRC (Quit: Meh) 2010-10-13T20:46:36 *** Mathnerd314 has quit IRC (Ping timeout: 240 seconds) 2010-10-13T20:46:51 *** bhasker has quit IRC (Ping timeout: 252 seconds) 2010-10-13T20:48:50 *** jukkerknott has quit IRC (Quit: Page closed) 2010-10-13T20:54:22 *** Azrathud has joined #aichallenge 2010-10-13T21:01:04 Top 10 players: rsergio(3922), Hazard(3788), felixcoto(3785), dmj111(3761), davidjliu(3707), ruilov(3686), MariusCG(3677), savchenko(3659), asavis(3658), sequoh(3637) 2010-10-13T21:03:43 *** kaerun has joined #aichallenge 2010-10-13T21:04:28 *** tapwater has quit IRC (Remote host closed the connection) 2010-10-13T21:04:47 *** tapwater has joined #aichallenge 2010-10-13T21:06:52 *** Meatkat has quit IRC (Quit: Leaving) 2010-10-13T21:09:59 *** bhasker_ has quit IRC (Ping timeout: 250 seconds) 2010-10-13T21:13:37 *** Mathnerd314 has joined #aichallenge 2010-10-13T21:14:30 *** ratatata has quit IRC (Quit: Page closed) 2010-10-13T21:14:54 Does anyone else find it far easier to plan out ideas on paper than on a computer? 2010-10-13T21:20:09 *** mager has quit IRC () 2010-10-13T21:20:37 *** janzert_ has joined #aichallenge 2010-10-13T21:21:26 depends on the idea 2010-10-13T21:21:47 jmcarthur, there are ideas that are easier to type up rather than write down? 2010-10-13T21:22:17 obviously minor changes aren't worth writing out if you can just change a few lines and fix something 2010-10-13T21:22:26 i think so yes 2010-10-13T21:22:31 But I find big things (like refactoring) are best planned out on paper 2010-10-13T21:22:45 i typically only resort to paper when the idea is very visual (in my head) already 2010-10-13T21:22:49 I'm pretty much completely rewriting how my bot thinks 2010-10-13T21:23:11 wow... did I just somehow type a bunch of names? 2010-10-13T21:23:15 i do a lot of graph algorithms on paper. also anything relating to geometry 2010-10-13T21:23:34 *a lot of things related to geometry 2010-10-13T21:23:45 but even then... 2010-10-13T21:23:55 it's usually just to work an example to test my idea, not to write the idea itself out 2010-10-13T21:24:05 sometimes it's to show my idea to somebody else 2010-10-13T21:24:21 it's rarely to help me actually come up with the idea itself 2010-10-13T21:24:46 do you normally write text or draw diagrams when you do this? 2010-10-13T21:25:26 atm I'm mostly writing text, I never write actual code on paper 2010-10-13T21:25:35 well sure 2010-10-13T21:25:48 i do write math equations on paper sometimes, actually 2010-10-13T21:26:15 I agree it's very useful to test out an idea, but I also use it for general planning 2010-10-13T21:26:18 sometimes even in a pseudocode... but it's not really what most people would call pseudocode 2010-10-13T21:26:50 i tend to write more type signatures than code 2010-10-13T21:26:58 occasionally I'll have some pseudocode, but that's not too often 2010-10-13T21:35:06 javagamer: I have a notebook which I take with me on the train to work to jot ideas down on 2010-10-13T21:35:09 it helps a lot 2010-10-13T21:36:27 most of my planning is done then :) 2010-10-13T21:37:46 Heh. At first I was thinking you were talking about a notebook laptop. 2010-10-13T21:38:29 pen and paper :P 2010-10-13T21:39:15 Do you really use a pen? 2010-10-13T21:39:19 Hehe. 2010-10-13T21:40:33 as opposed to? 2010-10-13T21:40:43 *** bhasker has joined #aichallenge 2010-10-13T21:40:47 * sigh feels like a joke just flew over his head 2010-10-13T21:42:19 Pencil. 2010-10-13T21:44:14 with stuff like this I don't like erasing anything, so pencil doesn't give me any advantage 2010-10-13T21:44:28 *** rebelxt_ has joined #aichallenge 2010-10-13T21:44:40 sigh is dijkstra after all 2010-10-13T21:44:58 what did dijkstra do? 2010-10-13T21:45:06 he never erased 2010-10-13T21:45:25 he would take the time to make sure he never made a mistake when writing, basically 2010-10-13T21:45:43 he never started writing without knowing exactly what he was going to write and without knowing that it was correct 2010-10-13T21:46:00 haha, I'm nothing like that 2010-10-13T21:46:03 :P 2010-10-13T21:46:06 I just like a record of my mistakes :P 2010-10-13T21:46:19 when I was in uni I used pencil for assignments and exams so I could correct myself 2010-10-13T21:46:36 but stuff like this is different 2010-10-13T21:50:49 It sounds like he's missing (at least part of) the point of writing stuff down if he always knows exactly what he's going to write beforehand 2010-10-13T21:51:29 This writing is really helping me brainstorm, I just got a bunch of new ideas and I think that a lot of behaviors I was specifically coding in will become emergent with my new idea 2010-10-13T21:52:04 javagamer: he wrote to remember later or to explain to others, not to sketch out and idea and see if it works 2010-10-13T21:52:10 *an idea 2010-10-13T21:52:35 *** Naktibalda has quit IRC (Remote host closed the connection) 2010-10-13T21:53:13 I get that, I'm just saying sketching out ideas can be really useful, but I guess it really can vary from person to person 2010-10-13T21:53:51 I much prefer not to have physical documents, but I can plan so much better this way, so whenever I resort to paper I can tell I'm really involved in something 2010-10-13T21:54:06 yeah i don't really work that way 2010-10-13T21:54:23 on the other hand, i'm not like dijkstra either 2010-10-13T21:54:48 i can't just not write or type things out before i have completed the idea, that is 2010-10-13T21:55:18 but the act of writing isn't to help with my idea, it's so i can get parts of it out of my head to recall later while i think about other parts in the meantime 2010-10-13T21:56:24 i have about 20 pages in a cheap notebook filled with scratch notes on bot ideas. 2010-10-13T21:56:31 lots of whitespace :) 2010-10-13T21:56:46 i have a single page on it 2010-10-13T21:56:50 very dense 2010-10-13T21:57:09 it was only for a single algorithm i was coming up with though, and that because i was going through it with my wife 2010-10-13T21:57:17 cool. 2010-10-13T21:59:12 Whenever I just want to remember something I just type it up, the only benefit I really see in handwriting something is to brainstorm 2010-10-13T21:59:37 Unless of course I don't have access to a computer 2010-10-13T22:01:04 Top 10 players: rsergio(3917), Hazard(3783), felixcoto(3780), dmj111(3757), davidjliu(3703), MariusCG(3683), ruilov(3672), savchenko(3655), asavis(3653), sequoh(3632) 2010-10-13T22:03:02 yup, unless paper is easier... like for diagrams 2010-10-13T22:03:58 here we go again, another major algorithm improvement, and bot becomes weaker overall against other opponents on tcp server 2010-10-13T22:04:27 *** qotile has joined #aichallenge 2010-10-13T22:04:46 is it stronger against your other versions? 2010-10-13T22:05:17 sigh, yes it beats previous versions 2010-10-13T22:05:37 yeah, hate when that happens 2010-10-13T22:08:38 sigh: it's pretty interesting. my bot uses it's own strategy to figure out most likely enemy's response to its actions. if i relax this assumption, my bot does better overall... great! 2010-10-13T22:09:32 yeah, I haven't gotten around to implementing that yet because I was worried the same thing would happen to me 2010-10-13T22:10:19 are you doing some sort of heavily pruned minimax? 2010-10-13T22:10:36 or else, how are you using the enemie's predicted moves? 2010-10-13T22:12:08 dunno how I thought "enemie's" was a valid thing to type 2010-10-13T22:12:20 rebelxt_: i was wondering about that ... but haven't tried it yet. 2010-10-13T22:12:58 yeah, using the same scoring function to predict enemies next move.. but it's not working out well at all so far 2010-10-13T22:13:47 my old bot is best when the enemy makes mistakes, so I was thinking of using _its_ logic instead of the new logic for that part. 2010-10-13T22:14:01 s/is best/is at its best/ 2010-10-13T22:14:17 maybe it's a bug, i'm not sure yet 2010-10-13T22:15:38 dmj111 and sigh, you guys are both doing better than i am at the moment, so maybe i'm not moving in the right direction 2010-10-13T22:16:07 honestly, I've just gotten lucky with my heuristics 2010-10-13T22:16:26 I've been veyr lazy with implementing more sophisticated algorithms 2010-10-13T22:17:49 sigh: in my case heuristics still beat the crap out of more advanced calculations ;-) 2010-10-13T22:18:16 my guesses on heuristics have been lucky. i have had to work through plenty of bugs, though, that originally led me astray from the main ideas. 2010-10-13T22:19:33 dmj111: you are doing very well! i spent good amount of time trying to reverse engineer some of your heuristics ;-) 2010-10-13T22:20:25 yeah, when I started I thought I had a handle on this game because I got near the top very quickly... got knocked off my pedestal when "obvious" improvements kept failing and my bot kept dropping 2010-10-13T22:20:37 ahhh... thanks. i changed one recently, that seems like it _should_ be better, and am getting wallopped. 2010-10-13T22:21:41 do any of you guys have galcon 1vs1 experience? 2010-10-13T22:21:48 ignoring rage bot has made my code easier, and it appears to give other people some joy too. 2010-10-13T22:21:52 never played galcon :P 2010-10-13T22:22:06 no. I played the flash game a few hours before starting to get the gist. 2010-10-13T22:22:34 amazing! and i thought my background in galcon would be helpful... 2010-10-13T22:23:21 stuff computers are good at is very different from what humans are good at 2010-10-13T22:23:29 easy becomes hard and hard becomes easy :P 2010-10-13T22:24:05 *** hornairs has joined #aichallenge 2010-10-13T22:24:07 haha. hard to admit but i've played 30K 1on1 Galcon games, and i can't formalize my heuristics well enough to get into top10 2010-10-13T22:24:24 *** janzert_ has quit IRC (Quit: Ex-Chat) 2010-10-13T22:24:25 agree with sigh. also, I think the ability to predict the future exactly changes things a lot. 2010-10-13T22:25:13 rebelxt_: nothing to be ashamed of, I wouldn't expect a chess grandmaster to be able to write a good chess program off the bat 2010-10-13T22:25:51 hell, I play a decent game of Go and I have *no idea* how I would start writing a bot for that, other than what I've read of the other bots out there 2010-10-13T22:26:09 sigh: good point, but i'm also a programmer. this just proves how many really smart ppl there are 2010-10-13T22:29:18 dmj111: what's ur current avg turn time? 2010-10-13T22:30:08 (assuming you are still using python/psyco) 2010-10-13T22:30:53 sigh: you are c++, right? what's ur turn time like when playing locally? 2010-10-13T22:31:51 a couple of ms per move 2010-10-13T22:32:23 NEED MORE COMPUTATION TIIIIIIME 2010-10-13T22:32:27 shit >_> 2010-10-13T22:32:30 sigh, very nice.. really fast test cycles 2010-10-13T22:33:00 rebelxt_: in the one log I have handy, the first move is 10ms, and the rest are mostly < 7ms. 2010-10-13T22:33:06 yeah, although I'm going to start using up my time soon 2010-10-13T22:33:23 sorry... factor of 10. 100ms, and 70ms. 2010-10-13T22:33:24 I'm starting to explore approaches that generate a lot of game states to compare them 2010-10-13T22:33:39 same here 2010-10-13T22:34:02 does every bot gets its own core btw? 2010-10-13T22:34:13 no idea.... 2010-10-13T22:34:17 dmj111: ok, i feel a little bit better now.. 150ms average 2010-10-13T22:35:22 that sounds good to me. My logging is on there too... so i am not sure how much difference that makes. 2010-10-13T22:35:34 sigh: hope you have better luck than my first attempt at this 2010-10-13T22:36:07 yeah, I'm so hesitant because I *know* that I'm gonna come across a lot of dead ends 2010-10-13T22:36:27 so far waiting has paid off, because I hear about all the mistakes that others have made :) 2010-10-13T22:37:13 haha 2010-10-13T22:38:04 honest, I'm most interested in what McLeopold comes up with 2010-10-13T22:38:10 he has some really interesting ideas 2010-10-13T22:38:33 and his bot in the main contest is a month old and still at 21 2010-10-13T22:38:39 yeah his current bot is Sep 16th, so he must have made good progress 2010-10-13T22:39:09 s/honest/honestly 2010-10-13T22:39:50 *** qotile has quit IRC (Quit: Page closed) 2010-10-13T22:40:39 he said he's waiting until he has a bot that will "punch everyone in the face" 2010-10-13T22:42:34 wow, my recent bot is right next to his month-old bot... /scared of his new bot 2010-10-13T22:45:39 *** deepblue has joined #aichallenge 2010-10-13T22:46:30 simulating the future is so version 1.0. i'm going to try simulating the past! ;-) 2010-10-13T22:47:16 If you can do that with a stateless bot, count me impressed 2010-10-13T22:47:45 mine so far is stateless 2010-10-13T22:47:53 i don't care about past :) 2010-10-13T22:48:03 mine too, stateless is easier to manage and test 2010-10-13T22:48:29 i agree 2010-10-13T22:48:32 and I see no point 2010-10-13T22:48:42 all important information is in the current state of the game 2010-10-13T22:49:03 same. i was just joking since i ran out of easy to implement ideas 2010-10-13T22:49:32 well, opponent behavior matters, especially this isn't a game with perfect information 2010-10-13T22:49:41 yes it is 2010-10-13T22:49:46 ? 2010-10-13T22:50:02 how is this not with perfect information? 2010-10-13T22:50:02 ok, don't know the correct term, but you don't know your opponent's current move 2010-10-13T22:50:11 unlike a turn based game 2010-10-13T22:50:25 so you can't find a specific optimal move 2010-10-13T22:50:56 would be boring otherwise :) 2010-10-13T22:51:20 plenty of interesting games where turns alternate :P 2010-10-13T22:51:47 sigh, exactly, my current bot reverses game's state and uses it's own algorithm to calculate opponent's current move.. then uses this to optimize my own current move. not working as expected though 2010-10-13T22:51:52 i mean this game :) 2010-10-13T22:52:20 that's quite an elaborate plan 2010-10-13T22:52:29 not sure if it's worth 2010-10-13T22:52:51 sounds like one deviation from its normal behaviour that the opponent does leads to misery ;) 2010-10-13T22:53:26 deepblue: unless i have a bug somewhere, it's much weaker than my previous bot which doesn't try to be smart 2010-10-13T22:53:45 i have the same problem 2010-10-13T22:53:56 my new ideas have a hard time beating my old versions 2010-10-13T22:54:44 i'm kinda glad to hear this ;-)... 2010-10-13T22:55:52 from hanging out here, you realise that everyone has the same problems 2010-10-13T22:56:13 haha 2010-10-13T22:56:15 indd 2010-10-13T22:56:48 right, but breakthroughs do happen periodically 2010-10-13T22:57:10 btw about the perfect information 2010-10-13T22:57:24 couldn't you argue that you don't know the opponent's next move? 2010-10-13T22:58:16 the point is that when your opponent is making a move, he knows what your move was 2010-10-13T22:58:28 sorry if perfect information is not the correct term 2010-10-13T22:58:28 yes 2010-10-13T22:59:01 well afaik its perfect info in that the game is fully observable 2010-10-13T22:59:22 but the action space is insane 2010-10-13T22:59:24 ok, easier way to put it... you don't know the exact state of the map when your orders take effect 2010-10-13T22:59:38 because your opponent made a move simultaneously 2010-10-13T23:00:14 indeed 2010-10-13T23:01:03 just because turns happen simultaneously I don't think disqualifies from being a game of perfect information 2010-10-13T23:01:05 Top 10 players: rsergio(3895), felixcoto(3762), Hazard(3759), dmj111(3722), davidjliu(3687), MariusCG(3686), ruilov(3670), savchenko(3660), asavis(3624), sequoh(3612) 2010-10-13T23:01:28 I'm basing my algorithm on that assumption :P 2010-10-13T23:02:40 janzert: yeah, I'm not sure of the exact definition 2010-10-13T23:03:35 "So for example, one may have complete information in the context of a Prisoner's Dilemma, but nonetheless this is a game of imperfect information since one does not know the action of the other player." - http://en.wikipedia.org/wiki/Complete_information 2010-10-13T23:03:57 best line to support my stance that I could find 2010-10-13T23:03:58 http://en.wikipedia.org/wiki/Perfect_information 2010-10-13T23:04:10 :) 2010-10-13T23:04:23 yeah, but that page doesn't clear up this issue 2010-10-13T23:04:39 I would argue that "In game theory, a game is said to have perfect information if all players know all moves that have taken place" does not hold 2010-10-13T23:04:43 "moves that have taken place" 2010-10-13T23:04:57 have is past tense so it does 2010-10-13T23:05:12 your move and opponent happen at same time so doesn't invalidate it 2010-10-13T23:05:17 in that case, is the prisoner's dilemma a game of perfect information? 2010-10-13T23:05:28 that would be my interpretation 2010-10-13T23:05:43 hmm, ok 2010-10-13T23:06:30 yeah, I guess when it's boiled down to that I would go the other way 2010-10-13T23:06:30 hehe 2010-10-13T23:06:38 sigh: i think that this would qualify as perfect information in game theory. but, i also get your intent that it is _not_ alternating turns. 2010-10-13T23:06:40 you can model it to be alternating moves, as long as the simulation runs after both players have made the move 2010-10-13T23:07:22 Error323: if you model it as alternating then it is clearly not perfect information as the "second" player has information hidden from them 2010-10-13T23:07:24 it's much easier to write a bot for that 2010-10-13T23:07:58 although prisoner's dilemna may not qualify because the game ends before you find out the other's move? 2010-10-13T23:08:27 so is iterated prisoner's dilemna considered 2010-10-13T23:08:30 janzert: iterated prisoner's dilemma then 2010-10-13T23:08:33 :P 2010-10-13T23:08:34 just looked it up... the difference is simultaneous vs. sequential games. 2010-10-13T23:08:46 sigh: not when you expose the first player's move after the second has made his move aswell 2010-10-13T23:09:22 after the first player has made his move, then that becomes part of the game state 2010-10-13T23:09:40 we should include a gambling part in the game 2010-10-13T23:09:41 dmj111: ahh good point. http://en.wikipedia.org/wiki/Game_theory#Simultaneous_and_sequential 2010-10-13T23:10:00 your bot learns about the opponents move as soon as he locks them in 2010-10-13T23:10:04 so you can gamble 2010-10-13T23:10:22 ok, that page settles it 2010-10-13T23:10:26 finish early and be safe or wait until the opponent decides but risk to time out 2010-10-13T23:10:27 :D 2010-10-13T23:10:32 this game is complete information but not perfect information 2010-10-13T23:10:43 yay 2010-10-13T23:10:46 everybody is right 2010-10-13T23:11:23 yep complete info 2010-10-13T23:13:00 simultaneous games are fundamentally different from sequential games, so that there is *some* distinction between the type of information available 2010-10-13T23:13:09 *so it makes sense 2010-10-13T23:13:16 *** davidjliu has joined #aichallenge 2010-10-13T23:13:28 but enough about davidjliu for now 2010-10-13T23:13:31 ^_^ 2010-10-13T23:13:39 lol? 2010-10-13T23:13:49 :P 2010-10-13T23:13:54 * deepblue says nothing 2010-10-13T23:13:55 ;-) 2010-10-13T23:14:25 luckily I'm very good at checking logs 2010-10-13T23:14:38 haha 2010-10-13T23:14:40 davidjliu: omg, erasing logs.. 2010-10-13T23:15:06 *** jukkerknott has joined #aichallenge 2010-10-13T23:15:21 *** hellman has joined #aichallenge 2010-10-13T23:15:29 will he find complete information? 2010-10-13T23:15:39 ooh, and apparently games have started up again! 2010-10-13T23:15:49 davidjliu: you won't believe this. we've just figured out that planet wars is simultaneous game with complete information ;-) 2010-10-13T23:16:03 they stopped? 2010-10-13T23:16:06 yeah, sounds true I guess 2010-10-13T23:16:11 yeah 2010-10-13T23:16:16 well, I mean, there was only one server for like a day 2010-10-13T23:16:17 not to be confused with perfect 2010-10-13T23:16:18 hadn't had a game for hours 2010-10-13T23:16:29 omg, 3 servers up! 2010-10-13T23:16:41 why is the elo gain on the tcp server so slow as well ? not enough bots up? 2010-10-13T23:16:55 bhasker: number of bots 2010-10-13T23:16:58 hi fives all around, wikipedia and three servers! 2010-10-13T23:17:06 hmm k 2010-10-13T23:17:30 my bot has been climbing like 4 elo ratins an hour or something 2010-10-13T23:17:31 :-\ 2010-10-13T23:17:38 i'll let mine run in a few minutes 2010-10-13T23:17:50 wondering if our discussion inspired cloud servers to start back up... 2010-10-13T23:18:24 oh hmm 2010-10-13T23:18:32 there are only 4 bots rated higher than mine running on tcp at this time 2010-10-13T23:19:45 people start up your bots 2010-10-13T23:19:53 i can start one of mine if you wan 2010-10-13T23:19:57 want* 2010-10-13T23:20:41 i started up one of mine 2010-10-13T23:21:07 on my single core machine... I have to run it when I am not around :) 2010-10-13T23:22:08 ok here you go 2010-10-13T23:22:10 mine is running too now 2010-10-13T23:22:18 ok i started my last official bot too 2010-10-13T23:22:34 (mine could be buggy) 2010-10-13T23:23:06 let's see whose is the buggiest bot.. wth 2010-10-13T23:23:50 dmj, gg 2010-10-13T23:24:48 hehe 2010-10-13T23:25:09 rebelxt that would be my bot 2010-10-13T23:25:40 bhasker: which one? oldman? 2010-10-13T23:25:51 no malazan 2010-10-13T23:26:16 ahh, i just played dmj, then oldman 2010-10-13T23:26:18 @34 now 2010-10-13T23:26:19 bhasker: I have no idea what you mean. 2010-10-13T23:26:37 doh 2010-10-13T23:26:38 contestbot: think 2010-10-13T23:26:39 rebelxt_: I'm sorry Dave, I cannot think. 2010-10-13T23:26:57 nice rebelxt :) i really need to work on the bot instead of reading the chat channel. 2010-10-13T23:27:03 i just beat deepblue 2010-10-13T23:27:06 :D 2010-10-13T23:27:13 http://72.44.46.68/canvas?game_id=477112 2010-10-13T23:27:14 hehe 2010-10-13T23:27:17 good job :) 2010-10-13T23:27:33 as a Dave, i always freak out a little at the "I'm sorry Dave" comments. 2010-10-13T23:27:41 dmj111: i tried improving my bot, and it's a dead end! ;-) 2010-10-13T23:27:45 dmj111: I'm sorry, Dave. 2010-10-13T23:28:21 how come i beat 5th ranked bot 2010-10-13T23:28:22 oh well 2010-10-13T23:28:23 rebelxt_: been there. a lot. since I use mercurial, I have a funny looking graph. 2010-10-13T23:28:27 but gain 1 elo point 2010-10-13T23:28:29 :( 2010-10-13T23:28:31 i messes something up 2010-10-13T23:28:37 Zannick: ha. 2010-10-13T23:29:15 meh lost by one ship 2010-10-13T23:29:16 http://72.44.46.68/canvas?game_id=477154 2010-10-13T23:30:40 bhasker: why did you attack that 10 neutral on first turn? ;-) 2010-10-13T23:31:01 well i jjust uncommented the code that ignores planets closer to enemy on first turn 2010-10-13T23:31:13 i saw that on some maps i lose because of that 2010-10-13T23:31:21 where he goes after my planets 2010-10-13T23:31:23 and i ignore his 2010-10-13T23:31:36 so i thought i will try it without that code 2010-10-13T23:31:53 though i am wondering if i should add some range 2010-10-13T23:31:59 if its further than a certain distance then ignore 2010-10-13T23:32:07 or something 2010-10-13T23:32:50 yeah it seems first turn is just ur half of the world and potentially same distance as enemy (depending on cost) 2010-10-13T23:33:07 who is oldman btw? 2010-10-13T23:33:10 no idea 2010-10-13T23:33:14 his bot is quite good 2010-10-13T23:33:21 yeah 2010-10-13T23:33:59 *** jukkerknott_ has joined #aichallenge 2010-10-13T23:35:22 oh well i want to see how high this bot can go 2010-10-13T23:35:30 still have a few planet selection bugs to fix 2010-10-13T23:36:39 I was having a decent battle with deepblue up to a point... my bot *forgot* to capture a 5 growth during the middle of battle: http://72.44.46.68/canvas?game_id=477233 2010-10-13T23:37:17 my bot does that if it gets behind too much 2010-10-13T23:37:24 it just doesn't try for the neutral after that 2010-10-13T23:37:30 and then its a losing battle from that point on 2010-10-13T23:37:57 basically i have some code to not attack neutrals if i am behind 2010-10-13T23:38:01 deepblue: gg. your expansion was faster than mine, as usual 2010-10-13T23:38:06 but that works like a two edged sword 2010-10-13T23:38:10 yeah 2010-10-13T23:38:12 double edged rather* 2010-10-13T23:38:19 i still have some issues with it though 2010-10-13T23:39:03 deepblue: i can't believe my bot did not go on attack when you took two smallest neutrals 2010-10-13T23:39:23 hihi 2010-10-13T23:39:27 *** Blkt has quit IRC (Quit: Error: do not makunbound t please!) 2010-10-13T23:39:42 *** jukkerknott has quit IRC (Quit: Page closed) 2010-10-13T23:40:02 lol my staging code bites me 2010-10-13T23:40:11 my bot decided it was better to send ships to the corner of the map 2010-10-13T23:40:16 http://72.44.46.68/canvas?game_id=477298 2010-10-13T23:40:17 :( 2010-10-13T23:40:35 and not sure what it was doing sending ships between planets and not capturing anything 2010-10-13T23:40:42 its great how we all have our small issues to iron out 2010-10-13T23:40:42 deepblue: you were down 160 ships for a moment and my bot did not profit. great ;) 2010-10-13T23:40:56 lucky me! 2010-10-13T23:41:15 bhasker: just beat me 2010-10-13T23:41:31 i did ? 2010-10-13T23:41:37 yeah 2010-10-13T23:41:39 i think... 2010-10-13T23:41:42 you are malazan? 2010-10-13T23:41:44 oh yea 2010-10-13T23:41:45 :D 2010-10-13T23:41:49 haha ok 2010-10-13T23:41:59 i thought i read that earlier... wasn't sure if it was you 2010-10-13T23:42:13 yea thats me 2010-10-13T23:42:21 at least that game i can see that my defense code is mostly working 2010-10-13T23:42:27 lots of planets going to zero 2010-10-13T23:43:38 bhasker: it's a nice feeling when that happens :) 2010-10-13T23:43:49 yea 2010-10-13T23:43:50 :D 2010-10-13T23:45:07 deepblue: was this your new bot? or ur last official one? 2010-10-13T23:45:50 who has the bot CCCF? 2010-10-13T23:45:56 no 2010-10-13T23:45:56 it's a little modified 2010-10-13T23:45:56 but partly sucks partly is awesome 2010-10-13T23:46:06 i mean 2010-10-13T23:46:13 modified from what is running on the official 2010-10-13T23:46:16 it's full of awesome! 2010-10-13T23:46:39 cccf is felixcoto afaik 2010-10-13T23:47:56 ahhh ok. 2010-10-13T23:48:11 thanks 2010-10-13T23:48:58 good to know :) 2010-10-13T23:50:15 my bot climbed 20 elo points since you guys started up your bots 2010-10-13T23:50:18 thank you ! 2010-10-13T23:50:31 it was annoying to see it move 3 points every hour 2010-10-13T23:51:36 glad to help! ;-) 2010-10-13T23:52:02 i see two disconnects here of malanzan 2010-10-13T23:52:22 !@#!@#!@# 2010-10-13T23:52:23 bhasker: User error -- Replace user. 2010-10-13T23:52:26 modifying code 2010-10-13T23:52:30 while running bot 2010-10-13T23:52:32 == bad idea 2010-10-13T23:52:35 syntax error 4tl 2010-10-13T23:52:36 :( 2010-10-13T23:52:38 hehe 2010-10-13T23:52:41 lost 8 games in a row 2010-10-13T23:52:42 lol 2010-10-13T23:52:57 bye bye elo points ;) 2010-10-13T23:52:57 make it self-modifying 2010-10-13T23:53:05 doh lol 2010-10-13T23:53:07 i was playing with the distance heuristic 2010-10-13T23:53:16 to make it not go for far off planets 2010-10-13T23:53:19 and had a syntax error in there 2010-10-13T23:53:21 :-\ 2010-10-13T23:53:34 i lost only like 6 elo points though 2010-10-13T23:53:35 so thats good 2010-10-13T23:54:07 bhasker: what language are you using? 2010-10-13T23:54:12 python 2010-10-13T23:54:27 so no compilation 2010-10-13T23:54:30 yeah 2010-10-13T23:55:22 python rocks.. and is slow as hell ;) 2010-10-13T23:55:24 *** Mathnerd314 has quit IRC (Ping timeout: 240 seconds) 2010-10-13T23:55:53 rebelxt_: slow for the computer, fast for us 2010-10-13T23:57:29 deepblue! finally beat you in a game hehe 2010-10-13T23:58:02 amstan: oh yeah! almost done re-writing my bot back to c++, and it's much longer (LOC)