2010-10-22T00:01:55 *** narnach has joined #aichallenge 2010-10-22T00:03:10 *** perestrelka has quit IRC (Ping timeout: 276 seconds) 2010-10-22T00:04:23 *** davidjliu has quit IRC (Quit: Page closed) 2010-10-22T00:06:00 Freaking hell 2010-10-22T00:06:01 http://72.44.46.68/canvas?game_id=620682 2010-10-22T00:06:07 i was winning and hazard wiped me out 2010-10-22T00:06:17 he hopped through every planet of mine 2010-10-22T00:06:19 !@#!@# 2010-10-22T00:06:20 bhasker: User error -- Replace user. 2010-10-22T00:07:00 *** Titankiller has joined #aichallenge 2010-10-22T00:07:41 *** hornairs has joined #aichallenge 2010-10-22T00:08:21 *** dorpairs has joined #aichallenge 2010-10-22T00:08:40 bhasker: yep good game though, you had a chance. my bot also likes to give up a better planet cluster sometimes instead of defending ;-) 2010-10-22T00:08:55 yea i lost it because my bot tried to defend 2010-10-22T00:08:59 after it got his base planet 2010-10-22T00:09:01 i sent ships back 2010-10-22T00:09:06 *** airhorns has quit IRC (Ping timeout: 240 seconds) 2010-10-22T00:09:08 thats because i give my planets a higher priority 2010-10-22T00:09:11 than enemy or neutrals 2010-10-22T00:09:15 i guess thats a bad idea :-\ 2010-10-22T00:10:13 i give my/enemy planets the same priority higher than neutrals though (2x) 2010-10-22T00:10:30 *** dorpairs has quit IRC (Remote host closed the connection) 2010-10-22T00:11:06 Top 10 players: Accoun(3953), rsergio(3938), savchenko(3935), sin_sun(3932), cfaftw(3910), deepblue(3908), dmj111(3890), Hazard(3849), chotting(3842), felixcoto(3834) 2010-10-22T00:12:25 *** hornairs has quit IRC (Ping timeout: 276 seconds) 2010-10-22T00:13:12 *** krokkrok has joined #aichallenge 2010-10-22T00:16:16 yea i have been playing around with my value function 2010-10-22T00:18:50 dudes.. he's first 2010-10-22T00:18:53 *** foucist has joined #aichallenge 2010-10-22T00:20:34 *** amstan_ is now known as amstan 2010-10-22T00:20:49 who is first? 2010-10-22T00:20:56 accoun 2010-10-22T00:20:58 antimatroid http://72.44.46.68/canvas?game_id=620869 :-\ 2010-10-22T00:20:59 ah 2010-10-22T00:21:13 i don't feel so bad about my bot getting stomped by his now 2010-10-22T00:21:21 infact i have like 40-50% success rate against his 2010-10-22T00:31:03 @ranking 2010-10-22T00:31:06 pgpaskar_: Top 10 players: Accoun(3943), rsergio(3942), savchenko(3926), sin_sun(3923), cfaftw(3904), deepblue(3901), dmj111(3884), Hazard(3841), chotting(3837), felixcoto(3820) 2010-10-22T00:35:38 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2010-10-22T00:44:53 bhasker :P 2010-10-22T00:45:37 malazan 2010-10-22T00:45:41 I think the problem was 2010-10-22T00:45:56 Your attacking force went defensive instead of steam rolling his original position 2010-10-22T00:46:08 Ie ideally you would have both swapped sides, wiping out and taking each other 2010-10-22T00:46:25 my current bot wont even run on the main site though 2010-10-22T00:46:27 Instead, it reverted back and allowed him to continue building up ships on the bottom, and newly on the top 2010-10-22T00:46:33 going to need to do an extensive clean up at some point 2010-10-22T00:48:06 nvm just read up :P 2010-10-22T00:51:45 *** Kineman has joined #aichallenge 2010-10-22T00:51:53 *** Appleman1234 has joined #aichallenge 2010-10-22T00:54:04 *** krokkrok has quit IRC (Quit: Page closed) 2010-10-22T00:56:28 *** Kineman has left #aichallenge 2010-10-22T01:02:43 *** seisatsu has quit IRC (Quit: Leaving) 2010-10-22T01:08:34 *** Titankiller has quit IRC (Read error: Connection reset by peer) 2010-10-22T01:09:13 *** rebelxt_ has quit IRC (Quit: Page closed) 2010-10-22T01:09:26 *** delt0r___ has joined #aichallenge 2010-10-22T01:09:56 *** delt0r has quit IRC (Read error: Operation timed out) 2010-10-22T01:11:06 Top 10 players: Accoun(3967), rsergio(3964), savchenko(3947), sin_sun(3938), dmj111(3923), deepblue(3921), cfaftw(3921), Hazard(3864), chotting(3842), felixcoto(3839) 2010-10-22T01:17:42 sorry was afk 2010-10-22T01:28:17 *** narnach has quit IRC (Quit: Leaving.) 2010-10-22T01:31:50 hey 2010-10-22T01:32:17 anyone reckon the sum of all the ships (ie player score) would be a good fitness function for training a neural net? 2010-10-22T01:42:15 *** seisatsu has joined #aichallenge 2010-10-22T01:44:59 *** hellman has quit IRC (Remote host closed the connection) 2010-10-22T01:48:52 *** perestrelka has joined #aichallenge 2010-10-22T01:59:24 *** McLeopold has joined #aichallenge 2010-10-22T02:00:43 amstan: I've been coding... 2010-10-22T02:00:58 McLeopold: sorry about that, i hope i didn't offend you 2010-10-22T02:01:03 McLeopold: it was a joke 2010-10-22T02:01:16 I know, so was my quiting. 2010-10-22T02:01:26 *** Palmik has joined #aichallenge 2010-10-22T02:01:31 But I did need to get off to concentrate. 2010-10-22T02:02:08 @rankings 2010-10-22T02:02:10 McLeopold: Top 10 players: Accoun(3976), rsergio(3967), savchenko(3957), sin_sun(3947), dmj111(3931), deepblue(3930), cfaftw(3928), Hazard(3873), felixcoto(3852), chotting(3846) 2010-10-22T02:03:09 Accoun builds fast and secure. 2010-10-22T02:03:29 Sometimes the maps just don't allow for any other strategy. 2010-10-22T02:03:43 He loses when the good stuff is in the middle. 2010-10-22T02:04:45 So the fact that he is #1 means the maps are biased toward fast growth. 2010-10-22T02:08:42 McLeopold: cool 2010-10-22T02:11:07 Top 10 players: Accoun(3976), rsergio(3967), savchenko(3957), sin_sun(3948), dmj111(3931), deepblue(3930), cfaftw(3928), Hazard(3874), felixcoto(3853), chotting(3846) 2010-10-22T02:12:42 *** mega1 has joined #aichallenge 2010-10-22T02:20:58 Anyone got a C++ bot that I can train against? 2010-10-22T02:21:49 I'd need the source ;P 2010-10-22T02:22:14 Might just use the dual bot... but seems like that might be a bit too easy 2010-10-22T02:26:47 rommel: by "train against" do you mean you have a bot that learns? 2010-10-22T02:32:18 its in pieces a little bit at the moment, but yes 2010-10-22T02:32:52 atm I'm training it agains't myself using JBotManager 2010-10-22T02:36:10 i'm quite happy with my macro method 2010-10-22T02:36:21 i now need to refine the micro side of everything 2010-10-22T02:36:34 currently that is hopeless now that i've thought about the problem more 2010-10-22T02:42:56 *** Cyndre has quit IRC (Quit: Leaving) 2010-10-22T03:01:46 rommel: did you look at the tcp server? 2010-10-22T03:03:46 yeah, was using it quite a bit, but mainly for backend testing, haven't gone near it with my bot yet 2010-10-22T03:04:17 rommel: hmm 2010-10-22T03:04:37 rommel: you can look on the forums, there's some people which released code/pseudocode 2010-10-22T03:04:45 maybe reimplement those 2010-10-22T03:06:00 rommel: this comes to mind: http://ai-contest.com/forum/viewtopic.php?f=17&t=886&hilit=top+50+open+source 2010-10-22T03:06:11 *** Titankiller has joined #aichallenge 2010-10-22T03:06:35 yeah was looking at that 2010-10-22T03:07:05 I need to get the training algorithm working 100% first... still a few kinks in my implemented system :) 2010-10-22T03:07:34 rommel: doing neural nets? 2010-10-22T03:08:23 *** vizier has joined #aichallenge 2010-10-22T03:11:08 Top 10 players: Accoun(4020), rsergio(3997), sin_sun(3990), cfaftw(3952), dmj111(3952), savchenko(3950), deepblue(3928), Hazard(3914), felixcoto(3876), chotting(3861) 2010-10-22T03:16:56 *** Titankiller has quit IRC (Read error: Connection reset by peer) 2010-10-22T03:17:52 lol @ Accoun 2010-10-22T03:28:54 ^^ 2010-10-22T03:29:58 oh no 3 russians on the top :P, first chess, now this! 2010-10-22T03:31:23 *** perror has joined #aichallenge 2010-10-22T03:31:34 *** websmith has quit IRC (Quit: Page closed) 2010-10-22T03:45:22 *** Titankiller has joined #aichallenge 2010-10-22T03:58:18 *** McLeopold has quit IRC (Quit: Leaving) 2010-10-22T03:59:07 hahaha 2010-10-22T04:00:03 gah, this problem is very difficult 2010-10-22T04:00:12 i'm back to rethinking the basics, and it gets so damn messy 2010-10-22T04:05:33 *** Itkovian has quit IRC (Quit: Itkovian) 2010-10-22T04:11:08 Top 10 players: Accoun(4046), sin_sun(4021), rsergio(3991), dmj111(3985), savchenko(3963), cfaftw(3963), deepblue(3946), Hazard(3929), felixcoto(3895), Manwe56(3845) 2010-10-22T04:16:47 *** mega1 has quit IRC (Ping timeout: 240 seconds) 2010-10-22T04:17:58 *** Accoun has quit IRC (Ping timeout: 276 seconds) 2010-10-22T04:19:34 *** mega1 has joined #aichallenge 2010-10-22T04:24:48 *** Bobng has joined #aichallenge 2010-10-22T04:27:04 haha raedwulf 2010-10-22T04:27:21 @amstan, yeah neural nets 2010-10-22T04:27:22 rommel: No! 2010-10-22T04:27:32 :O 2010-10-22T04:27:37 What'd I do? haha 2010-10-22T04:27:51 *** wvdschel has joined #aichallenge 2010-10-22T04:27:52 *** wvdschel has joined #aichallenge 2010-10-22T04:27:59 rommel: don't start your messages with @ 2010-10-22T04:28:17 how do you do the text directed at specific people? 2010-10-22T04:28:43 amstan: yes I am doing a neural net (that how?) 2010-10-22T04:29:38 *** noid05826 has joined #aichallenge 2010-10-22T04:29:57 *** noid05826 is now known as Accoun 2010-10-22T04:40:39 *** vizier has quit IRC (Quit: Page closed) 2010-10-22T04:44:59 *** Naktibalda has joined #aichallenge 2010-10-22T04:55:58 *** wvdschel has quit IRC (Quit: Leaving) 2010-10-22T05:00:31 *** Bobng has quit IRC (Read error: Connection reset by peer) 2010-10-22T05:01:03 *** Bobng has joined #aichallenge 2010-10-22T05:05:42 *** Bobng has quit IRC (Read error: Connection reset by peer) 2010-10-22T05:06:13 *** Bobng has joined #aichallenge 2010-10-22T05:09:13 *** wvdschel has joined #aichallenge 2010-10-22T05:11:09 Top 10 players: Accoun(4086), sin_sun(4010), rsergio(4006), cfaftw(4005), Hazard(3996), savchenko(3989), dmj111(3956), Manwe56(3938), felixcoto(3913), deepblue(3905) 2010-10-22T05:12:06 shit, i think i have had a bug in generate future states all this time 2010-10-22T05:12:30 a planet only receives growth on a turn if the owner isn't neutral and it's the same as last turn yeah? 2010-10-22T05:12:35 i missed that last part in my code 2010-10-22T05:13:39 wait no, you get the growth before fighting, nevermind 2010-10-22T05:14:14 isn't growth done before sending ships 2010-10-22T05:14:27 so could you send, for example, 105 ships on the first turn 2010-10-22T05:14:36 given that you start with 100 ships and have 5 growth rate 2010-10-22T05:17:09 you can't send that extra 5 2010-10-22T05:17:23 you send fleets, update growth, then take incoming fleets 2010-10-22T05:18:05 future states gets really messy once you try to consider future moves 2010-10-22T05:25:40 *** Meatkat has joined #aichallenge 2010-10-22T05:25:53 !cpu 2010-10-22T05:25:53 Meatkat: I have taken 241.03 seconds of user time and 8.12 seconds of system time, for a total of 249.15 seconds of CPU time. My children have taken 0.00 seconds of user time and 0.00 seconds of system time for a total of 0.00 seconds of CPU time. I'm taking up 16024 kB of memory. 2010-10-22T05:28:43 *** Tobu has quit IRC (Quit: Leaving.) 2010-10-22T05:30:43 antimatroid: ah ok 2010-10-22T05:30:51 thanks antimatroid 2010-10-22T05:41:41 *** Bobng has quit IRC (Read error: Connection reset by peer) 2010-10-22T05:41:44 I'm testing on the original 100 maps in the starter pack and notice that losses and victories are not evenly distributed. 2010-10-22T05:41:52 :) 2010-10-22T05:41:56 Am I seeing patterns where there is none? 2010-10-22T05:42:20 *** Bobng has joined #aichallenge 2010-10-22T05:42:22 Or the maps are not really random[ly ordered]. 2010-10-22T05:44:33 even distribution isn't random 2010-10-22T05:45:42 I know. 2010-10-22T05:45:54 random is hard to define 2010-10-22T05:45:55 *** Tobu has joined #aichallenge 2010-10-22T05:47:02 *** nigg_ has joined #aichallenge 2010-10-22T05:47:03 *** Bobng has quit IRC (Read error: Connection reset by peer) 2010-10-22T05:50:06 *** Tobu has quit IRC (Ping timeout: 240 seconds) 2010-10-22T05:52:38 *** Tobu has joined #aichallenge 2010-10-22T05:56:24 *** HakanD has joined #aichallenge 2010-10-22T06:06:57 *** Titankiller has quit IRC (Read error: Connection reset by peer) 2010-10-22T06:11:10 Top 10 players: Accoun(4069), sin_sun(4040), Hazard(4040), cfaftw(4033), rsergio(3998), savchenko(3992), felixcoto(3938), dmj111(3927), deepblue(3913), Manwe56(3910) 2010-10-22T06:15:50 *** seisatsu has quit IRC (Quit: Leaving) 2010-10-22T06:20:57 *** krokkrok has joined #aichallenge 2010-10-22T06:25:32 *** Florian95 has joined #aichallenge 2010-10-22T06:35:00 *** tapwater has quit IRC (Quit: tapwater) 2010-10-22T06:35:57 *** Meatkat has quit IRC (Ping timeout: 245 seconds) 2010-10-22T06:40:32 *** Utkarsh has joined #aichallenge 2010-10-22T06:47:27 *** foucist has quit IRC (Ping timeout: 240 seconds) 2010-10-22T06:48:28 *** foucist has joined #aichallenge 2010-10-22T07:04:10 *** BtbN_ has joined #aichallenge 2010-10-22T07:11:09 Top 10 players: cfaftw(4066), Accoun(4054), Hazard(4054), rsergio(4023), sin_sun(4022), savchenko(3989), Manwe56(3963), dmj111(3949), felixcoto(3949), deepblue(3930) 2010-10-22T07:13:23 *** aerique has joined #aichallenge 2010-10-22T07:17:38 are the stats cumulative? Ie someone starting out will never catch the front ?\ 2010-10-22T07:20:17 the stats are not cumulative. In fact, as it works now, a new bot tends to go higher than its actual rating at the start, so coming late is not a handicap 2010-10-22T07:25:02 *** BtbN_ has quit IRC (Quit: Verlassend) 2010-10-22T07:27:03 *** krokkrok has quit IRC (Quit: Page closed) 2010-10-22T07:28:25 *** Titankiller has joined #aichallenge 2010-10-22T07:31:14 *** Accoun has quit IRC (Ping timeout: 265 seconds) 2010-10-22T07:34:04 they probably will reset it all for the final ranking anyway 2010-10-22T07:36:23 *** yasith has quit IRC (Ping timeout: 245 seconds) 2010-10-22T07:36:38 bots are just overshooting with the way it picks out your opponents, that's alright, elos are to be reset when the final contest runs 2010-10-22T07:38:47 *** HakanD has quit IRC (Ping timeout: 240 seconds) 2010-10-22T07:49:44 *** yasith has joined #aichallenge 2010-10-22T07:51:56 *** JensTi has joined #aichallenge 2010-10-22T07:53:54 *** noid05826 has joined #aichallenge 2010-10-22T07:54:03 *** noid05826 is now known as Accoun 2010-10-22T07:54:40 *** gerald5122k has joined #aichallenge 2010-10-22T07:54:41 hi 2010-10-22T07:55:48 woah my bot is 15th i am so happy :) but not very representative he should be 50 ~~ 60 i think 2010-10-22T08:03:40 *** Utkarsh has quit IRC (Ping timeout: 276 seconds) 2010-10-22T08:05:25 *** gerald5122k has quit IRC () 2010-10-22T08:07:54 *** hellman has joined #aichallenge 2010-10-22T08:08:59 *** Utkarsh has joined #aichallenge 2010-10-22T08:11:10 Top 10 players: cfaftw(4058), Accoun(4046), sin_sun(4027), Hazard(4024), rsergio(4017), savchenko(3983), Manwe56(3967), felixcoto(3946), dmj111(3944), deepblue(3926) 2010-10-22T08:12:26 *** HakanD has joined #aichallenge 2010-10-22T08:14:20 *** Tobu has quit IRC (Quit: Leaving.) 2010-10-22T08:15:25 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-10-22T08:24:25 *** gerald512kk has joined #aichallenge 2010-10-22T08:24:44 *** boegel has joined #aichallenge 2010-10-22T08:30:36 *** Olathe has joined #aichallenge 2010-10-22T08:36:06 *** Palmik has quit IRC (Ping timeout: 252 seconds) 2010-10-22T08:40:16 *** rommel has quit IRC (Quit: Page closed) 2010-10-22T08:43:44 *** n9thbit has joined #aichallenge 2010-10-22T08:47:18 *** Titankiller2 has joined #aichallenge 2010-10-22T08:49:46 *** Titankiller has quit IRC (Ping timeout: 252 seconds) 2010-10-22T08:51:13 *** Titankiller2 has quit IRC (Read error: Connection reset by peer) 2010-10-22T08:58:05 *** aerique has quit IRC (Quit: ..) 2010-10-22T08:59:02 *** p4p4 has joined #aichallenge 2010-10-22T09:05:08 *** HakanD has quit IRC (Ping timeout: 245 seconds) 2010-10-22T09:06:37 is the short lifespan of the extra servers a stability issue or sth? 2010-10-22T09:10:51 *** delt0r has joined #aichallenge 2010-10-22T09:11:06 *** delt0r___ has quit IRC (Ping timeout: 240 seconds) 2010-10-22T09:11:10 Top 10 players: cfaftw(4055), Accoun(4043), sin_sun(4025), Hazard(4022), rsergio(4016), savchenko(3980), Manwe56(3961), felixcoto(3944), dmj111(3940), deepblue(3921) 2010-10-22T09:16:04 *** Utkarsh has joined #aichallenge 2010-10-22T09:22:17 *** n9thbit has quit IRC (Ping timeout: 265 seconds) 2010-10-22T09:23:18 *** cfaftw has joined #aichallenge 2010-10-22T09:24:22 dang my bot is doing surprisingly well on the official server right now 2010-10-22T09:24:35 it's had a lucky series of games 2010-10-22T09:27:26 * Naktibalda have noticed winning and lossing streaks 2010-10-22T09:29:56 last week it went down to 20-something 2010-10-22T09:29:59 now it's at #1 2010-10-22T09:30:06 incredible 2010-10-22T09:31:06 clearly this means i should just leave that bot as my official submission, and just hope that in the final showdown my bot gets all favorable map-bot matchups 2010-10-22T09:31:22 8) 2010-10-22T09:41:05 cfaftw> hi 2010-10-22T09:41:19 cfaftw> CCCF-ex12 is u ? 2010-10-22T09:42:31 Accoun, no... my bot isn't playing on the TCP 2010-10-22T09:42:56 Accoun, CCCF is felixcoto 2010-10-22T09:43:27 cfaftw> 100% information? 2010-10-22T09:43:53 what information? 2010-10-22T09:44:43 *** JensTi has quit IRC (Quit: Leaving.) 2010-10-22T09:45:28 cfaftw> about CCCF is felixcotos bot 2010-10-22T09:45:46 Accoun, oh, yeah i'm positive 2010-10-22T09:46:34 positive sounds more like 51% than 100% :) 2010-10-22T09:46:34 positiv nrgativ... 2010-10-22T09:46:59 use please more sample english 2010-10-22T09:48:06 and now i very need 4 cores processor 2010-10-22T09:49:14 *** boegel has quit IRC (Read error: Operation timed out) 2010-10-22T09:55:32 Accoun: you have a more manly beard than me 2010-10-22T09:55:35 i bow before thee 2010-10-22T09:59:26 beards are for kids... it's all about having a lucious Burt-Reynolds-esque mustache 2010-10-22T10:00:30 ... which i don't have :'( 2010-10-22T10:00:38 luscious* 2010-10-22T10:00:42 lol 2010-10-22T10:00:47 i have full caveman style atm 2010-10-22T10:00:51 it's been a very busy semester 2010-10-22T10:00:51 *** jeff_ has joined #aichallenge 2010-10-22T10:01:14 i think im going to write an ai program to write my galcon ai program 2010-10-22T10:01:18 *** jeff_ is now known as Guest63276 2010-10-22T10:01:25 yo dog 2010-10-22T10:01:27 >.< 2010-10-22T10:01:28 :o why didn't i think of that?! 2010-10-22T10:01:37 why didn't my BOT think of that ?! 2010-10-22T10:01:40 :D 2010-10-22T10:02:13 antimatroid, can you explain to me the "yo dawg" meme? 2010-10-22T10:02:41 mr zibit likes to put everything inside something 2010-10-22T10:02:43 Hey, I'm seeing lots of fail-on-first turn with my robot on the ai-contest server. see http://ai-contest.com/forum/viewtopic.php?f=18&t=911 . I did extensive debugging, on many platform, valrgind'ed my program, there seems to be no way it would fail repeatedly like that, always on the first move. THis happens on the server for many games in a row, then my program can win like 20-30 games in a row, and repeats 2010-10-22T10:02:55 Is there an issue with the server failing to load some programs ? 2010-10-22T10:03:03 ohhhhhhhh 2010-10-22T10:03:13 ok 2010-10-22T10:03:20 (yeah, I know it looks like a program crash... but I really was thorough in investigating it) 2010-10-22T10:03:53 Guest63276: maybe some particular configuration of map? 2010-10-22T10:03:53 seeing the pictures from doing a "yo dawg" google search made things much more clear 2010-10-22T10:04:08 raed: I played over 10000 games on the TCP server 2010-10-22T10:04:15 never seen anyting like it 2010-10-22T10:04:20 hmmmm 2010-10-22T10:04:39 it comes and go, one day is fine, then the next day only failures.... really feel server side 2010-10-22T10:04:51 another random idea: does it run close to a second running time? 2010-10-22T10:05:01 someone mentioned having it bail out at 0.8s 2010-10-22T10:05:16 It works for me at 0.8s 2010-10-22T10:05:20 nope. 2010-10-22T10:05:22 timed out at 0.95s 2010-10-22T10:05:47 official server works mostly fine for me if I put the time cap at 0.01s/turn. raise it to 0.1s though and it starts failing 2010-10-22T10:05:49 first turn is really quick, about all I do is clear a 600Kb array, send one move, nothing really intesive 2010-10-22T10:05:52 mega1: sorry i meant the program bails out at 0.8s :P 2010-10-22T10:06:02 rather than the engine bailing out the program for you 2010-10-22T10:06:13 fluff: using C++ how do you check the time ? Clock() ? 2010-10-22T10:06:29 guest: clock() and gettimeofday() 2010-10-22T10:07:13 I'm at a loss 2010-10-22T10:08:30 Any chance I'd get the admin to send a core file (god forbid it would be a crash) or explanation where it timed out or did an invalid move (impossible, but yeah, you never know...) 2010-10-22T10:09:13 Other question, is it forbiddn to open a socket from my programm and send debuggin info back ? 2010-10-22T10:09:38 it would suggest its a security issue 2010-10-22T10:09:54 so probably not 2010-10-22T10:10:26 guest, you can't open sockets 2010-10-22T10:10:32 me too. I kinda think it may be related to overburdened servers. if the bot process is put on hold for a second then you're screwed 2010-10-22T10:10:56 well just see - the main serve rdoes 4 games a minute and then you get the other second server back up 2010-10-22T10:11:00 and it does 40 games a minute 2010-10-22T10:11:11 Top 10 players: cfaftw(4046), Accoun(4035), sin_sun(4019), Hazard(4013), rsergio(4007), savchenko(3972), Manwe56(3955), felixcoto(3937), dmj111(3933), deepblue(3912) 2010-10-22T10:11:20 ok. So, I just make the best bot I can using the benzedrine TCP server, and hope that by the time the real contest runs, this doesn't happen ? 2010-10-22T10:11:37 guest: sounds like a good enough plan for now 2010-10-22T10:11:38 no don't do that 2010-10-22T10:11:51 i mean, don't depend on it magically working 2010-10-22T10:13:18 *** fluffluff1 has joined #aichallenge 2010-10-22T10:13:39 i can't believe my bot even works after going through it today 2010-10-22T10:13:42 cfaftw: I try ubuntu+gcc, win+VS, mac OS X, ran valgrind, ran VS debug checks, ran 10000 games on the TCP server, am checking clock()... with server-side info on *why* it fails, that seems impossible. I realise the admin can't really debug everything for contestant, but the profile of how this happens really seems to imply a server issue 2010-10-22T10:14:09 i can't believe how messy things get once you try to make future moves and account for information 2010-10-22T10:14:26 >> i mean, don't depend on it magically workin. Any suggestion, from here ? 2010-10-22T10:14:39 somehow i'm throwing all of that (i think somewhat buggily) into a tree and it doens't always lose 2010-10-22T10:15:05 guest63276, i'm just saying... i wouldn't want to take that risk personally. i would keep digging. maybe you're using some non-standard library? 2010-10-22T10:15:42 oh wait, that would just fail to compile 2010-10-22T10:15:43 guest63276: your bot running on tcp is no different to you running a game locally, i have had similar problems with it not running on the official server 2010-10-22T10:15:44 nvm 2010-10-22T10:15:48 it wont run on there atm in fact 2010-10-22T10:16:16 *** fluffluff has quit IRC (Ping timeout: 276 seconds) 2010-10-22T10:16:50 craft: I went so far as removing my use of memset and went with a loop, for this reason. Any thing suspiscious with: 2010-10-22T10:17:34 cfaft 2010-10-22T10:17:45 i don't know C++, but i think those are all standard, so no 2010-10-22T10:17:55 enough people seem to be having problems with it that a serverside problem is not all that unlikely imho 2010-10-22T10:18:24 fluffluff1, didn't you just quit? lol 2010-10-22T10:18:40 oh wait, that was vanilla fluffluff 2010-10-22T10:18:42 cfa: just timed out and reconnected with another name 2010-10-22T10:19:08 *** fluffluff1 is now known as fluffluff 2010-10-22T10:19:40 i guess a server-side issue isn't out of the question... i just think it's weird that python and (as far as i know) java users haven't had problems with this 2010-10-22T10:19:52 my bot is in python and has never had mysterious problems 2010-10-22T10:19:53 *** Azrathud has joined #aichallenge 2010-10-22T10:20:07 *** JamesMG has joined #aichallenge 2010-10-22T10:20:58 *** nigg_ has quit IRC (Remote host closed the connection) 2010-10-22T10:21:14 ok. well, got to get to work. thanks for the suggestion 2010-10-22T10:21:50 oh Guest63276 2010-10-22T10:21:58 there maybe one more problem 2010-10-22T10:21:59 yeah? 2010-10-22T10:22:05 yeah, I know nothing about how they're handling the time issues on the server. maybe there's differences between the languages 2010-10-22T10:22:23 i had some initial issues with my reader (but with the java testing tool, not the real backend) 2010-10-22T10:22:44 if for some reason your backend continuously read stdin... somehow falls into a loop, it might lock stdin 2010-10-22T10:22:46 and timeout 2010-10-22T10:22:55 thats just what happened with my initial backend code 2010-10-22T10:23:12 well, that would not make my submission fail, though ? 2010-10-22T10:23:23 *** Bobng has joined #aichallenge 2010-10-22T10:23:43 maybe its not normally the case, i did also notice that the C++ starter package code doesn't terminate on EOF on stdin 2010-10-22T10:23:51 whereas the java code does 2010-10-22T10:24:27 so that is the reason for application crash at end of the game? 2010-10-22T10:24:28 but once it gets EOF, game is over so the program can crash all it wants to 2010-10-22T10:24:32 do you have your own C++ code for reading stdin, or just the starter package 2010-10-22T10:24:44 *** Frontier has joined #aichallenge 2010-10-22T10:25:24 the python starter package lacks the try...except around raw_input call 2010-10-22T10:25:36 just the starter package 2010-10-22T10:25:42 rabidus_: the starter program hung for me when running through the java tool 2010-10-22T10:25:49 there's a thought... I'm using starter package input handling too 2010-10-22T10:26:09 *** Tobu has joined #aichallenge 2010-10-22T10:26:15 i rewrote mine in C 2010-10-22T10:26:21 well my whole bot is in C 2010-10-22T10:26:23 :P 2010-10-22T10:26:25 :o I just realized... BlizzCon is this weekend!!! 2010-10-22T10:26:37 * cfaftw holds breath in anticipation of Diablo 3 release date announcement 2010-10-22T10:26:57 raed: ok. thanks, I'll go work now, but I'll try that, rewriting the input handler 2010-10-22T10:27:30 *** Guest63276 has quit IRC (Quit: Page closed) 2010-10-22T10:31:27 cfaftw: raw_input doesn't raise any exceptions 2010-10-22T10:31:59 EOFError on EOF 2010-10-22T10:37:40 *** Titankiller has joined #aichallenge 2010-10-22T10:39:21 decided to make an ANN bot yet cfaftw? 2010-10-22T10:39:54 JamesMG, haha... well i didn't get a chance to read any ANN stuff last night, so not yet :) 2010-10-22T10:42:57 I thought about it a bit myself after our conversation yesterday 2010-10-22T10:43:53 haven't come up with anything particularly workable yet though 2010-10-22T10:44:52 have you thought about how you'd train it? 2010-10-22T10:47:02 game tape :) 2010-10-22T10:48:02 I don't know if I could have it run the bot proper, but as a filter for good moves it may be useful 2010-10-22T10:48:14 *** bhasker has quit IRC (Quit: bhasker) 2010-10-22T10:48:16 *** Itkovian has joined #aichallenge 2010-10-22T10:48:25 yeah, i was thinking of showing it games existing games, and classify all the moves of the winner as "good moves" and moves of the loser as "bad moves" 2010-10-22T10:49:49 *** rebelxt_ has joined #aichallenge 2010-10-22T10:49:57 you'll have to feed it the game map at the time the move was made as well 2010-10-22T10:50:02 alongside each move 2010-10-22T10:50:13 yeah 2010-10-22T10:50:22 try to teach it to be situational 2010-10-22T10:51:27 but again... i know virtually nothing about ANNs, so i'm thinking very abstractly 2010-10-22T10:56:54 *** mceier has joined #aichallenge 2010-10-22T11:00:41 *** Appleman1234 has quit IRC (Quit: Leaving) 2010-10-22T11:00:59 not Appleman! noooooooooo 2010-10-22T11:04:11 *** wvdschel has quit IRC (Remote host closed the connection) 2010-10-22T11:09:07 *** rebelxt_ has quit IRC (Ping timeout: 265 seconds) 2010-10-22T11:11:12 Top 10 players: cfaftw(4057), Accoun(4042), sin_sun(4027), rsergio(4023), Hazard(4022), savchenko(4000), Manwe56(3963), felixcoto(3944), dmj111(3942), deepblue(3905) 2010-10-22T11:12:00 hey, you're in first again 2010-10-22T11:13:45 yeah, it's crazy 2010-10-22T11:14:18 it had a great streak of matches 2010-10-22T11:15:09 that bot's success gives me high hopes for the one i'm writing now 2010-10-22T11:16:25 that bot that is #1 has bugs... one particularly glaring one is that if any number of enemy planets are in line with eachother but perpendicular to my planet, my planet won't attack any of those planets no matter how cheap they are 2010-10-22T11:16:47 *** Azrathud has quit IRC (Ping timeout: 240 seconds) 2010-10-22T11:17:15 I'm not even sure how you accomplished that bug 2010-10-22T11:17:45 lol 2010-10-22T11:17:52 it's best not to think about it 2010-10-22T11:18:22 buridan's donkey's problem :) 2010-10-22T11:19:50 haha, buridan's ass... that's my new favorite paradox 2010-10-22T11:20:09 that's pretty excellent, and applicable to this contest 2010-10-22T11:20:26 and it definitely explains what's going on with my submission bot 2010-10-22T11:20:45 mostly 2010-10-22T11:21:37 OMG, ass is a synonim to donkey 2010-10-22T11:21:57 haha, yes 2010-10-22T11:22:14 "ass" as a word originated to a type of animal 2010-10-22T11:22:21 not sure how it came to mean "butt" 2010-10-22T11:22:24 *** n9thbit has joined #aichallenge 2010-10-22T11:22:26 :) 2010-10-22T11:22:52 cfaftw: probably because when you see someone riding a donkey it looks like they have a big ass from behind? ;) 2010-10-22T11:23:10 * Naktibalda have never seen someone riding donkey 2010-10-22T11:23:11 haha, you know what? that actually makes sense. 2010-10-22T11:23:26 *** Utkarsh has quit IRC (Ping timeout: 264 seconds) 2010-10-22T11:23:28 Naktibalda: national geographic.. broaden your world ;) 2010-10-22T11:24:01 foucist, i was thinking City Slickers, but that works too 2010-10-22T11:24:52 yeah plenty of donkey riding in movies 2010-10-22T11:25:38 I bet at least half of Clint Eastwood movies have donkey riding 2010-10-22T11:26:15 hell, i won't even watch a movie if it doesn't have at least one donkey riding scene. 2010-10-22T11:26:38 that's called porno, right? 2010-10-22T11:26:39 ... i take that back 2010-10-22T11:27:01 now this conversation is going places 2010-10-22T11:27:23 :E 2010-10-22T11:27:47 *** Itkovian has quit IRC (Quit: Itkovian) 2010-10-22T11:27:49 *** Utkarsh has joined #aichallenge 2010-10-22T11:33:12 *** wh1teside has quit IRC (Ping timeout: 252 seconds) 2010-10-22T11:38:31 *** wvdschel has joined #aichallenge 2010-10-22T11:50:28 *** lavalamp_ has joined #aichallenge 2010-10-22T11:50:59 a game that made me lol: http://72.44.46.68/canvas?game_id=627778 2010-10-22T11:53:18 LOL 2010-10-22T11:53:21 that is amazing 2010-10-22T11:53:34 isn't it awesome? >.< 2010-10-22T11:53:46 infinite loop bot 2010-10-22T11:54:11 thanks for sharing that one 2010-10-22T11:54:15 :) 2010-10-22T11:54:16 haha 2010-10-22T11:54:41 ship routing code gone bad... 2010-10-22T11:55:04 haha, wow 2010-10-22T11:55:18 is that your minimax bot lavalamp_? 2010-10-22T11:56:56 it's fun watching the fleet sizes getting larger as the loop goes around the second time 2010-10-22T11:57:30 yeah that's my bot in a woefully incomplete state 2010-10-22T11:57:58 *** Titankiller has quit IRC (Ping timeout: 252 seconds) 2010-10-22T11:58:11 I got it to stop crashing and everything sorta working so I had it play a few games 2010-10-22T11:58:41 I need better move generation, right now it can't attack and defend at the same time, hehe 2010-10-22T11:59:57 yeah that's a problem 2010-10-22T12:00:26 *** Sarin has joined #aichallenge 2010-10-22T12:02:38 nice, I like the ship explosion attacks it does 2010-10-22T12:03:54 thanks ^^ 2010-10-22T12:04:12 cfaftw: are you using game trees structures aswell? 2010-10-22T12:04:49 Error323, no 2010-10-22T12:06:10 i might have to abandon this 2010-10-22T12:06:16 *** fawek has joined #aichallenge 2010-10-22T12:06:26 jmcarthur, abandon your bot or the contest? 2010-10-22T12:06:31 i need more time to work on things that are actually likely to make money 2010-10-22T12:06:35 the contest 2010-10-22T12:06:42 :'( say it ain't so 2010-10-22T12:07:01 i haven't touched my bot in over week anyway, i think 2010-10-22T12:07:04 are you in school? 2010-10-22T12:07:07 me neither :( 2010-10-22T12:07:24 nah, and i'm even unemployed, which is all the more reason that i need to focus on making money 2010-10-22T12:07:47 i had a little web service that i've been working on and this bot is cutting into it too much 2010-10-22T12:07:50 *** p4p4 has quit IRC (Read error: Connection timed out) 2010-10-22T12:08:10 i'm out of school but employed... i would agree that you should focus on projects that have a higher likelihood of making money ;) 2010-10-22T12:09:18 if i wasn't doing this bot, i'd be working on project euler problems 2010-10-22T12:09:22 i hate pulling out like this. i spent so much time on it already... 2010-10-22T12:09:44 jmcarthur: bummer but yeah money is good... 2010-10-22T12:09:44 oh well. gotta do what i gotta do 2010-10-22T12:09:51 *** p4p4 has joined #aichallenge 2010-10-22T12:10:11 wait a second... this is just a trick 2010-10-22T12:10:38 he wants us to THINK he's gone... then on the last day of submissions, BAM, he sweeps the contest with a surprise bot 2010-10-22T12:11:03 dang my plan failed 2010-10-22T12:11:12 Top 10 players: cfaftw(4044), Accoun(4031), sin_sun(4013), rsergio(4010), Hazard(4010), savchenko(3987), Manwe56(3951), felixcoto(3933), dmj111(3930), deepblue(3890) 2010-10-22T12:11:20 Accoun is up there? 2010-10-22T12:11:31 yep 2010-10-22T12:11:34 huh 2010-10-22T12:12:02 he was pretty high on the tcp server the other day 2010-10-22T12:12:14 i wish he was easier to talk to. he could have the potential to have interesting discussions with about strategy... 2010-10-22T12:15:24 interesting discussions about êàíàäñêèõ ïèåäèêîâ 2010-10-22T12:15:44 lol 2010-10-22T12:16:02 LOL 2010-10-22T12:18:44 *** Titankiller has joined #aichallenge 2010-10-22T12:19:45 juh? 2010-10-22T12:23:16 meh, thats not utf8 is it 2010-10-22T12:23:29 *** Utkarsh has quit IRC (Ping timeout: 272 seconds) 2010-10-22T12:23:45 * raedwulf gets back to his bot 2010-10-22T12:23:56 *** wvdschel has quit IRC (Quit: Leaving) 2010-10-22T12:24:00 yeah... it's bot workin-on time! 2010-10-22T12:24:16 (also, lunch time) 2010-10-22T12:26:17 * raedwulf has spent too much time on this, i need to do well lol ^^ 2010-10-22T12:26:33 1.3k lines of code so far, if thats any measure 2010-10-22T12:26:44 o.O 2010-10-22T12:27:00 do you have a bot on the TCP or in the official rankings? 2010-10-22T12:27:33 not yet, but its on official rankings just testing my base code 2010-10-22T12:27:36 which appears to work fine 2010-10-22T12:28:07 the bot there is a rage-bot similar (it seems to behave slightly differently to the java version due to the way i reimplemented it) 2010-10-22T12:28:08 the bot i'm working on now is 954 lines 2010-10-22T12:28:10 total 2010-10-22T12:28:23 i have quite a few comment lines tho 2010-10-22T12:28:34 *** fawek has quit IRC (Read error: Connection reset by peer) 2010-10-22T12:28:48 meh im going up in rankings with a crap bot 2010-10-22T12:28:50 http://www.ai-contest.com/profile.php?user_id=11556 2010-10-22T12:29:18 *** Utkarsh has joined #aichallenge 2010-10-22T12:29:19 my current bot on the official server is ~940 lines 2010-10-22T12:30:04 nice 2010-10-22T12:30:07 though we are comparing apples to oranges 2010-10-22T12:30:10 C vs Python 2010-10-22T12:30:15 oh yea mine's also in C, so its a bit more verbose 2010-10-22T12:30:17 yep 2010-10-22T12:30:32 lol, i love your account name 2010-10-22T12:30:34 for example i have a file just dealing with memory management :P 2010-10-22T12:30:41 haha, Amanieu chose it :P 2010-10-22T12:30:59 its a team effort, though so far im the team 2010-10-22T12:31:07 he's a bit busy at uni atm 2010-10-22T12:31:13 but he's going to have free time next week 2010-10-22T12:31:24 partially why i have to comment it 2010-10-22T12:31:45 because im using lots of acronyms to save typing massively long lines 2010-10-22T12:32:23 *** _cfaftw <_cfaftw!~cfaftw@d121070.ncnr.nist.gov> has joined #aichallenge 2010-10-22T12:32:37 <_cfaftw> got disconnected there 2010-10-22T12:33:05 *** bhasker has joined #aichallenge 2010-10-22T12:33:18 oh ok, what was the last line you saw? 2010-10-22T12:33:30 <_cfaftw> i checked the logs, i didn't miss anything 2010-10-22T12:33:34 ok 2010-10-22T12:33:57 <_cfaftw> you two go to school together? 2010-10-22T12:34:26 no, we work on similar game projects, i've known him for quite a while 2010-10-22T12:34:43 *** rebelxt has joined #aichallenge 2010-10-22T12:34:45 <_cfaftw> cool 2010-10-22T12:34:48 also went to visit him in manchester about a week ago 2010-10-22T12:35:09 erm, but more than a week 2010-10-22T12:35:11 hehe 2010-10-22T12:35:20 6 servers up, 90 GPM! did miracle just happen? 2010-10-22T12:35:39 we like each other's code styles as well, and we tend to understand each other's code, so that helps a lot 2010-10-22T12:35:40 <_cfaftw> apparently 2010-10-22T12:35:41 I think janzert happened 2010-10-22T12:36:15 janzert: kudos! 2010-10-22T12:36:22 *** cfaftw has quit IRC (Ping timeout: 255 seconds) 2010-10-22T12:36:25 <_cfaftw> raedwulf, that's cool... this contest would be fun was a team 2010-10-22T12:36:27 *** _cfaftw is now known as cfaftw 2010-10-22T12:36:28 just guessing though 2010-10-22T12:36:49 though im prefer C and he now uses more C++, i started in C++, but i wasn't happy because i felt a bit uncomfortable 2010-10-22T12:36:51 hehe 2010-10-22T12:36:59 *** p4p4 has quit IRC (Read error: Connection timed out) 2010-10-22T12:37:25 i have never touched C++ 2010-10-22T12:37:53 haven't had a reason to, really 2010-10-22T12:38:28 i had too many cyclic references in my classes, for convenience, whereas a simpler design without cycles would have been better 2010-10-22T12:39:12 *** p4p4 has joined #aichallenge 2010-10-22T12:39:15 possibly just adding a few methods to my C structs now would be good enough, but i wouldn't need to 2010-10-22T12:39:16 cfaftw: tree bot is in python, and even if i re-write it in c++ and get 10x more nodes evaluated in 1s, it's not going to help it play any better 2010-10-22T12:39:50 rebelxt, i don't doubt that ;) 2010-10-22T12:40:19 yea, i don't know python :P 2010-10-22T12:40:23 if anyone resubmits can you let me know if you get the compilation success/fail email back 2010-10-22T12:40:32 actually i don't know any on that list except C, C++ and C# 2010-10-22T12:40:32 I believe I have it fixed now 2010-10-22T12:40:49 janzert: how come the servers are so 'unstable' ? 2010-10-22T12:40:58 and i've been working with c# 3.0, so dropping to an old mono version wouldn't cut it for me 2010-10-22T12:41:27 if they had lua, i would have hacked a bot up in lua 2010-10-22T12:41:28 the worker instances? I think it's mostly because of the old version of ubuntu being used 2010-10-22T12:41:36 raedwulf, they do have lua 2010-10-22T12:41:37 well 2010-10-22T12:41:46 they do :o nice 2010-10-22T12:41:50 is it in the backend? 2010-10-22T12:41:51 i know somebody is doing a lua bot, where "mybot" is some C++ bootstrapper 2010-10-22T12:42:15 not sure if there is official server support for it 2010-10-22T12:42:31 yea, i suppose i could embed lua :P 2010-10-22T12:42:41 but the effort of that would defeat the purpose 2010-10-22T12:42:44 now that so many servers are online, ratings should stabilize in a day or two.. hopefully ;) 2010-10-22T12:43:11 contestbot: rankings 2010-10-22T12:43:11 oh, it does have lua in the backend 2010-10-22T12:43:13 interesting 2010-10-22T12:43:13 rebelxt: Top 10 players: Hazard(4051), sin_sun(4031), cfaftw(4031), Accoun(4020), rsergio(4008), Manwe56(3991), felixcoto(3979), savchenko(3978), dmj111(3952), RebelXT(3902) 2010-10-22T12:43:46 cfaftw: saw you in first position again just a while ago 2010-10-22T12:44:05 not gonna rewrite my bot into lua though 2010-10-22T12:44:07 haha 2010-10-22T12:44:44 rebelxt, yeah, i think my bot got a lucky string of matches :) 2010-10-22T12:45:03 rebelxt, but again, that bot is dead to me... i've learned from it. now i'm focusing on my new bot. 2010-10-22T12:45:03 boringbot1, boringbot2, boringbot3 2010-10-22T12:45:11 neat 2010-10-22T12:45:31 cfaftw: mine had a lucky streak too. 2010-10-22T12:46:22 wow, that is quite a winning streak 2010-10-22T12:46:24 oh amazon is offering 750 free hours on EC2 starting November 1 2010-10-22T12:46:31 Oh, nice 2010-10-22T12:46:41 just make every single contestant sign up for that and run their bot from there :D 2010-10-22T12:46:49 http://aws.amazon.com/free/ 2010-10-22T12:47:29 does anyone remember Killer's bot being on top? 2010-10-22T12:47:40 his bot has been quiet for a while 2010-10-22T12:48:12 good idea :), they're only micro instances though 2010-10-22T12:57:20 haha, this bot is surprisingly violent: http://72.44.46.68/canvas?game_id=631400 2010-10-22T12:58:08 all it can do is nothing, or send all of the available ships from one planet to another, each turn 2010-10-22T13:00:07 *** hellman has quit IRC (Remote host closed the connection) 2010-10-22T13:00:24 lol 2010-10-22T13:00:50 hopefully my ai technique would wipe those sorts of bots out 2010-10-22T13:01:03 lol it sounds all fancy "my ai technique" 2010-10-22T13:01:10 my AI kung fu 2010-10-22T13:01:23 my AI CHUCK NORRIS! 2010-10-22T13:01:32 *** p4p4 has quit IRC (Read error: Connection timed out) 2010-10-22T13:01:49 oh man i hope Chuck Norris doesn't enter the contest 2010-10-22T13:01:56 damn should have called my bot the chucknorris 2010-10-22T13:01:57 or even find out about it 2010-10-22T13:01:58 JamesMG, is that your minimax one? 2010-10-22T13:02:03 yeah lavalamp_ 2010-10-22T13:02:13 the one I hacked together to test McLeopold's map yesterday 2010-10-22T13:02:20 nice 2010-10-22T13:02:24 I made it 1 ply, but with a final state estimation 2010-10-22T13:02:29 im considering minimax if this one fails 2010-10-22T13:03:08 i have some ideas to prune the tree to make it workable 2010-10-22T13:03:10 *** p4p4 has joined #aichallenge 2010-10-22T13:03:33 that's why it's so limited right now raedwulf, when it had a deeper tree with more move options it crawled along 2010-10-22T13:03:35 raedwulf, what kind of bot are you doing right now? 2010-10-22T13:04:03 cfaftw: it just does a lot of analysis on the current state 2010-10-22T13:04:09 no particular algorithm 2010-10-22T13:04:13 oh cool 2010-10-22T13:04:16 that's what my bot does 2010-10-22T13:04:46 well, its the *safe bet* 2010-10-22T13:05:14 im designing it so i *could* almost plug it into an algorithm that uses cost etc. 2010-10-22T13:05:15 JamesMG, interesting. Mine is doing 2 ply, and I think I'm using the word wrong because that's 2 moves from each player that mine does... 2010-10-22T13:05:59 I may be using it wrong too, this is my first exposure to minimax 2010-10-22T13:06:15 raedwulf, that's a good idea 2010-10-22T13:06:35 yeah, traditionally, ply is one turn by one player 2010-10-22T13:06:44 all mine is considering right now is this turn's potential moves for itself 2010-10-22T13:06:44 bbl 2010-10-22T13:06:48 but here, of course, turns are simultaneous decisions... 2010-10-22T13:06:50 I think that's 1 ply? 2010-10-22T13:08:03 i need to find out why my sniping logic is so terrible 2010-10-22T13:08:40 JamesMG: in my implementation 1 ply is one turn for both players 2010-10-22T13:10:03 *** boegel has joined #aichallenge 2010-10-22T13:10:25 rebelxt: ur using some form of minimax? 2010-10-22T13:10:40 phreeza: free EC2 micro instances - pretty cool. could run a TCP server for free starting Nov 1st 2010-10-22T13:11:13 Top 10 players: Hazard(4049), cfaftw(4033), sin_sun(4030), Accoun(4018), rsergio(4007), felixcoto(3994), Manwe56(3993), savchenko(3976), dmj111(3952), davidjliu(3902) 2010-10-22T13:11:14 *** boegel has quit IRC (Client Quit) 2010-10-22T13:11:28 Error323: yes but my tree bot has no chance against my heuristics bot 2010-10-22T13:11:44 i see 2010-10-22T13:12:18 i've been trying it aswell but just... no way to prune the action space without loss of good actions 2010-10-22T13:13:18 exactly 2010-10-22T13:14:11 its a similar problem to go 2010-10-22T13:14:14 "Go" 2010-10-22T13:14:19 of course felixcoto had to break my bot's winning streak once again.. 2010-10-22T13:14:21 its way too much branching 2010-10-22T13:14:42 probably more branching if you consider each size of fleet 2010-10-22T13:14:44 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-10-22T13:15:00 raedwulf: you are right, Go AI still has no chance against decent human players... amazing 2010-10-22T13:15:04 there are 122 choose 22 possible sets of orders on the first turn. 2010-10-22T13:15:16 I fear the best bot will be a delecate heuristic, which is a pitty 2010-10-22T13:15:31 Error323, O:) 2010-10-22T13:15:32 kinda of inevitable as game complexity rises, isn't it? 2010-10-22T13:15:37 yep 2010-10-22T13:15:43 well minimax is kinda cheating 2010-10-22T13:15:45 :P 2010-10-22T13:15:46 i think there will still be some use for minimax 2010-10-22T13:15:57 I'm beginning to doubt that 2010-10-22T13:16:02 but it'd be shallow, yes, and for narrow purposes 2010-10-22T13:16:50 minimax would probably only work on a hybrid 2010-10-22T13:17:06 but heh, no idea how that would work - you'dprobably lose the benefits of minimax 2010-10-22T13:17:37 ill find out once i implement my first bot, then my minimax 2010-10-22T13:18:01 rewriting my bot code now because i actually sat down with a piece of paper and scribbled my method on it 2010-10-22T13:18:10 and actually dry ran it through my brain 2010-10-22T13:18:11 well i wasted a week on a fast alpha beta 2010-10-22T13:18:17 imho, best bots already seem to be doing single ply analysis (combinations of moves for both players) plus lots of heuristics. scoring function matters the most 2010-10-22T13:18:18 :( 2010-10-22T13:18:24 and then coded a heuristik bot in 1 day 2010-10-22T13:18:31 which is now doing fairly well 2010-10-22T13:18:32 :P 2010-10-22T13:18:37 nice 2010-10-22T13:19:20 well my rage bot seems to win 1/4 games 2010-10-22T13:19:23 3/4* 2010-10-22T13:19:26 i meant 2010-10-22T13:19:28 lol 2010-10-22T13:19:39 which is a bit silly lol 2010-10-22T13:19:50 i guess lots of folk still have sample bots running? 2010-10-22T13:20:14 probably at least a third of everyone 2010-10-22T13:20:17 interestingly my bot loses against the original rage bot, but wins a lot better against other sample bots... go figure 2010-10-22T13:20:18 unfortunetly.. ppl never take down starter package uploads (some changed one line and re-uploaded) 2010-10-22T13:21:26 raedwulf: try to make sure you can beat rage bot on all maps. you are either not defending or over-expanding and unable to defend 2010-10-22T13:21:45 hmmm, well thats just a test bot still 2010-10-22T13:21:55 my planned heuristics should tear it apart 2010-10-22T13:22:11 though its only theoretical thus far 2010-10-22T13:24:32 theoretical butt-kicking 2010-10-22T13:24:35 the best kind 2010-10-22T13:24:35 :D 2010-10-22T13:24:44 ;-) 2010-10-22T13:25:23 i started like a week ago hehe :P 2010-10-22T13:25:31 i only decided on my algorithm today :P 2010-10-22T13:26:00 last night i had a dream where my bot unlocked uber production mode when each planets starts generating ships twice as fast 2010-10-22T13:26:19 yay :D 2010-10-22T13:26:25 it was also able to analyze 200-ply deep in 50 milliseconds! 2010-10-22T13:26:34 wooohooooo :D 2010-10-22T13:26:41 so, not a single game lost! 2010-10-22T13:26:42 *** Azrathud has joined #aichallenge 2010-10-22T13:26:43 you need to take a break 2010-10-22T13:26:45 lol 2010-10-22T13:26:47 hahaha 2010-10-22T13:26:58 lol, agreed 2010-10-22T13:27:04 *** Azrathud has quit IRC (Client Quit) 2010-10-22T13:27:07 did it start finding it too poor a challenge, and so took over all computers in the world 2010-10-22T13:27:23 and sent fleets to each PC and mainframe 2010-10-22T13:27:44 massive fleets attacking NASA to launch space shuttles into space! 2010-10-22T13:27:59 to spread rebelxt's bot everywhere in the galaxy 2010-10-22T13:28:02 then the universe! 2010-10-22T13:28:09 i sometimes feel like Ender 2010-10-22T13:28:10 then i woke up :( maybe next night it can continue spreading around 2010-10-22T13:28:14 haha 2010-10-22T13:28:26 *** boegel has joined #aichallenge 2010-10-22T13:28:32 Focus, raedwulf, FOCUS 2010-10-22T13:28:37 it turns out our bots are actually controlling some galactic force in battle somewhere in the universe 2010-10-22T13:28:51 we just think it's a game 2010-10-22T13:28:52 ^^ 2010-10-22T13:28:54 *** Frontier has quit IRC (Read error: Operation timed out) 2010-10-22T13:29:22 ^^^ 2010-10-22T13:29:51 maybe, there's some secret americo-canadian inter-galactic automated force that we're helping write the AI to coordinate their attack 2010-10-22T13:30:06 *** jmpespxoreax has left #aichallenge 2010-10-22T13:30:12 *** bhasker_ has joined #aichallenge 2010-10-22T13:30:15 *** tapwater has joined #aichallenge 2010-10-22T13:31:55 so the winner gets to have his life ruined, like Ender's? 2010-10-22T13:32:09 *** p4p4 has quit IRC (Read error: Connection timed out) 2010-10-22T13:32:27 JamesMG: no everyone gets to have their life ruined 2010-10-22T13:32:32 Awesome! 2010-10-22T13:32:37 winning bot will probably have opening book with 3,141,592,653,589,793,238 positions 2010-10-22T13:32:45 haha 2010-10-22T13:32:59 with a magic compression into 2Mb! 2010-10-22T13:33:06 *** bhasker has quit IRC (Ping timeout: 252 seconds) 2010-10-22T13:33:08 of course 2010-10-22T13:33:11 *** p4p4 has joined #aichallenge 2010-10-22T13:33:25 http://72.44.46.68/canvas?game_id=631843 2010-10-22T13:33:27 so, which would be more exciting, the compression algorithm or the bot? 2010-10-22T13:33:28 this thing is beomcing fun 2010-10-22T13:33:55 well, it might just generate opening book during the 2-3 second period allocated for the first turn by analyzing every possible game 2010-10-22T13:33:56 raedwulf: you can send larger files to organizers and they upload files for you 2010-10-22T13:33:59 meh, it'll become fun when i actually finish my bot's algorithm xd 2010-10-22T13:34:06 Naktibalda: neat ^^ 2010-10-22T13:34:17 * raedwulf sends up his 300Gb entry 2010-10-22T13:34:30 errr i better start uploading now, my upload speed is slow 2010-10-22T13:34:35 ;-) 2010-10-22T13:34:39 JamesMG, you need some "will this work" math in there :) 2010-10-22T13:34:52 *** nann has joined #aichallenge 2010-10-22T13:35:00 haha no kidding 2010-10-22T13:35:11 voodooo maths 2010-10-22T13:35:18 *** rebelxt has left #aichallenge 2010-10-22T13:36:03 I think it's interesting the ways it takes the limited options and does things I had to think about quite a bit on my first bot 2010-10-22T13:36:08 can i assume the code will always be run on a posix system? 2010-10-22T13:36:12 i need to use alloca 2010-10-22T13:36:14 maybe when I get my bot working I'll let it chug away on an opening position and see what it comes up with, and how different it is from just a couple ply analysis 2010-10-22T13:36:43 meh, ill use a malloc/free 2010-10-22T13:37:20 lavalamp_: you should run it on McLeopold's limited map7 from yesterday 2010-10-22T13:37:32 less chugging, interesting options 2010-10-22T13:38:08 my new bot beat rage bot locally: it spent 195 moves doing nothing, then it took a 1 ship +3 neutral-- because it knew that rage bot's incoming fleet wouldn't land before the end of the game... 2010-10-22T13:38:23 *** bhasker has joined #aichallenge 2010-10-22T13:38:24 haha 2010-10-22T13:38:26 *** Apophis has quit IRC (Ping timeout: 276 seconds) 2010-10-22T13:38:26 hahaha 2010-10-22T13:38:30 that's awesome 2010-10-22T13:38:39 schoolyard bully planetwars AI 2010-10-22T13:38:51 JamesMG, yeah, I'll be trying those maps once I get a few more things working 2010-10-22T13:39:28 lol 2010-10-22T13:40:45 *** bhasker_ has quit IRC (Ping timeout: 272 seconds) 2010-10-22T13:42:38 * raedwulf had a duh moment 2010-10-22T13:45:07 *** yasith has quit IRC (Ping timeout: 240 seconds) 2010-10-22T13:48:32 What is Zepp's forum username? 2010-10-22T13:52:58 *** p4p4 has quit IRC (Read error: Connection timed out) 2010-10-22T13:53:10 nm, looks like it's Zepp :) 2010-10-22T13:53:31 it smells like Zepp 2010-10-22T13:53:38 it tastes like Zepp... it is Zepp! 2010-10-22T13:54:11 *** p4p4 has joined #aichallenge 2010-10-22T13:54:58 JamesMG, your bot is rated worse than mine now, lol... if I were at home I'd start mine so it could play you... 2010-10-22T13:56:40 I broke it 2010-10-22T13:56:52 but it should be getting *slightly* smarter about now 2010-10-22T13:58:21 I'd like to have them fight though, would be fun 2010-10-22T13:58:42 etc 2010-10-22T13:58:44 er 2010-10-22T13:59:20 *** yasith has joined #aichallenge 2010-10-22T14:00:13 I'll throw it on a 2 minute timer over the weekend if you want to try and get a few games in lavalamp_ 2010-10-22T14:00:32 Yeah, that'd be cool, JamesMG 2010-10-22T14:00:49 i just put my latest devel bot on TCP 2010-10-22T14:00:53 jb.exp.something 2010-10-22T14:01:48 jb.exp.88f4 2010-10-22T14:03:42 wow, my minimax one does not handle being outnumbered at all well 2010-10-22T14:03:48 cfaftw, works better than mine :) 2010-10-22T14:04:02 dang, first game and i see a bug 2010-10-22T14:04:03 lol 2010-10-22T14:04:20 JamesMG: does it go "eeek, RETREAT RETREAT!" 2010-10-22T14:04:29 or "AIIIIIIYYYYY, KAMIKAZEEEEE" 2010-10-22T14:05:01 neither, it stays in bed 2010-10-22T14:05:20 lol 2010-10-22T14:05:40 maybe it sees that the best move is not to play 2010-10-22T14:05:40 "just don't feel like fightin today guys, good luck" 2010-10-22T14:05:57 "THEY'RE ATTACKING, THEY'RE ATTACKING..." ... "James go back to sleep..." 2010-10-22T14:06:39 JamesMG, I see you improved your move generation, what do you consider now? 2010-10-22T14:06:55 I think I need to evaluate the negative impact and minimize it in situations where there is no entirely positive move 2010-10-22T14:07:06 which is what is making it throw the white flag more or less :) 2010-10-22T14:07:33 lavalamp_: subset of moves from pairs of friendly planets 2010-10-22T14:08:25 interesting 2010-10-22T14:08:30 I don't know any algorithm to efficiently get subsets from multiple sets 2010-10-22T14:08:38 that'll be my next step I think 2010-10-22T14:08:46 with one item from each set 2010-10-22T14:10:03 *** janzert has quit IRC (Read error: Connection reset by peer) 2010-10-22T14:11:16 Top 10 players: Hazard(4110), cfaftw(4079), Accoun(4063), rsergio(4006), sin_sun(3993), felixcoto(3983), Manwe56(3965), savchenko(3964), dmj111(3964), davidjliu(3903) 2010-10-22T14:11:35 *** janzert has joined #aichallenge 2010-10-22T14:12:35 * raedwulf implements a heap sort 2010-10-22T14:20:03 yay, i won against RageBot on the TCP 2010-10-22T14:20:49 *** p4p4 has quit IRC (Read error: Connection timed out) 2010-10-22T14:20:51 ^^ 2010-10-22T14:20:51 *** Florian95 has quit IRC (Quit: Leaving...) 2010-10-22T14:21:16 *** perror has quit IRC (Quit: Bye all !) 2010-10-22T14:24:53 *** jmpespxoreax has joined #aichallenge 2010-10-22T14:27:18 cfaftw: you're going up again :P 2010-10-22T14:27:45 on official server or TCP? 2010-10-22T14:28:10 official server, you're second 2010-10-22T14:29:11 *** p4p4 has joined #aichallenge 2010-10-22T14:35:31 oh 2010-10-22T14:35:34 yeah another good streak 2010-10-22T14:35:38 nice 2010-10-22T14:35:53 im really not in a programming mood 2010-10-22T14:36:01 probably need to sleep early and wake about 2am 2010-10-22T14:36:04 to get anything done 2010-10-22T14:36:31 ... blast the music into my ears... always helps 2010-10-22T14:44:32 hi all 2010-10-22T14:44:39 sup 2010-10-22T14:45:07 hi 2010-10-22T14:45:15 we're taking over the universe 2010-10-22T14:45:18 cfaftw: how goes the new bot on TCP? 2010-10-22T14:45:28 it's ok 2010-10-22T14:45:33 cf is your bot cccf9ext12? 2010-10-22T14:45:41 261 current elo 2010-10-22T14:45:47 no, that is felixcoto 2010-10-22T14:46:07 http://72.44.46.68/getplayer?player=jb.exp.88f4 2010-10-22T14:46:09 ah ok 2010-10-22T14:46:11 I'm hoping to see it play my old code 2010-10-22T14:46:15 my current devel bot 2010-10-22T14:46:37 ok i see 2010-10-22T14:46:56 *** choas has joined #aichallenge 2010-10-22T14:47:07 finally getting to top 100 2010-10-22T14:47:28 one week ago i didnt knew anything from programation, i am quite hapy :) 2010-10-22T14:48:00 this challenge is definitely very good for learning java 2010-10-22T14:48:14 *** Apophis has joined #aichallenge 2010-10-22T14:48:31 nice! 2010-10-22T14:48:37 what's your bot? 2010-10-22T14:48:44 the same name then mine 2010-10-22T14:48:46 i mean... link to your account page 2010-10-22T14:49:05 it is http://ai-contest.com/profile.php?user_id=11041 2010-10-22T14:49:28 but i dont have programmed any defense, if i do this i think i will win a lot of ranks 2010-10-22T14:49:46 my goal is to go top 20 but i am not sure i can. 2010-10-22T14:50:12 hmmm, got timed out/crashed against JamesMG this night, http://www.ai-contest.com/visualizer.php?game_id=6002348, but my profile shows it as a win. shure, i can need some cheering up, but hmmm, not this way. 2010-10-22T14:50:41 that's a bug, it always shows as the first player's error, but the win/loss is correct 2010-10-22T14:51:29 which is strange since that code has been running for thousands of games on the TCP server without issue 2010-10-22T14:51:34 i noticed another bug too. When two players have same ranking in ranking page ( with same elo ), if you click the profile page for both player, they have not the same ranking 2010-10-22T14:51:43 *** nann has quit IRC (Ping timeout: 276 seconds) 2010-10-22T14:51:56 strange, too, that the whole affair does not show up in his profile, at all 2010-10-22T14:52:22 whoa, yeah 2010-10-22T14:53:59 well 2010-10-22T14:54:11 that game JamesMG's submission is 166468 2010-10-22T14:54:17 his current submission is 166469 2010-10-22T14:55:31 same code, was fighting with the C# compiler :) 2010-10-22T14:55:32 i work in java. For each of mine planets, i need to create a list of neighbours ordered by distance. Do you think adding a atribute to planet class ( tab of planets, that lenghts 'pw.myplanets.size()-1' ) is the best choice i can do ? 2010-10-22T14:55:37 interesting... my lowly bot just won against CCCF-ex9 2010-10-22T14:55:51 but yeah that would be why it doesn't display p4p4 2010-10-22T14:56:37 you submitted twice without waiting for a response? or you hadn't gotten any responses? 2010-10-22T14:57:11 hadn't gotten any 2010-10-22T14:57:30 was impatient I guess 2010-10-22T14:57:41 well, i dunno what race caused that, but it was definitely a race 2010-10-22T14:58:16 likely one of the servers tried to get 166468 but it only had your current submission of 166469 2010-10-22T14:58:28 I can re-up if that'll help sort it out 2010-10-22T14:58:43 i don't know 2010-10-22T14:59:19 we still need to address the cause of the bug rather than just addressing symptoms 2010-10-22T14:59:21 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-10-22T14:59:24 the case of the phantom entry :) 2010-10-22T14:59:39 *** n9thbit has quit IRC (Ping timeout: 265 seconds) 2010-10-22T15:07:18 *** p4p4 has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2010-10-22T15:11:16 Top 10 players: Hazard(4099), cfaftw(4041), Accoun(4041), rsergio(3991), felixcoto(3984), sin_sun(3967), asavis(3954), Manwe56(3938), dmj111(3922), MariusCG(3909) 2010-10-22T15:14:00 *** RainCT has joined #aichallenge 2010-10-22T15:14:13 that's it i'm pulling the plug... i've seen enough. need to do more work on the bot later 2010-10-22T15:16:46 322 Elo is pretty respectable... 2010-10-22T15:18:33 I'd be ok with a positive ELO 2010-10-22T15:18:42 *** Accoun has quit IRC () 2010-10-22T15:18:50 it's decent 2010-10-22T15:18:52 but it's not winning 2010-10-22T15:19:31 actually, i've been testing incrementally 2010-10-22T15:19:34 so some things aren't implemented 2010-10-22T15:19:43 my elo has been going up each time, which is a good thing 2010-10-22T15:21:39 *** wh1teside has joined #aichallenge 2010-10-22T15:22:23 accoun just resubmitted 2010-10-22T15:23:39 did he changed his programming language? 2010-10-22T15:26:46 from C++? 2010-10-22T15:26:48 no 2010-10-22T15:27:10 @rankings 2010-10-22T15:27:11 amstan: Top 10 players: Hazard(4104), rsergio(4037), cfaftw(4000), felixcoto(4000), sin_sun(3989), asavis(3956), dmj111(3937), MariusCG(3913), davidjliu(3910), bix0r4ever(3889) 2010-10-22T15:27:12 he's gone again 2010-10-22T15:27:21 he resubmitted 2010-10-22T15:27:40 anyone have his id? 2010-10-22T15:28:04 jsut go to my account page, i lost to him recently 2010-10-22T15:28:09 3 games ago 2010-10-22T15:28:28 he resubmitted again 2010-10-22T15:28:33 and lost 2010-10-22T15:29:05 yeah 2010-10-22T15:29:12 but Raschi is looking strong 2010-10-22T15:29:18 never seen him before 2010-10-22T15:29:26 http://ai-contest.com/visualizer.php?game_id=6019117 2010-10-22T15:29:27 hmmmm 2010-10-22T15:29:27 Current Rank: 252 2010-10-22T15:29:31 cfaftw: not really 2010-10-22T15:29:32 lol 2010-10-22T15:29:37 but it's all wins 2010-10-22T15:29:47 but... he definitely submitted like 30 times today 2010-10-22T15:29:52 oh 2010-10-22T15:29:53 yeah 2010-10-22T15:29:55 that's why.. 2010-10-22T15:30:33 there's definitely a race somewhere with submissions 2010-10-22T15:30:45 umm 2010-10-22T15:30:46 watching Raschi's bot, it looks strong; it makes good decisions 2010-10-22T15:30:46 http://ai-contest.com/profile_submissions.php?user_id=10464&page=102 2010-10-22T15:30:48 i don't like this 2010-10-22T15:30:49 Raschi played himself 2010-10-22T15:30:59 he's resubmitting every minute 2010-10-22T15:31:18 gah, I'm starting to think we should impose a rate limit on resubmission 2010-10-22T15:31:21 how is he successful so often? C# shouldn't compile that well! 2010-10-22T15:31:28 LOL 2010-10-22T15:31:30 amstan - training? 2010-10-22T15:31:40 automated submitter? 2010-10-22T15:31:42 how did he play himself? 2010-10-22T15:31:44 ...that was way back before the serious problems started 2010-10-22T15:31:46 raedwulf: he should use the tcp for this 2010-10-22T15:31:54 he played an old version of himself 2010-10-22T15:32:11 yea, he should - was wondering if that was a plausible explanation 2010-10-22T15:32:12 it's a race condition that easily occurs when someone submits too close together 2010-10-22T15:32:12 once every 10 minutes for submission should be plenty 2010-10-22T15:32:24 janzert: exactly 2010-10-22T15:32:37 please don't do that unless C# submissions actually compile :( 2010-10-22T15:32:38 k, i'm emailing him 2010-10-22T15:32:44 good 2010-10-22T15:32:45 stop it or DIE 2010-10-22T15:32:46 *** Accoun has joined #aichallenge 2010-10-22T15:32:52 stop it AND DIE 2010-10-22T15:32:56 this is just not cool, like it's unfair to everyone 2010-10-22T15:32:56 :P 2010-10-22T15:32:56 http://ai-contest.com/profile_submissions.php?user_id=10464&page=73 2010-10-22T15:33:00 yep 2010-10-22T15:33:04 Accoun: how are you doing? 2010-10-22T15:33:12 7:28:00, :08, :18, :32, :50 2010-10-22T15:33:16 just add it to the contest rules too :P 2010-10-22T15:33:18 that's a lot of pages of submissions 2010-10-22T15:33:19 five submissions in one minute 2010-10-22T15:33:51 http://ai-contest.com/profile_submissions.php?user_id=10527&page=2 2010-10-22T15:34:03 I'm no better when I'm trying to get it to catch the C# compiler working though 2010-10-22T15:34:16 amstan, is it possible to set a hard limit on rate of submissions? 2010-10-22T15:34:25 oh, he switched to java on oct 21 2010-10-22T15:34:28 cfaftw: probably, but someone needs to code it 2010-10-22T15:34:45 lool like cloud severs very slow mashines 2010-10-22T15:34:48 cfaftw, someone meaning a contest organizer? 2010-10-22T15:34:50 look like cloud severs very slow mashines 2010-10-22T15:34:53 I'll try and look at this evening 2010-10-22T15:34:58 *** Naktibalda has joined #aichallenge 2010-10-22T15:35:00 Accoun: yeah, we're still working on them, sorry for that 2010-10-22T15:35:03 dang i just talked to myself again 2010-10-22T15:35:15 haha cfaftw 2010-10-22T15:35:16 cfaftw: someone meaning anyone that wants to do it 2010-10-22T15:35:36 cloud servers still about 3x faster than main 2010-10-22T15:37:00 *yawn* 1441 lines of ok and still no working bot >.< 2010-10-22T15:37:05 1 sec on cloud servers like .5 sec on my core 2 2.333 2010-10-22T15:37:19 of code* 2010-10-22T15:37:36 readwulf, clearly you need more lines of code 2010-10-22T15:37:39 yes 2010-10-22T15:37:41 :P 2010-10-22T15:37:44 raedwulf, otherwise it would be working already 2010-10-22T15:37:49 ^^ 2010-10-22T15:37:52 can someone search on the forums, someone was talking about automated code uploads, i forgot who though: http://ai-contest.com/forum/search.php 2010-10-22T15:37:59 probably will do something at 2k 2010-10-22T15:41:22 amstan: i couldn't find anything 2010-10-22T15:41:26 oh well 2010-10-22T15:41:28 w/e 2010-10-22T15:41:34 maybe it was in here 2010-10-22T15:41:51 like.. it's a useful thing, to have an autoupload script, but don't put it in cron 2010-10-22T15:43:13 it's probably on purpose 2010-10-22T15:43:29 game http://ai-contest.com/visualizer.php?game_id=6029519 in my games list and absent in enemy games list 2010-10-22T15:43:54 Accoun: yeah, he's resubmitting like crazy 2010-10-22T15:43:57 Accoun: i gave him a warning 2010-10-22T15:43:58 we know about Raschi 2010-10-22T15:46:39 *** Titankiller has quit IRC (Read error: Connection reset by peer) 2010-10-22T15:48:11 last year contest masine be faster than me 2010-10-22T15:48:19 in this yaer slover 2010-10-22T15:48:54 perl bot version play well on TCP but anytame timeout on contest server 2010-10-22T15:48:55 maybe put in the FAQ about the tcp server? 2010-10-22T15:49:19 or even on the submit page 2010-10-22T15:49:32 so people know about it, and maybe won't submit so often 2010-10-22T15:50:19 "Save the world, recycle... Save the cycles, use TCP server." 2010-10-22T15:50:43 amstan: think it would be ok for me to add in a hard limit of no new submission for 10 minutes after a successful submit? 2010-10-22T15:50:57 janzert: yeah 2010-10-22T15:52:04 raedwulf: tcp is not exactly a right, it's a privilege, we have no official binds with dhmatei 2010-10-22T15:52:11 ah ok 2010-10-22T15:53:10 how did Raschi play himself :o? 2010-10-22T15:53:29 new submission against old 2010-10-22T15:53:41 ah ok 2010-10-22T15:53:48 old submissions can stay in play for a short time after a new one is in 2010-10-22T15:55:50 janzert: I think I have an old submission still in play from Oct 13th 2010-10-22T15:56:13 ouch, hmm. 2010-10-22T15:56:23 got the id for it by any chance? 2010-10-22T15:56:30 JamesMG: did you do anything weird with submitting? 2010-10-22T15:56:40 JamesMG: can you link us a game? 2010-10-22T15:56:56 [11:58] likely one of the servers tried to get 166468 but it only had your current submission of 166469 2010-10-22T15:57:10 http://www.ai-contest.com/visualizer.php?game_id=6002348 2010-10-22T15:57:19 this game isn't on my current submission's game list 2010-10-22T15:57:36 JamesMG: is it in the other guy's list? 2010-10-22T15:57:39 yes 2010-10-22T15:57:50 JamesMG: it will always stay there 2010-10-22T15:58:10 JamesMG: unless that game happened a while after you submitted your new bot 2010-10-22T15:58:17 then it's a bug 2010-10-22T15:58:28 it happened today, haven't submitted since the 13th 2010-10-22T15:58:33 hmm 2010-10-22T15:58:35 yeah 2010-10-22T15:58:37 that's not cool 2010-10-22T15:59:17 hazard's games are nice 2010-10-22T16:01:19 answer from raschi: I did no automated upload, I did it manually. I have a lot of time :) 2010-10-22T16:01:27 lol :P 2010-10-22T16:01:32 O.o 2010-10-22T16:01:35 lol 2010-10-22T16:02:02 heh, automated or manual doesn't really matter to me. Just please slow down :) 2010-10-22T16:02:23 @where tcp 2010-10-22T16:02:23 amstan: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/ 2010-10-22T16:04:14 well.. 2010-10-22T16:04:23 this was nicely solved without banning and stuff 2010-10-22T16:04:28 just communication is enough 2010-10-22T16:04:52 he agree to hammer TCP instead? 2010-10-22T16:05:09 there's nothing to hammer in tcp, but to play normally 2010-10-22T16:05:13 *** nann has joined #aichallenge 2010-10-22T16:05:16 there's no concepts of submissions there 2010-10-22T16:05:31 yeah I know, just exaggerating 2010-10-22T16:05:36 i sent him a link 2010-10-22T16:05:42 what he does, his problem 2010-10-22T16:06:35 *** Gorovkhaia2 has quit IRC (Ping timeout: 240 seconds) 2010-10-22T16:06:39 *** Gorovkhaia2 has joined #aichallenge 2010-10-22T16:08:04 *** Apophis has quit IRC (Ping timeout: 240 seconds) 2010-10-22T16:11:09 *** rebelxt has joined #aichallenge 2010-10-22T16:11:15 Top 10 players: Hazard(4099), rsergio(4028), felixcoto(4009), cfaftw(3972), asavis(3966), sin_sun(3963), dmj111(3957), RoadToKommunizm(3955), MariusCG(3935), davidjliu(3893) 2010-10-22T16:12:31 *** aerique has joined #aichallenge 2010-10-22T16:12:47 *** hornairs has joined #aichallenge 2010-10-22T16:14:43 amstan: oh dear, you sound like an administrator 2010-10-22T16:15:09 i am an admin, lol 2010-10-22T16:15:28 yes and i know what one sounds like, because i was one :P 2010-10-22T16:15:44 for game servers and stuff 2010-10-22T16:15:46 hehe 2010-10-22T16:15:51 oh yeah, those 2010-10-22T16:16:01 it's enjoyable kicking noobs off game servers 2010-10-22T16:16:05 ^^ 2010-10-22T16:16:17 i'm an admin for 3 sauerbraten servers 2010-10-22T16:16:21 nice 2010-10-22T16:17:01 i used to run 1 jk2, 2 wolf:et, 2 CSS on a dedicated server 2010-10-22T16:17:26 * raedwulf really should sleep now 2010-10-22T16:17:28 nn 2010-10-22T16:17:33 good night 2010-10-22T16:18:36 *** boegel has quit IRC (Quit: Leaving) 2010-10-22T16:26:21 *** Bobng has quit IRC (Read error: Connection reset by peer) 2010-10-22T16:44:39 Ooops, just wrote a 557MB logfile 2010-10-22T16:44:43 gee I wonder why it timed out. 2010-10-22T16:47:01 haha, just one game? or one turn? ;-) 2010-10-22T16:47:11 one game, 105ish turns 2010-10-22T16:47:21 was spitting out my entire decision tree each turn 2010-10-22T16:47:26 ouch 2010-10-22T16:47:29 haha nice 2010-10-22T16:47:30 anyone know does php or apache take care of automatically deleting uploaded files if the script leaves them laying around? 2010-10-22T16:47:35 *** Utkarsh has quit IRC (Ping timeout: 272 seconds) 2010-10-22T16:47:57 janzert, I'd be surprised if it did 2010-10-22T16:48:13 I'm impressed it didn't time out sooner, that's a lot of data to be writing mid-turn 2010-10-22T16:48:15 because the current upload handler just leaves them laying around if it encounters any error 2010-10-22T16:48:46 *** rebelxt has left #aichallenge 2010-10-22T16:51:24 lavalamp_: http://72.44.46.68/canvas?game_id=634189 2010-10-22T16:51:31 it still has it's issues, but it was going to win one! 2010-10-22T16:51:39 until it timed out 2010-10-22T16:52:08 *** Utkarsh has joined #aichallenge 2010-10-22T16:52:20 Nice :) 2010-10-22T16:53:29 *** cfaftw has quit IRC (Quit: going home) 2010-10-22T16:54:00 it's attack ship count estimations are way too optimistic still 2010-10-22T16:54:18 how many ply/nodes is it doing? 2010-10-22T16:54:47 1 ply, planets owned * planets * 2 nodes 2010-10-22T16:55:21 cool 2010-10-22T16:57:01 that's wrong actually, it got up to 4508 nodes last game, forgot I uncommented some other options 2010-10-22T16:58:04 *** narnach has joined #aichallenge 2010-10-22T16:58:51 was taking 2 seconds per turn to process that, I let it run out because it's on TCP so it can get away with that sort of murder 2010-10-22T16:59:04 *** Appleman1234 has joined #aichallenge 2010-10-22T16:59:04 I need to optimise a bit 2010-10-22T16:59:18 4500, nice 2010-10-22T17:02:16 oh but 2 seconds! haha 2010-10-22T17:02:27 yeah :( 2010-10-22T17:03:15 I think mine was getting 2500 a second this morning 2010-10-22T17:04:06 *** wvdschel has joined #aichallenge 2010-10-22T17:04:12 *** wvdschel has quit IRC (Changing host) 2010-10-22T17:04:12 *** wvdschel has joined #aichallenge 2010-10-22T17:04:40 *** antimatroid has quit IRC (Read error: Connection reset by peer) 2010-10-22T17:04:57 *** antimatroid has joined #aichallenge 2010-10-22T17:09:40 *** nann has quit IRC (Quit: Verlassend) 2010-10-22T17:09:57 *** Apophis has joined #aichallenge 2010-10-22T17:10:04 *** delt0r___ has joined #aichallenge 2010-10-22T17:11:16 Top 10 players: Hazard(4071), felixcoto(4051), rsergio(4045), cfaftw(3995), asavis(3994), dmj111(3982), RoadToKommunizm(3980), sin_sun(3973), RebelXT(3889), davidjliu(3887) 2010-10-22T17:11:42 *** delt0r has quit IRC (Ping timeout: 265 seconds) 2010-10-22T17:14:37 *** davidjliu has joined #aichallenge 2010-10-22T17:27:31 got it down to 200ms to construct 2250 nodes 2010-10-22T17:33:48 *** Naktibalda has quit IRC (Remote host closed the connection) 2010-10-22T17:36:20 *** Naktibalda has joined #aichallenge 2010-10-22T17:38:11 *** bhasker_ has joined #aichallenge 2010-10-22T17:40:37 *** bhasker has quit IRC (Ping timeout: 252 seconds) 2010-10-22T17:50:06 *** lavalamp_ has quit IRC (Quit: bye all) 2010-10-22T17:51:54 *** aerique has quit IRC (Quit: ...) 2010-10-22T17:53:39 Ok, 10 minutes cool of time between successful submissions in place 2010-10-22T17:53:59 nice 2010-10-22T17:54:12 *** Naktibalda has quit IRC (Remote host closed the connection) 2010-10-22T17:56:12 janzert: what branch? 2010-10-22T17:56:23 trunk/server :/ 2010-10-22T17:56:42 http://code.google.com/p/ai-contest/source/detail?r=445 2010-10-22T17:56:54 that is trunk and server 2010-10-22T17:57:51 I suppose I should post something to the forums as well 2010-10-22T17:59:23 announcements are good ideas 2010-10-22T18:00:17 *** choas has quit IRC (Quit: leaving) 2010-10-22T18:01:24 *** rebelxt has joined #aichallenge 2010-10-22T18:01:41 amstan: You ok with it or would you rather I back it out for now? 2010-10-22T18:02:04 janzert: as long as it works there should be no reason for it to not be there 2010-10-22T18:02:06 so keep it 2010-10-22T18:02:26 janzert: yeah.. post an announcement, in the announcements forums, we need a few of those\ 2010-10-22T18:02:27 yep, tested it locally first then tested it again after putting it live 2010-10-22T18:02:53 i should go through some of the site code and fix spelling errors 2010-10-22T18:03:10 successfully is spelled wrong in two different ways now :/ 2010-10-22T18:03:23 which reminds me that I should resubmit an invalid bot to stop from adding to the starter bot pool on the server :) 2010-10-22T18:03:46 oops, what did I add? 2010-10-22T18:03:55 *** bhasker has joined #aichallenge 2010-10-22T18:03:57 'succesfully' 2010-10-22T18:04:12 bah 2010-10-22T18:04:16 twice :) 2010-10-22T18:04:31 and 'sucessfully' already existed 2010-10-22T18:04:49 I got the correct form of to/too at least :) 2010-10-22T18:05:19 *** bhasker__ has joined #aichallenge 2010-10-22T18:06:33 ok, I've got the comment in submission.php and the one in submit.php where else is it wrong? 2010-10-22T18:06:37 *** bhasker_ has quit IRC (Ping timeout: 265 seconds) 2010-10-22T18:07:27 hmm, variable named successfull but I think I'll leave that 2010-10-22T18:07:44 heh 2010-10-22T18:08:09 if I touch that code though it'll probably bug me enough to change it now 2010-10-22T18:08:15 *** bhasker has quit IRC (Ping timeout: 240 seconds) 2010-10-22T18:08:50 trunk/planet_wars/www/profile_submissions_widget.php has "sucessfully" on lines 90, 94 2010-10-22T18:10:25 ok and doing a couple greps that looks like all of 'em 2010-10-22T18:10:36 cool 2010-10-22T18:10:49 *** Titankiller has joined #aichallenge 2010-10-22T18:10:54 at least in the php code 2010-10-22T18:11:10 yay it's chugging along at 36 games/hour now 2010-10-22T18:11:17 Top 10 players: Hazard(4090), rsergio(4059), felixcoto(4052), asavis(4030), cfaftw(4001), sin_sun(3977), dmj111(3972), bocsimacko(3917), davidjliu(3893), RebelXT(3889) 2010-10-22T18:14:33 * janzert debates misspelling successfully in his spell correction commit message :P 2010-10-22T18:15:28 ಠ_ಠ 2010-10-22T18:16:27 a1k0n: where's your bot? 2010-10-22T18:24:31 announcement posted 2010-10-22T18:25:26 *** Titankiller has quit IRC (Read error: Connection reset by peer) 2010-10-22T18:25:48 *** JamesMG has quit IRC (Quit: Leaving.) 2010-10-22T18:26:16 *** smellyhippy has quit IRC (Ping timeout: 252 seconds) 2010-10-22T18:27:33 *** jaspervdj has quit IRC (Quit: NEVER GONNA GIVE YOU UP NEVER GONNA LET YOU DOWN) 2010-10-22T18:32:02 looks like official instances are dying, time to bring up mine 2010-10-22T18:33:00 *** Naktibalda has joined #aichallenge 2010-10-22T18:34:58 *** wvdschel has quit IRC (Quit: Leaving) 2010-10-22T18:36:59 *** rebelxt has left #aichallenge 2010-10-22T18:39:34 *** smellyhippy has joined #aichallenge 2010-10-22T18:39:35 *** smellyhippy has joined #aichallenge 2010-10-22T18:43:15 *** Naktibalda has quit IRC (Read error: Connection reset by peer) 2010-10-22T18:43:17 *** amstan_ has joined #aichallenge 2010-10-22T18:43:18 *** ChanServ sets mode: +o amstan_ 2010-10-22T18:46:23 *** amstan has quit IRC (Ping timeout: 245 seconds) 2010-10-22T18:52:45 *** Azrathud has joined #aichallenge 2010-10-22T18:53:23 *** amstan_ is now known as amstan 2010-10-22T19:04:58 it looks like the official server deals much more game then last days 2010-10-22T19:09:52 janzert fixed a bunch of stuff 2010-10-22T19:10:07 amstan: in secret laboratory 2010-10-22T19:10:40 in a secret laboratory, where it storms every night 2010-10-22T19:11:16 Top 10 players: Hazard(4100), rsergio(4069), felixcoto(4059), asavis(4036), cfaftw(4003), dmj111(3984), sin_sun(3964), bocsimacko(3910), MariusCG(3901), davidjliu(3894) 2010-10-22T19:16:59 *** mceier has quit IRC (Quit: leaving) 2010-10-22T19:20:19 *** robtheblob has joined #aichallenge 2010-10-22T19:20:23 *** Sarin has quit IRC (Remote host closed the connection) 2010-10-22T19:22:45 *** RainCT has quit IRC (Remote host closed the connection) 2010-10-22T19:25:34 *** robtheblob has left #aichallenge 2010-10-22T19:25:55 *** robtheblob has joined #aichallenge 2010-10-22T19:26:04 *** robtheblob has left #aichallenge 2010-10-22T19:29:52 *** hornairs has quit IRC (Ping timeout: 276 seconds) 2010-10-22T19:42:07 *** adg has joined #aichallenge 2010-10-22T19:43:54 *** smellyhippy has quit IRC (Ping timeout: 240 seconds) 2010-10-22T19:46:19 *** amriedle has joined #aichallenge 2010-10-22T19:49:39 *** smellyhippy has joined #aichallenge 2010-10-22T19:51:32 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-10-22T19:51:34 *** iFire` has joined #aichallenge 2010-10-22T19:53:01 *** Titankiller has joined #aichallenge 2010-10-22T19:59:27 *** iFire` has quit IRC (Read error: Connection reset by peer) 2010-10-22T20:00:39 *** iFire has joined #aichallenge 2010-10-22T20:00:42 *** justin_pdx has joined #aichallenge 2010-10-22T20:11:17 Top 10 players: Hazard(4108), asavis(4090), rsergio(4079), felixcoto(4053), cfaftw(4029), dmj111(3981), sin_sun(3971), MariusCG(3918), bocsimacko(3916), davidjliu(3903) 2010-10-22T20:19:32 *** bhasker__ has quit IRC (Ping timeout: 265 seconds) 2010-10-22T20:20:36 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-10-22T20:32:50 *** amstan_ has joined #aichallenge 2010-10-22T20:32:50 *** ChanServ sets mode: +o amstan_ 2010-10-22T20:36:09 *** amstan has quit IRC (Ping timeout: 265 seconds) 2010-10-22T20:39:54 are any of you allowing your states to make moves in the future? i would love to know how to the hell someone would have managed that 2010-10-22T20:40:33 it gets very messy in that i might then make other moves rendering it useless, so i need a way to remove it when that happens 2010-10-22T20:41:04 i did that but abandoned it. 2010-10-22T20:41:22 i don't attempt to make future moves anymore. 2010-10-22T20:42:02 but i figured it'd be useful in a minimax scenario for detecting places where your planets are snipable. but you can see that without the elaborate machinery. 2010-10-22T20:43:13 it's useful for minimax for many reasons 2010-10-22T20:43:17 just very tricky to get right 2010-10-22T20:43:30 yeah. i had to have a separate outgoing column in each frame of each planet 2010-10-22T20:43:40 i'm thinking i need to store outgoing fleets again, and incrementally make them every time i generate future states, but that seems very costly 2010-10-22T20:43:53 getting all the arithmetic right required a stupid hack where fleets already in flight had to exist in 'negative time' columns 2010-10-22T20:44:12 tiemout control on TCP server just maked harder? 2010-10-22T20:44:23 yeah, it would have been easier if i didn't have neutral resolutions handled the way they are 2010-10-22T20:44:30 i doubt it would have changed, Accoun 2010-10-22T20:44:42 Accoun: are you making future fleets in your bot yet? 2010-10-22T20:49:15 i think i have ping probles in this cause 2010-10-22T20:49:21 i think i have ping problems in this cause 2010-10-22T20:49:36 timeout any time 2010-10-22T20:54:39 *** rebelxt_ has joined #aichallenge 2010-10-22T21:08:29 a1k0n: are you coordinating attacks? it gets even more tricky to handle "moves" then :( 2010-10-22T21:10:59 was considering it. it's way too hard to come up with all the potential combinations; i was thinking of just doing "closest N planets attack, and each sends a constant proportion of their own ships to the destination" 2010-10-22T21:11:17 Top 10 players: Hazard(4097), rsergio(4085), asavis(4077), felixcoto(4045), cfaftw(4019), Raschi(3954), dmj111(3953), bocsimacko(3939), sin_sun(3936), davidjliu(3893) 2010-10-22T21:12:10 it's not too hard if you do it when you are creating potential moves 2010-10-22T21:12:25 it's a whore to handle them into the future with states though 2010-10-22T21:14:57 *** Accoun has quit IRC () 2010-10-22T21:18:04 for what its worth, I coordinate attacks by decreasing num ships on a planet, and adding to the fleets list a new fleet that will arrive at the future time. If that makes sense. 2010-10-22T21:18:20 *** Accoun has joined #aichallenge 2010-10-22T21:18:29 dmj111: what if you then consider moves that render them useless? 2010-10-22T21:18:35 are you leaving the false information in there? 2010-10-22T21:19:02 they might not even be sendable anymore with the planets future state 2010-10-22T21:19:09 Then, for the rest of the current turn, those ships are accounted for, so I don't double up on them. On the next move, I may change my mind because of something the enemy has done, and then wasted the ships I already sent. 2010-10-22T21:19:25 i am making moves by the enemy in my turns 2010-10-22T21:19:50 they can render my "moves" useless, so i need a way to add/remove them while generating future states, and it just gets ridiculously messy 2010-10-22T21:20:00 *** bhasker has joined #aichallenge 2010-10-22T21:20:13 i see. I am not doing any tree type searches, so i hasn't mattered to me. 2010-10-22T21:20:21 it will :P 2010-10-22T21:20:22 _ it hasn't _ 2010-10-22T21:20:24 true. 2010-10-22T21:20:39 i am convinced now that minimax will work quite well 2010-10-22T21:20:53 if one sets it up properly 2010-10-22T21:21:06 *** yasith has quit IRC (Ping timeout: 240 seconds) 2010-10-22T21:21:14 will be fun to see. 2010-10-22T21:21:30 my bot atm has so many bugs i can't believe it works at all 2010-10-22T21:22:03 yeah.. got to go back to work on mine. later. 2010-10-22T21:24:40 antimatroid: cool! you must be way ahead of me since my tree bot sucks big time 2010-10-22T21:24:42 i wasted my entire swot vac week, haven't done a bit of study for exams yet and my first one is tuesday 2010-10-22T21:24:51 rebelxt_: mine sucks atm too 2010-10-22T21:25:09 but i had it better a few days ago when it was way more general in it's thinking 2010-10-22T21:25:23 and now i know of many a bugs which explains it having no fucking idea 2010-10-22T21:25:57 lol 2010-10-22T21:26:25 heh - i'm messing with move generators. maybe there is still some light at the end of the tunnel 2010-10-22T21:29:56 *** rrommel has joined #aichallenge 2010-10-22T21:30:17 *** rommel has joined #aichallenge 2010-10-22T21:30:53 g'day all :) 2010-10-22T21:31:06 afternoon 2010-10-22T21:31:21 good night here ;-) 2010-10-22T21:35:22 *** yasith has joined #aichallenge 2010-10-22T21:35:47 hello 2010-10-22T21:36:07 everybody's workin'? 2010-10-22T21:36:08 :) 2010-10-22T21:36:58 yeah ;-) trying to beat the crap out of this bot 2010-10-22T21:37:11 this game is drugs lol 2010-10-22T21:37:34 it's worse than drugs, imho 2010-10-22T21:38:00 james cameron once said in a blog post he does his best to make it more addictive than crack 2010-10-22T21:38:22 antimatroid: james cameron made avatar 2010-10-22T21:38:29 jeff cameron made this contest 2010-10-22T21:38:33 oh aha 2010-10-22T21:38:35 jeff * :P 2010-10-22T21:38:52 happened the same to me yesterday. I added 15 characters in my source code, and my bot went from 400 to 100. I was under drug effect but it helped me a lot ! 2010-10-22T21:39:04 since this challenge is based on Galcon, i'm actually curious if it was Phil Hassey's original idea or based on some really old game 2010-10-22T21:39:07 http://jpcameron.com/blog/?p=83 2010-10-22T21:39:29 rebelxt_: it is based on phil hassey's, he was hanging around for a while 2010-10-22T21:39:34 there were game like this on amstrad if i remember good 2010-10-22T21:39:46 gerald512kk: lol man. this is funny. my bot went from 50 to 10 after changing 3 characters last week 2010-10-22T21:40:12 "This means that if it’s 3:00 AM and you just submitted your code, there is a fairly large incentive to go to bed. This is clearly unacceptable." 2010-10-22T21:40:14 lol 2010-10-22T21:40:54 amstan_: rofl! this is awesome post 2010-10-22T21:41:22 ho my god this guy is crazy 2010-10-22T21:42:04 antimatroid: ahh, yeah, that problem you describe is what made my previous bot attempt useless 2010-10-22T21:42:33 (future move choices getting invalidated by the other player's earlier choices) 2010-10-22T21:42:48 jeff is clearly an awesome dude (too bad his job has him occupied all the time) 2010-10-22T21:44:41 a1k0n: it's completely ruining my bot atm 2010-10-22T21:44:47 pretty sure that is a large part of it anyway 2010-10-22T21:44:59 the payoff function is also dependent on the future states being calculated correctly 2010-10-22T21:47:01 i can't wait for proper srtategy sharing at the end 2010-10-22T21:47:02 somebody would help me in query for a java advice ? ( it is about modifying planet class ) 2010-10-22T21:47:18 move selection is what i'm most interested in, but i'm also least interested in sharing my thoughts on it 2010-10-22T21:50:15 antimatroid: payoff function is tricky: i used to just calculate profit from capturing a planet in terms of number of ships gained within a number of turns. not anymore - there are better ways! 2010-10-22T21:51:21 my payoff function is currently all about expected ships at final time 2010-10-22T21:53:26 antimatroid: i see - my official bot does the same. now i'm trying to think in terms of attacking planet clusters instead of single planets 2010-10-22T21:53:54 i expect minimax to deal with that 2010-10-22T21:54:14 it's great how much i can just lump onto minimax resolving 2010-10-22T21:54:24 well, that's true if you can get more than 1 ply deep ;) 2010-10-22T21:55:02 i am usually getting at least 4-6 layers atm 2010-10-22T21:55:20 i'm not picking huge numbers of turns at each node though 2010-10-22T21:55:29 a minimax bot that knows when to fast forward time = sexy 2010-10-22T21:55:42 i have functions for running nodes atm 2010-10-22T21:56:03 but i'm only running my terminal nodes before getting their payoff rather than strategically picking decision nodes to run 2010-10-22T21:56:08 i shall try to do that at some point 2010-10-22T21:56:37 actually, i have a really good idea for that, thanks 2010-10-22T21:57:28 awesome! whoever finds a way to solve this will probably have a great chance of winning this 2010-10-22T21:58:19 since you prompted that i pmd you the idea ;) 2010-10-22T21:58:55 *** Olathe has quit IRC (Ping timeout: 276 seconds) 2010-10-22T22:01:23 *** McLeopold has joined #aichallenge 2010-10-22T22:03:36 hey McLeopold 2010-10-22T22:03:41 sup 2010-10-22T22:03:41 *** amstan_ is now known as amstan 2010-10-22T22:04:00 i like what ur latest rambot is doing.. i.e. putting pressure 2010-10-22T22:04:17 really? it does badly 2010-10-22T22:04:46 well, at least you are attempting to do the right thing! kinda what all good galcon players do. it's a key strategy in galcon 2010-10-22T22:06:07 of course in galcon you can re-direct fleets at any time so it's a very different game 2010-10-22T22:06:33 I'd like to create maps with 0 growth planets and 0 ships in between growth planets. 2010-10-22T22:06:45 I'd call them "jump" maps. 2010-10-22T22:07:33 McLeopold: I haven't done much with the network stuff, but I have a playgame.py up on bitbucket now. I am a little distracted trying to figure out my bot though. 2010-10-22T22:08:11 I don't like how any of the code looks, but it is running ... 2010-10-22T22:08:26 coo;l 2010-10-22T22:09:12 That means we just need a python version of dhartmei server? 2010-10-22T22:09:23 dmj111: is that python engine impl or tcp server client? (don't have the link) 2010-10-22T22:09:51 rebelxt_: http://bitbucket.org/dmj111/planetwars_test_tools 2010-10-22T22:09:56 both. 2010-10-22T22:10:46 both are equally cruddy at the moment. 2010-10-22T22:11:00 dmj111: sweet 2010-10-22T22:11:17 Top 10 players: Hazard(4127), asavis(4102), rsergio(4091), felixcoto(4039), cfaftw(4038), dmj111(3950), Raschi(3945), bocsimacko(3942), sin_sun(3932), davidjliu(3888) 2010-10-22T22:11:18 *** seisatsu has joined #aichallenge 2010-10-22T22:11:53 i like the disclaimer: "THIS CODE IS PROBABLY BUGGY" 2010-10-22T22:12:08 i's day this is always true 2010-10-22T22:12:16 i'd say, lol 2010-10-22T22:12:24 i get to be captain obvious, again. 2010-10-22T22:12:37 ;-) 2010-10-22T22:12:59 # FIX THIS! 2010-10-22T22:13:00 rebelxt_: User error -- Replace user. 2010-10-22T22:13:02 # FIX WHAT? 2010-10-22T22:13:03 rebelxt_: Run as fast as you can and don't look back. 2010-10-22T22:13:25 contestbot: botdoc 2010-10-22T22:13:26 amstan: botdoc = (#1) http://ubottu.com/stdin/supydocs/plugins/, or (#2) http://supybook.fealdia.org/devel/, or (#3) http://sourceforge.net/apps/mediawiki/gribble/index.php?title=Supybot_Resources 2010-10-22T22:13:46 config reply.whenAddressedBy.chars @ 2010-10-22T22:13:50 #test 2010-10-22T22:13:51 amstan: I do not know about 'test', but I do know about these similar topics: 'tcp' 2010-10-22T22:13:51 @test 2010-10-22T22:13:52 amstan: I do not know about 'test', but I do know about these similar topics: 'tcp' 2010-10-22T22:14:03 wait.. 2010-10-22T22:14:05 gar 2010-10-22T22:14:41 @config reply.whenAddressedBy.chars @ 2010-10-22T22:14:41 amstan: Ready to serve, my lord. 2010-10-22T22:14:45 @test 2010-10-22T22:14:46 amstan: I do not know about 'test', but I do know about these similar topics: 'tcp' 2010-10-22T22:14:48 #test 2010-10-22T22:21:49 amstan: tnx ;-) 2010-10-22T22:24:53 *** rebelxt_ has quit IRC (Quit: Page closed) 2010-10-22T22:31:55 *** Appleman1234 has quit IRC (Ping timeout: 276 seconds) 2010-10-22T22:51:05 *** Appleman1234 has joined #aichallenge 2010-10-22T22:53:36 morning 2010-10-22T23:06:30 *** rommel has quit IRC (Quit: Page closed) 2010-10-22T23:10:14 *** pjreddie has quit IRC (Ping timeout: 265 seconds) 2010-10-22T23:11:17 Top 10 players: Hazard(4115), asavis(4095), rsergio(4082), felixcoto(4032), cfaftw(4029), bocsimacko(3953), dmj111(3942), Raschi(3937), sin_sun(3924), MariusCG(3881) 2010-10-22T23:30:05 *** JCS^ has joined #aichallenge 2010-10-22T23:36:21 *** JCS^ has joined #aichallenge 2010-10-22T23:41:33 *** rebelxt_ has joined #aichallenge 2010-10-22T23:42:00 wow, now that Raschi isn't resubmitting every couple minutes he's shot right up there :) 2010-10-22T23:42:59 *** amriedle has quit IRC (Quit: Lost terminal) 2010-10-22T23:43:10 *** seisatsu has quit IRC (Quit: Leaving) 2010-10-22T23:44:40 maybe all games since max(1 hour, time since last submission) should be kept around for players? 2010-10-22T23:45:57 s/players/bayeselo/ 2010-10-22T23:47:16 and game-choosing should be the same way (people resubmitting all the time still wait an hour or so between games) 2010-10-22T23:49:03 and having 104 pages of submissions is another bug (why would anyone care?) 2010-10-22T23:51:09 *** gerald512kk has quit IRC () 2010-10-22T23:58:28 *** davidjliu has quit IRC (Quit: Page closed)