2011-01-19T00:01:41 ah, amstan posted the xkcd link 2011-01-19T00:17:03 *** delt0r___ has joined #aichallenge 2011-01-19T00:18:32 heh, it still says 1 upvote 2011-01-19T00:18:38 on the other hand, 100% like it 2011-01-19T00:19:22 *** delt0r_ has quit IRC (Ping timeout: 276 seconds) 2011-01-19T00:20:29 I think that's an automatic upvote from me 2011-01-19T00:20:42 yes 2011-01-19T00:23:14 I think I can downvote myself too... 2011-01-19T00:24:21 yes 2011-01-19T00:24:48 works for comments too: downvote self, complain about downvotes, ???, profit 2011-01-19T01:00:13 *** Mathnerd314 has quit IRC (Quit: ChatZilla 0.9.86-rdmsoft [XULRunner 1.9.2.12/20101026210630]) 2011-01-19T01:14:19 *** chris__0076 has joined #aichallenge 2011-01-19T01:16:56 *** chris_0076 has quit IRC (Ping timeout: 240 seconds) 2011-01-19T01:24:40 *** antimatroid has joined #aichallenge 2011-01-19T01:25:42 *** antimatroid1 has quit IRC (Ping timeout: 246 seconds) 2011-01-19T01:33:33 *** jbroman has quit IRC (*.net *.split) 2011-01-19T01:33:33 *** dmj111 has quit IRC (*.net *.split) 2011-01-19T01:33:35 *** amstan has quit IRC (*.net *.split) 2011-01-19T01:33:35 *** McLeopold has quit IRC (*.net *.split) 2011-01-19T01:33:36 *** politik has quit IRC (*.net *.split) 2011-01-19T01:33:38 *** delt0r___ has quit IRC (*.net *.split) 2011-01-19T01:33:39 *** rabidus has quit IRC (*.net *.split) 2011-01-19T01:33:39 *** smellyhippy has quit IRC (*.net *.split) 2011-01-19T01:33:40 *** javagamer has quit IRC (*.net *.split) 2011-01-19T01:33:54 *** rabidus has joined #aichallenge 2011-01-19T01:39:09 *** smellyhippy has joined #aichallenge 2011-01-19T01:39:09 *** smellyhippy has joined #aichallenge 2011-01-19T01:40:38 *** javagamer has joined #aichallenge 2011-01-19T01:43:53 *** sigh has quit IRC (Ping timeout: 240 seconds) 2011-01-19T01:53:55 *** antimatroid has quit IRC (Read error: Connection reset by peer) 2011-01-19T02:20:45 *** Palmik has joined #aichallenge 2011-01-19T02:37:54 *** sigh has joined #aichallenge 2011-01-19T02:53:09 *** chris__0076 is now known as chris_0076 2011-01-19T03:14:11 *** Palmik has quit IRC (Remote host closed the connection) 2011-01-19T03:27:51 *** mega1 has joined #aichallenge 2011-01-19T03:29:03 *** aerique has joined #aichallenge 2011-01-19T03:35:41 *** Frontier has joined #aichallenge 2011-01-19T03:35:41 *** McLeopold has joined #aichallenge 2011-01-19T03:35:41 *** antimatroid has joined #aichallenge 2011-01-19T03:35:41 *** delt0r___ has joined #aichallenge 2011-01-19T03:35:41 *** amstan has joined #aichallenge 2011-01-19T03:35:41 *** dmj111 has joined #aichallenge 2011-01-19T03:35:41 *** jbroman has joined #aichallenge 2011-01-19T03:35:41 *** politik has joined #aichallenge 2011-01-19T03:35:41 *** card.freenode.net sets mode: +o amstan 2011-01-19T03:38:54 *** boegel has joined #aichallenge 2011-01-19T03:51:58 *** dr`away is now known as weechat_user 2011-01-19T04:03:41 *** weechat_user is now known as dr- 2011-01-19T04:13:22 *** Naktibalda has joined #aichallenge 2011-01-19T04:37:59 *** mega1 has quit IRC (Ping timeout: 265 seconds) 2011-01-19T04:40:03 *** mega1 has joined #aichallenge 2011-01-19T04:49:18 *** Anthony__Dune has joined #aichallenge 2011-01-19T05:01:05 *** Kingpin13 has joined #aichallenge 2011-01-19T05:01:24 The decision making for the next contest needs a deadline. I suggest to set a date for game ideas to be 'submitted' that are proven to be technically feasible. 2011-01-19T05:01:43 Those can be voted for on the forums afterwards. 2011-01-19T05:03:51 *** Meatkat has joined #aichallenge 2011-01-19T05:16:07 i somewhat agree with frontier 2011-01-19T05:16:14 Frontier: did you catch the 2 player game of life idea 2011-01-19T05:16:23 i'm unsure about it, it seems awesome 2011-01-19T05:17:53 lol 2011-01-19T05:18:23 that is at least challenging, but little hard to start with 2011-01-19T05:18:49 i would think a starter strategy would be to place a glider in their region each turn 2011-01-19T05:19:02 and send it off to battle 2011-01-19T05:19:07 antimatroid: No I haven't seen it. 2011-01-19T05:20:48 Frontier: basically, start with a grid separated into regions where you can fill up to (n + previously unused) blank squares, then iterate using game of life rules (for creating a new live cell, you need 3 adjacent squares,so there is no ties) 2011-01-19T05:21:07 whoever contains has the most squares after m turns filled in wins 2011-01-19T05:21:15 - contains 2011-01-19T05:22:12 sigh: ^^ what do you think? 2011-01-19T05:22:43 i'm worried about the fact that you basically always need to play out all m rounds then 2011-01-19T05:22:57 game of life is complicated and well studied 2011-01-19T05:23:14 is multiplayer game of life studeid extensively? 2011-01-19T05:23:57 dunno, but either way I think that a lot of the good formations will be just copied from what is already known 2011-01-19T05:24:19 unless multiplayer is sufficiently different 2011-01-19T05:25:32 i would expect people to, but if you could only place 5 things per turn for example, it would be pretty hard to form a glider gun 2011-01-19T05:26:31 and you also have the problem of how to place your final squares (with all reserves) so that you maximise your filled squares the next turn 2011-01-19T05:27:05 i would prefer something where it's possible to actually kil them 2011-01-19T05:27:34 my other idea was to allow people to place their squares at the start, then on following moves they can walk their squares to adjacent blank squares, reverting collisions 2011-01-19T05:27:50 that would allow shorter games, but maybe not be as game of lifey 2011-01-19T05:28:58 Frontier: like/hate? 2011-01-19T05:29:30 hmm... I still can't see it being that interesting 2011-01-19T05:29:42 I just read the rules for game of life. I'm still undecided. It is easy to implement so it is a sure candidate at lease :) 2011-01-19T05:29:47 *least 2011-01-19T05:29:57 I can imagine a lot of offline processing finding patterns, if the existing ones don't work so well 2011-01-19T05:29:58 Frontier: what cs world have you been living in? 2011-01-19T05:30:09 so it comes down to who hard codes the best patterns 2011-01-19T05:30:27 I have always been living in my very own world ^^ 2011-01-19T05:31:01 sigh: that's why i'm unsure about it 2011-01-19T05:31:04 sigh: And this is something I cannot comment on so... I remain undecided. 2011-01-19T05:31:12 it was just something that came to mind earlier from a comment rebelxt made 2011-01-19T05:31:29 what he say? 2011-01-19T05:31:45 something about the simplified ants reminding him of game of life 2011-01-19T05:32:06 i think I remember that 2011-01-19T05:32:38 I think someone disagreed and I agree with that disagreement 2011-01-19T05:32:50 :) 2011-01-19T06:00:34 *** rabidus has quit IRC (Ping timeout: 240 seconds) 2011-01-19T06:12:45 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-01-19T06:18:31 *** Anthony__Dune has quit IRC (Ping timeout: 255 seconds) 2011-01-19T06:21:23 *** rabidus has joined #aichallenge 2011-01-19T06:40:08 *** dr- is now known as dr`away 2011-01-19T07:01:39 jmcarthur: I just read a little of your talk with McLeopold. In the offical jumbo rules scouts can move and attack in the same move. In the classic version they could not 2011-01-19T07:02:05 as for spys/marshal or scout movement... i am not really fussed either way 2011-01-19T07:04:38 *** Accoun has quit IRC () 2011-01-19T07:08:40 *** asdffsa has joined #aichallenge 2011-01-19T07:13:45 *** sigh has quit IRC (Remote host closed the connection) 2011-01-19T07:14:20 *** UncleVasya has joined #aichallenge 2011-01-19T07:18:40 *** rabidus has quit IRC (Ping timeout: 246 seconds) 2011-01-19T07:23:39 Why do scouts always die on their missions? I would not want to be a scout in this game :( 2011-01-19T07:26:32 *** |UncleVasya| <|UncleVasya|!~UncleVasy@193.239.129.233> has joined #aichallenge 2011-01-19T07:30:17 *** UncleVasya has quit IRC (Ping timeout: 272 seconds) 2011-01-19T07:31:40 *** |UncleVasya| <|UncleVasya|!~UncleVasy@193.239.129.233> has quit IRC (Quit: Changing server...) 2011-01-19T07:34:30 *** Palmik has joined #aichallenge 2011-01-19T07:35:30 *** UncleVasya has joined #aichallenge 2011-01-19T07:48:00 *** boegel has quit IRC (Quit: *poof!*) 2011-01-19T07:48:10 *** Accoun has joined #aichallenge 2011-01-19T07:55:00 *** dr`away is now known as weechat_user 2011-01-19T07:55:16 *** weechat_user is now known as dr- 2011-01-19T08:17:31 *** tapwater has joined #aichallenge 2011-01-19T08:17:40 *** delt0r_ has joined #aichallenge 2011-01-19T08:19:47 *** delt0r___ has quit IRC (Ping timeout: 276 seconds) 2011-01-19T08:21:16 *** Anthony__Dune has joined #aichallenge 2011-01-19T08:41:02 *** Mathnerd314 has joined #aichallenge 2011-01-19T10:24:40 *** dr- is now known as dr`away 2011-01-19T10:29:54 *** UncleVasya has quit IRC (Ping timeout: 265 seconds) 2011-01-19T10:30:03 i have seen multiplayer games of life before. i thought they were boring, personally 2011-01-19T10:30:26 delt0r_: ah okay, i'm only familiar with the classic version 2011-01-19T10:46:18 *** |UncleVasya| <|UncleVasya|!~UncleVasy@193.239.129.216> has joined #aichallenge 2011-01-19T10:50:15 *** delt0r has joined #aichallenge 2011-01-19T11:26:46 http://ai-contest.com/forums/viewtopic.php?f=20&t=1258, help! 2011-01-19T11:31:33 *** dr`away is now known as dr- 2011-01-19T11:36:07 *** aerique has quit IRC (Quit: ...) 2011-01-19T11:39:02 bot wins! :) 2011-01-19T11:47:09 *** Anthony__Dune has quit IRC (Quit: Quitte) 2011-01-19T12:09:45 amstan_i, most of the spam on the other forums i am on are from humans 2011-01-19T12:09:54 ie people paid to post spam 2011-01-19T12:10:07 javagamming.org 2011-01-19T12:10:19 they now as a programing question 2011-01-19T12:10:26 this has worked.. 2011-01-19T12:10:40 but this makes it hard for non programmers to join 2011-01-19T12:11:08 but in seriousness... 3 per day is not so bad... 2011-01-19T12:11:30 oh and they stopped permiting links in their first few posts.. 2011-01-19T12:12:21 or first few days... posts with links are rejected and emailed back to them, and they are asked to repost without links 2011-01-19T12:12:51 since spam doods don't want' to wait... this is pretty effective ... so far 2011-01-19T12:16:14 delt0r: oh.. good idea! 2011-01-19T12:16:24 i can do that too 2011-01-19T12:17:11 ugh.. nvm, i can't 2011-01-19T12:17:15 what is this, i don't even 2011-01-19T12:23:56 well our guy reprogrammed the forum code to do it... 2011-01-19T12:28:19 *** _UncleVasya <_UncleVasya!~UncleVasy@193.239.129.236> has joined #aichallenge 2011-01-19T12:28:23 so whats a "python exceptions" for bot errors about? 2011-01-19T12:28:29 amstan_i:^ 2011-01-19T12:31:11 *** rabidus has joined #aichallenge 2011-01-19T12:31:44 *** |UncleVasya| <|UncleVasya|!~UncleVasy@193.239.129.216> has quit IRC (Ping timeout: 240 seconds) 2011-01-19T12:54:49 *** _UncleVasya <_UncleVasya!~UncleVasy@193.239.129.236> has quit IRC () 2011-01-19T13:12:56 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2011-01-19T13:18:51 amstan_i: i've also been deleting spam without saying anything, so the problem may be even worse than you have personally observed 2011-01-19T13:19:11 jmcarthur: it's horrible, yeah 2011-01-19T13:26:04 amstan_i: Aside from fighting spam, I really think you should set a deadline for new game ideas for next term. We can then sort out what is technically feasible and post the candidates on the forum where everyone can cast a vote on their favourite idea. 2011-01-19T13:26:52 An assignment to tasks would be useful as well. I.e.: Ant game idea: visualisation - Frontier 2011-01-19T13:27:22 But that is another story 2011-01-19T13:30:15 I think that before putting anything up for "vote" the organizers have to be happy with the list of games... 2011-01-19T13:30:46 after all --its not a democracy...and at the end of the day, they are the ones that foot the bill in terms of work 2011-01-19T13:34:18 delt0r: I wish it was a democracy, but if someone is unhappy with the organization they will probably move away to other competitions. Then again I'm fine with amstan having the last word on the list of games. 2011-01-19T13:36:30 With a timeline we could estimate how much time we can put into working on the contest. As I read it there could be quite some community effort in the prerun, like what has been done during the contest last time (bot tools with human interaction for example). 2011-01-19T13:37:34 But this time work might go into backend components as well. Someone from the community might write the game server code. 2011-01-19T13:46:43 If you need me I'm in the pub: http://www.irishpub-hannover.de/irishpub-hannover/home.html 2011-01-19T13:51:58 unfortunately, i think j3camero gets the last word 2011-01-19T13:52:11 thats really what i mean 2011-01-19T13:52:16 i find that i disagree with j3camero quite often 2011-01-19T13:52:18 if a vote is not listen too 2011-01-19T13:52:21 don't have one 2011-01-19T13:53:13 well, i do think that j3camero would pay attention if his preference is, say, the most hated choice in a straw poll 2011-01-19T13:53:22 but you know... if thats the person responsible for running the comp... then they should have the last word/// 2011-01-19T13:53:39 as long as he's well-informed i'm okay with it 2011-01-19T13:53:47 sometimes i think he is not, though 2011-01-19T13:53:54 he makes decisions without really thinking about it that much 2011-01-19T13:54:30 it's not bad enough to revolt or something though :) 2011-01-19T13:54:52 jmcarthur, I get the feeling that j3camero is just a little out of touch with most of the community ? 2011-01-19T13:55:04 that is probably the main issue 2011-01-19T13:55:54 and whats the role? His he a professor at the uni... a student or what? 2011-01-19T13:56:05 he was a student there 2011-01-19T13:56:14 now he's out i think 2011-01-19T13:56:37 that's my understanding, at least 2011-01-19T13:56:42 right 2011-01-19T13:56:54 so this started as more a student thing 2011-01-19T13:57:20 i'm certainly not going to argue that j3camero doesn't deserve a say. after all, he's the one putting money into it, too 2011-01-19T13:58:13 well the thing is... if one is too out of touch, you end up with a community of one 2011-01-19T14:02:11 yeah, which is what happens after a revolt... which again i think is way beyond what's called for 2011-01-19T14:02:23 *** mceier has joined #aichallenge 2011-01-19T14:02:26 jmcarthur, what is your role? are you a student at the uni or... just a hard worker 2011-01-19T14:02:32 i think he just needs a stern "listen to us" lecture 2011-01-19T14:02:34 the latter 2011-01-19T14:02:41 not even that hard, honestly :) 2011-01-19T14:02:47 lol 2011-01-19T14:03:12 I don't even think he would need that... 2011-01-19T14:03:34 just a "you realize that you are way out of touch here" 2011-01-19T14:03:41 not that i know him 2011-01-19T14:04:12 just most of the time... these things are more about a lack of communication... in my experience 2011-01-19T14:04:42 or miss communication 2011-01-19T14:18:18 i believe it has been mentioned to him a couple times already 2011-01-19T14:22:11 *** asdffsa has quit IRC (*.net *.split) 2011-01-19T14:22:14 *** jbroman has quit IRC (*.net *.split) 2011-01-19T14:22:15 *** Accoun has quit IRC (*.net *.split) 2011-01-19T14:22:15 *** Palmik has quit IRC (*.net *.split) 2011-01-19T14:22:15 *** Frontier has quit IRC (*.net *.split) 2011-01-19T14:22:15 *** dmj111 has quit IRC (*.net *.split) 2011-01-19T14:22:17 *** antimatroid has quit IRC (*.net *.split) 2011-01-19T14:22:17 *** politik has quit IRC (*.net *.split) 2011-01-19T14:22:18 *** mceier has quit IRC (*.net *.split) 2011-01-19T14:22:19 *** mega1 has quit IRC (*.net *.split) 2011-01-19T14:22:19 *** McLeopold has quit IRC (*.net *.split) 2011-01-19T14:22:20 *** Mathnerd314 has quit IRC (*.net *.split) 2011-01-19T14:22:21 *** AlliedEnvy has quit IRC (*.net *.split) 2011-01-19T14:22:21 *** Error323 has quit IRC (*.net *.split) 2011-01-19T14:22:22 *** jmpespxoreax has quit IRC (*.net *.split) 2011-01-19T14:22:22 *** BtbN has quit IRC (*.net *.split) 2011-01-19T14:22:23 *** davidd has quit IRC (*.net *.split) 2011-01-19T14:22:24 *** Zannick has quit IRC (*.net *.split) 2011-01-19T14:22:24 *** ssalbiz has quit IRC (*.net *.split) 2011-01-19T14:22:25 *** pgpaskar_ has quit IRC (*.net *.split) 2011-01-19T14:26:24 *** mceier has joined #aichallenge 2011-01-19T14:26:24 *** Mathnerd314 has joined #aichallenge 2011-01-19T14:26:24 *** Accoun has joined #aichallenge 2011-01-19T14:26:24 *** Palmik has joined #aichallenge 2011-01-19T14:26:24 *** asdffsa has joined #aichallenge 2011-01-19T14:26:24 *** mega1 has joined #aichallenge 2011-01-19T14:26:24 *** politik has joined #aichallenge 2011-01-19T14:26:24 *** jbroman has joined #aichallenge 2011-01-19T14:26:24 *** dmj111 has joined #aichallenge 2011-01-19T14:26:24 *** antimatroid has joined #aichallenge 2011-01-19T14:26:24 *** McLeopold has joined #aichallenge 2011-01-19T14:26:24 *** Frontier has joined #aichallenge 2011-01-19T14:26:24 *** Zannick has joined #aichallenge 2011-01-19T14:26:24 *** AlliedEnvy has joined #aichallenge 2011-01-19T14:26:24 *** Error323 has joined #aichallenge 2011-01-19T14:26:24 *** ssalbiz has joined #aichallenge 2011-01-19T14:26:24 *** jmpespxoreax has joined #aichallenge 2011-01-19T14:26:24 *** davidd has joined #aichallenge 2011-01-19T14:26:24 *** BtbN has joined #aichallenge 2011-01-19T14:26:24 *** pgpaskar_ has joined #aichallenge 2011-01-19T14:29:55 Frontier, delt0r, jmcarthur: I don't have the last word, Jeff does 2011-01-19T14:31:08 and i don't think Jeff would do a vote on this 2011-01-19T14:31:14 like.. he would be fine with it 2011-01-19T14:31:22 but he would not use the result of the vote as the game 2011-01-19T14:31:52 he's thinking is more like "what the comunity wants is not necessarly what's best" 2011-01-19T14:34:32 yeah 2011-01-19T14:38:28 j3camero used to be a student at waterloo, like me 2011-01-19T14:38:31 but he finished 2011-01-19T14:38:43 now he works in downtown toronto in high frequency trading 2011-01-19T14:38:49 i'm doing my coop term in toronto as well 2011-01-19T14:39:12 yes, i do believe j3camero is out of the loop 2011-01-19T14:39:18 but like.. 2011-01-19T14:39:27 his loop is the main one.. 2011-01-19T14:39:45 so this channel/forums is the one out of the loop 2011-01-19T14:39:58 he should be online more 2011-01-19T14:40:00 sharing more ideas 2011-01-19T14:40:05 i tried to make him do this 2011-01-19T14:40:07 but it's hard 2011-01-19T14:40:12 i'll accept any ideas that you guys have 2011-01-19T14:40:26 ugh.. 2011-01-19T14:40:32 i should probably post this on the forums 2011-01-19T14:40:46 jmcarthur, delt0r, Frontier: read ^^^ 2011-01-19T14:41:03 in terms of tasks.. 2011-01-19T14:41:07 there aren't many 2011-01-19T14:41:12 we have to decide on a game 2011-01-19T14:41:27 one of the things that is important right now is a better rankings algo 2011-01-19T14:41:34 that can handle more than 1v1 games 2011-01-19T14:41:57 and i already posted that on the forums 2011-01-19T14:43:23 interesting that j3camero is working on HFT. bocsimacko also seems to be working in stock market research field 2011-01-19T14:44:02 a headhunter just contacted me about the financial sector today, too. i'm not convinced i would enjoy it yet 2011-01-19T14:44:58 i already interviewed for a job in finance but wasn't all that enthused about it 2011-01-19T14:51:16 i'm especially not sure i'd want to move to singapore for the job he thinks would fit me best 2011-01-19T14:51:48 singapore, lol 2011-01-19T14:53:33 supposed to be fairly nice, actually. good food, english is the main language, they have a lot of gov. programs to encourage foreigners to move there, and the pay would probably be insanely high compared to the typical cost of living there 2011-01-19T14:54:13 trying to understand singlish would be entertaining for only about a day, though 2011-01-19T14:55:28 and it gets hot and humid 2011-01-19T14:55:33 even in the winter 2011-01-19T14:56:13 and i'm not a huge fan of very dense cities 2011-01-19T15:03:49 *** rabidus has quit IRC (Ping timeout: 272 seconds) 2011-01-19T15:04:23 jmcarthur: yeah.. dense cities.. are weird 2011-01-19T15:05:54 when i'm in a dense city i feel like i just can't get a moment to breathe 2011-01-19T15:14:46 *** mceier has quit IRC (Quit: leaving) 2011-01-19T15:14:48 yep 2011-01-19T15:14:54 the public transit is always full 2011-01-19T15:15:01 pedestrians are racing while walking 2011-01-19T15:15:10 everyone is rude 2011-01-19T15:15:26 cars are useless 2011-01-19T15:15:38 but there are good things 2011-01-19T15:15:46 interesting job opportunities 2011-01-19T15:15:50 wifi everywhere 2011-01-19T15:16:01 and everything is within walking distance 2011-01-19T15:28:05 *** Accoun has quit IRC () 2011-01-19T15:29:05 *** bhasker has joined #aichallenge 2011-01-19T15:29:12 * pgpaskar_ explodes 2011-01-19T15:32:21 I am in a "dense" city... for Europe... 2011-01-19T15:32:39 but not compared to asia or south america 2011-01-19T15:32:56 Vienna... its doesn't feel like a city... 2011-01-19T15:33:08 but you get all the benefits... 2011-01-19T15:33:23 24 hour public transport thats real cheap... etc 2011-01-19T15:33:54 but not supper busy like london 2011-01-19T15:34:15 or worse Tokyo --Tokyo is insane 2011-01-19T15:36:59 * antimatroid can't stand big cities 2011-01-19T15:37:06 i am extremely averse to queues, they make me rage 2011-01-19T15:37:15 even a 3 minute red light will have me wanting to end humanity 2011-01-19T15:38:20 *** McLeopold has left #aichallenge 2011-01-19T15:38:48 *** McLeopold has joined #aichallenge 2011-01-19T15:39:38 delt0r_: you say it doesn't feel like a city, but your city is over four times as dense as mine 2011-01-19T15:41:51 http://en.wikipedia.org/wiki/Huntsville,_Alabama 2011-01-19T15:42:49 jmcarthur: i know 2011-01-19T15:42:55 but its a vienna thing 2011-01-19T15:43:04 even for europe 2011-01-19T15:43:10 its a excpetion to the rule 2011-01-19T15:43:44 Munchic is less dense, but feels much more rushed... you cram into u bahn etc 2011-01-19T15:43:53 Munchen 2011-01-19T15:44:11 *** antimatroid has quit IRC (Ping timeout: 240 seconds) 2011-01-19T15:50:35 *** Accoun has joined #aichallenge 2011-01-19T15:52:27 *** sigh has joined #aichallenge 2011-01-19T15:59:35 *** bhasker has quit IRC (Ping timeout: 265 seconds) 2011-01-19T16:03:45 *** rebelxt has left #aichallenge 2011-01-19T16:05:11 *** sigh has quit IRC (Remote host closed the connection) 2011-01-19T16:12:13 *** UndiFineD has quit IRC (Ping timeout: 272 seconds) 2011-01-19T16:17:01 *** delt0r___ has joined #aichallenge 2011-01-19T16:17:17 *** UndiFineD has joined #aichallenge 2011-01-19T16:19:50 *** delt0r_ has quit IRC (Ping timeout: 272 seconds) 2011-01-19T16:35:32 *** mega1 has quit IRC (Read error: Connection reset by peer) 2011-01-19T16:35:45 *** mega1 has joined #aichallenge 2011-01-19T16:41:48 *** Palmik has quit IRC (Remote host closed the connection) 2011-01-19T16:45:35 *** choas has joined #aichallenge 2011-01-19T16:49:28 amstan_i: I'm drunk now ... and I read. You are not in control. ok. Jeff is, but he is busy creating the next finance crisis. ok. He should really show up more often, I really thought you were responsible amstan, just because you seem to be an insider and you are on the irc channel often. 2011-01-19T16:51:24 A propos dense cities, I live in Hannover's central and it is really quiet here. Mainly because it is the historic part of the city and a lot of areas can't be reached with a car. 2011-01-19T16:52:49 ...and we've got this amazing city hall here ^^: http://www.heise.de/foto/galerie/Rathaus-Hannover-a88e40c30a9d2fdd135b073175a5ab40/ 2011-01-19T16:52:49 It is really huge and impressive when you drive by. 2011-01-19T16:53:54 Oh shit this green field to the right makes me really sentimental now. I laid there with a girl friend. 2011-01-19T17:01:16 *** bhasker has joined #aichallenge 2011-01-19T17:02:47 i ought to post to the forums the game of life idea we had here yesterday 2011-01-19T17:04:03 *** sigh has joined #aichallenge 2011-01-19T17:14:29 Zannick: what time did that conversation take place? 2011-01-19T17:14:57 21 hours ago? 2011-01-19T17:15:12 *** dr- is now known as dr`away 2011-01-19T17:16:13 now I have to do timezone math :( 2011-01-19T17:17:41 hmm... I don't like the game of life idea 2011-01-19T17:17:51 It's just one big alpha-beta tree 2011-01-19T17:18:16 add 3 hours, subtract a day. 2011-01-19T17:18:19 Plus, have you ever tried tweaking the rules for the normal game of life? 2011-01-19T17:18:23 i didn't give a time on purpose. :P 2011-01-19T17:18:35 i have, yes. they do wild things 2011-01-19T17:18:42 It only really works well 1 way. 2011-01-19T17:19:54 yeah 2011-01-19T17:24:21 It's interesting, but I think fails the fun requirement. 2011-01-19T17:36:22 fun is hard to measure 2011-01-19T17:36:42 i implemented the game of life twice in high school 2011-01-19T17:36:56 so i'd really enjoy it despite being an admin 2011-01-19T17:38:24 *** Zannick has quit IRC (*.net *.split) 2011-01-19T17:38:25 *** ssalbiz has quit IRC (*.net *.split) 2011-01-19T17:38:25 *** pgpaskar_ has quit IRC (*.net *.split) 2011-01-19T17:41:00 *** Zannick has joined #aichallenge 2011-01-19T17:41:00 *** ssalbiz has joined #aichallenge 2011-01-19T17:41:00 *** pgpaskar_ has joined #aichallenge 2011-01-19T17:44:43 *** mega1 has quit IRC (Ping timeout: 276 seconds) 2011-01-19T17:48:10 *** Frontier has left #aichallenge 2011-01-19T17:50:57 *** Mathnerd314 has quit IRC (Ping timeout: 260 seconds) 2011-01-19T17:51:55 *** antimatroid has joined #aichallenge 2011-01-19T17:52:08 *** choas has quit IRC (Quit: leaving) 2011-01-19T17:55:11 *** bhasker has quit IRC (Quit: bhasker) 2011-01-19T17:55:24 *** Migi32 has joined #aichallenge 2011-01-19T17:57:55 *** Accoun has quit IRC (Ping timeout: 264 seconds) 2011-01-19T18:28:31 *** Cold-Phoenix has joined #aichallenge 2011-01-19T18:32:57 the game of life is fun to implement, but it's not such a fun game to play, IMO 2011-01-19T18:34:47 jmcarthur: it is, but i have no idea if 2 player would be fun or not 2011-01-19T18:34:53 it really was just an idea that came to min 2011-01-19T18:35:07 i'm not really liking any ideas atm, so i don't mind what gets chosen 2011-01-19T18:41:08 *** Kingpin13 has quit IRC (Quit: You're a kitty!) 2011-01-19T18:41:43 I'm kind of down on the ant ideas. 2011-01-19T18:42:11 The original idea was to be 1 process per ant, which I think would have been fun. 2011-01-19T18:52:39 1 process per ant has too much overhea 2011-01-19T18:52:44 it just wont work 2011-01-19T18:52:54 well would, but the number of games played would go down a lot 2011-01-19T18:53:06 and tcp games would be really painful 2011-01-19T18:53:23 *** BtbN has quit IRC (Ping timeout: 260 seconds) 2011-01-19T18:53:58 *** AlliedEnvy has quit IRC (Ping timeout: 260 seconds) 2011-01-19T18:55:12 *** BtbN has joined #aichallenge 2011-01-19T18:57:34 tcp games would be painful. 2011-01-19T18:57:42 I hadn't thought of that. 2011-01-19T18:58:00 But turning it into a single process game takes all the fun out of it for me. 2011-01-19T18:59:38 antimatroid: what if it were 50ms per move per process with a population limit? 2011-01-19T18:59:54 of 20, so a total of 1 second per side. 2011-01-19T19:00:00 *** Cold-Phoenix has quit IRC () 2011-01-19T19:00:15 You'd have to change the win condition to objectives rather than population. 2011-01-19T19:01:08 The engine could just spin up 20 processes for each side, then only communicate with them when needed. 2011-01-19T19:01:50 the only way i can think of for 1 process per ant to work would be to have a special programming language just for the ants 2011-01-19T19:02:00 to work *well* 2011-01-19T19:02:13 with reliable, fair resource distribution, etc. 2011-01-19T19:02:33 It would be easier. 2011-01-19T19:03:17 personally i think it would be cool if the AI's job was to "generate" programs for the ants based on some other inputs throughout the game 2011-01-19T19:03:32 but that might be above the skill level of some of our target audience 2011-01-19T19:03:54 i meant for "generate" to be in asterisks, not quotation marks 2011-01-19T19:04:02 The problem that sounded fun to solve in the original idea was ant communication. 2011-01-19T19:04:11 ah 2011-01-19T19:04:29 well, that would also be doable with the program generation idea 2011-01-19T19:04:56 So, kind of like what that one PlanetWars team did with php? 2011-01-19T19:05:02 not quite 2011-01-19T19:05:14 so here's the idea in full: 2011-01-19T19:05:46 *** AlliedEnvy has joined #aichallenge 2011-01-19T19:05:53 *** Blkt has joined #aichallenge 2011-01-19T19:06:38 *** bhasker has joined #aichallenge 2011-01-19T19:06:55 each ant has a limited amount of computational resources to run a program of your choosing. the limitations are tight enough that you would probably be unable to hard code an effective strategy into them before the game. as ants spawn, it's the bot's job to assign a program to each new ant, after which the ants are autonomous 2011-01-19T19:07:46 the php bot i believe was just generated before uploading, statically, so it's not really the same at all 2011-01-19T19:07:56 and this isn't necessarily something you would want to use genetic programming with 2011-01-19T19:08:02 probably not, in fact 2011-01-19T19:08:14 So, I could create a few ant types, then control which ones are spawned? 2011-01-19T19:08:25 that would be a coarse-grained strategy, sure 2011-01-19T19:08:50 Or modify the ant types as the game progresses, then try them out? 2011-01-19T19:08:52 a more fine-grained strategy might be instructing new ants to go to particular areas of the board, etc. based on how the game is going 2011-01-19T19:08:55 right 2011-01-19T19:09:04 Kind of like evolving ant in the game. 2011-01-19T19:09:11 it's kind of like rts, but you only get to give one, complex order to each unit 2011-01-19T19:09:21 rather than being able to change simple orders over time 2011-01-19T19:09:34 evolving the ants is probably not the best strategy 2011-01-19T19:09:44 the game would be too short 2011-01-19T19:09:54 Then a language would need to be defined for the ants. 2011-01-19T19:10:25 probably the best strategy would be to hardcode a set of parameterizable, composable programs into your bot and choose which ones to use 2011-01-19T19:10:29 yes 2011-01-19T19:11:06 this would probably be, at best, infeasible for february 2011-01-19T19:11:08 *** bhasker has quit IRC (Ping timeout: 264 seconds) 2011-01-19T19:11:13 totaly 2011-01-19T19:11:23 and at worst, not feasible for newbies at all 2011-01-19T19:11:46 that would allow for separation of ants, but still have just one process per player 2011-01-19T19:11:53 right 2011-01-19T19:12:37 McLeopold: the problem with giving them almost no time to counter the additional overhead is it just detracts from the degree of sophistication strategies can be 2011-01-19T19:12:55 and really the idea of these contests are for interesting strategies/ideas to come out yeah? 2011-01-19T19:13:09 for extra difficulty it could be such that you only get input about the game if you have ants come back to the anthill to report to the AI 2011-01-19T19:13:12 well, yes, but I don't think it limits it too much 2011-01-19T19:13:23 planet wars was awesome because it was so different, everything considered atm seems to be on a grid 2011-01-19T19:13:25 you could maybe consider the AI to be a queen ant that can program her workers or something 2011-01-19T19:13:39 i would be down for a game on randomly generated graphs 2011-01-19T19:13:42 but not sure what game to use 2011-01-19T19:13:53 antimatroid: like planet wars? :P 2011-01-19T19:14:05 that was completely connected with full information 2011-01-19T19:14:14 ah 2011-01-19T19:14:16 jmcarthur: let's shoehorn this idea into planet wars. 2011-01-19T19:14:20 you could do an ants game on a graph without full information, then you would have sink holes etc. 2011-01-19T19:14:25 Just to think about it. 2011-01-19T19:14:33 and you wouldn't have it fully connected, but people thought that was too much like planet wars 2011-01-19T19:14:35 into planet wars? 2011-01-19T19:15:16 The AI could create a fleet give it a set of orders, that could loop and it would follow indefinatly, until it dies. 2011-01-19T19:15:30 But control of the fleet never returns to the ai 2011-01-19T19:15:35 what options are there in 2d? just grids, graphs and continous? 2011-01-19T19:15:50 antimatroid: graphs are the only option :P 2011-01-19T19:15:58 :P 2011-01-19T19:16:05 even continuous is discrete due to the constraints of a reliable simulation 2011-01-19T19:16:10 yeah 2011-01-19T19:16:24 okay, well i think using a non-grid graph would be cool 2011-01-19T19:16:27 well, unless we do everything with inefficient rational numbers or computable reals 2011-01-19T19:16:47 Or, the AI give a planet an order to support another planet, and then all growth continuosly goes to that planet and control is not returned. 2011-01-19T19:17:21 So basically, you can give 1 complex order to an ant and you can't change it. 2011-01-19T19:17:28 it would be fun to go through and class games as to similarity of the underlying mechanics 2011-01-19T19:17:40 *** bhasker has joined #aichallenge 2011-01-19T19:17:41 essentially yeah 2011-01-19T19:18:06 that sounds much more fun 2011-01-19T19:18:55 maybe you could change it if the ant comes back home at some point in the future 2011-01-19T19:19:08 essentially as though the ant died and you get the ability to spawn a new one 2011-01-19T19:19:20 after throwing out single process, i stil like my ats variation best 2011-01-19T19:19:39 well, this is kind of single process 2011-01-19T19:19:43 it had lots of variation, was easy to implement the engine and strategies, easy to iterate states etc. 2011-01-19T19:19:51 it's just that the simulation itself has many processes within it 2011-01-19T19:20:43 so, a default ending order could be "return to hill" 2011-01-19T19:20:52 and you had a few objectives to fill (explore map, harvest to increase army, defend base, invade theirs) 2011-01-19T19:20:54 sure 2011-01-19T19:21:17 of course, we would have to pick the level of abstraction to provide with this internal programming language 2011-01-19T19:21:29 "return to hill" is probably itself too high level 2011-01-19T19:21:48 why not just allow each ant to share a string with a max size 2011-01-19T19:21:50 yeah, that would be difficult. every order would need to be implemented in the engine, and everyone would want more orders. 2011-01-19T19:21:55 if it's too long that's invalid 2011-01-19T19:22:19 "share a string"? you mean all ants have the same program? 2011-01-19T19:22:23 that was the older idea 2011-01-19T19:22:32 you could have a key/value store for each and and give basic operations to use it 2011-01-19T19:23:04 i don't see anything wrong with a simple string representation (and the engine could use an AST or something) 2011-01-19T19:23:14 what's the objective of the game 2011-01-19T19:23:17 I could see people wanting to define specific pathfinding 2011-01-19T19:23:22 yeah 2011-01-19T19:23:32 antimatroid: kill the other anthill? 2011-01-19T19:23:36 dunno yet 2011-01-19T19:24:05 but killing the other anthill would make some sense because that's "where" the bots actually are 2011-01-19T19:24:13 It needs to be objectives, not population, otherwise it's just PlanetWars. 2011-01-19T19:24:35 you can use relative coordinates in a single process to try and make the ants not know which is where/which, but that's incompatible with ant hills / home bases 2011-01-19T19:24:57 i'm a big fan of the spawn base with a win condition being invading that 2011-01-19T19:25:08 i also would prefer attacking to be an actual required move 2011-01-19T19:25:19 yeah i'm thinking something simple like you just destroy the opposing anthill and you win, and if neither is destroyed or both are destroyed at once it's just a tie 2011-01-19T19:25:29 you are all probably wanting collisions reverted? i liked fusion / forced fights to death 2011-01-19T19:25:47 I prefer control points that earn you 1 point per turn. 2011-01-19T19:26:04 what about 1 point per turn for each ant you own? 2011-01-19T19:26:10 didn't you say that? i liked that 2011-01-19T19:26:19 No, it can't be population, to close to PlanetWars. 2011-01-19T19:26:21 capture the flag? 2011-01-19T19:26:35 Kinda like CTF 2011-01-19T19:26:41 invade base is pretty much capture the flag, you just don't have to take it anywhere 2011-01-19T19:26:44 i meant that as a separate idea 2011-01-19T19:26:49 right 2011-01-19T19:27:02 CTF is a slightly more complex goal than simple assault 2011-01-19T19:27:03 oh actually, that could work 2011-01-19T19:27:09 The difference is that the game always goes the same amount of time, and you have to maintain control. 2011-01-19T19:27:16 the flag moves if it's on the same square as a live enemy ant? 2011-01-19T19:27:28 No moving, to complicated. 2011-01-19T19:27:43 and they have to get it back to their half, you get to place it at the start? 2011-01-19T19:27:48 But if there is fog, you do have to discover. 2011-01-19T19:27:50 McLeopold: i don't think we're talking about your idea 2011-01-19T19:28:08 we're talking about pretty much pure CTF 2011-01-19T19:28:14 what about just actual 2d grid form flags with x agents? 2011-01-19T19:28:26 ? 2011-01-19T19:29:12 Okay, I follow. 2011-01-19T19:29:59 With control points that don't move, you force the player with more control to split up forces. With CTF, it becomes too much of a fast growth game. 2011-01-19T19:30:20 McLeopold: how about the goal is to destroy the opposing ant hill, but there are a very limited number of food sources on the map (like 2 or 4) that you have to get food from to create more ants? 2011-01-19T19:30:22 I would prefer a game that doesn't favor fast growth. 2011-01-19T19:30:51 where it takes some time to go back and forth between and hill and food source 2011-01-19T19:30:57 That would be like control points, but it still favors population over objectives. 2011-01-19T19:30:58 *ant hill 2011-01-19T19:31:14 the food source could eventually run out 2011-01-19T19:31:16 would just plain flags be too boring? it'd be cool to watch 2011-01-19T19:31:25 and the objective is still to destroy the other hill 2011-01-19T19:31:30 not to have the most population 2011-01-19T19:31:50 so you have to choose whether to invest time in food or whether to go straight for the kill (which would also affect your opponent's choices) 2011-01-19T19:31:55 How about each map has it's own set of objectives? 2011-01-19T19:32:10 Maybe it's CTF, maybe it's control points? 2011-01-19T19:32:16 ouch, complicated 2011-01-19T19:33:13 Have you ever played paintball? 2011-01-19T19:33:17 yeah 2011-01-19T19:33:26 I'm thinking of rules likes the SPPL 2011-01-19T19:34:51 i do like it 2011-01-19T19:34:57 i'm just going over our options 2011-01-19T19:35:12 mainly just going to through various game modes from Halo and thinking about whether it would work here ;) 2011-01-19T19:35:30 Multi-Hill 2011-01-19T19:35:33 yeah 2011-01-19T19:35:59 assault is cool 2011-01-19T19:36:09 both sides start with a bomb and they have to carry it to the other base 2011-01-19T19:36:28 bomb resets if dropped and left alone for long enough 2011-01-19T19:36:33 In Halo, you have an offense and a defense. 2011-01-19T19:36:46 *** asdffsa has quit IRC (Quit: Page closed) 2011-01-19T19:36:51 not in multibomb 2011-01-19T19:36:57 in that mode both teams have a bomb 2011-01-19T19:36:58 ah 2011-01-19T19:37:20 But you get 1 point per "explosion" and keep going. 2011-01-19T19:37:23 there's also neutral bomb 2011-01-19T19:37:24 right 2011-01-19T19:37:32 neutral bomb has just one bomb that both teams try to get 2011-01-19T19:37:38 and then bring to enemy base 2011-01-19T19:38:06 So, if you get a random map, with a random game type (assault, ctf, hills), I think that would be fun. 2011-01-19T19:38:28 Especially with your ant language idea. 2011-01-19T19:38:33 yeah that would be fun, but now we're talking *really* complicated :) 2011-01-19T19:38:57 my idea alone was hard enough. multi game modes makes it yet harder 2011-01-19T19:39:23 But, it does make it so your ai needs to know how to program an effective ant. 2011-01-19T19:40:06 well, or at least how to use a predefined library of higher level functions 2011-01-19T19:40:10 Each starter pack could have some simple ants. One to find a bomb/flag. One to carry it to an objective. One to defend a hill. 2011-01-19T19:40:16 The tactics cross over. 2011-01-19T19:40:20 yeah 2011-01-19T19:41:18 we'd need to be careful with the language. would be easy to make some people upset if they don't like it, since it would be so integral to the game 2011-01-19T19:41:48 The ai could create an explore ant that would return with map info, because everyone starts out with no info? 2011-01-19T19:41:53 need to make it easy to write composable functions and such 2011-01-19T19:42:01 yeah 2011-01-19T19:42:10 except for the objective 2011-01-19T19:42:14 Give them monads. :) 2011-01-19T19:42:23 they would kill me 2011-01-19T19:42:29 I would kill me. 2011-01-19T19:42:34 i'll just give them monads without telling them 2011-01-19T19:42:39 nobody would know to complain then 2011-01-19T19:42:55 Make it functional. 2011-01-19T19:43:13 i actually think that would be a wise idea. like i said, composability 2011-01-19T19:43:14 Make it lisp and tell them it's a special "ant" language. 2011-01-19T19:43:22 it would just have to be... not annoying to imperative programmers 2011-01-19T19:43:50 obviously it should be SQL 2011-01-19T19:43:56 everybody loves SQL 2011-01-19T19:44:12 I've seen mandlebrot draw with sql... 2011-01-19T19:44:19 *drawn* 2011-01-19T19:44:23 yeah me too. it's horrible 2011-01-19T19:44:43 http://thedailywtf.com/Articles/Stupid-Coding-Tricks-The-TSQL-Madlebrot.aspx 2011-01-19T19:44:52 <3 daily wtf 2011-01-19T19:44:59 *** Migi32 has quit IRC (Remote host closed the connection) 2011-01-19T19:45:00 well, "<3" might be a bit much 2011-01-19T19:45:43 I think there might need to be an instruction count and a limit per ant per turn. 2011-01-19T19:45:53 yes 2011-01-19T19:46:04 Then you don't need to worry about clock time at all. 2011-01-19T19:46:24 i was thinking that for each step of the game each ant is given exactly X instructions, and if they don't decide what to do in that time they don't get to move on that turn 2011-01-19T19:46:26 Which, IMHO, is much nicer, since I timed out on half my games in the last one. 2011-01-19T19:47:23 there may be a higher level way to think of the limitation though, depending on the language used 2011-01-19T19:47:42 "instructions," for example, is too low level to accurately guage most functional languages 2011-01-19T19:48:11 well, if we design the ant language, then each function can have it's own weight. 2011-01-19T19:48:13 this would be a balancing act, but maybe each bit of functionality would simply come with a "cost" 2011-01-19T19:48:16 yeah that 2011-01-19T19:48:35 moving cost 10, sending a scent costs 1? 2011-01-19T19:49:00 Pushing a value to a key/value store costs 1? 2011-01-19T19:49:09 maybe something like that. not sure 2011-01-19T19:49:14 balancing act 2011-01-19T19:49:24 it would be hard 2011-01-19T19:49:36 and nobody would agree on what's fair 2011-01-19T19:50:06 especially considering different coding styles and the fact that different strategies would favor different functions 2011-01-19T19:50:19 well, it would favor particular tactics no matter how it's done 2011-01-19T19:50:42 but, that would be part of the problem spec 2011-01-19T19:50:43 well yeah, but i mean some of that would be artifically induced based on the costs in addition to whatever other forces are in play 2011-01-19T19:51:18 i'd rather it be open-ended on what strategies can be used, where possible 2011-01-19T19:53:04 Okay, even without the ant language (we need a better tag name) I like different objectives for each map. 2011-01-19T19:53:41 i do too, overall 2011-01-19T19:54:36 moving a flag or bomb would be a very simple add on to the current simple ant engine... 2011-01-19T19:54:41 antml 2011-01-19T19:54:53 it's not a markup language :P 2011-01-19T19:55:02 metalanguage! 2011-01-19T19:55:04 >.> 2011-01-19T19:55:18 heh, okay then i like it :) 2011-01-19T19:55:19 AntML 2011-01-19T19:55:23 yaal (yet another ant language) 2011-01-19T19:55:36 Yaal! Youall! 2011-01-19T19:55:54 ia ia 2011-01-19T19:56:10 ANAL is Not A Language 2011-01-19T19:56:38 hmm... 2011-01-19T19:56:47 joking 2011-01-19T19:57:04 antl's not the language 2011-01-19T19:57:47 okay, what are the functions? 2011-01-19T19:57:59 move, look, pickup, attack 2011-01-19T19:58:18 fart (for scents) 2011-01-19T19:58:31 what is hills? is that randomly assigning heights to each grid and only being able to move within a certain height change? 2011-01-19T19:58:34 that would be cool 2011-01-19T19:58:40 each square* 2011-01-19T19:59:03 antimatroid: that would be an interesting way to limit movement throughout the map 2011-01-19T19:59:06 hills would be locations on the map, perhaps spawn points, perhaps flags? 2011-01-19T19:59:24 yeah that was the idea, but i kind of like elevation. yet more complicated though :) 2011-01-19T19:59:42 that way walls aren't needed, every square is accessible 2011-01-19T19:59:45 i really like that actually 2011-01-19T19:59:53 antimatroid: http://phonons.wordpress.com/2010/06/01/cells-a-massively-multi-agent-python-programming-game/ 2011-01-19T20:00:02 That was Todd200's suggestion. 2011-01-19T20:00:19 it has elavations 2011-01-19T20:00:19 ah yeah 2011-01-19T20:00:27 identical agents, elevations 2011-01-19T20:00:52 huh, manipulating terrain 2011-01-19T20:00:57 didn't catch that the last time i saw it 2011-01-19T20:01:06 interesting way to add fortification or something 2011-01-19T20:01:10 yeah i have seen that, sort of got the idea with that 2011-01-19T20:01:33 simplest would be no editing terrain, just have the maps be generated on terrain where every square can be reached 2011-01-19T20:01:38 pretty easy to ensure 2011-01-19T20:01:48 not every square *must* be reachable 2011-01-19T20:01:59 walls are elegantly a special case of elevation, after all 2011-01-19T20:02:03 ant's should be able to dig holes into the enemy territory without being detected 2011-01-19T20:02:21 oh my 2011-01-19T20:02:21 oh, now we're talking, 3d grid ants? 2011-01-19T20:02:29 hard to visualise 2011-01-19T20:02:32 it hurts my brain 2011-01-19T20:02:33 AntFarmWars 2011-01-19T20:02:56 heh, so it's 2D from the side? :P 2011-01-19T20:03:03 no, just a simple tunnel connecting 1 grid unit to another. 2011-01-19T20:03:24 it'd be a bit like simultaneous worms 2011-01-19T20:04:04 the tunnel has a distance which is less than the actual distance, but takes longer to dig 2011-01-19T20:04:39 you could hid your ants in a corner, and then pop them up for defense 2011-01-19T20:04:41 okay, well, you guys keep thinking up hard ideas. i'll play minecraft 2011-01-19T20:04:49 :) 2011-01-19T20:04:58 multiplayer discrete minecraft we need 2011-01-19T20:05:11 what do you do in minecraft? 2011-01-19T20:05:14 i'm sure the guy would be cool with it for the publicity (not that he needs it at this point) 2011-01-19T20:05:23 it's lego for adults 2011-01-19T20:05:45 giant legos? 2011-01-19T20:06:26 big 3d discrete world where you can add/remove blocks 2011-01-19T20:06:36 the actual game has more and more features being added 2011-01-19T20:06:42 and at night zombies come out and fuck with you 2011-01-19T20:06:49 so you have to built forts and shit 2011-01-19T20:06:54 build* 2011-01-19T20:07:07 are there points or character levels? 2011-01-19T20:07:12 you can play a crap version online 2011-01-19T20:07:15 no, you do what you want 2011-01-19T20:07:33 http://www.minecraft.net/play.jsp 2011-01-19T20:07:49 No, I like my free time. 2011-01-19T20:08:05 Okay, one last shot at a totally different idea. 2011-01-19T20:08:11 Without ants... 2011-01-19T20:08:21 Anyone here play Munchkins? 2011-01-19T20:08:59 Anyone here? 2011-01-19T20:09:18 munchkin the card game? 2011-01-19T20:09:41 yes 2011-01-19T20:09:52 That would be a fun AI to write. 2011-01-19T20:09:58 But difficult. 2011-01-19T20:10:36 indeed 2011-01-19T20:10:53 jmcarthur: AntML needs the ability to define routes, or a list of points for an ant to follow. 2011-01-19T20:11:21 Then everyone can have fun posting A* implementations. 2011-01-19T20:12:14 Okay, I'm out. 2011-01-19T20:12:15 *** McLeopold has left #aichallenge 2011-01-19T20:14:49 *** Meatkat has quit IRC (Ping timeout: 265 seconds) 2011-01-19T20:18:03 *** Blkt has quit IRC (Quit: Error: do not makunbound t please!) 2011-01-19T20:58:23 antimatroid has started minecraft too, now? 2011-01-19T21:00:18 *** amstan has joined #aichallenge 2011-01-19T21:00:18 *** ChanServ sets mode: +o amstan 2011-01-19T21:07:01 sigh: no 2011-01-19T21:24:31 minecraft is digital crack 2011-01-19T21:25:06 yo 2011-01-19T21:25:36 haha... amstan, do you have irc notify you whenever someone says minecraft? :P 2011-01-19T21:25:55 zomg! i could make contestbot do that 2011-01-19T21:26:09 -_- 2011-01-19T21:48:18 we certainly don't want this to happen http://news.ycombinator.com/item?id=1538256 2011-01-19T21:56:54 *** AlliedEnvy has quit IRC (Ping timeout: 240 seconds) 2011-01-19T22:00:31 minecraft? 2011-01-19T22:00:44 jmcarthur: tldr, what? 2011-01-19T22:01:37 tl;dr? it's like three sentences 2011-01-19T22:02:16 irc breeds short attention spa 2011-01-19T22:02:35 basically the danger that a game is chosen for which an optimal simple strategy is discovered before the end of the contest 2011-01-19T22:02:45 sigh: you accidentally forgo a letter 2011-01-19T22:02:49 well, not only that 2011-01-19T22:03:03 that said optimal strategy is stupid 2011-01-19T22:03:05 amstan: it wasn't an accident :P 2011-01-19T22:03:23 oh you said simple 2011-01-19T22:03:27 n/m 2011-01-19T22:03:49 yes, the example given didn't sound like a stupid strategy quite elegant actually, but still ruins the contest 2011-01-19T22:04:47 this scenario was actually my worry when first hearing about planetwars 2011-01-19T22:05:02 amstan: http://contestbot.hypertriangle.com/channellogs/%23aichallenge/%23aichallenge.2011-01-19.log <-- in-page search for "2011-01-19T19:06:55" and start reading there 2011-01-19T22:05:15 at least planet wars had the advantage of being based of an existing game 2011-01-19T22:05:21 amstan: i think we got a little carried away, but there's an idea in there somewhere :) 2011-01-19T22:06:05 it wouldn't be for the next contest though. too hard, too little time 2011-01-19T22:06:22 I searched for "2011-01-19T19:06:55" and found the line you just typed -_- 2011-01-19T22:06:29 lol 2011-01-19T22:06:55 sigh: yep, that is a huge advantage to using an existing game in my opinion. You know at least some of the state space has already been explored and no degenerate cases turned up 2011-01-19T22:06:57 there should be at least one other occurrence 2011-01-19T22:07:30 two now, since I said it as well 2011-01-19T22:07:31 i highly doubt that just because a human-playable version of a game exists that the state space had been properly explored at all 2011-01-19T22:08:06 sigh: that's it? i even copied what i said here, refreshed, and re-searched to make sure 2011-01-19T22:08:20 don't worry it's there 2011-01-19T22:08:26 I meant two other occurances 2011-01-19T22:08:28 ah 2011-01-19T22:08:44 janzert: i had the same thought with planet wars, but it was possibly a good thing 2011-01-19T22:09:00 it had many people (at least me) trying to work out unsucessfully how to do it optimally 2011-01-19T22:09:17 i hate pages of algebra trying to solve just a general case with 3 planets, it was tricky 2011-01-19T22:09:29 :) 2011-01-19T22:10:12 well, planet wars had the advantage of being an insanely huge search space 2011-01-19T22:10:13 "I don't play WoW, for much the same reason that I've never tried heroin" 2011-01-19T22:10:18 *** AlliedEnvy has joined #aichallenge 2011-01-19T22:10:33 as long as it isn't truly solvable the appearance of resolvability is indeed probably a good thing for the contest :P 2011-01-19T22:11:49 it's good to be able to make a decent strategy with very little effort and improve it significantly with very little effort, but it's also good for there to be no foreseeable end to it 2011-01-19T22:12:07 so go find a game that is interesting, has good visuals appears simple to solve and boils down to the halting problem :) 2011-01-19T22:12:18 jmcarthur: planet wars was a bit like that 2011-01-19T22:12:24 yeah, everyone keeps wanting simple games for people to get started with, i want it to be interesting at the top of the leaderboard 2011-01-19T22:12:36 provided people can start programming easily, i don't mind how complicated the game is 2011-01-19T22:12:37 sigh: everybody kept saying it was hard to improve on though 2011-01-19T22:12:59 like, you could occasionally make a few baby steps, but then any further "improvements" ended up being negative 2011-01-19T22:13:13 i didn't run out of ideas to try with planet wars, i'm with sigh 2011-01-19T22:13:24 jmcarthur: well true, you had to put it a lot of work, and there were a lot of dead ends 2011-01-19T22:13:47 move selection was what killed me I think 2011-01-19T22:13:50 but I think that to get to say the top 500-200 there was a clear progression plan 2011-01-19T22:13:52 mine was so horrible and inefficient 2011-01-19T22:13:52 that's a good way to put it. "dead ends" 2011-01-19T22:14:29 and my completely broken tree that was submitted, but i'm still not entirely convinced that i wouldn't have done worse with it unbroken 2011-01-19T22:14:34 i would have rather there been a wider selection of viable strategies, but it turned out that most paths were dead ends and there were just a few that actually worked 2011-01-19T22:14:52 jmcarthur: when starting out it was fairly simple to improve the ai 2011-01-19T22:15:07 granted, within those few there was a lot of... err... "stylistic" variations 2011-01-19T22:15:25 yes 2011-01-19T22:15:30 which was good 2011-01-19T22:15:35 like the min spanning tree bot 2011-01-19T22:15:40 cool idea, sucked for the game 2011-01-19T22:15:48 antimatroid: yeah 2011-01-19T22:15:54 that worked for bartwe because he was the first to implement redistribution 2011-01-19T22:16:03 but it sucked compared to other redist methods 2011-01-19T22:16:05 or an ever increasing convex hull 2011-01-19T22:17:11 i like the idea of not having explicit walls inside grid squares, and just creating maps with squares of different height 2011-01-19T22:17:31 i think that could be nice for a shooter type of game 2011-01-19T22:17:43 i guess that snowball game does that, right 2011-01-19T22:17:44 ? 2011-01-19T22:17:57 i think mit battlecode does that 2011-01-19T22:18:02 snowball game lets you create walls like that, but you just can't move onto anything higher than 5 units 2011-01-19T22:18:09 everything starts at 3 except trees 2011-01-19T22:18:16 ah 2011-01-19T22:18:40 i've given up on snowballs, i go away next saturday and get back like 2 days before it ends 2011-01-19T22:18:44 wasn't going to really work 2011-01-19T22:19:11 plus i hate their movement collision detection, it's intuitively retarded 2011-01-19T22:19:22 intuitively retarded? 2011-01-19T22:19:26 that's an interesting phrase 2011-01-19T22:19:42 you can swap squares but not if there is an empty sqare between 2011-01-19T22:19:57 http://i.imgur.com/Z8UPO.png 2011-01-19T22:20:02 that is another retarded thing 2011-01-19T22:20:52 it sort of had to be done like that for snowballs to be done as moves were, but i don't like it 2011-01-19T22:21:06 they should have done snowballs, then moves and reverted collisions once moves were done 2011-01-19T22:21:49 yeah, it's annoying when rules seems arbitrary 2011-01-19T22:22:52 well, I guess this proves that you actually do value having a story and not just a formal specification :P 2011-01-19T22:23:20 not really, it just made it much harder to work out what moves were possible :P 2011-01-19T22:35:24 *** pyro- has joined #aichallenge 2011-01-19T22:51:03 *** albert_ has joined #aichallenge 2011-01-19T23:34:11 *** elimisteve has quit IRC (Quit: Later) 2011-01-19T23:40:28 *** tapwater has quit IRC (Quit: tapwater)