2011-06-25T00:20:13 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-06-25T00:37:25 *** zaphus has joined #aichallenge 2011-06-25T00:54:45 *** sigh has joined #aichallenge 2011-06-25T01:25:50 *** sigh has quit IRC (Ping timeout: 252 seconds) 2011-06-25T01:55:53 *** Apophis_ has joined #aichallenge 2011-06-25T02:00:09 *** nann has quit IRC (Ping timeout: 276 seconds) 2011-06-25T02:16:58 contestbot: later tell McLeopold every time a compile fails it stops the worker now http://pastebin.com/wRnKvVS3 it looks to be the changes on line 524-526 of worker.py in this commit https://github.com/aichallenge/aichallenge/commit/76eb30e847af2c49ac896ac86cd7f4c10410fd5f 2011-06-25T02:16:58 janzert: I think that worked... 2011-06-25T02:19:52 aichallenge: janzert epsilon * r40c327c / worker/worker.py : Make compilation task succeed even if the compilation itself fails - http://bit.ly/jXlUMW 2011-06-25T02:20:42 contestbot: later tell McLeopold I'm changing it to always return True for now if an exception isn't raised but I don't fully understand the code path so you should fix it the right way later 2011-06-25T02:20:42 janzert: I come to serve. 2011-06-25T02:46:05 aichallenge: janzert epsilon * r2d60c87 / sql/2_generate_matchup.sql : Add missing parentheses to get correct order of comparison - http://bit.ly/kmvqMS 2011-06-25T02:53:46 contestbot: later tell McLeopold I think I found the problem with unbalanced number of games for the ends of the rankings as well http://bit.ly/kmvqMS assuming I didn't screw up the database update that should be in place from around game 36950 2011-06-25T02:53:46 janzert: As you wish. 2011-06-25T03:08:54 *** nux67 has joined #aichallenge 2011-06-25T03:15:33 *** Kingpin13 has joined #aichallenge 2011-06-25T03:36:40 sir_macelon is demonstrating it's a level above the other bots at the moment :-) 2011-06-25T03:39:34 *** AlliedEnvy has quit IRC (Remote host closed the connection) 2011-06-25T03:41:17 *** AlliedEnvy has joined #aichallenge 2011-06-25T03:47:09 *** amstan has quit IRC (Ping timeout: 255 seconds) 2011-06-25T03:50:47 *** zaphus has quit IRC (Quit: Page closed) 2011-06-25T03:50:51 *** foRei has joined #aichallenge 2011-06-25T04:01:02 *** gobang has joined #aichallenge 2011-06-25T04:01:23 *** Palmik has joined #aichallenge 2011-06-25T04:20:03 *** choas has joined #aichallenge 2011-06-25T05:11:45 *** mr_borod has joined #aichallenge 2011-06-25T05:11:51 hello 2011-06-25T05:16:40 simple question: is it possible to participate in beta? (actually tried to create an account but no confirmation e-mail has been sent by the server :( ) 2011-06-25T05:37:13 go back to the registration page and click the activate link at the bottom. theres no email yet 2011-06-25T05:45:23 *** Naktibalda has joined #aichallenge 2011-06-25T06:04:35 *** stoch has joined #aichallenge 2011-06-25T06:06:13 mr_borod: there is link on the loging page you need to click. No sent email. No need for them yet. 2011-06-25T06:06:30 mr_borod: it *is* a bit confusing :) 2011-06-25T06:06:44 it should be made more obvious i suppose. 2011-06-25T06:07:38 mr_borod: so basically, you log-in and click the link. 2011-06-25T06:10:45 Bot features can be embarassing. There was this one idea, that i now despise as being bat. Yet the pure use of it is my strongest bot to date :'( 2011-06-25T06:11:17 a 2011-06-25T06:11:25 *** stoch has quit IRC (Quit: Page closed) 2011-06-25T06:15:14 *** Accoun has quit IRC () 2011-06-25T07:28:40 *** carlop has joined #aichallenge 2011-06-25T07:30:29 carlop: got any ideas for how to make nice maze maps? 2011-06-25T07:32:38 antimatroid: no, actualy i'm trying to enumerate all possible symmetry that a map can have, reducing the problem only to desing a player zone (that can be triangular, like current 8-player symmetric map, or square/rectangular/rombo) 2011-06-25T07:33:01 yeah, i've taken a look at your pastebin and wiki link 2011-06-25T07:33:07 so the map generator can be a part that generate a pattern, and a part that replicate the pattern in simmetric way 2011-06-25T07:33:40 for the rotational maps i made yesterday, i have 5 types of symmetry for 2 players, 3 types of symmetry for 4 players and 1 type for 8 players 2011-06-25T07:33:53 i'm not sure why i don't have 7 for 2 players, 2 of them don't seem to work 2011-06-25T07:34:04 on build interesting pattern (eg maze one) i don't have idea 2011-06-25T07:34:20 one can use a dfs to build a "perfect maze" 2011-06-25T07:34:24 i think that there are at least 7 pattern for 2player, maybe more 2011-06-25T07:34:27 but the players regions are going to be disconnected 2011-06-25T07:34:35 really i want multiple paths leading to places 2011-06-25T07:34:50 there should be 7 for 2 players i think 2011-06-25T07:35:03 based on the rotations of an nxn grid 2011-06-25T07:35:30 *** UncleVasya has joined #aichallenge 2011-06-25T07:35:41 i have an 8th (a maybe a 9th simmetric) that i'm not sure can be reduced to one of the previous one 2011-06-25T07:36:38 Good afternoon (or whatever else). 2011-06-25T07:36:45 *** Accoun has joined #aichallenge 2011-06-25T07:36:49 i think the 6th and 7th that you miss are this one: http://en.wikipedia.org/wiki/Wallpaper_group#Group_pg and the same pattern but vertical 2011-06-25T07:37:26 http://pastebin.com/Md2LfGpr 2011-06-25T07:37:30 those are the 7 i expect 2011-06-25T07:37:49 i don't have the vertical/horizontal flip then diagonal flip rotations 2011-06-25T07:38:02 when i generate it the way i'm doing all the others, it doesn't come out symmetrically :s 2011-06-25T07:38:16 even though i use the functions for that in one of the 4 player symmetric types 2011-06-25T07:38:58 carlop: do you know much group theory? 2011-06-25T07:39:11 i love groups, they're awesome 2011-06-25T07:40:28 *** racko has joined #aichallenge 2011-06-25T07:41:15 antimatroid: i have study only basical thing, but i think i can understand fast math thing 2011-06-25T07:41:48 i have not understand your paste with the symmetry.... sorry 2011-06-25T07:42:05 my paste has 2 player symmetric games 2011-06-25T07:42:22 done with the 7 non-identity rotations of the grid 2011-06-25T07:42:48 now i get it 2011-06-25T07:43:09 but 2nd and 7th can't work, cause they are not symmetric 2011-06-25T07:44:12 they have only a 90degree turn, that can work only with 4x player 2011-06-25T07:45:08 carlop: that would explain why 2 didn't work for me :) 2011-06-25T07:46:09 http://pastebin.com/0hAmMMMP 2011-06-25T07:46:19 those rotations leave the starting positions the same 2011-06-25T07:46:39 but i think as you mentioned, you can construct examples where a horizontal/vertical then diagonal flip does not leave the map unchanged 2011-06-25T07:47:08 like this one: http://pastebin.com/gSbeWeEE 2011-06-25T07:48:36 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-06-25T07:48:43 maybe we are talking of the same thing, i should need what orientation have the two ant in your vertical then diagonal flip example 2011-06-25T07:49:08 i do a transpose like a matrix when diagonally flipping 2011-06-25T07:49:32 *** UncleVasya has joined #aichallenge 2011-06-25T07:49:32 you can take h,v,d flips as generators for the group of rotations on a square 2011-06-25T07:54:04 whoo, gold in intermediate lotus challenge! 2011-06-25T07:54:12 i'm thinking.... i'm not sure this generetor work, cause they have always a fixed point 2011-06-25T07:54:40 carlop: sorry what? 2011-06-25T07:55:37 there are symmetry without a fixed point (eg traslation, a type of symmetry), but all flips preserve the center of the square in the same point 2011-06-25T07:55:59 yes 2011-06-25T07:56:15 if i had to do a diagonal flip to get to my rotation, i have had to use square maps 2011-06-25T07:56:22 otherwise they can be rectangles 2011-06-25T07:57:02 h and v would have needed it also 2011-06-25T07:57:19 Rabidus: what is that? and well done :P 2011-06-25T07:58:42 10 mins of interval between submissions must be on now, right? If so it might be broken because I've made two success uploads with 5 mins of interval. 2011-06-25T07:59:21 15:44:37 and 15:49:37 2011-06-25T07:59:33 it might have been changed 2011-06-25T07:59:44 actually, i think it's meant to be really short once your bot has compiled 2011-06-25T08:00:08 antimatroid: the problem is that i think we are not looking for composition of rotation and mirroring, but for composition of rotation, mirroring and translation 2011-06-25T08:00:22 *** onensora has joined #aichallenge 2011-06-25T08:00:58 no, i think you want all three of those along with tiling rotations/mirrows/transposes of those 2011-06-25T08:01:06 a translation isn't going to increase the size of the map 2011-06-25T08:02:06 look of this two example: they are made of mirroring and translation, and without translation they can be done 2011-06-25T08:02:21 can-> can't 2011-06-25T08:04:01 *** mleise has quit IRC (Ping timeout: 250 seconds) 2011-06-25T08:09:13 *** mleise has joined #aichallenge 2011-06-25T08:12:40 i go to lunch, then i try to understand where is the problem... 2011-06-25T08:16:30 *** Zaphus has joined #aichallenge 2011-06-25T08:20:11 i think i know how to do more mazey maps 2011-06-25T08:20:32 i'll make the random walks follow random lengthed straight line paths 2011-06-25T08:24:40 another possible idea for maze generation is this: http://en.wikipedia.org/wiki/Self_avoiding_walk, try to build a water path like the one, then adding brige somewhere 2011-06-25T08:25:57 the problem is i want a maze with noisy lines and wider than 1 width corridors 2011-06-25T08:26:15 and more than one way connecting stuff up 2011-06-25T08:26:40 i think maybe if i do that random sized linear lines with some noise 2011-06-25T08:26:46 then just random walk those all over the place 2011-06-25T08:26:48 that might work 2011-06-25T08:28:11 yes, your idea can work 2011-06-25T08:31:52 *** olexs has joined #aichallenge 2011-06-25T08:32:10 olexs: evening :) 2011-06-25T08:32:31 antimatroid: afternoon :) 2011-06-25T08:32:42 i made more map types :) 2011-06-25T08:33:09 uh, I see my bot has been beaten quite a lot since I've been gone 2011-06-25T08:33:23 time to get to work again, it seems 2011-06-25T08:35:07 hm, some very buggy behavior I observe from my bot in last games. have any parameters been changed, like viewradius? 2011-06-25T08:36:06 yes parameters have changed 2011-06-25T08:36:11 have you been hard coding stuff already? :P 2011-06-25T08:36:17 not really 2011-06-25T08:36:21 the battle resolution has changed also 2011-06-25T08:36:29 but my hexgrid seems broken with new viewradius :) 2011-06-25T08:36:32 it's now the occupied method 2011-06-25T08:36:49 it's actually computed dynamically, but it seems it doesn't work for the new radius... 2011-06-25T08:37:12 antimatroid: it's called the "Focus Method" on the front page - same thing ? 2011-06-25T08:37:35 Zaphus: same thing 2011-06-25T08:39:35 how's the general progress on the beta? release planned any time soon, or still indeterminate? :) 2011-06-25T08:40:38 wow - just checked out this battle: http://aichallengebeta.hypertriangle.com/visualizer.php?game=37336&user=221&turn=1173 thats the turn I 'won' - but notice the graph of ants, I'm red, the lower entry on the graph - must be winning the fights faster than food arrived! won't happen once olexs fixes his bot though 2011-06-25T08:40:55 olexs: still indeterminate, but we're getting much closer 2011-06-25T08:42:23 *** Zaphus has quit IRC (Quit: Page closed) 2011-06-25T08:44:24 antimatroid: it was gt5 most challenging event 2011-06-25T08:52:29 *** Accoun has quit IRC (Ping timeout: 258 seconds) 2011-06-25T08:53:11 *** Accoun has joined #aichallenge 2011-06-25T08:53:11 *** Accoun is now known as Acooun 2011-06-25T08:53:23 *** Acooun is now known as Accoun 2011-06-25T09:13:36 *** onensora has quit IRC (Ping timeout: 240 seconds) 2011-06-25T09:14:20 *** onensora has joined #aichallenge 2011-06-25T09:25:07 I can't get playgame.py to log my stderr anymore... -E worked before, only logs illegal moves but no stderr output 2011-06-25T09:41:20 *** me0w has joined #aichallenge 2011-06-25T09:42:33 *** me0w has quit IRC (Remote host closed the connection) 2011-06-25T09:43:31 *** onensora has quit IRC () 2011-06-25T09:51:57 aichallenge: Marco Leise epsilon * r7496a9d / ants/visualizer/js/Application.js : visualizer: fixed attack range cirles not drawing over map borders when zoomed out - http://bit.ly/k4oAIu 2011-06-25T10:37:06 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-06-25T10:43:38 *** olexs1 has joined #aichallenge 2011-06-25T10:44:42 *** olexs has quit IRC (Ping timeout: 276 seconds) 2011-06-25T10:52:39 *** davidjliu has joined #aichallenge 2011-06-25T10:53:04 *** Asmoo has joined #aichallenge 2011-06-25T11:01:20 *** olexs1 has quit IRC (Quit: Leaving.) 2011-06-25T11:04:25 I have an idea. It's an additional work for mleise :) 2011-06-25T11:05:57 To add a new mode to the visualizer: 'show game without info'. 2011-06-25T11:06:38 aichallenge: Marco Leise epsilon * r891bd8f / ants/visualizer/js/Application.js : visualizer: wait half a turn after spawning ants before attack lines are drawn - http://bit.ly/il8dEn 2011-06-25T11:06:58 And make it possible to give a link on the game with this mode on. 2011-06-25T11:07:54 UncleVasya: that sounds completely useless :p 2011-06-25T11:08:27 I'll explain the use of it now. 2011-06-25T11:09:26 And then add to channel bot the ability to send PM with that link. 2011-06-25T11:10:41 I subscribe to channel bot like to RSS. 2011-06-25T11:11:25 Asking him to send me PM with link when my bot plays a new game. 2011-06-25T11:12:08 Just like he says about github changes. 2011-06-25T11:14:21 When I update browser waiting for new games, I see the outcome of the game before I watch the game itself. No intrigue :( 2011-06-25T11:15:33 And with this feature of channel bot and visualizer watching games can be much more interesting. 2011-06-25T11:17:04 mleise: what do you think? 2011-06-25T11:19:46 hmm can you explain what the visualizer should look like? 2011-06-25T11:20:47 a RSS feed with the latest games and results ok, but what is the role of the visualizer? 2011-06-25T11:22:46 In RSS feed with latest game a link comes like that: http://aichallengebeta.hypertriangle.com/visualizer.php?game=37498 2011-06-25T11:23:50 I follow link and before I start watching game I see who is winner. 2011-06-25T11:24:40 I suggest to sent link like that http://aichallengebeta.hypertriangle.com/visualizer.php?game=37498&mode=no_info 2011-06-25T11:25:43 where mode=no_info means that all info about outcome isn't shown. 2011-06-25T11:26:39 Just player names in normal order. 2011-06-25T11:31:27 Also it can be useful if someone wants to share a cool replay with a very unexpected final to his friends. 2011-06-25T11:53:56 *** davidjliu has quit IRC (Ping timeout: 252 seconds) 2011-06-25T12:10:08 UncleVasya: Ok, now I understand it. I'll implement this only if I get more requests for it though. 2011-06-25T12:12:12 aichallenge: Marco Leise epsilon * r253c984 / ants/visualizer/js/Application.js : visualizer: hopefully fixed some attack lines not showing - http://bit.ly/lDta3Y 2011-06-25T12:21:56 mleise: more requests means that 1. more people come with this idea from their own heads or 2. I ask people what they think about this and they says 'yeah, it's cool' ? 2011-06-25T12:47:03 *** carlop has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 4.0.1/20110413222027]) 2011-06-25T12:47:51 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-06-25T12:56:55 *** Gplusplus has joined #aichallenge 2011-06-25T12:57:15 hi 2011-06-25T12:57:39 does anyone know what does mean "turn 1 bot 0 ignored actions" 2011-06-25T12:57:42 ? 2011-06-25T12:58:18 i put it on the forum too 2011-06-25T12:58:28 hope that's no problem 2011-06-25T13:07:55 *** gobang has quit IRC (Read error: Connection reset by peer) 2011-06-25T13:08:23 *** gobang has joined #aichallenge 2011-06-25T13:11:40 *** Gplusplus has quit IRC (Ping timeout: 252 seconds) 2011-06-25T13:26:57 Gplusplus: I believe it means the bot issued some illegal orders (such as an ant into water) 2011-06-25T13:27:09 that the engine then ignored 2011-06-25T13:31:30 aichallenge: janzert epsilon * rf28f044 / worker/worker.py : post_result will delete the last_game file on successful submission - http://bit.ly/jLeSj2 2011-06-25T13:36:02 *** stocha has joined #aichallenge 2011-06-25T13:36:10 hi in there. 2011-06-25T13:37:39 who is flagcapper builder ? 2011-06-25T13:39:11 <_flag> stocha: here 2011-06-25T13:45:16 *** Accoun has quit IRC (Read error: Connection reset by peer) 2011-06-25T13:46:46 *** Accoun has joined #aichallenge 2011-06-25T13:49:10 *** okayzed is now known as okay 2011-06-25T14:02:22 I am still still confused about account activation... :S I do not see any activation link neither on the account creation page nore on the login one :( any help? 2011-06-25T14:04:28 <_flag> As soon as the account is created there's a small link under or near the confirmation message 2011-06-25T14:04:40 <_flag> That's the only time you can click on it 2011-06-25T14:05:04 <_flag> Ignore the part about it sending you an email, that's only for the actual contest 2011-06-25T14:09:05 *** McLeopold has joined #aichallenge 2011-06-25T14:09:56 *** okay is now known as okayzed 2011-06-25T14:32:42 *** stocha has quit IRC (Ping timeout: 252 seconds) 2011-06-25T14:47:46 _flag, thank you 2011-06-25T14:48:14 _flag, but I did not recieve any confirmation message :S 2011-06-25T14:48:58 <_flag> I mean after you register, the conformation link is on the page 2011-06-25T14:49:40 ahhh 2011-06-25T14:50:14 does it mean that I have to make another account creation attempt? (the page I got after registration is already closed and for sure expired) 2011-06-25T14:50:48 <_flag> Yes, it's beta anyway so it's not a big deal 2011-06-25T14:51:02 _flag, thanks again, your help is rly appreciated :) 2011-06-25T14:51:20 * mr_borod afk *cheers* 2011-06-25T14:56:29 mr_borod: welcome to the 'nickname2' league 2011-06-25T14:57:45 *** stocha has joined #aichallenge 2011-06-25T14:58:43 _flag: well in fact, if you just login, i think you can still activate your thing. I'm under the impression i didn't saw it on first time, and i activated the account later. 2011-06-25T14:59:31 Okay, today i want to defend the position "Maps making process (or static maps) should be settled BEFORE the end of the beta" 2011-06-25T14:59:47 1/ People don't want to write code that will never be used. 2011-06-25T15:00:38 So let's take an example. Let's say you want to write blocking code. So you detect good point for blocking on map, clever algorithm for puting ant in there etc ... but let's say it take 3000 line of code and 3 weeks of tuning/debugging. 2011-06-25T15:01:30 Now, imagine you are already in the real contest, and sudenly, there is this new features of maps never having any blocking point (corridors are always 3 square wide). 2011-06-25T15:01:37 how would you feel ? isn't that bad ? 2011-06-25T15:02:00 (now maybe you had fun during those three week, but well) 2011-06-25T15:02:35 I have other example, involving time management etc. But i hope i already made the point clear. So i won't insist. 2011-06-25T15:02:40 what do you guys think ? 2011-06-25T15:05:24 stoch, your main point is to have a 'feature freeze', once it's beyond beta ? 2011-06-25T15:06:46 *** olexs has joined #aichallenge 2011-06-25T15:06:59 gobang: yes. Or as close as possible. I don't know, pay the participants if something change :p 2011-06-25T15:07:19 start tcpclient.py pedro 2081 "E:\code\aichallenge\ants\p\gobang0.exe" gobang_0 100 2011-06-25T15:07:27 *** gobang has quit IRC (Read error: Connection reset by peer) 2011-06-25T15:07:38 gobang: features freeze, and also edges values for parameters freeze. 2011-06-25T15:08:38 *** gobang has joined #aichallenge 2011-06-25T15:09:07 sorry for the noize 2011-06-25T15:10:08 rankings are pretty hard to interpret. It changes quite often. 2011-06-25T15:10:45 davidjliu only made 1 submission. I'm yet to understand what rank it should have though. 2011-06-25T15:14:47 have you seen the snake bot. Who can be stupid enough to build a snake bot really. It will always be very weak. 2011-06-25T15:15:20 Maybe he made a strong bot, and the thing bugged out to being a snake. 2011-06-25T15:15:33 once debugged, it will rule the challenge. maybe. 2011-06-25T15:16:05 http://aichallengebeta.hypertriangle.com/visualizer.php?game=37613&user=226 2011-06-25T15:18:23 *** stocha has quit IRC (Quit: Page closed) 2011-06-25T15:19:10 you'll end up playing 'tron' instead of 'ants' 2011-06-25T15:19:46 *** stocha has joined #aichallenge 2011-06-25T15:19:56 oh the snake bot is mine. silly me. 2011-06-25T15:20:02 *** stocha has quit IRC (Client Quit) 2011-06-25T15:22:51 i'm still waiting for the bot, that shits tetris like bricks, just for fun 2011-06-25T15:28:54 how comes, that for the last 24 h i get 'In Game: You are playing in a game right now.' on my profile ? rank & skill seem to be frozen there, too. 2011-06-25T15:35:26 *** olexs has left #aichallenge 2011-06-25T15:40:00 *** stocha has joined #aichallenge 2011-06-25T15:40:17 gobang i don't know. but games are played. after a while. As far as can see. 2011-06-25T15:40:26 gobang: 2011-06-25T15:41:44 now davidjlui never updated bot is 2d . O_o does it learn by itself or what. 2011-06-25T15:41:48 true, it's just those 2 line top on the profile page, that seem to get never updated 2011-06-25T15:42:38 if we wait a few week more, davidjliu never updated bot will take the lead. 2011-06-25T15:42:54 a python bot with that. 2011-06-25T15:43:18 *** gobang has quit IRC (Read error: Connection reset by peer) 2011-06-25T15:43:38 *** gobang has joined #aichallenge 2011-06-25T15:44:44 people were complaining about negative time for: your next game will be in .. seconds. then came a fix. 2011-06-25T15:45:58 i think i just need to build a strong bot. Sadly i have some bot features i want to explore first :p 2011-06-25T15:46:50 i aggree with the ranking system saying sir_macelon has the lead 2011-06-25T15:47:20 it's not very hard to agree with that, mind you 2011-06-25T15:48:55 *** Zeiris_ is now known as Zeiris 2011-06-25T15:49:02 i think, you're good, if you don't have to scroll up to see the top on the ranking page .. 2011-06-25T15:49:59 gobang: ? at this point i would say only sir_macelon has a good bot. 2011-06-25T15:50:13 definitely 2011-06-25T15:50:20 it sure is fighting bot. 2011-06-25T15:50:33 seem there room for improvement of the other features though. 2011-06-25T15:51:48 i don't understand how he manage to feeds itself though. clustered as he is. 2011-06-25T15:51:55 http://aichallengebeta.hypertriangle.com/visualizer.php?game=37739&user=7 2011-06-25T15:52:32 *** gobang has quit IRC (Read error: Connection reset by peer) 2011-06-25T15:52:53 *** gobang has joined #aichallenge 2011-06-25T15:58:15 he seems to have found a good cluster size, though. snakes (or blobs) longer that ~6 are not so effective, when they could like split up & explaore more space 2011-06-25T15:58:53 flagcapper went up to 2d place again. 2011-06-25T16:00:09 let's take sir_macelon back to 3rd. 2011-06-25T16:00:17 who's with me ? :) 2011-06-25T16:01:31 who's sir_macelon (in here ?) 2011-06-25T16:04:45 he calls himself sir_macelon most of the time i believe 2011-06-25T16:04:49 @seen sir_macelon 2011-06-25T16:04:49 stocha: sir_macelon was last seen in #aichallenge 4 days, 9 hours, 20 minutes, and 18 seconds ago: Zaphus: depends on the fighting skill, better food collection gives you an advantage and without good battling it might still be not enough to win 2011-06-25T16:05:23 mm, maybe i should go and see the log from 4 days ago :) 2011-06-25T16:05:42 *** gobang has quit IRC (Read error: Connection reset by peer) 2011-06-25T16:05:58 *** gobang has joined #aichallenge 2011-06-25T16:07:18 @seen lavalamp 2011-06-25T16:10:23 *** McLeopold has quit IRC (Quit: Leaving.) 2011-06-25T16:12:03 stocha: UbuRoi, king of poland 2011-06-25T16:12:36 gobang: plait t'il ? 2011-06-25T16:13:38 sorry my frech is too bad, i'll onla remember swear words of any given language 2011-06-25T16:13:38 ist dass ein Theaterstück des französischen ? 2011-06-25T16:13:46 merdre 2011-06-25T16:14:48 i certainly never heard of this one. how do you know it ? 2011-06-25T16:15:18 *** gobang has quit IRC (Read error: Connection reset by peer) 2011-06-25T16:15:40 *** gobang has joined #aichallenge 2011-06-25T16:15:40 fat man trying to conquer the world 2011-06-25T16:16:43 gobang: 1/do you go often to theatre ? Does everyone have seen it in deutchland ? 2011-06-25T16:17:26 gobang: 2/ i 'm not sure i got the point ? Do you think sir_macelon is fat ? :p 2011-06-25T16:17:52 i gui'm just teasing youess not 2011-06-25T16:18:45 and read a book, instead of watching all that television .. 2011-06-25T16:20:40 *** McLeopold has joined #aichallenge 2011-06-25T16:21:41 gobang: 1/ i don't always watch it. Nor listen to it. i only need the ambiance, you know. Then i have read a book, when i was young once. 2011-06-25T16:22:28 butmet some poeple, who never touched paper since school 2011-06-25T16:22:32 The first few pages at least. Or was it the synopsys ? 2011-06-25T16:22:59 gobang: it's because they have an ipad, or whatever e-reader. 2011-06-25T16:23:32 still then, there's project gutemberg 2011-06-25T16:25:04 you can get your free copay of'ullysses', and train some irc-markoff bot on that 2011-06-25T16:25:37 *** davidjliu has joined #aichallenge 2011-06-25T16:30:17 ahh, shucks, learning a new language, don't know, how to sort stuff in go right now. but then.. a* 2011-06-25T16:40:03 *** choas has quit IRC (Quit: leaving) 2011-06-25T16:49:24 actully, i'm curious to learn, how many ppl here picked up an 'obscure' language like OcmaMl, Lol, or Haske�� just for this 2011-06-25T16:49:32 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-06-25T16:49:49 i want to learn haskell, but i'm too lazy/distracted 2011-06-25T16:50:12 *** gobang has quit IRC (Read error: Connection reset by peer) 2011-06-25T16:50:33 *** gobang has joined #aichallenge 2011-06-25T16:50:33 seems, i can't even spell it 2011-06-25T16:51:40 *** stocha has quit IRC (Ping timeout: 252 seconds) 2011-06-25T16:56:05 *** Accoun has quit IRC () 2011-06-25T16:57:22 *** UncleVasya has joined #aichallenge 2011-06-25T16:58:41 thing, that fascinateds me about go, is that it's able sonme how to run without all that 'runtime polymorphism', compiler decides: been there, seen that, will do ( or not ). compiler is the most picky thing i've ever encounted before 2011-06-25T16:59:56 Stocha reminds me agnt Smith in Matrix 3. It's going to be a fight between us and thousands of his bots :D 2011-06-25T17:00:27 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 5.0/20110622232440]) 2011-06-25T17:04:44 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-06-25T17:05:57 if there's more_of_us: attack else: retreat 2011-06-25T17:07:37 *** Accoun has joined #aichallenge 2011-06-25T17:08:16 *** stocha has joined #aichallenge 2011-06-25T17:10:34 and: agent smith is't plainf by the rules.. 2011-06-25T17:12:16 mgngnbl ? 2011-06-25T17:12:42 what's the difference between the skill on the main rankings page and the individual profile pages? 2011-06-25T17:13:00 Okay so sto_greed_002 is a sort of anchor. 2011-06-25T17:13:21 basically if you are under in ranking, you really have to worry. if you are above .. well you have to worry also, but you are above. 2011-06-25T17:13:38 *** gobang has quit IRC (Read error: Connection reset by peer) 2011-06-25T17:13:49 StoSnake : well this a snake. yeah. 2011-06-25T17:14:15 StoRandom : another anchor. if you beat it ... you have beaten the reference weakest bot i could imagine :) 2011-06-25T17:14:26 (not that i can't make weaker bot, but well) 2011-06-25T17:14:38 *** gobang has joined #aichallenge 2011-06-25T17:14:46 So we are left with three others 2011-06-25T17:15:25 Sto03RepliCrush currently try to run shy bot code. It happens that this my highest ranked bot >_< 2011-06-25T17:15:27 *** UncleVasya has quit IRC () 2011-06-25T17:15:44 the stoha02 is simply a test. 2011-06-25T17:16:15 i wanted to put in there an aggressive bot, to see if it would disturb the strong bots. 2011-06-25T17:16:47 then the stocha bot, i waiting for something i'd like to put in first place of the ranking. Wich i haven't made my mind upon yet. 2011-06-25T17:18:06 Although the Sto03Repli and stoha02 i wanted to have as anchor. 2011-06-25T17:18:41 for some strange reason, I feel like my bot is red in 95% of its games 2011-06-25T17:18:55 davidjliu: it's always red. 2011-06-25T17:19:02 davidjliu: if you look from your profile. 2011-06-25T17:19:18 basically, the bot you look at the profile of, will be red. 2011-06-25T17:19:49 aah, ok.. That makes sense now 2011-06-25T17:19:54 so if you want to see your bot not red, you find one of your game, and you look at one of the opponent in there, and from there, if you find the same game, you won't be red. 2011-06-25T17:20:13 it happens, that i have trouble seeing red ... so it's not easy for me. 2011-06-25T17:20:21 I'm surprised that my bot's still holding up though... I think it's rather silly 2011-06-25T17:20:23 whenever i'm interested in a bot, i just can't see it's ants :p 2011-06-25T17:20:28 heh :) 2011-06-25T17:20:36 davidjliu: i think it was down to like 8th place 2011-06-25T17:21:00 and probably because of some strange interaction, it now goes as high as 3rd from time to time :) 2011-06-25T17:21:16 it has a nack at being right where it needs to :p 2011-06-25T17:21:36 my hypothesis though, is that it has started to learn the game by itself. 2011-06-25T17:21:47 so in a few days, it should be ranked 1st :) 2011-06-25T17:22:03 lol... I'm quite impressed with FlagCapper and Sir Macelon though 2011-06-25T17:22:09 and in a few month, it should have control of the whole internet. 2011-06-25T17:22:20 i think both fight. 2011-06-25T17:22:25 I'll have to upload a new bot to replace it then 2011-06-25T17:22:29 i didn't look to hard at FlagCapper though. 2011-06-25T17:22:37 yeah 2011-06-25T17:22:48 <_flag> Well I'm disappointed in mine ATM, so I'm glad you like it :P 2011-06-25T17:22:51 davidjliu: yeah, if we don't want to loose man kind, it would be advisable :p 2011-06-25T17:23:12 _flag: i'm disapointed at it too. why is it not first ? 2011-06-25T17:23:29 <_flag> It was for awhile, until macelon submitted a new version 2011-06-25T17:23:45 the funny thing about my bot is that it would currently actually work under the rules where a different engine was generated for each ant 2011-06-25T17:23:49 <_flag> Now it's crashing and timing out a bit too often, but I actually should be winning most of my games against macelon 2011-06-25T17:24:13 _flag: would you ? 2011-06-25T17:24:21 _flag: yeah, I watched one of those games... it was sad 2011-06-25T17:24:27 can you point me to a game ? 2011-06-25T17:24:33 <_flag> See this game for example: http://aichallengebeta.hypertriangle.com/visualizer.php?game=37812&user=7 2011-06-25T17:24:48 <_flag> At the height of my strength, my bot starts losing due to a timeout problem 2011-06-25T17:24:59 <_flag> I think I've fixed it now though, but I'm not entirely sure 2011-06-25T17:25:39 <_flag> And another example: http://aichallengebeta.hypertriangle.com/visualizer.php?game=37498&user=7 2011-06-25T17:25:41 sir_macelon start is leagues above yours in this game 2011-06-25T17:25:49 let's talk about the first first :p 2011-06-25T17:26:08 <_flag> Look at the most recent game, I beat him :) 2011-06-25T17:26:28 <_flag> If I give myself the timeout games, I'm currently 4-2 against him with this bot 2011-06-25T17:26:40 <_flag> So I just need to fix some bugs 2011-06-25T17:27:24 <_flag> stocha: The starts are affected largely by where food spawns, so it's not uncommon for games to be largely influenced by luck 2011-06-25T17:27:39 <_flag> I'm not claiming I'm far better than him though, he does fighting much better than I 2011-06-25T17:28:02 *** gobang has quit IRC (Read error: Connection reset by peer) 2011-06-25T17:29:35 _flag: i really simply understand nothing at those game. It seems you do fighting. But really i don't understand why it works so well for your isolated ants :p 2011-06-25T17:29:57 bah bot writing 2011-06-25T17:30:05 maps are the current interest imo :P 2011-06-25T17:30:27 davidjliu, _flag, stocha: any ideas on making maze like maps? 2011-06-25T17:30:45 the best idea i've had is random walks that do random lengthed linear lines with a bit of noise 2011-06-25T17:30:52 but it's not very maze generation like 2011-06-25T17:31:36 antimatroid: the thing is, you need dead-ends. 2011-06-25T17:31:49 but McLeopold had a maze generator. why not to use this ? 2011-06-25T17:32:01 i'm not sure what he made it in 2011-06-25T17:32:07 it would be good to have it all in the one file 2011-06-25T17:32:53 maybe that : get a tile. make a maze inside it. (with one in, one out, and a lot of dead ends). then connect the different tiles in/out 2011-06-25T17:33:00 dead ends could happen from the walks back tracking though 2011-06-25T17:33:05 or i could randomly kill off walks 2011-06-25T17:33:41 i think maze are good, for complex branched dead ends. And for the few connex path. 2011-06-25T17:33:54 like there should be only one way to round the map for each directions. 2011-06-25T17:34:00 the problem with making a "perfect maze" is that the players are going to be in separate regions (need to connect them somehow?) and the walls are going to all be straight 2011-06-25T17:34:04 i'd prefer noisy walls 2011-06-25T17:34:09 so maybe you do just that. build the two round ways. 2011-06-25T17:34:41 then generate arbitrary tree, that do not overlap (all dead ends) that grow from the initial connex cross. 2011-06-25T17:34:52 i want everything to be connected with more than one path 2011-06-25T17:34:58 obviously, the initial cross should be twisted also. 2011-06-25T17:35:04 you should be able to navigate your way around stuff if need be 2011-06-25T17:35:10 what is everything ? 2011-06-25T17:35:16 starting points ? 2011-06-25T17:35:30 not everything, most, yes starting points 2011-06-25T17:35:37 i don't mind little caves with one entrance 2011-06-25T17:35:39 then really, with the map so small, and if you also want to avoid blocking ... there is not much room for maze. 2011-06-25T17:35:44 but i want them to be wide enough to be non-blockable 2011-06-25T17:36:01 at least, you can still grow dead end fractal trees here and there. 2011-06-25T17:36:11 hm... I like the idea of mazes more in the sense of well-defined passageways 2011-06-25T17:36:16 if i enfore the rule that an nxn block can traverse the map, then i can still get 1 width openings, but i don't get long really thin corridors etc. 2011-06-25T17:36:18 but not so much as in mazes that are difficult to navigate 2011-06-25T17:36:20 i quite like that 2011-06-25T17:36:46 in my own maps, there are a lot of dead ends. In the server maps, there are very very few. 2011-06-25T17:37:32 i think that some long twisted tiny corridor would be fun. 2011-06-25T17:37:38 if only to loose the bots a bit :) 2011-06-25T17:38:28 the map are so small, that twisting is a good way to make them a little bigger. 2011-06-25T17:38:52 try to maximize the distance between the starting positions maybe. 2011-06-25T17:39:22 *** amstan has joined #aichallenge 2011-06-25T17:39:22 *** ChanServ sets mode: +o amstan 2011-06-25T17:39:51 stocha: http://pastebin.com/YTtYSCFD 2011-06-25T17:40:05 so : you generate 1000 000 map, and you keep those with the maxim distance between starting place :p 2011-06-25T17:40:12 a 2x2 block can traverse all land squares of that map, while there still existing 1 width corridors in places 2011-06-25T17:40:32 stocha: i have a variable min_starting_distance in the map generator 2011-06-25T17:40:47 it forces the ants to be at least that far appart without any water even being considered 2011-06-25T17:41:01 not much room for twists though. 2011-06-25T17:42:23 twists? 2011-06-25T17:42:30