2011-06-08T00:06:33 *** davidjliu has quit IRC (Quit: Page closed) 2011-06-08T00:07:39 *** janzert has left #aichallenge 2011-06-08T00:07:47 *** janzert has joined #aichallenge 2011-06-08T00:32:31 *** contestbot has joined #aichallenge 2011-06-08T00:48:16 *** contestbot has joined #aichallenge 2011-06-08T01:02:52 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-06-08T01:28:55 aichallenge: janzert epsilon * r2aba547 / manager/mass_skill_update.py : Short script to quickly update skill information for games missing it - http://bit.ly/jbMMqt 2011-06-08T01:52:07 *** Ice_Harley has joined #aichallenge 2011-06-08T01:55:21 *** Dlayne has joined #aichallenge 2011-06-08T01:57:45 *** Palmik has joined #aichallenge 2011-06-08T02:17:00 *** locutus2 has joined #aichallenge 2011-06-08T02:25:39 aichallenge: Marco Leise epsilon * r809ca0c / (12 files in 4 dirs): visualizer: playback of local games in progress - http://bit.ly/jExdPE 2011-06-08T02:29:00 *** jmpespxoreax has quit IRC (Ping timeout: 240 seconds) 2011-06-08T02:30:20 *** mcstar has joined #aichallenge 2011-06-08T02:30:30 *** amstan has quit IRC (Remote host closed the connection) 2011-06-08T02:46:00 *** stocha has joined #aichallenge 2011-06-08T02:47:01 it would be interesting to see how things go with 1vs1 also. Especially because there are not a lot of bots right now. Wen can then better understand how the multiplayer assymetry affect the game. 2011-06-08T02:48:33 *** stocha has quit IRC (Client Quit) 2011-06-08T02:53:27 *** locutus2 has quit IRC (Ping timeout: 250 seconds) 2011-06-08T02:54:17 *** locutus2 has joined #aichallenge 2011-06-08T02:56:51 *** Kingpin13 has joined #aichallenge 2011-06-08T03:01:11 *** locutus2 has quit IRC (Ping timeout: 240 seconds) 2011-06-08T03:18:49 *** aerique has joined #aichallenge 2011-06-08T03:22:50 stocha: the scoring doesn't work for 1v1 2011-06-08T03:22:53 we've been over this 2011-06-08T03:23:24 mcstar: oh god, tired! 2011-06-08T03:24:12 *** boegel has joined #aichallenge 2011-06-08T03:25:37 *** okay is now known as okayzed 2011-06-08T03:25:39 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-06-08T03:29:46 *** Eruonen has joined #aichallenge 2011-06-08T03:30:13 *** Eruonen has quit IRC (Client Quit) 2011-06-08T03:36:05 *** sir_macelon has joined #aichallenge 2011-06-08T03:36:40 hi 2011-06-08T03:44:09 *** locutus2 has joined #aichallenge 2011-06-08T04:02:18 *** berak has joined #aichallenge 2011-06-08T04:09:59 *** qacek has quit IRC (Quit: WeeChat 0.3.5) 2011-06-08T04:22:27 *** Eruonen has joined #aichallenge 2011-06-08T04:31:26 *** olexs has joined #aichallenge 2011-06-08T04:56:09 *** Naktibalda has joined #aichallenge 2011-06-08T04:57:07 *** locutus2 has left #aichallenge 2011-06-08T04:57:15 *** locutus2 has joined #aichallenge 2011-06-08T05:11:29 *** UncleVasya has joined #aichallenge 2011-06-08T05:14:05 *** olexs has quit IRC (Quit: Leaving.) 2011-06-08T05:18:26 *** stocha has joined #aichallenge 2011-06-08T05:18:50 i can't seem to get to the log 2011-06-08T05:18:58 i can't get to the log. 2011-06-08T05:19:09 chat log i mean. 2011-06-08T05:22:33 you're a log :P 2011-06-08T05:22:53 am i ? 2011-06-08T05:22:59 any hope of a tcp server yet ? 2011-06-08T05:23:08 don't think so sorry 2011-06-08T05:23:10 some official words on the topic ? 2011-06-08T05:23:32 i really think a tcp server should be an official goal. 2011-06-08T05:23:41 my guess is that mcleopold is the only person currently up to working on it, and he's busy with other stuff 2011-06-08T05:23:50 it is a goal, but not a top one atm i don 2011-06-08T05:23:51 t think 2011-06-08T05:23:56 it will happen 2011-06-08T05:24:10 so there will be a tcp server before the end of the beta. good then. 2011-06-08T05:24:21 i didn't say that 2011-06-08T05:24:30 it would be nice, but i'm not claiming anything 2011-06-08T05:25:03 what about the issus of top bots crashing on start 50% of the time ? (especially the java bot seems quite strong seeing they are in the top 10 and crash half of the time :p ) 2011-06-08T05:25:21 Will there be a beta with a lot of 1vs1 matches ? 2011-06-08T05:25:33 are at least some. 2011-06-08T05:25:42 the game isn't 1v1 2011-06-08T05:25:49 the scoring doesn't work well with 1v1 2011-06-08T05:26:02 it 1 vs Many. many raging from 1 to 10. 2011-06-08T05:26:26 atm we can handle 1-25 enemies 2011-06-08T05:26:27 what makes you think the scoring work well with one vs 2+ ? 2011-06-08T05:26:41 but i suspect it'll be more like 3-5 enemies or so in the end 2011-06-08T05:26:57 why does the jvm always crash ? 2011-06-08T05:27:00 i can't say it will, but i can say it doesn't work for 1v1 2011-06-08T05:27:10 because the beta server doesn't have enough memory? 2011-06-08T05:27:19 how does it not work with 1vs1 ? 2011-06-08T05:27:21 i think that's what i've seen be said, i'm in the middle of exams atm :P 2011-06-08T05:27:29 why not start with 3 player games then ? 2011-06-08T05:27:38 3 player games i would be fine with 2011-06-08T05:27:57 ah you are not active then :p 2011-06-08T05:28:11 no, i have been watching mostly for a month or so 2011-06-08T05:28:23 how is the exam doing fine ? 2011-06-08T05:28:32 handed the first one in today 2011-06-08T05:28:42 happy ? 2011-06-08T05:28:45 confident 2011-06-08T05:28:51 i thought it went not too well, but then i found out i was only meant to do 4/5 questions, so i'm much more confident 2011-06-08T05:28:58 cause then my work question just gets thrown out 2011-06-08T05:29:08 i enjoyed being able to study for an exam while i took it 2011-06-08T05:29:16 worst** 2011-06-08T05:29:22 ah okay :) 2011-06-08T05:29:27 i was wondering what you meant :p 2011-06-08T05:29:36 48 hour take home exams <3 2011-06-08T05:29:46 i have another one of those next week, then a normal 3 hour exam 2011-06-08T05:29:48 You do some sport inbetween ? 2011-06-08T05:29:57 sport? 2011-06-08T05:30:07 i play (field) hockey, and ride a bike for commuting/wherever 2011-06-08T05:30:10 yeah sport. football, tenis, water crawling etc. 2011-06-08T05:30:16 except today, it's was windy, cold and raining, so i drove 2011-06-08T05:30:28 it's said that it help to do some sport, before the exams. 2011-06-08T05:31:04 I did a lot of jungling before the exams :). When i cared. 2011-06-08T05:31:29 i usually get slack and then have to study on the night before 2011-06-08T05:32:42 i've read it's not the best thing to do. Yet a lot of sucessfull exam people do that. 2011-06-08T05:33:05 probably not the best, but it's the path of least amount of work :P 2011-06-08T05:33:17 maybe exams are simply not worth optimizing too much the sucess rate :p 2011-06-08T05:33:34 yeah, exams are a painful necessity 2011-06-08T05:33:51 that seems weird then. 2011-06-08T05:33:54 what are they for ? 2011-06-08T05:34:13 testing ones knowledge of what was taught in a class, and knowing that you're testing that student 2011-06-08T05:35:00 i don't like exams that the average person struggles to finish on time, or exams where it's just about memorising a bunch of crap, and multiple choice questions ridiculous 2011-06-08T05:35:09 but otherwise they're not too bad 2011-06-08T05:36:04 *** olexs has joined #aichallenge 2011-06-08T05:36:18 it's probably even worst not to have them. But it doesn't sound like very much effort was put into the subject. Although it seems quite important. 2011-06-08T05:37:01 meh, i find if one goes to lectures and does the assignments one learns the material 2011-06-08T05:37:10 Millions if not billions of people will be spending fifteen years of their life strugling with exam logic. 2011-06-08T05:37:18 subjects i haven't cared about i just haven't gone to class for and done everything hte night before 2011-06-08T05:37:29 i usually don't do well on those classes, but i get through 2011-06-08T05:37:54 Was that the kind of behavior the exam builder are trying to encourage ? 2011-06-08T05:38:12 depends on who writes the exam 2011-06-08T05:38:28 i guess some simply do that to earn their salary :p 2011-06-08T05:38:50 i'm at a small uni, so the lecturers are all actually pretty keen on the teaching side 2011-06-08T05:38:53 which is nice 2011-06-08T05:38:55 The world is a ship, with no one at the wheel. really. 2011-06-08T05:39:18 they are very concerned with the subject they teach ? 2011-06-08T05:39:19 bad thing is not much selection for pure maths, we've got one poor guy doing 3 classes for us this year on top of his undergrad classes 2011-06-08T05:39:37 on that ship comment, you know of swarm intelligence yes? 2011-06-08T05:39:48 i have heard of it. obviously. 2011-06-08T05:39:51 i like to attribute a lot of the worlds problems to what i call swarm stupidity 2011-06-08T05:40:09 the collective self interested actions of many individuals leading to worse outcomes for all 2011-06-08T05:40:14 antimatroid: can you show some questions or tasks examples? 2011-06-08T05:40:21 and also about large group of human being dumness. Panic dynamic and such. 2011-06-08T05:40:46 It seems that when you put together dum ant or fish, you end up with some emerging cleverness. 2011-06-08T05:41:05 and when you put together quite intelligent human being, you end up with emerging dumness. 2011-06-08T05:41:38 maybe this is an oversimplification. 2011-06-08T05:42:40 because there are some kind of clerverness to large group of human being. It is just not in par with expectations. And technological power. 2011-06-08T05:44:16 *** antimatroid has quit IRC (Ping timeout: 240 seconds) 2011-06-08T05:46:46 *** antimatroid has joined #aichallenge 2011-06-08T05:46:50 bah my internet died :( 2011-06-08T05:47:01 pfff i hate windows 2011-06-08T05:47:16 yeah, i'm not going back to windows 2011-06-08T05:47:23 i had to clean them all 2011-06-08T05:47:45 anyone want to pastebin from when i last said something for me? 2011-06-08T05:47:49 logs are down D: 2011-06-08T05:48:48 Frontier: I think that dark-blue color for ants is really ugly: http://aichallengebeta.hypertriangle.com/visualizer.php?game=17625 2011-06-08T05:49:08 aerique: are you there? 2011-06-08T05:50:11 antimatroid: can you show some questions or tasks examples? 2011-06-08T05:50:20 *** locutus2 has quit IRC (Ping timeout: 240 seconds) 2011-06-08T05:50:22 It seems that when you put together dum ant or fish, you end up with some emerging cleverness. 2011-06-08T05:50:24 stocha: got it :P 2011-06-08T05:50:30 cheers 2011-06-08T05:50:33 mcstar: yup (almost lunch time though :) 2011-06-08T05:50:40 and when you put together quite intelligent human being, you end up with emerging dumness. 2011-06-08T05:50:47 UncleVasya: examples of what? :P 2011-06-08T05:51:08 just a quick question: how did you compile your bot? i uploaded mine but it cant compile 2011-06-08T05:51:19 are you on windows? 2011-06-08T05:51:24 to whom the question mcstar ? 2011-06-08T05:51:27 you're lisp aye? 2011-06-08T05:51:31 aerique: http://aichallengebeta.hypertriangle.com/profile.php?user=17 2011-06-08T05:51:31 mcstar: check MyBot.lisp in the starter package 2011-06-08T05:51:33 of your exams tasks 2011-06-08T05:51:50 oh, i don't have the exam questions, i'll find an assignment question 2011-06-08T05:51:51 i did, it isnt really any different though 2011-06-08T05:51:57 mcstar: the worker does "sbcl --script MyBot.lisp to compile it 2011-06-08T05:52:02 mcstar: okey, i'll take a look 2011-06-08T05:52:06 it work locally perfectly 2011-06-08T05:52:09 s 2011-06-08T05:53:49 is it possible that the worker doesnt update my files properly? 2011-06-08T05:54:02 mcstar: it might just be crapping out because it sees the compile warnings 2011-06-08T05:54:09 Let x be a non-zero element of a torsion-free abelian group G, and let _* = {g \in G: ng \in for some n \in Integers}. 2011-06-08T05:54:09 (i) Prove that _* is a pure subgroup of G 2011-06-08T05:54:10 (ii) Prove that _* is isomorphic to a subgroup of the rationals 2011-06-08T05:54:17 but i supressed it for the last upload 2011-06-08T05:54:17 mcstar: can you try muffling them? 2011-06-08T05:54:20 oh 2011-06-08T05:54:21 i did 2011-06-08T05:54:37 for the 5th submission it was supposed to be muffled 2011-06-08T05:54:47 but it does show them 2011-06-08T05:54:58 yep :D 2011-06-08T05:55:22 i added (declaim (sb-ext:muffle-conditions sb-ext:compiler-note)) 2011-06-08T05:55:25 what about redirecting stderr to stdout?: (setf *error-output* *standard-output*) 2011-06-08T05:55:33 in te compile script MyBot.lisp 2011-06-08T05:55:42 UncleVasya: have you done much maths? 2011-06-08T05:55:46 why would that help? 2011-06-08T05:55:57 we needed to do that for planet wars 2011-06-08T05:56:10 ill try 2011-06-08T05:56:29 oh and the memory limit wasnt a porblem? 2011-06-08T05:56:29 i assume (but haven't checked in the code) that when the worker sees output in stderr it assumes it fails 2011-06-08T05:56:46 *** Dlayne is now known as InsaneMalkavian 2011-06-08T05:57:02 ofcourse it is. allocate less memory ;-) 2011-06-08T05:57:10 antimatroid: oh, Good Lord. It's scaring me even more then mcstar's function callbacks and return values... 2011-06-08T05:57:21 *than 2011-06-08T05:57:23 :P 2011-06-08T05:57:48 mcstar: https://github.com/aichallenge/aichallenge/blob/epsilon/worker/compiler.py#L229 2011-06-08T05:57:51 maths <3 2011-06-08T05:57:51 aerique: what about those messages of heap exhaustion? 2011-06-08T05:58:13 i see you had them too 2011-06-08T05:58:20 for the first submissions 2011-06-08T05:58:21 mcstar: https://github.com/aichallenge/aichallenge/blob/epsilon/worker/compiler.py#L320 2011-06-08T05:59:11 mcstar: your bot only gets so much memory to use on the server, we were trying to find a good value for --dynamic-space-size 2011-06-08T05:59:15 sbcl says that it these options can be saved to the memory image 2011-06-08T05:59:32 and how much is it now? 2011-06-08T05:59:54 mcstar: yeah, but i'm not sure if that's available in the sbcl used by the challenge (which is 1.0.40?) 2011-06-08T06:00:02 mcstar: 256mb i think 2011-06-08T06:00:39 it isnt that old is it? 2011-06-08T06:00:44 i have 1.0.48 2011-06-08T06:00:49 and thats up to date 2011-06-08T06:01:21 all right, ill try redirecting 2011-06-08T06:01:30 thx aerique bon apetite 2011-06-08T06:01:57 mcstar: it's not too old but the option to save the d-s-s limit to your image is pretty recent as well 2011-06-08T06:03:15 *** moondust has joined #aichallenge 2011-06-08T06:05:57 UncleVasya: than: i love her more than him, then: i eat something then ill drink too 2011-06-08T06:06:59 yep 2011-06-08T06:09:24 janzert: you had your own beta server up ? 2011-06-08T06:13:12 contestbot: later tell McLeopold Please look at the line #344 of playgame.py. I believe it must be True there instead of False. 2011-06-08T06:13:12 UncleVasya: As you wish. 2011-06-08T06:16:08 @later tell janzert you had your own ranking server up ? 2011-06-08T06:16:08 stocha: Aye, aye, sir 2011-06-08T06:16:47 stocha: yes, he had. For one day. 2011-06-08T06:23:20 UncleVasya: what was the purpose ? to let people test their bot, or to test the server ? 2011-06-08T06:23:35 2nd 2011-06-08T06:24:00 I think it would be nice to have a place for people to test their bots also :) 2011-06-08T06:24:31 tipically, people need to test serveral version of their bots at once. Trying to sort out features from one another. 2011-06-08T06:24:36 He was testing a server technology that will be used during the official contest. 2011-06-08T06:24:39 and they need different kind of strong opposition for that. 2011-06-08T06:25:02 stocha: write your own engine 2011-06-08T06:25:28 the trouble with tcp server (beside that there is none) is that people will need a machine to host their bot. Wich some won't be able to afford. 2011-06-08T06:25:43 the contest isnt started yet 2011-06-08T06:25:45 why write my own engine, when there is already the official stuff ? 2011-06-08T06:26:20 i dont really understand whats the problem 2011-06-08T06:26:27 what if my engine doesn't conform 100% with the official one (bug on either sides). Conclusions will be wrong then. 2011-06-08T06:26:52 The testing different version of the bot problem ? 2011-06-08T06:28:17 the only real problem is the green banner/sidebar on that rust colored background 2011-06-08T06:36:56 you run your bot locally for tcp 2011-06-08T06:42:39 this is getting ridiculous 2011-06-08T06:43:03 why does it keep compiling mybot??? 2011-06-08T06:47:39 It found you bot's logic very interesting to learn. 2011-06-08T06:48:29 *** locutus2 has joined #aichallenge 2011-06-08T06:50:50 i mean it shouldnt compile it 2011-06-08T06:51:07 the instructions are to load the files to memory and then to dump the image 2011-06-08T07:03:33 it needs to compile to do that. sbcl always compiles afaik 2011-06-08T07:03:48 Success: ready to play 2011-06-08T07:03:56 at last 2011-06-08T07:04:09 killing stderr helped 2011-06-08T07:04:32 the problem is that in that version the muffling had no effect 2011-06-08T07:04:48 it worked for me though 2011-06-08T07:05:44 and yes i read sbcl`s man page and it really always compiles 2011-06-08T07:07:14 so the conclusion is that it is not a good practice to muffle the warnings, because it might not work on the server, and the user is deceived 2011-06-08T07:07:57 and it should be noted somewhere explicitely to try to redirect stderr 2011-06-08T07:08:33 aerique: how about you put it in the .md for your starter pack? 2011-06-08T07:11:12 mcstar: will do 2011-06-08T07:14:29 brrr, i hate winter 2011-06-08T07:22:55 *** guyverthree has joined #aichallenge 2011-06-08T07:22:57 good afternoon guys 2011-06-08T07:23:19 that movie is quite lame 2011-06-08T07:23:41 anyone else notied that the starter bots in the python samples use a differnt ants.py ??? 2011-06-08T07:48:45 m 2011-06-08T07:56:43 hello out there 2011-06-08T08:01:17 hello 2011-06-08T08:01:50 computerwiz used to say hello as well and nobody replies, must be frustrating 2011-06-08T08:04:21 it is quite 2011-06-08T08:04:59 it usually means we have nothing to say to that 2011-06-08T08:05:18 and i really dont know why is the different ants.py is a problem 2011-06-08T08:05:35 they are just sample bots 2011-06-08T08:07:07 george1914: are you the george? 2011-06-08T08:08:23 he is 2011-06-08T08:08:33 #7 2011-06-08T08:09:49 it's not a problem persay just that for newbies (I.E. me to python it was confusing) 2011-06-08T08:10:12 I use these challeneges to work in new languages and it was confusing 2011-06-08T08:10:22 I'm from a Straight C background :) 2011-06-08T08:17:29 http://ai-contest.com/visualizer.php?game_id=9485825 2011-06-08T08:17:30 yeah :P 2011-06-08T08:20:45 george1914: you crashed against me a LOT!? 2011-06-08T08:20:48 Today everyone remembers PW games with George :) 2011-06-08T08:21:13 http://ai-contest.com/visualizer.php?game_id=9356999 2011-06-08T08:21:14 http://ai-contest.com/visualizer.php?game_id=9434338 2011-06-08T08:22:24 *** xQuasar has joined #aichallenge 2011-06-08T08:23:38 http://ai-contest.com/visualizer.php?game_id=9118743 2011-06-08T08:23:38 http://ai-contest.com/visualizer.php?game_id=8984275 2011-06-08T08:23:39 yeah :) 2011-06-08T08:23:42 i forgot how cool the games were 2011-06-08T08:24:56 I think people would not remember Ants games like PW ones. 2011-06-08T08:26:57 I am stiil watching my bot's PW games sometimes and give a links to an interesting ones to my friend. 2011-06-08T08:28:16 Who came with idea to use Galcon as a contest game that time? 2011-06-08T08:28:24 someone on reddit 2011-06-08T08:28:40 tron was awesome too 2011-06-08T08:30:41 I will try tron if zeta comes. 2011-06-08T08:30:57 yep 2011-06-08T08:31:04 i expect a tron server to be quite popular 2011-06-08T08:31:09 it's a great place to get started 2011-06-08T08:31:14 antimatroid: Did you compete the very first contest? 2011-06-08T08:31:35 i "competed" in tron (:( bad memories) 2011-06-08T08:31:42 the contests before that were waterloo based 2011-06-08T08:31:48 Stone-paper-scissors if I wrote it correctly. 2011-06-08T08:32:00 i was sitting around 10th in tron before getting disqualified for timing out on one turn in one game 2011-06-08T08:32:12 i didn't shut up until that rule was changed for pw, it was unecessarily harsh 2011-06-08T08:32:49 really sucked working on a bot then not even getting a final ranking, especially when it looked so good :( 2011-06-08T08:33:14 *** FireFly has joined #aichallenge 2011-06-08T08:33:22 paper-rock-scissors is what i call it 2011-06-08T08:33:44 pro tip, most people always start with scissors 2011-06-08T08:34:01 that's my anecdotal evidence, and i've played it a disproportionate amount as someone who used to skate 2011-06-08T08:36:07 UncleVasya: http://csclub.uwaterloo.ca/contest/profile.php?user_id=2788 :( 2011-06-08T08:36:14 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-06-08T08:38:19 oh man, this is just cruel 2011-06-08T08:38:20 http://csclub.uwaterloo.ca/contest/profile_games.php?user_id=2788 2011-06-08T08:38:29 3 losses prior to disqualification :( 2011-06-08T08:39:26 i wish i could view the games :\, do they load for anyone else? 2011-06-08T08:40:07 *** Kingpin13 has quit IRC (Ping timeout: 250 seconds) 2011-06-08T08:41:32 *** UncleVasya has joined #aichallenge 2011-06-08T08:43:53 UncleVasya: http://csclub.uwaterloo.ca/contest/profile_games.php?user_id=2788 :( 2011-06-08T08:44:19 that was my performance in tron up to when i got disqualified 2011-06-08T08:44:26 i did that bot with a friend 2011-06-08T08:44:49 I was disconnected. Did you answered my question? Or did you at least get it? About age. 2011-06-08T08:45:08 what? 2011-06-08T08:45:27 How old are you? 2011-06-08T08:45:29 22 2011-06-08T08:45:32 how old are you? 2011-06-08T08:45:52 20 2011-06-08T08:45:52 *** delt0r_ has joined #aichallenge 2011-06-08T08:46:00 what do you do? 2011-06-08T08:47:47 *** delt0r- has quit IRC (Ping timeout: 268 seconds) 2011-06-08T08:48:03 Just ended studying in the technical school as a junior programmer. Now I'll go to the university with the same profession. 2011-06-08T08:48:18 http://ai-contest.com/visualizer.php?game_id=9518797 2011-06-08T08:48:25 aha ^^ he would have hated that 2011-06-08T08:48:36 antimatroid: Why were you disqualified frmo tron? 2011-06-08T08:48:45 i timed out on a move 2011-06-08T08:48:53 that was a disqualification in tron :| 2011-06-08T08:49:03 What? 2011-06-08T08:49:06 i didn't shut up at the start of pw until that changed, it was unecessarily harsh 2011-06-08T08:49:16 Your bot gets disqualified if it times out once?? 2011-06-08T08:49:20 it did 2011-06-08T08:49:22 not anymore 2011-06-08T08:49:24 that was archaic 2011-06-08T08:49:29 That sounds so stupid... 2011-06-08T08:49:33 and devestating 2011-06-08T08:49:43 mostly devestating 2011-06-08T08:49:53 couldn't you make a new account and resubmit the bot under a new name 2011-06-08T08:49:54 i didn't even get to see the game it crashed on :( 2011-06-08T08:50:01 not once the final contest is running 2011-06-08T08:50:08 we hadn't timed out before that 2011-06-08T08:50:13 wow 2011-06-08T08:50:15 what the hell? 2011-06-08T08:50:16 yeah 2011-06-08T08:50:27 imagine what it was like for me :P 2011-06-08T08:50:31 look at that page, 3 losses :( 2011-06-08T08:50:37 whoever put that rule in place deserves a punch in the face 2011-06-08T08:50:51 ehe, it's mostly water under the bridge now 2011-06-08T08:50:56 it's depressing to go back to the page though 2011-06-08T08:51:01 haha 2011-06-08T08:51:02 :/ 2011-06-08T08:52:24 we didn't get our bot finsihed till like a week left, it was sitting in first for a day or so :\ 2011-06-08T08:52:31 that was rather cool so late in the contest 2011-06-08T08:52:48 pretty simple in the end too 2011-06-08T08:53:02 Hiding is coll: 2011-06-08T08:53:06 http://aichallengebeta.hypertriangle.com/visualizer.php?game=17708&debug=true 2011-06-08T08:53:06 http://aichallengebeta.hypertriangle.com/visualizer.php?game=16636&debug=true 2011-06-08T08:53:10 *cool 2011-06-08T08:54:13 iteratively make a game tree, recycling what you have at the start of each turn, deleting the rest, then you have edge cases of lose/separate with less squares/draw/separate with at least as many squares/win, then otherwise rank "game states" based on the difference between the number of squares you can access first to what they can access first 2011-06-08T08:54:32 then we had a "max path" heuristic thing for if the bots were in separated regions 2011-06-08T08:55:16 if you can reach less squares, put that below draw, otherwise above etc. 2011-06-08T08:56:59 lol i added a killer logic to my bot 2011-06-08T08:58:02 Error: replaydata.ants.10.6.length should be within [833 .. 834], but was found to be 832! 2011-06-08T08:58:06 i hate this 2011-06-08T08:58:17 mcstar: is Leviathan your bot? 2011-06-08T08:58:19 this really should be fixed, 2011-06-08T08:58:19 *** BtbN has quit IRC (Remote host closed the connection) 2011-06-08T08:58:22 no 2011-06-08T08:58:30 UncleVasya: it's mine 2011-06-08T08:58:44 mine is .... imagine 2011-06-08T08:59:14 "killer" was quite sarcastic 2011-06-08T08:59:33 *** Palmik has quit IRC (Read error: Connection reset by peer) 2011-06-08T08:59:57 oh man, i got beaten by so many java bots in pw 2011-06-08T09:00:00 it selects random directions, but the mean of the selection depends upon the turn, basically it cycles through the 4 directions 2011-06-08T09:00:58 I saw one bot make circles with all his ants in endgame, dont remember who it was 2011-06-08T09:01:10 is there a snake bot yet? 2011-06-08T09:01:14 hehe 2011-06-08T09:01:15 xD 2011-06-08T09:01:21 might do it 2011-06-08T09:01:26 thought about that 2011-06-08T09:01:26 please do 2011-06-08T09:01:27 i want to see it 2011-06-08T09:01:31 olexs: Green Tea's one 2011-06-08T09:01:35 I'm still struggling to get my bot running in python 2011-06-08T09:01:35 it wont be good, but it'll be nice to watch 2011-06-08T09:01:50 python's going to be pretty slow for this 2011-06-08T09:01:52 the lanuage is so frustrating first time around 2011-06-08T09:02:02 c++ <3 :P 2011-06-08T09:02:04 guyverthree: why didnt you try lisp? 2011-06-08T09:02:21 I like how out of the top 10 bots, 7 are in C++... 2011-06-08T09:02:44 look at that http://csclub.uwaterloo.ca/contest/rankings.php :P 2011-06-08T09:02:45 i could try lisp 2011-06-08T09:03:05 antimatroid: nice :) 2011-06-08T09:03:06 top 10 was C++ for tron 2011-06-08T09:03:13 mcstar i could try lisp 2011-06-08T09:03:26 yeah sure you could :D 2011-06-08T09:03:43 I took part in the second contest, couldn't do anything in last 2 months before submission end though. finished in 400-ish place 2011-06-08T09:03:48 guyverthree: use my c++ bot! 2011-06-08T09:03:48 11th was mega1 with lisp for tron :) 2011-06-08T09:03:52 *** BtbN has joined #aichallenge 2011-06-08T09:03:52 with java... 2011-06-08T09:03:58 hehe 2011-06-08T09:04:12 aerique: yep, he had the worlds best hex bot a few years ago apparently 2011-06-08T09:04:16 i said lisp first, antimatroid 2011-06-08T09:04:22 nah uh 2011-06-08T09:04:28 (23:02:01) antimatroid: c++ <3 :P 2011-06-08T09:04:28 (23:02:05) mcstar: guyverthree: why didnt you try lisp? 2011-06-08T09:04:42 guyverthree: I'm writing in functional language first time on this contest. Until that I even didn't see any code in OCaml. 2011-06-08T09:04:49 u said u love c++ 2011-06-08T09:04:54 but i told him to try lisp 2011-06-08T09:04:57 hehe 2011-06-08T09:04:58 it was implied 2011-06-08T09:05:03 ah sure implied 2011-06-08T09:05:06 >D 2011-06-08T09:05:08 LOL my random bot obliterated a rank 25 bot - http://aichallengebeta.hypertriangle.com/visualizer.php?game=17544 2011-06-08T09:05:24 have you guys all got good food collection yet? 2011-06-08T09:05:31 I am confused by the massive amonut of brackets 2011-06-08T09:05:43 i haven't done anything except make a basic starter bot :-| 2011-06-08T09:05:44 thats parantheses not brackets 2011-06-08T09:05:46 phython has no brackets 2011-06-08T09:05:47 xQuasar: you're australian? 2011-06-08T09:05:49 which state? 2011-06-08T09:05:54 Yup! NSW - Sydney 2011-06-08T09:05:59 yopu know what i mean.... 2011-06-08T09:06:00 tassie <3 2011-06-08T09:06:09 mine can still be improved, but midgame the system seems to work fine. early game needs a lot of work 2011-06-08T09:06:24 bloody everyone is nsw, you sigh and smiley 2011-06-08T09:06:35 smiley might have been melbs 2011-06-08T09:06:39 guyverthree: give it some time, u will be able to parse it faster than python or c 2011-06-08T09:07:11 antimatroid also my bot is completely random :P i'm gonna work on food collection tomorrow thoguh 2011-06-08T09:07:15 i've got my food collection algorithm worked out 2011-06-08T09:07:23 i just need to work out how to do fighting 2011-06-08T09:07:28 antimatroid, keeping track of the changes is a bitch when all the information is sent every turn 2011-06-08T09:07:44 worst is the implicit food information 2011-06-08T09:08:05 antimatroid: what do you mean? 2011-06-08T09:08:11 anyways i'm off 2011-06-08T09:08:12 antimatroid: Every my food do BFS to find closest ant. I see some ways to optimise food gathering but now I having rest and dealing with testing aplication. 2011-06-08T09:08:12 gnight 2011-06-08T09:08:18 in that, if there was food on a square last time it was visible and it no longer is the next time it becomes visible, the only way you're told this is by not being told there's food there 2011-06-08T09:08:33 UncleVasya: That's the way I'm planning to do my food collection! 2011-06-08T09:08:38 UncleVasya: BFS from food :) 2011-06-08T09:09:16 UncleVasya: are you restarting the search every time you collect a piece of food? 2011-06-08T09:09:23 and are you reusing already used ants? 2011-06-08T09:09:23 antimatroid: hi, yes, i was the player george - i didnt think i got that many timeouts... had timechecks almost everywhere... i guess i was just unfortunate those times ... 2011-06-08T09:09:27 well i'm off lunch now see you all later :) 2011-06-08T09:09:27 antimatroid: in a way, the removal of 'r' is justified though - because it might be out of sight when it gets removed 2011-06-08T09:09:39 and thus if 'r' was still there yo'd be effectively seeing out of your fog of war 2011-06-08T09:09:50 *** guyverthree has quit IRC (Quit: Page closed) 2011-06-08T09:09:54 xQuasar: you wouldn't have received the 'r' line until the square was visible 2011-06-08T09:10:03 bfs from food makes most sense. however, it needs to be optimized; otherwise long distance from food to an ant can be preferred, although the ant in question might have food closer by 2011-06-08T09:10:09 george1914: have you thought much about ants? 2011-06-08T09:10:20 antimatroid: Oh. in that case, i think they should've kept 'r' 2011-06-08T09:10:24 *** xQuasar has quit IRC () 2011-06-08T09:10:41 i argued for 'r', but meh, just extra messy programming 2011-06-08T09:11:01 olexs: there's better than simple bfs possible 2011-06-08T09:11:06 i've described this a few times 2011-06-08T09:11:48 antimatroid: i know, i also use something a bit different, that results in optimal distirbution of closest ants to all visible food 2011-06-08T09:11:55 antimatroid: a little - my current bot just chases food it sees, and move randomly otherwise, with weights to make it explore, spread out etc ... 2011-06-08T09:12:03 antimatroid: Currently yes. In future I'm planning not to stop but take attention to any food I'll find while do BFS and also mark ants in special way. So while doing one BFS I'll be able to deal with few pieces of food and few ants. 2011-06-08T09:12:17 start with an "availability" grid, with ant locations at 0 and other squares as not available, then iteratively do a bfs from all non-collected food squares simultaneously until you find the ant that can get within the spawn radius of a food square fastest, update the availability information of that ants location, it's target location and the food location, remove the food and continue collecting 2011-06-08T09:12:40 i have sketched a plan for a good one ... but want to wait and see what the final settings are on eg maps, combatradius, etc 2011-06-08T09:12:50 something like that, yes 2011-06-08T09:12:57 we're going for lots of variation on maps 2011-06-08T09:13:08 dont want to craft specific rules for fighting, and then have the combatradius change ... 2011-06-08T09:13:25 a good bot should work with any combat- or visibility-radius, taking those into account 2011-06-08T09:13:52 yes, but a good bot might not necessarily win with the parameters used in the final contest 2011-06-08T09:13:59 ok but eg corridors/maze would be worth having specific combat tactics for ... 2011-06-08T09:14:02 optimising for parameters is something people want to do 2011-06-08T09:14:08 ignoring it just places luck into the mix 2011-06-08T09:14:08 also true 2011-06-08T09:14:17 eg set 'trap' at some point - incoming ants always die 2011-06-08T09:14:27 george1914: have you considered playing around with map generation? 2011-06-08T09:14:32 Also I'm planning to pay attention to the difference beetwen distances of ants and food so if one ant can gather few pieces of food on his way, these pieces are marked processed and they does not need to search for another ants. 2011-06-08T09:14:36 but if those maps are rare/absent, then no need to spend time on that 2011-06-08T09:14:37 i made the current map generator, we're planning another kind of symmetry 2011-06-08T09:14:47 and if corridors become a problem, i will remove them 2011-06-08T09:14:55 for the sake of an interesting contest 2011-06-08T09:15:00 hm, map generation? no 2011-06-08T09:15:06 it's one of the funnest parts 2011-06-08T09:15:12 :) 2011-06-08T09:15:22 i was surprised when i could tesselate with odd/prime numbers of players 2011-06-08T09:15:32 again - maybe i can figure out the map type, if i know size and that it has symmetry ... 2011-06-08T09:15:49 the plan is to make that hard for you 2011-06-08T09:15:50 can you do Penrose Tilings also? :) 2011-06-08T09:15:58 aha 2011-06-08T09:15:58 hiding hte number of players and having more than one kind of symmetry 2011-06-08T09:16:10 * antimatroid goes to wikipedia 2011-06-08T09:16:20 i tried doing the water with perlin noise, but never got anything nice 2011-06-08T09:16:30 atm i just use a bunch of random walks to carve land out of water 2011-06-08T09:16:40 basically, the competition looks great - most rule sets look ok, but i'd like to know which one is being used before i write too much code ... 2011-06-08T09:16:45 corridors seems to be interesting to me. Bots will be able to defend only few ways to their territory. 2011-06-08T09:17:08 you should have been here when there was like 5 battle rules under consideration, let alone radiuses :P 2011-06-08T09:17:22 i did watch that on the forums ... :) 2011-06-08T09:17:29 we didn't get the best option 2011-06-08T09:17:34 apparently it wasn't intuitive enough 2011-06-08T09:18:31 that was, let enemies(A) = no. enemies within attack radius of ant A, A lives if |enemies(A)| < |enemies(B)| for all B \in enemies(A) 2011-06-08T09:18:32 in case the whole map are the corridors then it might be interesting to see how ants are filling them. 2011-06-08T09:18:50 it seems complicated, but once you understand it, it's the ultimate battle rule 2011-06-08T09:18:56 george1914: theres a limit of 2MB for an upload, pls dont exceed that limit with your code 2011-06-08T09:19:22 the one that i am most unsure about is the multiplayer aspect ... in eg risk - you gang up on the strong players, maintain balance of power - the best player often loses ... will things like that happen here also? :) 2011-06-08T09:19:28 that would be hard to plan for 2011-06-08T09:19:37 one option if corridors become a problem is doing random walks with random nxn blocks, so not as many corridors exist 2011-06-08T09:19:54 george1914: i don't expect people to be able to gang up that well 2011-06-08T09:20:11 and besides, you want to take out the person who is likely to possibly rank higher than you 2011-06-08T09:20:26 the battle rules allow for cooperation, but i don't plan to count on it 2011-06-08T09:20:52 it's going to be hard enough that you have n-players and non-zero sum, so you can't assume they're rational like you do with normal backward induction for example 2011-06-08T09:20:54 wont work like that 2011-06-08T09:21:22 antimatroid: no, I do not reuse ants for another search. I mark them busy until the end of turn. But if on the next turn some another food will find this ant faster than his previous destination, he can change his task. 2011-06-08T09:21:27 besides, game trees are pretty much out anyway, too many damn ants 2011-06-08T09:21:31 mcstar: hi, did i exceed that? i really thought not ... the zip files are a few k 2011-06-08T09:21:35 you were talking about languages, but is there some kind of flexibility about the language used or not ? 2011-06-08T09:22:01 UncleVasya: take a look at my description of mine, it's a pretty solid food collection algorithm 2011-06-08T09:22:05 * mcstar was joking, because you were worried about writing too much code 2011-06-08T09:22:10 I mean, if I use python rather than c, which is slower, does it count for the timeout or resources ? 2011-06-08T09:22:15 you can make minor tweaks in different places depending on how you want your bot to behave 2011-06-08T09:22:17 ah, i see, thanks :) 2011-06-08T09:22:37 george1914: i updated the c++ bot the other day too btw 2011-06-08T09:22:42 it was still processing 'r' lines 2011-06-08T09:22:54 also, any hate on how i've done that? 2011-06-08T09:23:04 ignoring the fact that i'm a communist programmer (everything is public) 2011-06-08T09:23:19 antimatroid: heheh, I've copypasted your description to the notepad a few days ago :) 2011-06-08T09:23:29 But didn't read it yet. 2011-06-08T09:23:37 dr-: all languages are treated the sae way 2011-06-08T09:23:39 same 2011-06-08T09:24:02 you dont get extra time just because you use python 2011-06-08T09:24:07 if that what u meant 2011-06-08T09:24:11 okay, like I tought ... 2011-06-08T09:24:17 you pick your language, you get the limitations/benefits that come with it 2011-06-08T09:24:19 that's the deal 2011-06-08T09:24:28 yeah, it was the question 2011-06-08T09:24:34 understandable 2011-06-08T09:24:37 yep, just stating hte official opinion :) 2011-06-08T09:25:09 but, my favourite saying, you get to pick your language then go to town doing whatever the hell you want :) 2011-06-08T09:25:13 and now comes the part when antimatroid asks you to check out his c++ starter bot 2011-06-08T09:25:24 it seemed ok to me, but then my c++ is not perfect ... i maybe did change some things, eg make some fns const, remove 'using namespace std', but mainly my own preferences 2011-06-08T09:25:28 mcstar: i shall convert everyone! 2011-06-08T09:25:39 godspeed 2011-06-08T09:25:52 All hail Stroustrop! 2011-06-08T09:25:53 yeah, i'm far from being a good programmer 2011-06-08T09:25:57 *** davidjliu has joined #aichallenge 2011-06-08T09:26:08 dr-: You can write a faster code in some language only if you can write a faster code in this language. 2011-06-08T09:26:19 i have my own "starter bot" that's a bit more complicated than the starter package one 2011-06-08T09:26:38 the main target for those are people who aren't capable/not willing to write a bot from scratch themselves 2011-06-08T09:26:53 i expect any serious competitor to pretty much start from scratch 2011-06-08T09:26:53 it would be interesting to compare c++ and python (for example) with the same behaviour 2011-06-08T09:27:27 antimatroid: I think Greedy bot is pretty good for these people to look at. 2011-06-08T09:28:16 the idea is to have it so easy that a complete beginner programmer (that can compile and run programs) can get started, yet still have enough depth that it's interesting for people at the top 2011-06-08T09:28:54 antimatroid: what does your starter bot that the basic one don't ? 2011-06-08T09:29:07 it doesn't do anything strategy wise more 2011-06-08T09:29:15 i just structure things different for myself 2011-06-08T09:29:19 more opeators, more structs etc. 2011-06-08T09:29:40 and i like to put in little hax functions that make my life easier 2011-06-08T09:29:58 Here does everything that starter bot does but in very complicated mathematician way:) 2011-06-08T09:30:21 His bot proofs theoretically every move he wants to do. 2011-06-08T09:30:28 you can all thank me for not having to use godawful x,y indexing too 2011-06-08T09:30:34 that would have been ever so painful 2011-06-08T09:30:50 One proof = 2-3 log pages 2011-06-08T09:31:02 UncleVasya: the state space for games is RIDICULOUSLY large, like larger than I would have though before playing around with ai game 2011-06-08T09:31:03 ss 2011-06-08T09:31:25 however, my honours thesis this year is on game theory :P 2011-06-08T09:31:27 but theory side 2011-06-08T09:32:08 godawful x,y indexing? 2011-06-08T09:32:34 you know when you write on a page/in a document you go to the line and then write/read along that? 2011-06-08T09:32:52 if you use x,y indexing, you need to transpose your grid every time you read/write it to file 2011-06-08T09:32:57 which is incredibly confusing 2011-06-08T09:33:18 if you use row,col, you never ever have you think about that 2011-06-08T09:33:23 it's just always consistent 2011-06-08T09:33:52 lol 2011-06-08T09:34:09 besides, that's how you index matrices :P 2011-06-08T09:34:10 you can always associate y with row x with col 2011-06-08T09:34:34 and then when you output your moves? 2011-06-08T09:34:40 i called my functions with (somefun rows cols y x) 2011-06-08T09:34:42 you need to translate your 'n' values 2011-06-08T09:34:52 i would advise you just use row,col 2011-06-08T09:35:04 it might take a bit of getting used to, but then there is no confusion/stupid bugs 2011-06-08T09:35:10 x and y always confused me ... I never know which corresponds to column :/ 2011-06-08T09:35:15 it was no problem but for reasons of symmetry i now use row for y and col for x 2011-06-08T09:35:37 yes, but are you also careful of which order you use them in? 2011-06-08T09:35:47 people use x,y but row,col, they are swapped 2011-06-08T09:35:54 which gives your translation 2011-06-08T09:35:55 yeah, but my friend cannot understand my code 2011-06-08T09:36:08 dr-: success then? :P 2011-06-08T09:36:18 In my starter pack I've renamed row variable to x and col to y (because it's a usual way i'm reading coordinates) but I didn't thought about the real difference beetwen these concepts. 2011-06-08T09:36:18 natural obfuscation 2011-06-08T09:36:37 ?? 2011-06-08T09:36:42 UncleVasya: I recommend you use row,col and start thinking like that 2011-06-08T09:36:45 it's confusing as hell otherwise 2011-06-08T09:36:46 wtf its the opposite 2011-06-08T09:36:56 x,y , col,row 2011-06-08T09:36:58 that is the same 2011-06-08T09:37:04 but people use x,y then row,col 2011-06-08T09:37:14 UncleVasya just wrote it the opposite way 2011-06-08T09:37:21 oh, then that's even more confusing :P 2011-06-08T09:37:50 think of it like indexing the line, then word/character on a page 2011-06-08T09:38:28 and matrix multiplication is easy, for A.B = C, C_ij = row i of A dot col j of B 2011-06-08T09:38:38 imagine nothing, thats my bots name, only think in rows/cols, u can really get rid of x,y 2011-06-08T09:39:11 i almost went with a robot unicorn/charlie candy mountain theme with my planet wars bot 2011-06-08T09:39:25 as in i renamed everything to fit that theme, then got sick of it within 2 days and swapped back 2011-06-08T09:39:41 :D 2011-06-08T09:39:46 it would be funny to win the contest and then release the code 2011-06-08T09:39:51 and everyone be like wtf are you on 2011-06-08T09:39:55 my bot-logic file is called illogical 2011-06-08T09:40:06 after spock, he would have found it illogical... 2011-06-08T09:40:51 *** AlliedEnvy has quit IRC (Ping timeout: 255 seconds) 2011-06-08T09:41:03 oh, something you guys might not realise, if you have ants not moving, food will wait to spawn there until you move 2011-06-08T09:41:04 once you finish the code, you can always run it through an obfuscator 2011-06-08T09:41:09 so leaving ants static can be a bad idea 2011-06-08T09:41:28 WHAT?! 2011-06-08T09:41:39 that's how symmetric spawning works 2011-06-08T09:41:51 you're welcome to come up with a fairer way to do it 2011-06-08T09:42:01 Frontier`s masterplan was based on sitting duck bots 2011-06-08T09:42:24 that's right. I have a sitting duck caste in mind 2011-06-08T09:42:26 *** olexs has quit IRC (Quit: Leaving.) 2011-06-08T09:42:28 my requirements for symmetry are that a deterministic bot (including coordinates) should always draw with itself 2011-06-08T09:43:20 you mean on average right? 2011-06-08T09:43:20 I thing I understood. In my county x always means horisontal coordinate of a matrix and y lways means vertical coordinate. So x in my country = row in your and my y = col of you. 2011-06-08T09:43:20 Is that behaviour documented antimatroid? 2011-06-08T09:43:33 Frontier: not sure, sigh and i devised it a while ago 2011-06-08T09:43:46 i for one would actually be interested in doing away with spawning of food 2011-06-08T09:44:01 that way you can explore the map and have information about where no food is anymore 2011-06-08T09:44:11 we don't really want long ass games anyway right? 2011-06-08T09:44:28 hmm, not sure :p 2011-06-08T09:44:34 but i suspect it's too late to make such a change 2011-06-08T09:44:36 UncleVasya: the first part is correct, the second part is the opposite 2011-06-08T09:45:16 UncleVasya: i don't believe any mathematicians anywhere index matrices with x,y 2011-06-08T09:45:29 maybe on another planet :\ 2011-06-08T09:46:14 веслування = row = y, кол = col = x 2011-06-08T09:46:28 hehe 2011-06-08T09:46:40 i dont even try to pronounce those words 2011-06-08T09:46:50 Frontier: Where is players names stored in 0.replay? Is this a "player_info": [{}, {}, {}, {}, {}, {}, {}] ? 2011-06-08T09:46:59 кол, would probably be pronounced... col ? :) 2011-06-08T09:47:09 wtf, i have people on reddit bitching at me for using . as multiplication operation 2011-06-08T09:47:24 *** amstan has joined #aichallenge 2011-06-08T09:47:24 *** ChanServ sets mode: +o amstan 2011-06-08T09:47:44 UncleVasya: no, they are stored somewhere else in a playernames variable, take a look at the wiki. It is documented there. 2011-06-08T09:48:33 my algebra lecturer will write like 1.3.5 2011-06-08T09:48:38 or even 1.3 2011-06-08T09:48:40 So replay file never stores it? 2011-06-08T09:48:47 although to be fair, i've never seen the man write a decimal 2011-06-08T09:49:35 UncleVasya: If you run a local game, no. On the test server the names are added after the game finishes, I think. 2011-06-08T09:50:11 UncleVasya: You can download a game from the server using /game/ as the url. 2011-06-08T09:50:45 I have a replay made by local game and before open it in browser I want to input player names instead of "Player1, Player2...". I thoght I can just edit replay-file. 2011-06-08T09:50:46 There you will find the playernames you are looking for. 2011-06-08T09:51:14 I'll try it now. 2011-06-08T09:52:09 You can add them to your local replay files in the same way. Maybe you want to use a JSON programming library for that. 2011-06-08T09:53:27 I am consider 0.replay as a textfile. 2011-06-08T09:54:47 i go offline, program some logic 2011-06-08T09:54:50 bb 2011-06-08T09:54:52 *** mcstar has quit IRC (Quit: WeeChat 0.3.5) 2011-06-08T09:56:23 * Frontier wants to program some logic, too. But there is always something to do with the visualizer. 2011-06-08T09:56:33 Frontier: I've downloaded it and there are no players names in it. So this some another part of a system. Thank you. 2011-06-08T09:56:38 *** guyverthree has joined #aichallenge 2011-06-08T09:56:44 is anyone delveoping under windows ??? 2011-06-08T09:56:51 guyverthree: I am 2011-06-08T09:57:08 UncleVasya: Every replay on the test server has player names in it! I'm sure. 2011-06-08T09:58:35 I am having trouble with nearly all the starter bots prolly because i'm stupid 2011-06-08T09:59:22 I heard that you can install MacOS X on a PC with some hacks. That's interesting since I could write cross-platform apps then without Java, testing them on my own PC under Windows/Linux/MacOS 2011-06-08T09:59:24 use c++ with mingw and code::blocks 2011-06-08T09:59:27 I searched with Ctrl+f, no results. 2011-06-08T09:59:40 you can download code::blocks with mingw 2011-06-08T09:59:51 guyverthree: You'll also need Python 2.x not 3.x by the way 2011-06-08T09:59:56 antimatroid: Is a c++ starter pack already fixed? 2011-06-08T10:00:03 i have 2.7 2011-06-08T10:00:05 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-06-08T10:00:12 it's fixed except the timer hasn't changed 2011-06-08T10:00:19 the 'r' line stuff is gone 2011-06-08T10:01:01 wow, that comment could mean either ("you can download code::blocks with mingw") 2011-06-08T10:01:07 but today's engine is still sending r lines. 2011-06-08T10:01:18 download code::blocks and that will come with mingw if you pick the right download 2011-06-08T10:01:27 berak: are you sure about that? 2011-06-08T10:01:38 the one i got from git pull 2011-06-08T10:01:50 it shouldn't be 2011-06-08T10:01:56 i don't think 2011-06-08T10:02:06 btw, you can now "python playgame.py ... | java -jar visualizer.jar". But since the orders are not in the streaming replay you see no movement yet. 2011-06-08T10:02:46 In other words: this is a live game preview that I had to write another 1000 lines of code for :p 2011-06-08T10:03:23 did someone say live game preview? 2011-06-08T10:03:27 :D 2011-06-08T10:03:44 now to integrate with tcp and i will be back in heaven 2011-06-08T10:03:54 i can't bear the thought of non-live tcp games 2011-06-08T10:04:12 *** smiley1983 has joined #aichallenge 2011-06-08T10:04:13 delt0r_ <3 2011-06-08T10:04:40 (that's not me bugging you to do something again, just a general sentiment for jbotmanager) 2011-06-08T10:04:48 antimatroid: You'll need sedatives in case you are watching game without knowing result :) 2011-06-08T10:05:18 it forces you to watch the game in slow motion if you watch it live 2011-06-08T10:05:27 and it's somehow way more exciting not knowing the result 2011-06-08T10:05:38 yes 2011-06-08T10:05:50 Hi Smiley! 2011-06-08T10:06:26 Hi UncleVasya :) 2011-06-08T10:07:53 antimatroid: Actually a hard part was implementing it in a way that the engine would not stall the visualizer by sending data faster than the JavaScript would process it 2011-06-08T10:08:29 *** stocha has quit IRC (Quit: Page closed) 2011-06-08T10:08:30 ... and that's why i don't like web programming :P 2011-06-08T10:08:35 antimatroid: i'm shure it does: ants.py 385 (github) 2011-06-08T10:08:48 I have now written the parser in Java and build the JS Objects from there. 2011-06-08T10:09:01 yum 2011-06-08T10:09:05 *** McLeopold has joined #aichallenge 2011-06-08T10:09:11 any chance you want to turn it into another jbotmanager? :P 2011-06-08T10:09:15 It is no web application antimatroid. The live preview works offline only 2011-06-08T10:09:18 ie. lemme play against my bot 2011-06-08T10:10:06 antimatroid: I wont do that, no. Someone else must step up to do that. 2011-06-08T10:10:07 @later tell UncleVasya yes, it should be True, it was updated 2 minutes later, so make sure you always pull the latest code 2011-06-08T10:10:07 McLeopold: OK 2011-06-08T10:10:21 :P fair call 2011-06-08T10:10:29 McLeopold: Hi 2011-06-08T10:10:48 hi 2011-06-08T10:10:56 McLeopold: is the engine still sending 'r' lines? 2011-06-08T10:11:01 no 2011-06-08T10:11:06 well, it shouldn't 2011-06-08T10:11:07 berak: ^? 2011-06-08T10:11:17 i'm shure it does: ants.py 385 (github) 2011-06-08T10:11:40 McLeopold: Yesterday you tried to tell me something about game id-s. 2011-06-08T10:12:08 yes, playgame can take a game_id parameter to make sure you don't overwrite replay files 2011-06-08T10:12:20 ./playgame.py --help 2011-06-08T10:12:48 McLeopold: Is there a place I can read about using it and to not waste your time? Can I found on wiki pages? 2011-06-08T10:12:56 ah, ok 2011-06-08T10:14:10 aichallenge: McLeopold epsilon * rb6159f4 / (4 files): Updated engine seeding. Split into engine and player seed values that are not connected. Removed 'r' lines. Closes #161 - http://bit.ly/mvMUcy 2011-06-08T10:14:41 i'm looking at the trueskill code, trying to find out, how it's used. do i have to keep *ALL* gameresults per player to use this ? 2011-06-08T10:15:21 UncleVasya: I just saw the replay you posted a link to. Thanks, and here's another where a similar thing happens with a different opponent: http://aichallengebeta.hypertriangle.com/visualizer.php?game=16113 2011-06-08T10:20:25 smiley1983: in the starter pack method direction p1 p2 returns `Stop if locations are adjacent. I thing it should return `Stop only if p1 = p2. Because someone can have necessity to get direction to the adjacent cell. 2011-06-08T10:21:26 *** rabidus has joined #aichallenge 2011-06-08T10:23:34 UncleVasya: I'm looking into it now, on the surface it seems that it should only return `Stop if p1 = p2, about to run tests... 2011-06-08T10:24:12 aichallengebeta didn't send me a confirmation email :( 2011-06-08T10:24:49 now i'm thinking that did i wrote my email correctly :( 2011-06-08T10:25:22 rabidus: it isn't sending email yet, look for the activation link 2011-06-08T10:25:34 on the page after you submit the registration form 2011-06-08T10:25:47 ffffffffffuuuuuuuuuuu... 2011-06-08T10:25:58 i didn't notice that 2011-06-08T10:26:29 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-06-08T10:26:33 did ANYONE get that right in the first place ? 2011-06-08T10:26:41 I did 2011-06-08T10:26:52 no one except McLeo 2011-06-08T10:27:10 maybe we need the tag? 2011-06-08T10:27:26 yes, and a 2011-06-08T10:27:34 and -remove "check your email" 2011-06-08T10:27:35 rabidus: all users activated 2011-06-08T10:27:56 oh, thanks :) 2011-06-08T10:28:12 *** McLeopold has left #aichallenge 2011-06-08T10:28:27 *** UncleVasya has joined #aichallenge 2011-06-08T10:29:08 McLeopold: thanks for checking my issue - so it was my error then... bit embarassing :/ my apologies for wasting your time, i guess i'd've lost that money ... 2011-06-08T10:37:06 *** Zaphus has joined #aichallenge 2011-06-08T10:37:31 *** sir_macelon has quit IRC (Quit: Page closed) 2011-06-08T10:38:17 *** Zaphus has quit IRC (Client Quit) 2011-06-08T10:47:41 *** smiley1983 has quit IRC (Ping timeout: 252 seconds) 2011-06-08T10:53:53 *** smiley1983 has joined #aichallenge 2011-06-08T10:58:51 *** amstan has joined #aichallenge 2011-06-08T10:58:51 *** ChanServ sets mode: +o amstan 2011-06-08T11:05:25 *** locutus2 has quit IRC (Ping timeout: 260 seconds) 2011-06-08T11:07:13 smiley1983: Let's join our forces in the fight for a death with locutuses. I'll take locutus2 by myself while you will deal with locutus2_copy. Or I can cover you while you are dealing with both of them. Do you like the plan?:) 2011-06-08T11:07:36 is there a tcp server running? 2011-06-08T11:07:55 think about slowness.. :D 2011-06-08T11:08:10 *** McLeopold has joined #aichallenge 2011-06-08T11:08:22 rabidus: no,it isn't. 2011-06-08T11:08:29 ok 2011-06-08T11:12:08 rabidus: maybe there will be an official TCP-server, but no one knows. They are really busy with other things now. 2011-06-08T11:12:23 yeah np. 2011-06-08T11:12:42 https://github.com/McLeopold/TCPServer 2011-06-08T11:12:44 having fun without 2011-06-08T11:13:11 needs some work, so if you know python, you can fork it 2011-06-08T11:15:46 aichallenge: Marco Leise epsilon * r0a920e7 / (6 files in 3 dirs): visualizer: smaller fog buttons, closes #169 - http://bit.ly/kOX62R 2011-06-08T11:16:21 *** aerique has quit IRC (Quit: ...) 2011-06-08T11:17:28 *** aeter has joined #aichallenge 2011-06-08T11:21:24 *** guyverthree has quit IRC (Quit: Page closed) 2011-06-08T11:21:54 *** Ice_Harley has quit IRC (Ping timeout: 252 seconds) 2011-06-08T11:27:02 UncleVasya: I had a look at the direction function - it's returning two directions, one for the N-S axis and one for the W-E axis. So for an adjacent tile to the East, it will specify `Stop, `E 2011-06-08T11:27:41 The `Stop representing the N-S axis 2011-06-08T11:28:21 ok, exuse me. 2011-06-08T11:28:38 *** janzert has quit IRC (Ping timeout: 244 seconds) 2011-06-08T11:28:54 *** janzert has joined #aichallenge 2011-06-08T11:29:20 smiley1983: Our plan was succesfull! :) 2011-06-08T11:29:24 No worries - let me know if you have any further problems with it, it's great that someone's testing it. :) 2011-06-08T11:29:28 plan? 2011-06-08T11:30:42 I hope there will be more OCaml users. 2011-06-08T11:31:18 smiley1983: is Stop an OCaml keyword? 2011-06-08T11:31:26 no 2011-06-08T11:31:48 McLeopold: `S, `N, `W, E, `Stop 2011-06-08T11:32:07 I don't think so - I've prepended the tick so it doesn't interfere with user's variables 2011-06-08T11:32:09 custom type or enum? 2011-06-08T11:32:10 it's a type for direction variable 2011-06-08T11:32:13 if they open Ants 2011-06-08T11:32:22 "open Ants" that is 2011-06-08T11:32:28 yes, something like this 2011-06-08T11:32:29 Custom type 2011-06-08T11:33:31 *** okayzed is now known as okay 2011-06-08T11:34:03 um, or maybe enum - I wasn't familiar with the term, reading about it now 2011-06-08T11:34:44 *** boegel has quit IRC (Ping timeout: 260 seconds) 2011-06-08T11:36:33 okay, it seems to match the definition of an enum from my understanding of the wikipedia page on enumerated types 2011-06-08T11:38:01 McLeopold: I see no separate food bonus in the games on the server. Is it a forgotten feature :) 2011-06-08T11:38:03 ? 2011-06-08T11:38:40 what's going on? 2011-06-08T11:44:22 contestbot: later tell stocha just for a day or so of testing 2011-06-08T11:44:22 janzert: OK 2011-06-08T11:45:23 *** Eruonen has joined #aichallenge 2011-06-08T11:48:47 amstan: I really want to launch the site. Does Jeff have a contest schedule in mind? 2011-06-08T11:51:23 *** amstan has quit IRC (Ping timeout: 246 seconds) 2011-06-08T11:53:54 Frontier: it should be gone... 2011-06-08T11:53:59 uh *shouldn't* 2011-06-08T11:55:31 * janzert thinks about going through and setting issues that have to be done before launch 2011-06-08T11:55:56 * McLeopold already started that 2011-06-08T11:56:28 janzert: there are 2 new milestones for launch and after launch 2011-06-08T11:56:49 ok 2011-06-08T11:57:33 redo favicon can be post launch ;) 2011-06-08T11:57:55 yeah, but it should be done now 2011-06-08T11:58:05 in either case I don't see that list happening within a week 2011-06-08T11:58:16 computerwiz_222 doesn't seem to be very active here 2011-06-08T11:58:17 that 2 a day :P 2011-06-08T11:58:52 I'm thinking we should just launch and deal with the fallout 2011-06-08T11:59:20 so does the engine write out the bonus to the replay? I have a hacked replay that works nicely: http://marco-leise.homedns.org/antgui/stable/visualizer.php?game=8 2011-06-08T11:59:24 btw, I ran an ec2 worker yesterday both without secure jails and disabling secure jail pause and resume seemed to be around the same speed as the current worker 2011-06-08T11:59:41 that's good 2011-06-08T12:00:01 so, then we need to try your other idea of a listener within the process? 2011-06-08T12:00:04 McLeopold: well we really should get the launch milestone down to the minimal set of things that need to be done before launching 2011-06-08T12:00:04 *** okay is now known as okayzed 2011-06-08T12:00:18 I know jails working and email are at least on that list 2011-06-08T12:00:29 invalid replays 2011-06-08T12:00:36 getting money? 2011-06-08T12:00:51 at least a temporary source from somewhere 2011-06-08T12:01:19 temporary source is commonly called a credit 2011-06-08T12:01:26 :) 2011-06-08T12:01:48 Frontier: if food bonus isn't working, can you fix it? or at least create an issue on github 2011-06-08T12:01:49 in this case I'm thinking of Jeff's pocket ;) 2011-06-08T12:02:18 He said that's what he would do, to motivate him to get money. 2011-06-08T12:02:26 McLeopold: You mean... I should edit Phyton code? 2011-06-08T12:02:43 yeah 2011-06-08T12:03:28 McLeopold: so if you want to push launch as soon as possible, mind if I get a little more ruthless on the milestone? 2011-06-08T12:03:40 yeah 2011-06-08T12:03:57 you do mind? 2011-06-08T12:05:21 well, I'm not going to work any faster, or be able to create more time 2011-06-08T12:05:38 by ruthless I mean remove things from it 2011-06-08T12:05:48 just move to after launch 2011-06-08T12:05:55 so it is a truer guide on when we can really launch 2011-06-08T12:06:02 yep, that's my plan 2011-06-08T12:06:49 there's a start :) 2011-06-08T12:08:15 3 month for the contest? 2011-06-08T12:08:40 2 is the number I've been hearing and rather favor except... 2011-06-08T12:09:05 3 would be rather nice to push us into the start of the north american school year 2011-06-08T12:09:20 we do we want that? 2011-06-08T12:09:54 I think trying to start a new contest after this in september would be better 2011-06-08T12:10:01 yeah, especially if we could get into say around the 3rd week of september 2011-06-08T12:10:08 you are very optimistic 2011-06-08T12:10:19 The game stuff is easy 2011-06-08T12:10:36 It's the process handling and trueskill that's been killing us 2011-06-08T12:11:23 I think there's very little chance we'll launch another before mid october 2011-06-08T12:11:26 I've got a few ideas for new games I can code while this contest is running 2011-06-08T12:11:33 more likely next january 2011-06-08T12:11:47 I guess I am optimistic 2011-06-08T12:12:50 McLeopold: issue 159 looks interesting. what the heck is happening there? 2011-06-08T12:13:13 I don't know 2011-06-08T12:13:40 I could be anything from a bug in the engine to a corrupted replay to bad visualizer code 2011-06-08T12:16:04 wait... these are all cases of ant collisions 2011-06-08T12:17:36 It just looks odd because there is just one big brown area. 2011-06-08T12:18:43 *** carlop has joined #aichallenge 2011-06-08T12:18:47 not looking at the game at the moment, but from memory aren't some of the ants close enough between the two teams they should be dieing in battle as well? 2011-06-08T12:19:10 no 2011-06-08T12:19:10 at least it seems like there were combat lines implying that 2011-06-08T12:20:12 *** george1914 has quit IRC (Ping timeout: 252 seconds) 2011-06-08T12:20:31 now that I open up the game it does look correct 2011-06-08T12:20:34 janzert: Take a close look. The ants that stand alone against two attackers die while the attackers (both receiving 1/2 damage) survive 2011-06-08T12:20:43 right 2011-06-08T12:21:55 oh, I see it now 2011-06-08T12:22:10 I wonder if there's a way to better show two ants in a block of ants are stepping onto each other 2011-06-08T12:22:16 Frontier: so, a collision x or something would be really helpful 2011-06-08T12:22:59 maybe some sort of crosshatching where ants are overlapping? 2011-06-08T12:23:15 Frontier: also, can you make the lines indicate if the attack worked both ways or not? Maybe arrow heads? 2011-06-08T12:23:57 Or light up ants that recieved over 1 damage, and something different for those that receive less? 2011-06-08T12:24:03 *** olexs has joined #aichallenge 2011-06-08T12:25:24 * McLeopold losing battery power... 2011-06-08T12:25:28 *** McLeopold has quit IRC (Read error: Connection reset by peer) 2011-06-08T12:26:44 I don't want to make too many assumptions about the battle resolution code and the rendering is also quite slow in the Java visualizer. The ants that received >1 damage are the ones that disappear obviously :p 2011-06-08T12:41:16 *** berak has quit IRC (Ping timeout: 240 seconds) 2011-06-08T12:44:28 *** _0x471 <_0x471!~slash@p57B3E9E1.dip.t-dialin.net> has joined #aichallenge 2011-06-08T12:44:58 *** _0x47 <_0x47!~slash@p5B12E016.dip.t-dialin.net> has quit IRC (Ping timeout: 260 seconds) 2011-06-08T12:54:58 hi folks, nice contest as usual.... i'm starting to write my bot, does anyone know the competition's starting/ending date ? 2011-06-08T12:57:39 there is no official start date yet, as far as I know, but it should be fairly soon 2011-06-08T13:06:30 ok, thank, I hope that can start soon 2011-06-08T13:11:55 contestbot: beta 2011-06-08T13:11:56 smiley1983: I do not know about 'beta', but I do know about these similar topics: 'http://aichallengebeta.hypertriangle.com/' 2011-06-08T13:17:31 *** McLeopold has joined #aichallenge 2011-06-08T13:19:28 *** InsaneMalkavian has left #aichallenge 2011-06-08T13:21:53 visualizer is giving errors to me 2011-06-08T13:22:19 rabidus: link? 2011-06-08T13:22:27 http://aichallengebeta.hypertriangle.com/visualizer.php?game=17928 2011-06-08T13:22:59 that error needs to be fixed, adding "debug=true" every time is annoying 2011-06-08T13:23:03 yeah, that's a know issues. we'll get to it sometime 2011-06-08T13:23:08 ok 2011-06-08T13:23:14 http://aichallengebeta.hypertriangle.com/visualizer.php?game=17928&debug=true 2011-06-08T13:23:35 olexs: care to fix it? do you know python? 2011-06-08T13:23:50 McLeopold: would have done it already if I did :) 2011-06-08T13:23:57 ah, thx 2011-06-08T13:23:58 .> 2011-06-08T13:24:05 too bad :( 2011-06-08T13:25:50 *** davidjliu has quit IRC (Ping timeout: 252 seconds) 2011-06-08T13:26:51 *** aeter has left #aichallenge 2011-06-08T13:28:02 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86.1 [Firefox 4.0.1/20110413222027]) 2011-06-08T13:31:03 Frontier: what's wrong with this replay? I can't figure it out.http://pastebin.com/JeSfwe3r 2011-06-08T13:32:18 is it the one from above? (17928) 2011-06-08T13:32:24 no 2011-06-08T13:32:31 I'm doing some local testing 2011-06-08T13:33:25 Frontier: I get this in the ant canvas: SyntaxError: Unexpected token ILLEGAL at Object.parse (native) 2011-06-08T13:34:51 At least I have a quick fix for http://aichallengebeta.hypertriangle.com/visualizer.php?game=12534&turn=589&row=1&col=42 now. 2011-06-08T13:35:15 Is this Firefox? 2011-06-08T13:35:37 chrome 2011-06-08T13:35:40 Do other browsers give better error messages? A line number would be nice. 2011-06-08T13:36:04 The quick fix looks nice, but you still need to distiguish between a collision death and an attack death. 2011-06-08T13:36:09 other than that maybe the Json has illegal characters in it, like " 2011-06-08T13:36:27 double quote is illegal? 2011-06-08T13:36:40 I don't know, never had to care 2011-06-08T13:36:49 it passes jsonlint.com 2011-06-08T13:37:07 *** Palmik has joined #aichallenge 2011-06-08T13:38:22 how about the null value? that seems new to me 2011-06-08T13:38:42 If it is really the JSON parser then I can't do anything about it. 2011-06-08T13:39:00 that's valid json 2011-06-08T13:39:07 and it is new 2011-06-08T13:39:29 Just try it once with -1 instead of null or something 2011-06-08T13:39:51 And try in a different browser that gives line numbers 2011-06-08T13:40:23 line 152 in firefox 2011-06-08T13:41:16 In Opera it loads 2011-06-08T13:42:59 opera says unterminated string 2011-06-08T13:43:48 aichallenge: Marco Leise epsilon * rbb3157c / (3 files): visualizer: changed death animation slightly - http://bit.ly/irFcSA 2011-06-08T13:44:11 it works fine for me in every browser, your clipboard sucks :p 2011-06-08T13:44:17 areg 2011-06-08T13:54:00 *** Cyndre has quit IRC (Ping timeout: 240 seconds) 2011-06-08T13:58:52 *** Frontier has quit IRC (Quit: Leaving.) 2011-06-08T14:01:02 *** Frontier has joined #aichallenge 2011-06-08T14:01:46 *** olexs has quit IRC (Quit: Leaving.) 2011-06-08T14:02:51 *** Frontier1 has joined #aichallenge 2011-06-08T14:03:04 *** Frontier has quit IRC (Client Quit) 2011-06-08T14:08:49 *** aeter has joined #aichallenge 2011-06-08T14:19:39 *** Eruonenenen has joined #aichallenge 2011-06-08T14:19:53 *** okayzed is now known as okay 2011-06-08T14:22:50 *** Eruonen has quit IRC (Ping timeout: 240 seconds) 2011-06-08T14:22:58 *** smiley1983 has quit IRC (Quit: leaving) 2011-06-08T14:25:33 *** okay is now known as okayzed 2011-06-08T14:26:50 *** Cyndre has joined #aichallenge 2011-06-08T14:34:22 game engine stopped? 2011-06-08T14:36:55 *** berak has joined #aichallenge 2011-06-08T14:39:03 *** Frontier1 has quit IRC (Quit: Leaving.) 2011-06-08T14:44:36 *** Mathnerd314 has joined #aichallenge 2011-06-08T14:49:34 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-06-08T14:51:39 *** locutus2 has joined #aichallenge 2011-06-08T14:56:19 antimatroid: remember the bonus we designed if there is more than one survivor? 2011-06-08T15:01:24 *** janzert has left #aichallenge 2011-06-08T15:01:40 *** janzert has joined #aichallenge 2011-06-08T15:01:40 *** _0x471 <_0x471!~slash@p57B3E9E1.dip.t-dialin.net> has quit IRC (Read error: Connection reset by peer) 2011-06-08T15:02:04 *** _0x47 <_0x47!~slash@p57B3E9E1.dip.t-dialin.net> has joined #aichallenge 2011-06-08T15:03:43 http://games.stanford.edu/ 2011-06-08T15:11:31 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-06-08T15:26:45 aichallenge: McLeopold epsilon * r21cba7d / ants/playgame.py : interject names into replay json - http://bit.ly/lQDA8c 2011-06-08T15:26:46 aichallenge: McLeopold epsilon * r54fb973 / (ants/ants.py ants/playgame.py): added bonus to replay, fixed some scoring issues. Closes #171 - http://bit.ly/kbtL0B 2011-06-08T15:26:46 aichallenge: McLeopold epsilon * r5c9810b / (3 files): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://bit.ly/mKOb2G 2011-06-08T15:44:42 *** Accoun has quit IRC () 2011-06-08T15:45:39 *** McLeopold has quit IRC (Ping timeout: 240 seconds) 2011-06-08T15:46:42 *** McLeopold has joined #aichallenge 2011-06-08T15:47:15 *** aerique has joined #aichallenge 2011-06-08T15:54:02 *** olexs has joined #aichallenge 2011-06-08T16:02:54 hey, if the top notch bot did not kill the only other survivor(1 ant left) in turn 200, it's pretty unlikely, it will happen anywhere from now till turn 20000. whykeep the game alive in this case till turn 2000 ? there was a clear winner earlier on, keeping it alive further on, will just cause more entropy/global_warming/khattam_shud 2011-06-08T16:02:54 *** Accoun has joined #aichallenge 2011-06-08T16:03:11 *** carlop has quit IRC (Ping timeout: 240 seconds) 2011-06-08T16:03:31 berak: you are correct 2011-06-08T16:03:42 that's why we have the food bonus to discourage this 2011-06-08T16:04:45 k 2011-06-08T16:05:34 But, I don't think it is going to solve the problem 2011-06-08T16:06:10 @later tell janzert do you think we could come up with a simple way to call a game over early? 2011-06-08T16:06:10 McLeopold: Job's done. 2011-06-08T16:06:25 isn't having like 300 turns enough ? 2011-06-08T16:07:01 no 2011-06-08T16:07:08 McLeopold: which values have the trueskill parameters beta and draw prob currently? 2011-06-08T16:07:22 I'm thinking 10 turns with spawning only going to 1 bot, or something like that 2011-06-08T16:07:46 because this game specific parameters, it might be useful to estimate this based on the played games yet 2011-06-08T16:08:16 locutus2: I think draw_probability is at 1%, beta is calculated from that 2011-06-08T16:09:41 *** Naktibalda has joined #aichallenge 2011-06-08T16:10:16 i would do the job if i get the appropiate tables of the database 2011-06-08T16:10:56 and it would be interesting to see sigma in the ranking, to see how variable the current skill is 2011-06-08T16:12:16 The displayed skill is the mean or mean-3*sigma (like the original trueskill version)? 2011-06-08T16:12:32 -3*sigma 2011-06-08T16:13:01 but the starter bots really throw off the draw probability 2011-06-08T16:13:36 *** aeter has left #aichallenge 2011-06-08T16:14:08 yes the draw prob should be near zero, so i think 2% percent is ok 2011-06-08T16:15:52 the beta value instead stands aprox. for the needed strength-difference to reache a 80%/20% win ratio 2011-06-08T16:16:15 sorry i mean <1% for draw prob 2011-06-08T16:17:25 the less random a game is the lower beta 2011-06-08T16:27:18 *** locutus2 has quit IRC () 2011-06-08T16:31:18 something nice about tron was that the usable map gradually got smaller, so we didn't really need to artificially end a game. we should just make the map get smaller over time to force ants to fight ;) 2011-06-08T16:31:41 tough to do with wrapping, i know 2011-06-08T16:31:47 i'm mostly joking 2011-06-08T16:32:36 I'm going to do it 2011-06-08T16:32:44 the water will fill up :) 2011-06-08T16:32:59 sudden death? :) 2011-06-08T16:33:13 as in all ants suddenly die? 2011-06-08T16:33:28 never player worms? 2011-06-08T16:33:31 *played 2011-06-08T16:33:39 once 2011-06-08T16:33:51 water begins rising, and all worms went down to 1hp, so that they die with any attack 2011-06-08T16:34:34 also happened at a specific point in the game to keep it within time limit 2011-06-08T16:43:30 don't make the water smaller. make the ants bigger ;) 2011-06-08T16:43:40 err 2011-06-08T16:43:50 don't make the water bigger. make the ants bigger 2011-06-08T16:45:57 *** delt0r- has joined #aichallenge 2011-06-08T16:47:24 *** delt0r_ has quit IRC (Ping timeout: 260 seconds) 2011-06-08T16:51:18 *** smippy has joined #aichallenge 2011-06-08T16:51:18 *** smippy has joined #aichallenge 2011-06-08T16:52:11 *** smellyhippy has quit IRC (Ping timeout: 240 seconds) 2011-06-08T16:57:12 gosh, there's a real ant sitting on the the top of my monitor, seems it's watching me! (5th floor) oh, that beauty !" // poor frontier, you'll never can take up with that, being restricted to 4 pixels per ant or smething 2011-06-08T16:57:13 *** Frontier has joined #aichallenge 2011-06-08T16:59:50 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 4.0.1/20110518052719]) 2011-06-08T17:07:46 jmcarthur: so attack radius increases slowly? 2011-06-08T17:08:56 @later tell janzert I updated the jskills on the beta server with default values given an initial mu of 50 and a draw probability of 1% 2011-06-08T17:08:56 McLeopold: As you wish. 2011-06-08T17:09:51 *** guyverthree has joined #aichallenge 2011-06-08T17:09:54 evening all 2011-06-08T17:12:07 *** berak has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2011-06-08T17:12:22 aichallenge: McLeopold epsilon * rac9c5b9 / manager/jskills/TSUpdate.java : update to trueskill calculations with draw prob at 1% - http://bit.ly/jJnLT5 2011-06-08T17:13:16 aichallenge: McLeopold epsilon * rcf69e5e / ants/playgame.py : added color if colorama is installed for python - http://bit.ly/mRWcHR 2011-06-08T17:15:10 aichallenge: McLeopold epsilon * r401c7c3 / worker/engine.py : Capturing head and tail for stderr. Closes #146 - http://bit.ly/iOi60F 2011-06-08T17:39:30 just to make things clear R for removal of food is no longer being send right ??? 2011-06-08T17:39:40 beign sent 2011-06-08T17:39:49 being sent 2011-06-08T17:48:52 *** ltriant has joined #aichallenge 2011-06-08T17:52:01 anyone in the house ??? 2011-06-08T17:52:08 yes 2011-06-08T17:52:11 that is correct 2011-06-08T17:52:34 cool thanks 2011-06-08T17:52:40 sorry for being a bit rude 2011-06-08T17:52:46 I'm just very eager 2011-06-08T17:58:43 *** aerique has quit IRC (Quit: ...) 2011-06-08T18:04:12 *** carlop has joined #aichallenge 2011-06-08T18:11:27 *** Cyndre_ has joined #aichallenge 2011-06-08T18:14:00 *** Cyndre has quit IRC (Ping timeout: 240 seconds) 2011-06-08T18:19:43 *** akranis has joined #aichallenge 2011-06-08T18:23:07 *** Palmik has quit IRC (Remote host closed the connection) 2011-06-08T18:26:00 *** carlop has quit IRC (Ping timeout: 255 seconds) 2011-06-08T18:33:10 *** antimatroid has quit IRC (Ping timeout: 258 seconds) 2011-06-08T18:38:04 *** xQuasar has joined #aichallenge 2011-06-08T18:38:13 When is a better server going to be purchased? D: 2011-06-08T18:38:54 Some languages are suffering terribly as a result of lack of RAM :( 2011-06-08T18:41:02 xQuasar: let jeff (j3camero) that you want to donate money, and well get one up and running 2011-06-08T18:41:22 *** smippy is now known as smellyhippy 2011-06-08T18:49:30 McLeopold: I'm a high school stupid, no money to donate :( 2011-06-08T18:49:34 STUDENT* 2011-06-08T18:49:35 LOL 2011-06-08T18:49:48 wtftypo 2011-06-08T18:50:52 haha 2011-06-08T18:51:07 point proven? 2011-06-08T18:51:13 Yeah, ok :P 2011-06-08T18:51:21 :) 2011-06-08T18:51:53 *** xQuasar has quit IRC () 2011-06-08T18:52:17 aichallenge: McLeopold epsilon * rb332005 / (ants/ants.py worker/engine.py): fix to incorrect ant length. Closes #145 - http://bit.ly/kkNkBw 2011-06-08T18:53:59 aichallenge: McLeopold epsilon * r47c69a4 / ants/playgame.py : fixed incorrectly named html file - http://bit.ly/iYBWYG 2011-06-08T19:02:59 *** bhasker has joined #aichallenge 2011-06-08T19:06:51 *** okayzed is now known as okay 2011-06-08T19:10:54 *** davidjliu has joined #aichallenge 2011-06-08T19:12:23 *** amstan has joined #aichallenge 2011-06-08T19:12:23 *** ChanServ sets mode: +o amstan 2011-06-08T19:13:27 *** AlliedEnvy has joined #aichallenge 2011-06-08T19:14:17 aichallenge: McLeopold epsilon * rb7b8cbc / (website/api_game_result.php website/sql.php): added worker_id, rank and skill to replay data. Closes #109 - http://bit.ly/k641qP 2011-06-08T19:14:30 McLeopold: what's going on? 2011-06-08T19:14:35 stuff 2011-06-08T19:14:55 playgame has pretty red text output now 2011-06-08T19:15:10 *** Kingpin13 has joined #aichallenge 2011-06-08T19:15:45 amstan: can you start fixing the tools make file? we need to sed play_one_game.* to remove the "dist" part of the paths 2011-06-08T19:16:34 idk, would it be bad if we just give them the whole repo? 2011-06-08T19:16:53 call it an sdk 2011-06-08T19:17:30 yes 2011-06-08T19:17:43 we could even link them to the github link download directly 2011-06-08T19:18:01 I don't think jeff would agree 2011-06-08T19:18:43 too much extraneous stuff that would just confuse too many people 2011-06-08T19:19:00 ants/ is full of that stuff anyway 2011-06-08T19:19:09 it's either that or make a whitelist of files 2011-06-08T19:19:15 a blacklist is really too much and ugly 2011-06-08T19:19:36 I thought a whitelist was already being used 2011-06-08T19:20:03 via a makefile yes, but still needs some work 2011-06-08T19:20:39 it's not really whitelisting, more like blacklisting 2011-06-08T19:20:44 moving "starter_bots" and "sample_bots" down one level and getting rid of "dist" would make it easier to go back and forth 2011-06-08T19:20:48 copying everything from ants/ then deleting other stuff 2011-06-08T19:22:19 *** olexs has quit IRC (Quit: Leaving.) 2011-06-08T19:24:03 @later tell Frontier you now have a "challenge_rank" and "challenge_skill" array in the replay json 2011-06-08T19:24:03 McLeopold: Yep. 2011-06-08T19:29:00 *** amstan has quit IRC (Ping timeout: 255 seconds) 2011-06-08T19:29:37 amstan/janzert: I don't want to code dead ant vision, so I'm going to just close the issue #111. Any objections? 2011-06-08T19:30:54 no objections? ok 2011-06-08T19:31:00 I'd just leave it alone for now 2011-06-08T19:31:28 if you don't want to implement it, maybe make a comment to that effect and see if someone else can step in to do it 2011-06-08T19:31:45 dead ant vision ???? 2011-06-08T19:31:55 a dead ant does not see 2011-06-08T19:32:12 but you should maybe see what it saw before it died right? :) 2011-06-08T19:32:31 *** Cyndre_ has quit IRC (Quit: Leaving) 2011-06-08T19:32:32 maybe 2011-06-08T19:33:00 but then it just like an area that you just left 2011-06-08T19:33:07 nah, let's just make it part of the spec, an ant must remain alive to report vision 2011-06-08T19:33:20 sounds fair to me 2011-06-08T19:33:30 heh, seemed like there was fairly broad support of the opposite though 2011-06-08T19:33:38 and it does seem to encourage scouts 2011-06-08T19:33:40 people can always code it if they want it 2011-06-08T19:33:50 from me, but I'm changing my mind 2011-06-08T19:33:53 guyverthree: right, so just let the issue set for now 2011-06-08T19:34:14 but, it needs to be define before launch 2011-06-08T19:34:19 *** guyverthree has quit IRC (Quit: Page closed) 2011-06-08T19:34:25 we shouldn't change the game 2011-06-08T19:34:26 McLeopold: yes 2011-06-08T19:34:41 okay, I'm not coding it, and sigh isn't coding it, so... 2011-06-08T19:34:43 I do disagree with your comment that it could wait until after beta 2011-06-08T19:34:50 so just leave it for the moment 2011-06-08T19:35:03 if it's the last thing holding us up from launching we ca throw it out 2011-06-08T19:35:10 ok 2011-06-08T19:38:03 amstan: did you see this actually cause a problem? https://github.com/aichallenge/aichallenge/issues/117 2011-06-08T19:38:27 it works for me everytime 2011-06-08T19:38:54 janzert: I just remembered, you need to pass -Xmx100m to the popen stuff in manager 2011-06-08T19:39:11 or less 2011-06-08T19:40:45 so long as I'm not running a second worker against the beta server it doesn't cause a problem since it is called in between games 2011-06-08T19:40:54 :/ 2011-06-08T19:41:31 is skill supposed to be an absolute measurement? 2011-06-08T19:41:40 not really 2011-06-08T19:41:54 are they greyed-out players former #1s? 2011-06-08T19:42:00 more absolute than the bayeselo 2011-06-08T19:42:02 it's not anchored in any way at all 2011-06-08T19:42:13 k 2011-06-08T19:42:14 anchored between 0 and 100 2011-06-08T19:42:17 bayeselo actually was anchored for the latter half of the contest 2011-06-08T19:42:29 it's trueskill 2011-06-08T19:42:33 not anchored it's bounded 2011-06-08T19:42:38 more or less at least 2011-06-08T19:42:53 bounded with a high probability to 0 through 100 2011-06-08T19:42:55 meh 2011-06-08T19:43:32 so, the rankings look off because we never adjusted beta or tau for the different starting mu 2011-06-08T19:43:32 it's a different thing though than anchored 2011-06-08T19:45:26 btw, where do you get tau from? in the original paper it seems to be gamma 2011-06-08T19:45:56 from here, aka, the dynamic http://www.moserware.com/2010/03/computing-your-skill.html 2011-06-08T19:47:56 ahh, in the comments? 2011-06-08T19:48:00 yeah 2011-06-08T19:48:05 I have a git issue... 2011-06-08T19:48:22 weird I wonder why the name got changed. ahh well 2011-06-08T19:48:33 git diff is showing no changes, but git status shows the files modified 2011-06-08T19:48:47 I know there are changes there, and I'd like to see them 2011-06-08T19:48:48 did git add the files already? 2011-06-08T19:48:53 git diff HEAD 2011-06-08T19:49:15 ding! problem solved 2011-06-08T19:49:40 just cause I've had the same problem more than once :) 2011-06-08T19:54:10 aichallenge: janzert epsilon * rb74e2c7 / (worker/engine.py worker/sandbox.py): Split secure and unsecure sandboxes into two classes in preparation of further divergence in implementation - http://bit.ly/jXXMC3 2011-06-08T19:54:44 *** akranis has quit IRC (Quit: Lmnar) 2011-06-08T19:59:24 aichallenge: McLeopold epsilon * ref107fc / (3 files): small changes to lists - http://bit.ly/kUMUvV 2011-06-08T19:59:50 *** amstan has joined #aichallenge 2011-06-08T19:59:50 *** ChanServ sets mode: +o amstan 2011-06-08T20:03:03 *** McLeopold has quit IRC (Read error: Connection reset by peer) 2011-06-08T20:03:06 McLeopold: regarding game length, a way to cut off games early would be nice but I'm not sure of a fair way to do it. Someone needs to run some statistics on the beta server games ;) 2011-06-08T20:03:37 e.g. average game length, average move the eventual winner took the lead, etc. 2011-06-08T20:03:56 and histograms of both 2011-06-08T20:05:29 histograms showing the move the eventual winner took the lead and the move that all bots were in their final ranks would be really great for determining an appropriate game length to use 2011-06-08T20:06:18 *** Eruonenenen has quit IRC () 2011-06-08T20:07:40 eyeballing it, it looks like many of the games involving top bots end in under 500 turns 2011-06-08T20:24:20 *** mceier has joined #aichallenge 2011-06-08T20:27:27 contestbot: later tell McLeopold now that the beta and the dynamic variance have better settings I wonder if we shouldn't reset mu and sigma again 2011-06-08T20:27:27 janzert: OK 2011-06-08T20:32:48 *** bhasker has quit IRC (Ping timeout: 240 seconds) 2011-06-08T20:40:29 *** jmpespxoreax has joined #aichallenge 2011-06-08T20:46:35 *** mceier has quit IRC (Quit: leaving) 2011-06-08T21:01:15 *** FireFly has quit IRC (Quit: swatted to death) 2011-06-08T21:16:29 *** Frontier has quit IRC (Quit: Leaving.) 2011-06-08T21:36:21 *** smellyhippy has quit IRC (Ping timeout: 260 seconds) 2011-06-08T21:53:50 *** okay is now known as okayzed 2011-06-08T21:59:51 *** antimatroid has joined #aichallenge 2011-06-08T22:32:50 *** Chris_0076 has joined #aichallenge 2011-06-08T22:33:01 *** Chris_0076 has quit IRC (Remote host closed the connection) 2011-06-08T22:38:22 *** Zaphus has joined #aichallenge 2011-06-08T22:45:57 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-06-08T23:06:03 *** qacek has joined #aichallenge 2011-06-08T23:21:26 *** computerwiz_222 has joined #aichallenge 2011-06-08T23:29:53 *** antimatroid has quit IRC (Read error: Connection reset by peer) 2011-06-08T23:30:02 *** antimatroid has joined #aichallenge 2011-06-08T23:34:24 *** qacek has quit IRC (Quit: WeeChat 0.3.5) 2011-06-08T23:36:03 *** qacek has joined #aichallenge 2011-06-08T23:37:38 *** antimatroid has quit IRC (Ping timeout: 268 seconds) 2011-06-08T23:44:42 *** davidjliu has quit IRC (Quit: Page closed) 2011-06-08T23:47:39 *** computerwiz_222 has quit IRC (Remote host closed the connection)