2011-11-12T00:07:27 *** tobym has quit IRC (Remote host closed the connection) 2011-11-12T00:11:42 *** datachomper has joined #aichallenge 2011-11-12T00:12:09 *** treeform has joined #aichallenge 2011-11-12T00:22:38 *** thestinger has quit IRC (Quit: WeeChat 0.3.6) 2011-11-12T00:24:43 *** luizribeiro has quit IRC (Read error: Operation timed out) 2011-11-12T00:24:52 *** tobym has joined #aichallenge 2011-11-12T00:27:29 *** luizribeiro has joined #aichallenge 2011-11-12T00:32:23 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-12T00:46:37 *** Guest19284 has quit IRC (Quit: Page closed) 2011-11-12T01:16:41 *** amstan has joined #aichallenge 2011-11-12T01:16:41 *** ChanServ sets mode: +o amstan 2011-11-12T01:24:03 <_flag> amstan: I'd like to push some changes to the C starter bot but I'm a bit new to git, would it just be "git push origin epsilon"? 2011-11-12T01:24:16 yes 2011-11-12T01:24:24 you might need to pull first 2011-11-12T01:24:36 to update your own version with the new changes 2011-11-12T01:25:04 <_flag> I think I have the most recent one 2011-11-12T01:29:30 *** JorgeB has joined #aichallenge 2011-11-12T01:32:29 aichallenge: FlagCapper epsilon * r579775a / (4 files): Updated C starter pack to support hills. - http://git.io/XVw_0Q 2011-11-12T01:36:15 *** u_ has quit IRC (Quit: u_) 2011-11-12T01:37:13 *** treeform has quit IRC (Remote host closed the connection) 2011-11-12T01:44:57 *** hacklash has quit IRC (Ping timeout: 252 seconds) 2011-11-12T02:13:38 *** antimatroid1 has joined #aichallenge 2011-11-12T02:13:42 *** antimatroid has quit IRC (Quit: Leaving.) 2011-11-12T02:15:25 *** Flort has quit IRC (Ping timeout: 260 seconds) 2011-11-12T02:16:31 *** b0rder has quit IRC (Ping timeout: 248 seconds) 2011-11-12T02:19:48 *** pairofdice has joined #aichallenge 2011-11-12T02:20:01 *** MiltonBradley has joined #aichallenge 2011-11-12T02:20:30 Hello I have a question. I am just getting started and wondering how I run the play_one_game.cmd 2011-11-12T02:20:31 *** delt0r_ has quit IRC (Ping timeout: 256 seconds) 2011-11-12T02:20:45 http://aichallenge.org/visualizer.php?game=87587 \o/ 2011-11-12T02:21:17 Hm. I'm having a lot of trouble getting CD to work for finding hives in the maze maps :/ 2011-11-12T02:21:29 MiltonBradley: What happens when you just run it? 2011-11-12T02:22:06 I love it how my ants steal that green hill from the middle of their group 2011-11-12T02:22:45 I double click on it and a black cmd window pops up then disappears in a split second 2011-11-12T02:23:17 Try running it from the command line 2011-11-12T02:23:31 how do I do that? 2011-11-12T02:23:48 I opened the command line. I'm a little embarassed. Never done this before... Sorry 2011-11-12T02:24:01 Basic command line tutorials are a bit out of scope for this channel, I think :) 2011-11-12T02:24:07 Try looking one up - "windows command line tutorial" 2011-11-12T02:24:18 ok i monkey fisted it 2011-11-12T02:24:32 I drug the play_one_game.cmd into the cmd window 2011-11-12T02:24:38 Oh 2011-11-12T02:24:46 Didn't know that works 2011-11-12T02:24:50 *** Fandekasp has joined #aichallenge 2011-11-12T02:24:58 me neither 2011-11-12T02:24:59 *** dorisabayon has joined #aichallenge 2011-11-12T02:25:18 ok now it says 'python' is not recognized as an internal or external command, operable program or batch file. 2011-11-12T02:25:33 I have both python 2.7 and 3.2 installed 2011-11-12T02:26:10 Bear in mind that play_one_game is just a demo script 2011-11-12T02:26:24 Try entering what it contains yourself, replacing 'python' with the path to your python interpreter 2011-11-12T02:26:43 ok I open it in a text file? 2011-11-12T02:26:55 You could modify it, yes 2011-11-12T02:27:03 You could also just type what it contains straight into the command line window 2011-11-12T02:27:04 *** Flort has joined #aichallenge 2011-11-12T02:27:17 Probably after changing to the directory it's in (this is where the basic command line tutorial comes in) 2011-11-12T02:27:54 Or you could edit the *.cmd in any gooey editor 2011-11-12T02:28:33 pairofdice: I believe that's what he meant by "open it in a text file?" 2011-11-12T02:28:43 I feel like my ants will only get worse if I touch them 2011-11-12T02:30:27 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 260 seconds) 2011-11-12T02:31:12 *** dvladim has joined #aichallenge 2011-11-12T02:31:34 Ok I replaced python with the path "C:\Python27\python.exe" -- IHere's a paste from my command line - C:\Users\Paul>C:\Python27\python"%~dp0playgame.py" --engine_seed 42 --player_see d 42 --end_wait=0.25 --verbose --log_dir game_logs --turns 500 --map_file "%~dp0 maps\maze\maze_04p_01.map" %* "python ""%~dp0sample_bots\python\GreedyBot.py""" "python ""%~dp0sample_bots\python\LeftyBot.py""" "python ""%~dp0sample_bots\pyth 2011-11-12T02:32:31 *** delt0r_ has joined #aichallenge 2011-11-12T02:32:37 500 turns takes a while 2011-11-12T02:36:36 well in any case it didn't work 2011-11-12T02:36:49 how do you get yours to run pairofdice? 2011-11-12T02:37:34 "Didn't work" is unhelpful. What happened? 2011-11-12T02:37:53 You're probably better off adding c:\python27 to your path, since it's referred to elsewhere in that command line too 2011-11-12T02:38:05 C:\Python32\python.exe tools/playgame.py "C:\Python32\python.exe MyBot.py" "python tools/sample_bots/python/GreedyBot.py" "python tools/sample_bots/python/GreedyBot.py" "python tools/sample_bots/python/GreedyBot.py" --map_file tools/maps/multi_hill_maze/maze_04p_02.map --log_dir game_logs --turns 500 -e --loadtime=3000 --turntime=2000 2011-11-12T02:38:46 ok so basically i need to drag my stuff into the python folder 2011-11-12T02:38:58 Err, no 2011-11-12T02:39:03 err I mean I need to drag my stuff into the python folder 2011-11-12T02:39:42 What I pasted is my "tutorial.cmd" that I got from reading the tutorial, on the ants site 2011-11-12T02:40:59 MiltonBradley: No, you don't. 2011-11-12T02:41:15 Add c:\python27 to your path. I'm sure Google has lots of resources on how to do that :) 2011-11-12T02:41:24 to what path 2011-11-12T02:41:48 settings, control panel, system 2011-11-12T02:42:06 advanced, environment variables 2011-11-12T02:42:10 *** tobym has quit IRC (Remote host closed the connection) 2011-11-12T02:42:15 system variables, PATH 2011-11-12T02:42:43 good lord what a pain in the ass this is. Ok i'll give it a go 2011-11-12T02:43:00 thanks for your help btw 2011-11-12T02:45:16 But if it's not in your path, you can reference it directly 2011-11-12T02:52:15 Ok how do I reference it directly? Type it enclosed in "'s such as "C:\python27\python.exe" 2011-11-12T02:55:02 No "' 2011-11-12T02:56:09 Adding it to the path is going to be by far the easiest option 2011-11-12T02:56:25 Typing the full path each time is going to be a pain 2011-11-12T03:01:09 can't I type the full path once in the cmd line file or something 2011-11-12T03:01:18 or add to the path 2011-11-12T03:01:20 w/e 2011-11-12T03:01:23 is easiest 2011-11-12T03:01:29 *** ejls has quit IRC (Ping timeout: 258 seconds) 2011-11-12T03:01:44 *** ejls has joined #aichallenge 2011-11-12T03:02:18 Add it to the path as pairofdice described 2011-11-12T03:02:25 That way you can just refer to it as 'python' from then on 2011-11-12T03:04:37 *** Fandekasp has quit IRC (Ping timeout: 240 seconds) 2011-11-12T03:04:54 *** dorisabayon has quit IRC (Ping timeout: 240 seconds) 2011-11-12T03:04:58 *** antimatroid1 has quit IRC (Ping timeout: 245 seconds) 2011-11-12T03:05:11 Each time? Why don't you se the *.cmd like the tutorial tells you to 2011-11-12T03:05:19 use* 2011-11-12T03:05:47 You can worry about paths later 2011-11-12T03:07:44 *** retybok_ has joined #aichallenge 2011-11-12T03:09:01 I am trying to use the *.cmd 2011-11-12T03:10:04 I am following the tutorial to the nose. 2011-11-12T03:10:07 C:\Users\Paul>C:\Python27\Tools\play_one_game.cmd 'python' is not recognized as an internal or external command, operable program or batch file. C:\Users\Paul> 2011-11-12T03:12:02 So where the *.cmd says python you put c:\python27\python 2011-11-12T03:12:13 Or something similar 2011-11-12T03:12:57 MiltonBradley: Seriously. Just add it to the path. 2011-11-12T03:12:57 Maybe even with an .exe at the end 2011-11-12T03:13:10 Then the .cmd will work, and so will any other time you want to use Python. 2011-11-12T03:13:26 Unless if you have multiples and you want a specific 2011-11-12T03:13:40 ;p 2011-11-12T03:14:17 But yea, adding it to path is painless 2011-11-12T03:14:27 nickjohnson: I can only do one thing at a time hang on 2011-11-12T03:15:32 awesome it's working 2011-11-12T03:15:38 thanks a lot 2011-11-12T03:15:59 btw what does %~dp0 mean 2011-11-12T03:21:41 *** antimatroid has joined #aichallenge 2011-11-12T03:21:55 *** dvladim has quit IRC (Ping timeout: 260 seconds) 2011-11-12T03:22:16 I have no I ... ALL GLORY TO THE HYPNOTOAD 2011-11-12T03:29:19 *** Jak_o_Shadows has joined #aichallenge 2011-11-12T03:42:42 *** g0llum has joined #aichallenge 2011-11-12T03:48:12 *** ztfw has joined #aichallenge 2011-11-12T03:51:21 *** Zaphus has joined #aichallenge 2011-11-12T03:55:55 *** retybok_ has quit IRC (Ping timeout: 258 seconds) 2011-11-12T04:01:03 *** JorgeB has quit IRC (Quit: Computer has gone to sleep.) 2011-11-12T04:11:55 *** kilae has joined #aichallenge 2011-11-12T04:23:08 My code is still way too inefficient 2011-11-12T04:25:44 *** foRei has joined #aichallenge 2011-11-12T04:26:21 *** treeform has joined #aichallenge 2011-11-12T04:26:27 *** merk has joined #aichallenge 2011-11-12T04:26:36 hallo 2011-11-12T04:28:01 anyone here 2011-11-12T04:28:07 coding in c++? 2011-11-12T04:30:39 *** antimatroid has quit IRC (Quit: Leaving.) 2011-11-12T04:30:49 *** antimatroid has joined #aichallenge 2011-11-12T04:31:11 *** dvladim has joined #aichallenge 2011-11-12T04:31:16 *** choas has joined #aichallenge 2011-11-12T04:31:52 *** MiltonBradley has quit IRC (Quit: Page closed) 2011-11-12T04:35:54 *** NoxiaZ^ has joined #aichallenge 2011-11-12T04:38:08 *** treeform has quit IRC (Remote host closed the connection) 2011-11-12T04:39:02 *** TestUser has joined #aichallenge 2011-11-12T04:39:35 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-11-12T04:39:51 *** Nescio has joined #aichallenge 2011-11-12T04:41:21 *** djr_ has joined #aichallenge 2011-11-12T04:41:50 *** yoden has quit IRC (Ping timeout: 260 seconds) 2011-11-12T04:44:53 *** TestUser has quit IRC (Remote host closed the connection) 2011-11-12T04:49:31 *** Fulairyn has joined #aichallenge 2011-11-12T04:50:24 Hello everyone 2011-11-12T04:50:31 I'm trying to start in vb.net 2011-11-12T04:51:33 *** ProperChaos has joined #aichallenge 2011-11-12T04:51:38 I tried to not let the ants bang into eachother, and checked it by this: If a.Passable(newloc) and a.Unoccupied(newloc) 2011-11-12T04:51:56 unfortunately, this doesn't work. Am I doing something wrong ? 2011-11-12T04:54:29 Fulairyn: "doesn't work"? 2011-11-12T04:55:26 well, they still walk into echother ;), so it works, but doesn't do what I want :( 2011-11-12T04:57:15 And now i'm a little confused, I thought it would be simple to start in vb, and I saw the function a.unoccupied and a.passable and I thought, with those two I can check if I can move to that spot. I saw other people doing it whit that to, but in c# or something 2011-11-12T04:57:46 you also need to check if two ants aren't getting the same order-destination 2011-11-12T04:59:05 You don't care about where the ants are but where they are going 2011-11-12T04:59:05 offcourse!, cause they check if it's free there, and it is free, but there are 2 ants going to the same destination at the same time. Thanks! 2011-11-12T04:59:09 *** praveen has joined #aichallenge 2011-11-12T04:59:17 np :) 2011-11-12T05:03:36 *** Fulairyn has quit IRC (Ping timeout: 265 seconds) 2011-11-12T05:06:34 *** djr_ has quit IRC (Quit: Leaving) 2011-11-12T05:07:30 *** antimatroid has quit IRC (Ping timeout: 260 seconds) 2011-11-12T05:13:12 guys 2011-11-12T05:13:18 anyone here coding in c++? 2011-11-12T05:13:29 *** antimatroid has joined #aichallenge 2011-11-12T05:14:47 *** grwip has joined #aichallenge 2011-11-12T05:16:33 *** treeform has joined #aichallenge 2011-11-12T05:20:02 *** treeform has quit IRC (Remote host closed the connection) 2011-11-12T05:20:23 *** Jak_o_Shadows1 has joined #aichallenge 2011-11-12T05:20:49 *** retybok_ has joined #aichallenge 2011-11-12T05:22:13 *** antimatroid has quit IRC (Ping timeout: 240 seconds) 2011-11-12T05:23:21 *** Jak_o_Shadows has quit IRC (Ping timeout: 252 seconds) 2011-11-12T05:23:59 merk: "don't ask to ask, just ask" applies to your question - even though you were asking something, it would probably help to just ask the specific question you would ask someone using C++ instead of waiting for someone to profess their use of it ;) 2011-11-12T05:27:05 merk: yes :-) 2011-11-12T05:46:35 *** merk has quit IRC (Ping timeout: 260 seconds) 2011-11-12T05:46:49 *** merk has joined #aichallenge 2011-11-12T05:50:48 *** merk has quit IRC (Client Quit) 2011-11-12T05:51:13 *** merk has joined #aichallenge 2011-11-12T05:57:49 *** ALplus has quit IRC (Ping timeout: 258 seconds) 2011-11-12T05:58:38 *** mj41 has joined #aichallenge 2011-11-12T06:04:40 *** kilae has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243]) 2011-11-12T06:04:47 *** mj41 has quit IRC (Ping timeout: 248 seconds) 2011-11-12T06:07:04 *** pm_ has joined #aichallenge 2011-11-12T06:10:11 *** pm_ has left #aichallenge 2011-11-12T06:12:05 *** Donnetje has joined #aichallenge 2011-11-12T06:17:53 *** dvladim has quit IRC (Ping timeout: 245 seconds) 2011-11-12T06:18:59 *** merk has quit IRC (Ping timeout: 258 seconds) 2011-11-12T06:26:03 *** NoxiaZ^ has quit IRC (Ping timeout: 252 seconds) 2011-11-12T06:27:04 *** Donnetje has quit IRC (Quit: Page closed) 2011-11-12T06:32:02 *** choas has quit IRC (Quit: leaving) 2011-11-12T06:33:46 *** Fandekasp has joined #aichallenge 2011-11-12T06:36:00 *** ikaros has joined #aichallenge 2011-11-12T06:37:31 *** dorisabayon has joined #aichallenge 2011-11-12T06:49:40 *** retybok_ has quit IRC (Ping timeout: 256 seconds) 2011-11-12T07:06:10 *** Zaphus has quit IRC (Quit: Page closed) 2011-11-12T07:08:13 *** mikewintermute has joined #aichallenge 2011-11-12T07:12:09 *** Flapadar has joined #aichallenge 2011-11-12T07:16:17 *** okan009 has joined #aichallenge 2011-11-12T07:16:49 Just wondering if it is possible to process ant hills in lua yet? 2011-11-12T07:16:56 It's not in the starter pack but is there a way to do it anyway? 2011-11-12T07:17:22 hi is there anyone who writing in C , i have a question in C code ? 2011-11-12T07:19:25 *** praveen has quit IRC (Ping timeout: 265 seconds) 2011-11-12T07:19:34 okan009: few people are coding in C, but C++ programmers can probably help you, too? 2011-11-12T07:20:05 *** sever_h has joined #aichallenge 2011-11-12T07:21:13 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-12T07:21:14 Flapadar: If it is missing in the starter pack, then it wasn't updated since hills were introduced into the game. You can look at other starters or at how (for example) enemy ants are handled and duplicate that behavoir for hills 2011-11-12T07:22:27 maybe , i have a problem with searching around for food , specially north and south , i write with that starter package , i think i am misunderstod that code , because i am looking for west and east of ant correctly 2011-11-12T07:23:15 and i cant figure out my problem 2011-11-12T07:26:52 *** Jak_o_Shadows1 has quit IRC (Remote host closed the connection) 2011-11-12T07:27:09 *** Nescio has quit IRC (Quit: Page closed) 2011-11-12T07:27:36 *** amstan has quit IRC (Ping timeout: 276 seconds) 2011-11-12T07:28:50 *** Conorach has joined #aichallenge 2011-11-12T07:29:15 *** sever_h has quit IRC (Quit: Page closed) 2011-11-12T07:37:35 Cheese mleise 2011-11-12T07:37:50 I'll probably just have to re-write the parser to insert hills into the structs. 2011-11-12T07:39:35 okan009: isn't all food in some 'food list' ? 2011-11-12T07:40:31 pick one ant and one food item and move the ant in the right direction for starters ^^ 2011-11-12T07:40:43 mm i guess no , i have a map matrix , and i'm searching in ant's range 2011-11-12T07:40:56 Are you searching row,col 8 by 8? 2011-11-12T07:41:11 You basically have to loop 64 times (two 1,8 loops) if you want to do it that way 2011-11-12T07:41:58 for r = -8,8 do 2011-11-12T07:41:58 for c = -8,8 do 2011-11-12T07:41:58 local deltaR,deltaC = wrap(row+r,0,ants.config.rows),wrap(col+c,0,ants.config.cols) 2011-11-12T07:42:28 My bad, much more than 64 lol 2011-11-12T07:42:53 You shouldn't hard code these numbers! 2011-11-12T07:43:18 there's a variable for the radius of line of sight of each ant? 2011-11-12T07:43:36 ...how about you look at the website 2011-11-12T07:43:38 yes, and you are supposed to use it :) 2011-11-12T07:44:12 Hm, isn't in the lua starter nor is it parsed when it loads 2011-11-12T07:44:15 Must have been added since? 2011-11-12T07:44:38 In the last contest we started with 23 planets (I think) and people started hard-coding that number. Their bots crashed when we added 2 more planets. So we added game variables. 2011-11-12T07:44:45 *** zeroen has joined #aichallenge 2011-11-12T07:45:23 Flapadar: no, these variables were there practically from the start. complain with the author of the lua starter :) 2011-11-12T07:46:02 Ah lol. Is there a place where I can read the input of io.stdin:lines() without running the bot? 2011-11-12T07:46:12 So as I can rewrite the parser to use everything that's given etc 2011-11-12T07:46:26 yes i am searching 8 to 8 square , and i am looking for distance of food , but i am searching differnt , i divided that square to corner to corner , 4 pieces , triangle , if it found in west triangle for example it will move west , but it doesnt see north and south triangle :D 2011-11-12T07:46:50 Flapadar: It *IS* in the Lua starter -> ants.config 2011-11-12T07:47:10 Flapadar, i have lua and hills implemented its no big deal 2011-11-12T07:47:30 I don't know the Lua syntax, but it may be something like ants.config[viewradius2] 2011-11-12T07:48:48 ants.config doesn't have all the members listed, and there's no way to output the members once they're loaded as stdout is used by python for directions. If there's a data file with the input to the bot, I can just read that and see what's being put in it. 2011-11-12T07:48:55 ikaros - is it just another index for a land type? 2011-11-12T07:49:10 maybe == 5 ? 2011-11-12T07:49:55 Flapadar, http://pastebin.com/4CzSdqVN 2011-11-12T07:50:00 have a look here how i do it 2011-11-12T07:50:04 Flapadar: you can use stderr to print debug info 2011-11-12T07:50:15 just dont copy everything because i have changed ants to state and other stuff 2011-11-12T07:50:34 *** robinski has joined #aichallenge 2011-11-12T07:50:36 Also the engine does offer logging the bot input, yes. The switch is -I, I think (?) 2011-11-12T07:51:03 Thanks ikaros, mleise 2011-11-12T07:51:33 to print to a file you can use logfile = io.open("bot.log", "w") , logfile:write("blabla") and dont forget logfile:close() in the end 2011-11-12T07:52:20 i changed some things in the former ants.lua file.. there were parameters which didnt exist and other were missing 2011-11-12T07:52:30 you should be able to look this up on the pastebin link 2011-11-12T07:53:38 Hope someone can help me out, after submitting my code, I keep getting "Test Error: compiled, but failed test cases" 2011-11-12T07:53:49 would anyone be able to point me to some more information on fixing this 2011-11-12T07:54:11 you can run the test locally perhaps you see whats wrong there 2011-11-12T07:54:43 its tools/test_bot.sh 2011-11-12T07:55:09 actually I did run the test_bot on my windows machine, no errors or problems arose 2011-11-12T07:56:21 I guess it could be a difference in Mono C# and Microsoft C#... guess I now have to figure out what difference 2011-11-12T07:57:54 doesn't it give you any extra information why it failed the test cases? 2011-11-12T07:58:51 you could also try running playgame.py with --strict flag 2011-11-12T07:59:18 That reminds me - is there any way to get a stack trace when a bot times out? Mine seems to time out when I hit a large number of ants, not sure whether its my use of astar being inefficient or other parts of my code being inefficient 2011-11-12T07:59:22 ProperChaos: Well it tells me the bot crashed during testing... but not as to why 2011-11-12T07:59:58 robinski - add -e --capture_errors to the command line file 2011-11-12T08:00:17 Flapadar: you could do some performance tests locally? 2011-11-12T08:00:28 *** yahim91 has joined #aichallenge 2011-11-12T08:00:51 I'm sure there's some kind of profiler for whatever language you're working in 2011-11-12T08:01:25 I think lua has a profiler, if not I can always just write a basic one lol 2011-11-12T08:02:05 Flapadar: well the bot is not crashing at my end, and the --capture_errors is already in... I somehow crashes after submission 2011-11-12T08:02:30 robinski: perhaps you could install Mono and test using that compiler? 2011-11-12T08:02:44 Try increasing the number of turns locally to something high like 1000. If it still doesn't crash, install mono :P 2011-11-12T08:04:26 I'll go install mono... guess that will be the best way of testing 2011-11-12T08:05:25 *** robinski has quit IRC (Quit: Page closed) 2011-11-12T08:05:58 *** Redgis has joined #aichallenge 2011-11-12T08:07:01 *** zaphod has joined #aichallenge 2011-11-12T08:07:27 *** zaphod is now known as Guest79727 2011-11-12T08:17:36 *** retybok_ has joined #aichallenge 2011-11-12T08:17:54 *** okan009 has quit IRC (Ping timeout: 265 seconds) 2011-11-12T08:24:49 *** Flapadar has quit IRC (Ping timeout: 260 seconds) 2011-11-12T08:25:40 *** Flapadar has joined #aichallenge 2011-11-12T08:29:11 *** zeroen has quit IRC (Quit: Page closed) 2011-11-12T08:32:22 *** marijnfs_ has joined #aichallenge 2011-11-12T08:40:36 *** NoxiaZ^ has joined #aichallenge 2011-11-12T08:55:14 *** antimatroid has joined #aichallenge 2011-11-12T09:00:22 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-11-12T09:01:34 *** dorisabayon has quit IRC (Ping timeout: 260 seconds) 2011-11-12T09:07:39 *** Fandekasp has quit IRC (Ping timeout: 258 seconds) 2011-11-12T09:12:38 *** dvladim has joined #aichallenge 2011-11-12T09:14:12 *** NoxiaZ^ has quit IRC (Ping timeout: 276 seconds) 2011-11-12T09:17:59 *** Flapadar has quit IRC (Remote host closed the connection) 2011-11-12T09:23:07 *** g0llum has joined #aichallenge 2011-11-12T09:23:35 *** pieces029_ has joined #aichallenge 2011-11-12T09:23:44 *** pieces029_ has quit IRC (Client Quit) 2011-11-12T09:25:30 *** jurgel has joined #aichallenge 2011-11-12T09:27:03 *** jurgel has quit IRC (Client Quit) 2011-11-12T09:31:45 *** xathis has joined #aichallenge 2011-11-12T09:32:48 *** Hexren has joined #aichallenge 2011-11-12T09:34:59 *** Hexren has left #aichallenge 2011-11-12T09:39:44 *** boegel has joined #aichallenge 2011-11-12T09:42:20 some people are way too paranoid.... 2011-11-12T09:42:25 *** Harpyon has joined #aichallenge 2011-11-12T09:43:30 *** underflow has joined #aichallenge 2011-11-12T09:43:40 hello there 2011-11-12T09:48:30 *** boegel has quit IRC (Ping timeout: 252 seconds) 2011-11-12T09:49:35 *** sigh has quit IRC (Remote host closed the connection) 2011-11-12T09:50:09 *** Fandekasp has joined #aichallenge 2011-11-12T09:50:13 *** dorisabayon has joined #aichallenge 2011-11-12T09:54:03 *** fufu has joined #aichallenge 2011-11-12T09:54:29 *** fufu is now known as Guest88128 2011-11-12T09:55:36 *** fufu94 has joined #aichallenge 2011-11-12T09:56:06 hi 2011-11-12T09:56:18 somebody speak french here ? 2011-11-12T10:00:54 oui oui but not me :) 2011-11-12T10:02:22 Je parle français via Google translate 2011-11-12T10:03:43 *** retybok_ has quit IRC (Ping timeout: 248 seconds) 2011-11-12T10:08:10 *** NoxiaZ^ has joined #aichallenge 2011-11-12T10:08:43 how do you guys find path and don't get timeout? 2011-11-12T10:09:22 I store short paths sections 2011-11-12T10:09:38 between some kind of crossroads that I set up while exploring 2011-11-12T10:09:59 My ants google theirs paths 2011-11-12T10:10:05 nice 2011-11-12T10:10:54 maps.google.com 2011-11-12T10:11:16 what algorithm do you use, bfs a *? 2011-11-12T10:11:21 I do BFS for the whole map and iterate 2011-11-12T10:11:25 a* 2011-11-12T10:11:45 scales a bit better than A* with lots of ants. 2011-11-12T10:12:13 bugnuts: what do you mean by 'scales'? 2011-11-12T10:12:50 You do it once if it's 5 ants, you do it once if it's 500 ants 2011-11-12T10:12:55 i suppose you find the path starting from food to ant 2011-11-12T10:12:55 Scales well 2011-11-12T10:13:14 yeah 2011-11-12T10:13:38 ah ok, I do it the other way around... 2011-11-12T10:15:08 so you find the path between the crossroad and the food because it is a smaller path ? 2011-11-12T10:16:46 you can easily create a distance map from all the food simultaneously 2011-11-12T10:16:50 should be fast and not time out 2011-11-12T10:17:01 if you do bfs smartly 2011-11-12T10:17:04 dijkstra style 2011-11-12T10:17:10 no for each ant I collect what is reachable within their view radius (food, enemies, hills) and then find the path for the first one around 2011-11-12T10:17:32 not the most efficient way I know but that kind of works well 2011-11-12T10:17:51 thanks grwip 2011-11-12T10:18:36 marijnfs, the map you said is just for an ant ? 2011-11-12T10:19:12 no for every square you get a distance to the closest food 2011-11-12T10:19:47 and your ant can then look at the squares around him and move to the square that has the shortest distance 2011-11-12T10:20:05 pretty simple 2011-11-12T10:20:28 *** delt0r_ has quit IRC (Ping timeout: 256 seconds) 2011-11-12T10:20:30 this makes close ants go for the same food though 2011-11-12T10:20:38 but it also keeps them moving 2011-11-12T10:21:03 and you are creating it every turn ? 2011-11-12T10:21:11 fufu94 moi oui 2011-11-12T10:21:55 Is there any way to debug my C# IA into the mono debugger while running with ./play_one_game.sh ? 2011-11-12T10:32:16 *** ProperChaos has quit IRC (Ping timeout: 265 seconds) 2011-11-12T10:33:16 *** delt0r_ has joined #aichallenge 2011-11-12T10:35:53 *** u_ has joined #aichallenge 2011-11-12T10:38:23 *** Fandekasp has quit IRC (Ping timeout: 248 seconds) 2011-11-12T10:38:30 *** dorisabayon has quit IRC (Ping timeout: 258 seconds) 2011-11-12T10:39:14 *** pedrosorio has joined #aichallenge 2011-11-12T10:39:48 yeah every turn 2011-11-12T10:39:50 its fast 2011-11-12T10:40:40 if you do it properly you have to do an evaluation for at most the number of tiles on the map 2011-11-12T10:40:49 which is not a lot of work for a computer 2011-11-12T10:41:41 except if you are doing it in Octave =P 2011-11-12T10:43:12 *** amstan has joined #aichallenge 2011-11-12T10:43:12 *** ChanServ sets mode: +o amstan 2011-11-12T10:45:16 *** Fandekasp has joined #aichallenge 2011-11-12T10:45:24 *** dorisabayon has joined #aichallenge 2011-11-12T10:47:07 *** retybok_ has joined #aichallenge 2011-11-12T10:52:31 *** KP13 has joined #aichallenge 2011-11-12T10:53:01 *** Kingpin13 has quit IRC (Ping timeout: 255 seconds) 2011-11-12T10:53:17 *** KP13 is now known as Kingpin13 2011-11-12T10:57:29 haha 2011-11-12T10:57:55 marijnfs_: true story =P 2011-11-12T10:58:35 Python could be just a tiny bit faster, I wouldn't mind.... 2011-11-12T10:59:19 *** dvladim has quit IRC (Read error: Operation timed out) 2011-11-12T11:02:06 How much memomory can we safely use? 2011-11-12T11:03:11 1.5GB AFAIK 2011-11-12T11:04:00 Oooh... I might end up using significantly more than that in the worst-case :/ 2011-11-12T11:07:33 Hmmm... I can definitely cut down on that though. 2011-11-12T11:08:56 hmmm 2011-11-12T11:09:11 what was the biggest map size? 2011-11-12T11:09:17 200x200 apparently 2011-11-12T11:09:25 ... only 200^2???? 2011-11-12T11:10:12 "Maps are limited to at most 200x200 squares" from the game spex 2011-11-12T11:10:14 *spec 2011-11-12T11:10:20 hehehe 2011-11-12T11:10:41 * p_l just decided to go with 3D array of 200x200x200 2011-11-12T11:11:24 * rofer was going to have a 4D array of 200x200x200x200 :P 2011-11-12T11:11:44 rofer: I don't need more than 200 variables per field ;-)( 2011-11-12T11:11:53 -( 2011-11-12T11:13:33 I can easily keep time-delayed data for few different views of map :) 2011-11-12T11:15:24 an array of 200^3 64bit words is apparently barely ~61 MiB 2011-11-12T11:17:11 *** boegel has joined #aichallenge 2011-11-12T11:19:06 Is there any way to debug my C# IA into the mono debugger while running with ./play_one_game.sh ? 2011-11-12T11:19:30 underflow: there are some extensions to gdb to debug mono binaries, I think 2011-11-12T11:19:43 you should be able to use them to attach to the process 2011-11-12T11:19:53 p_l: I've heard some say that you can use up to something like 1GB, so you could easily store 1MB per square 2011-11-12T11:20:12 p_l i'm using monodevelop, and i don't want do debug in console 2011-11-12T11:20:39 underflow: do you have a windows machine? You could always use C# 2010 Express 2011-11-12T11:20:41 Extrarius: I need space for GC, tactical unit cooperation algorithms and such 2011-11-12T11:21:27 Extarius: i use linux 2011-11-12T11:21:50 *** fufu94 has quit IRC (Quit: Page closed) 2011-11-12T11:21:59 debugging works fine, but i want to attach the debugging process with play_one_game.sh 2011-11-12T11:22:15 to test my bot in real conditions 2011-11-12T11:25:13 could you do it the other way around, and set your bot to open a couple of named pipes (one for in and one for out, if neccessary) and then make the game server open those pipes for your bot? Then you could debug your bot normally and launch the server afterwards 2011-11-12T11:27:16 Extarius i don't understand what you mean 2011-11-12T11:27:16 *** lorill has joined #aichallenge 2011-11-12T11:28:04 the way the server works is to launch your bot with stdin and stdout set to pipes and the server sends the updates to your bot and reads the output 2011-11-12T11:28:49 My solution to debugging is to have the bot listen on TCP and use netcat as the bot executable. 2011-11-12T11:29:28 Zao: I suppose sockets are probably the 'more linux' way of doing it 2011-11-12T11:30:47 Extrarius: I picked it mostly because netcat exists on just about any OS, including Windows. 2011-11-12T11:30:59 *** goffrie_ has joined #aichallenge 2011-11-12T11:31:03 *** goffrie has quit IRC (Ping timeout: 252 seconds) 2011-11-12T11:31:11 fifos and named pipes are much harder to do portably, especially to find a fake-bot proxy. 2011-11-12T11:31:24 ok, i've also to recode the playgame.py to be able to do that ? 2011-11-12T11:31:27 and to use sockets ? 2011-11-12T11:31:48 no, if you use sockets, you don't need to modify playgame.py 2011-11-12T11:31:49 *** yoden has joined #aichallenge 2011-11-12T11:32:00 playone 'samplebot.py' 'samplebot.py', 'netcat localhost 9001' -e -... 2011-11-12T11:32:13 (or whatever syntax the runner script has) 2011-11-12T11:33:02 so i've to add a TCP client to my bot ? 2011-11-12T11:33:54 TCP server, but yeah. 2011-11-12T11:34:01 Or just don't fret about it too much. 2011-11-12T11:45:48 *** underflow has quit IRC (Ping timeout: 245 seconds) 2011-11-12T11:54:50 *** boegel has quit IRC (Quit: Leaving) 2011-11-12T11:56:45 *** treeform has joined #aichallenge 2011-11-12T11:58:28 *** NotABug has joined #aichallenge 2011-11-12T12:02:14 *** Bob_ has joined #aichallenge 2011-11-12T12:03:08 *** NoxiaZ^ has quit IRC (Ping timeout: 258 seconds) 2011-11-12T12:04:14 can anyone tell me how to use the --food_rate parameter for playgame.py .. i tried --food_rate 5 and it complains: 2011-11-12T12:04:16 playgame.py: error: option --food_rate: invalid integer value: '--player_seed' 2011-11-12T12:04:30 where --player_seed is the parameter after food_rate 2011-11-12T12:04:35 its 2 numbers 2011-11-12T12:04:41 ahh 2011-11-12T12:04:46 --food_rate 5 11 2011-11-12T12:04:51 iirc 2011-11-12T12:05:26 works. thx 2011-11-12T12:05:33 i need to test with alot of food now :) 2011-11-12T12:07:33 aichallenge: = epsilon * rc4fe258 / (2 files): measure time at start of input and changes to README - http://git.io/gYu_Og 2011-11-12T12:07:33 aichallenge: Janzert epsilon * rc955b30 / (2 files): 2011-11-12T12:07:33 aichallenge: Merge pull request #365 from pedrosorio/epsilon 2011-11-12T12:07:33 aichallenge: Octave starter package - changes - http://git.io/2JuS2Q 2011-11-12T12:09:19 aichallenge: Janzert epsilon * r55578b2 / (2 files in 2 dirs): 2011-11-12T12:09:20 aichallenge: Merge pull request #359 from gcp/epsilon 2011-11-12T12:09:20 aichallenge: Fix Java/CoffeeScript bugs when input chunks end in the middle of a line. - http://git.io/tEYuiw 2011-11-12T12:12:39 *** Bob_ has quit IRC (Quit: Page closed) 2011-11-12T12:16:53 *** retybok_ has quit IRC (Read error: Connection reset by peer) 2011-11-12T12:17:40 *** lorill has quit IRC (Quit: Page closed) 2011-11-12T12:18:49 There has to be some incantantion I can do to make this faster 2011-11-12T12:21:26 *** NotABug has quit IRC (Quit: Page closed) 2011-11-12T12:33:32 *** goffrie_ is now known as goffrie 2011-11-12T12:42:13 *** datachomper has joined #aichallenge 2011-11-12T12:42:17 pairofdice: python? 2011-11-12T12:43:13 Yeah 2011-11-12T12:43:49 But I have an introductory book to c++ and a lot of free time. 2011-11-12T12:44:00 And I've been meaning to 'branch out' 2011-11-12T12:45:32 C++ can certainly make your bot faster 2011-11-12T12:46:01 but the price is a little piece of your soul for each line you write 2011-11-12T12:46:03 I was shocked by how much faster java was than Ruby. I switch and 100ms searches went to 0-2ms. I'm sure C++/C would have been even more mind blowing. 2011-11-12T12:46:20 ruby is one of the slowest languages 2011-11-12T12:46:23 ruby is notoriously slow 2011-11-12T12:46:49 my turns are all under 10ms... so java is too fast 2011-11-12T12:46:52 clearly xD 2011-11-12T12:46:56 no need c++ 2011-11-12T12:47:02 the difference between java and C is very small in comparison 2011-11-12T12:47:11 JIT/Hotspot does a lot to close the gap I think 2011-11-12T12:47:13 yeah 2011-11-12T12:48:52 My bot is at it's limits time-wise but there's still more things I want to do with it 2011-11-12T12:52:33 I've done a quick hack in C++ which was still at top200 using almost no time 2011-11-12T12:52:59 but now I've decided I want to challenge myself and do a top500 using Octave 2011-11-12T12:53:06 bergmark: Ruby has one of the slower implementations out there 2011-11-12T12:53:23 Java packs a shitload of optimizations that can make it run comparably to C 2011-11-12T12:53:41 (bear in mind octave can get you close to the 500ms with a single bfs :P) 2011-11-12T12:54:00 *** Hexren has joined #aichallenge 2011-11-12T12:54:34 First hurdle, compiling the starter bot... 2011-11-12T12:54:51 xD 2011-11-12T12:54:58 pairofdice: are you using linux? 2011-11-12T12:55:43 Nope, 'rusty old XP 2011-11-12T12:56:04 One of the things where Windows doesn't have the advantage 2011-11-12T12:57:22 install cygwin then 2011-11-12T12:57:34 or wubi and try out ubuntu 2011-11-12T12:57:35 :) 2011-11-12T12:59:31 *** Hexren has quit IRC (Quit: Leaving.) 2011-11-12T12:59:45 hmm, can we send octave bots? 2011-11-12T13:00:04 *** dvladim has joined #aichallenge 2011-11-12T13:00:05 pairofdice: grab some prepackaged VM image for VMware Player, use that 2011-11-12T13:00:10 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-12T13:00:31 pairofdice: no restarts, works quite fast, doesn't have the braind damaged horrors of cygwin 2011-11-12T13:02:44 though personally, I consider running XP nowadays to be equivalent of purposefully dropping the soap in high security prison showers and bending for it 2011-11-12T13:03:25 (when it comes to security) 2011-11-12T13:04:35 Hmph :) 2011-11-12T13:05:03 avdg: yep =) 2011-11-12T13:05:34 yes, you can do quite a lot if you're one of the rare wizards of NT security controls, but the UI message passing will kill you without UAC and secure desktop 2011-11-12T13:05:43 *** xathis has quit IRC () 2011-11-12T13:10:59 Holy fine structure constant, it compiled fine 2011-11-12T13:12:04 What kind of an inside-joke is --funroll-loops? 2011-11-12T13:15:00 *** dvladim has quit IRC (Ping timeout: 256 seconds) 2011-11-12T13:16:40 Ahh, c++ starteribotti käänty ongelmitta, hauskaa 2011-11-12T13:18:43 pairofdice: are you finnish? 2011-11-12T13:18:59 Guilty 2011-11-12T13:19:35 eheh nice 2011-11-12T13:19:42 student? 2011-11-12T13:21:09 *** willem has joined #aichallenge 2011-11-12T13:22:19 Nah, unemployed social outcast 2011-11-12T13:22:32 =X 2011-11-12T13:23:10 I noticed that because I was doing 1 erasmus semester at TUT last year 2011-11-12T13:23:44 I really enjoyed it 2011-11-12T13:24:43 Are there any plans to change the game settings for the final tournament or are they fixed (things like attackradius2, turntime etc.)? 2011-11-12T13:25:33 there's no plans 2011-11-12T13:26:55 amstan: ok thanks :) 2011-11-12T13:27:29 *** thestinger has joined #aichallenge 2011-11-12T13:27:42 Is this written down somewhere since some approaches to the problem could become invalid if they change them. 2011-11-12T13:28:10 if something is broken with the current game we might change it, but it doesn't look like it 2011-11-12T13:28:38 I'm actually using the dead ant info now xD 2011-11-12T13:29:17 amstan: Well if its before the tournament, its ok. Thanks. 2011-11-12T13:29:26 yeah, probably 2011-11-12T13:31:09 *** NoxiaZ^ has joined #aichallenge 2011-11-12T13:37:19 > but now I've decided I want to challenge myself and do a top500 using Octave 2011-11-12T13:37:24 thestinger: really???? 2011-11-12T13:37:28 Octave (and Matlab) are both pretty fast as long as you do vector operations 2011-11-12T13:37:39 pedrosorio: 2011-11-12T13:37:49 BenJackson_: have you tried it? ;) 2011-11-12T13:37:58 not for this contest 2011-11-12T13:38:05 Surprise us with eigenvectors 2011-11-12T13:38:36 in Octave if you have some vector of numbers and want to operate on them all in parallel, like a .* 5 .+ 7 2011-11-12T13:38:41 that's way way faster than writing a for loop in octave 2011-11-12T13:39:06 so as long as you can think/write your code in terms of all ants in parallel rather than one at a time you'll be fine 2011-11-12T13:39:12 BenJackson_ I know how to use Octave/Matlab and the general guidelines of optimization, otheriwse I wouldn't be trying to do it 2011-11-12T13:39:18 ah 2011-11-12T13:39:24 but it's not trivial like say, in C++ 2011-11-12T13:39:31 that's why it's a challenge for me 2011-11-12T13:39:42 I kind of think that Octave might be easier than python 2011-11-12T13:40:06 they're probably similar for interpreted speed but octave can do lots of the math "built in" 2011-11-12T13:40:21 BenJackson_: nah 2011-11-12T13:40:51 I've tried implementing BFS naively (i.e. "normal" programming) in octave you waste your turn time doing one of those 2011-11-12T13:42:49 well I think my goal for today is tossing out my variant of CD and doing something entirely different 2011-11-12T13:43:43 *** NoxiaZ^ has quit IRC () 2011-11-12T13:46:17 *** sigh has joined #aichallenge 2011-11-12T13:49:11 *** roflmao has joined #aichallenge 2011-11-12T13:52:45 hm I've got an idea for combat zapping 2011-11-12T13:53:06 when a friendly ant sees an enemy ant 2011-11-12T13:53:16 Run away untill you get rescued 2011-11-12T13:53:20 I know that tactic 2011-11-12T13:53:22 that is nearby but outside of shooting range 2011-11-12T13:53:26 pairofdice: xD 2011-11-12T13:53:36 lol pairofdice that works too :P 2011-11-12T13:53:38 run to the hill! 2011-11-12T13:53:51 you tell it to go to the outline of the combat radius + 1 2011-11-12T13:54:06 so it will be ONE square away from engaging in combat 2011-11-12T13:54:31 and you grab your ant's nearest friend 2011-11-12T13:54:32 and send him to the same outline 2011-11-12T13:54:34 once both ants are on the outline right outside of combat range 2011-11-12T13:54:40 send them in to zap the enemy ant 2011-11-12T13:54:50 two gang up on one, you don't lose an ant and he does 2011-11-12T13:55:00 i explained that horrible 2011-11-12T13:55:02 horribly* 2011-11-12T13:55:27 BTW kind of pro tip: 2011-11-12T13:55:39 if you hack ants.py you can make it so that maps don't have to be symmetrical 2011-11-12T13:55:47 so you can create test scenarios to test your algorithms 2011-11-12T13:56:00 i found some horribly wrong bugs in my algorithms that just weren't obvious in a full game with hundreds of ants 2011-11-12T13:56:12 but when you have a controlled environment you can really test your algorithms well 2011-11-12T13:56:36 so right now I have two ants close to one enemy ant and I want to code it so they zap that enemy ant without losing any myself 2011-11-12T13:57:47 pedrosorio: yeah, I'm actually using the dead ant info for something small :P 2011-11-12T13:58:18 for figuring out if the enemy ants are staying on their hill or moving off immediately when I'm trying to raze it 2011-11-12T13:58:43 so now I can avoid sacrificing all my ants when they are staying on their hill (by surround it and waiting for it to drain) 2011-11-12T13:58:54 but I'll still try to take it right away if they aren't staying on their hill 2011-11-12T13:59:24 hmmm interesting 2011-11-12T14:00:00 when implementing the octave starter package I thought no one would use them for anything :P 2011-11-12T14:02:24 pedrosorio: good surprise eh? 2011-11-12T14:02:33 thestinger: I once saw my bot set up a perfect killzone at a hill and zap dozens of ants 2011-11-12T14:02:46 but then there was a gap so one of the killers stepped up and then there was chaos 2011-11-12T14:03:45 yeah, I made it so my ants wouldn't just suicide onto hills with ants coming out of them - but they ended up doing the same thing by waiting next to hill with ants moving off it right away :P 2011-11-12T14:03:57 amstan: bad surprise because I didn't implement them as the way it was in C++ was too slow for Octave =X 2011-11-12T14:03:57 so they were all dying from the enemy ants colliding into them 2011-11-12T14:04:54 I just left a comment line saying you can implement it in the state grid if you want (something as simple as keeping a rows*columns*maxplayers tensor to keep count of dead ants for each player, should be enough) 2011-11-12T14:06:57 BenJackson_: http://ants.fluxid.pl/replay.30714 here's the game where I noticed the issue 2011-11-12T14:07:19 you can see all my ants dying to the hill near the end, but they're actually dying b/c they are waiting next to it and the enemy ant is colliding with them 2011-11-12T14:07:35 my ants are fairly good about that 2011-11-12T14:07:56 however I just watched a replay where I could have taken a couple decisive steps and instead let a DIFFERENT ant cap the hill 2011-11-12T14:08:02 left to myself I would have gotten it 2011-11-12T14:11:01 *** roflmao has quit IRC (Quit: Leaving.) 2011-11-12T14:16:55 *** thestinger has quit IRC (Quit: WeeChat 0.3.6) 2011-11-12T14:19:22 *** mikewintermute has left #aichallenge 2011-11-12T14:26:00 *** dvladim has joined #aichallenge 2011-11-12T14:27:02 *** Anilm3 has joined #aichallenge 2011-11-12T14:28:44 *** mikewintermute has joined #aichallenge 2011-11-12T14:30:19 *** Anilm3 has quit IRC (Client Quit) 2011-11-12T14:30:49 * BenJackson_ throws a bot up on the new tcpants.com 2011-11-12T14:33:17 is there some up to date information on the replaydata format? 2011-11-12T14:33:19 poor testbot 2011-11-12T14:36:19 *** mikewintermute has left #aichallenge 2011-11-12T14:44:53 *** jasox has joined #aichallenge 2011-11-12T14:45:47 Heh, my rank 1200 bot got paired with rank ~80 bots and swiftly destroyed 2011-11-12T14:46:20 that happens... 2011-11-12T14:46:38 Though I'm expeting it to get to atleast 200 2011-11-12T14:46:44 not that I'm doing something in Octave, I'm actually thinking it would be nice if I could submit to the competition bots made in 2 different languages 2011-11-12T14:46:50 oh well 2011-11-12T14:47:50 *** thestinger has joined #aichallenge 2011-11-12T14:53:17 *** jasox has quit IRC (Quit: Leaving) 2011-11-12T14:55:27 *** hacklash has joined #aichallenge 2011-11-12T14:59:12 *** hacklash has left #aichallenge 2011-11-12T14:59:51 you can just send your bot and they figure out your language right? 2011-11-12T15:02:31 *** Accoun has quit IRC () 2011-11-12T15:04:22 *** dvladim has quit IRC (Ping timeout: 256 seconds) 2011-11-12T15:04:59 marijnfs: It looks for a file named MyBot.XXX, where XXX is the typical file extension for your language. 2011-11-12T15:05:15 (c, rb, py, hs, …) 2011-11-12T15:08:27 *** KP13 has joined #aichallenge 2011-11-12T15:10:54 *** dom7b5 has quit IRC (Ping timeout: 240 seconds) 2011-11-12T15:11:16 *** jhannah has joined #aichallenge 2011-11-12T15:11:18 *** Kingpin13 has quit IRC (Ping timeout: 240 seconds) 2011-11-12T15:11:35 any Perl folk about? 2011-11-12T15:12:12 playing around with AI::Pathfinding::AStar :) 2011-11-12T15:14:44 Perl has that? 2011-11-12T15:15:07 want 2011-11-12T15:15:09 :P 2011-11-12T15:15:21 https://metacpan.org/module/AI::Pathfinding::AStar 2011-11-12T15:15:30 yep I've searched 2011-11-12T15:15:35 nice 2011-11-12T15:15:35 I assume AStar is pretty popular? 2011-11-12T15:15:42 yes 2011-11-12T15:15:49 do you mean in AI 2011-11-12T15:15:52 or in this contest? 2011-11-12T15:15:57 for this contest 2011-11-12T15:16:18 as far as I've seen, not really 2011-11-12T15:16:22 but who knows 2011-11-12T15:16:39 So many ants, not enough bullets 2011-11-12T15:16:39 oh? is there some other pathfinding algo that's more popular for this? 2011-11-12T15:16:49 not necessarily 2011-11-12T15:16:59 but for this contest doing bfs is proving more efficient 2011-11-12T15:17:13 what's bfs? 2011-11-12T15:17:15 A* allows you to get source->destination 2011-11-12T15:17:23 but you have too many sources(ants) 2011-11-12T15:17:29 breadth first search 2011-11-12T15:18:03 it searches everything from a given point outwards 2011-11-12T15:18:22 "breadth first" 2011-11-12T15:19:47 huh. thanks. :) this is my first AI attempt 2011-11-12T15:19:56 *** mleise has quit IRC (Quit: Leaving.) 2011-11-12T15:20:05 *** djr_ has joined #aichallenge 2011-11-12T15:20:09 Go forth and experiment 2011-11-12T15:20:16 my ants are getting stuck on walls, thought AStar was the way to go to fetch the food :) 2011-11-12T15:20:58 if there's a general "spread out, dudes" way to use bfs to successfully route around walls, i'm not aware of it 2011-11-12T15:22:11 *** choas has joined #aichallenge 2011-11-12T15:22:29 *** junky has joined #aichallenge 2011-11-12T15:22:58 The Markoc Decision Processes videos from the #aiclass I watched today seem relevant 2011-11-12T15:23:01 Markov 2011-11-12T15:23:31 pairofdice: URL? 2011-11-12T15:23:38 Some are calling it diffusion 2011-11-12T15:23:54 *** dom7b5 has joined #aichallenge 2011-11-12T15:23:58 https://www.ai-class.com 2011-11-12T15:25:23 jhannah BFS can certainly avoid walls 2011-11-12T15:25:38 A* is just a slight modification of BFS to get to a destination faster 2011-11-12T15:25:57 It's easy to 'reuse' the results of BFS for all the ants. A* is a bit trickier 2011-11-12T15:26:00 if you have a heuristic that tells you how close an intermediate step is 2011-11-12T15:27:25 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-12T15:27:36 when I first started looking at aichallenge I thought about using Floyd Warshall to compute all paths 2011-11-12T15:28:20 it's O(n^3) but that seemed ok for map sizes of 150.. until my brain kicked in and I realized that n is the area there so it's more like n^6 in the width of the map 2011-11-12T15:28:24 with similar storage requirements :) 2011-11-12T15:29:16 *** mcstar has joined #aichallenge 2011-11-12T15:29:37 *** g0llum has joined #aichallenge 2011-11-12T15:30:38 Floyd Warshall OMG 2011-11-12T15:30:52 they do give you a lot of storage though 2011-11-12T15:31:06 floyd warshall is probably overkill :P 2011-11-12T15:31:25 but i'm not writing a bot so i don't know ;) 2011-11-12T15:31:45 and knowing the shortest paths between all nodes is too much 2011-11-12T15:31:47 *** treeform has quit IRC (Remote host closed the connection) 2011-11-12T15:37:07 *** datachomper has joined #aichallenge 2011-11-12T15:37:46 *** age has joined #aichallenge 2011-11-12T15:38:12 *** age is now known as Guest37890 2011-11-12T15:39:39 *** datachomper has quit IRC (Client Quit) 2011-11-12T15:42:34 *** Guest37890 has quit IRC (Ping timeout: 265 seconds) 2011-11-12T15:42:38 Ahh, new bot by pguillory, no games yet 2011-11-12T15:42:43 Betting starts now! 2011-11-12T15:43:34 oh, no wonder i moved up to 16 2011-11-12T15:49:16 *** Extrarius has quit IRC (Read error: Connection reset by peer) 2011-11-12T15:49:29 *** Extrarius has joined #aichallenge 2011-11-12T15:53:30 *** Accoun has joined #aichallenge 2011-11-12T15:57:04 im confused.. how can those ants kill each other if the attackradius2==5 http://www.abload.de/image.php?img=screenshothq8d2.png 2011-11-12T15:57:52 <_flag> They kill each other next turn 2011-11-12T15:58:16 but the current attackradius ^2 is equal to 16 2011-11-12T15:58:19 i dont get it 2011-11-12T15:58:23 ikaros: that's just the way the visualizer works... it draws the fight lines during the move resulting in the fight 2011-11-12T15:58:27 <_flag> Next turn they both move towards each other 2011-11-12T15:58:31 ahhh 2011-11-12T15:58:32 ikaros: so the lines are already there before they actually fight 2011-11-12T15:58:37 ok thanks guys 2011-11-12T15:58:42 but otherwise it would be hard to see them 2011-11-12T15:58:45 was thinking i was going mad 2011-11-12T15:58:56 ikaros: the replay is divided into two steps per turn 2011-11-12T15:59:04 because i checked several games and it was like this everywhere :) 2011-11-12T15:59:07 yea 2011-11-12T15:59:12 watch the turn counter as you hit the step button 2011-11-12T15:59:19 my mistake because i used the arrow keys 2011-11-12T15:59:34 they just skip the half turns 2011-11-12T15:59:57 oh cool, I didn't know about arrow keys 2011-11-12T16:00:22 *** roflmao1 has joined #aichallenge 2011-11-12T16:00:24 they dont work with pentadactyl 2011-11-12T16:00:26 its nice but as you see it can confuse one :) 2011-11-12T16:00:32 but now that i know.. :) 2011-11-12T16:00:45 mcstar: in vimperator/pentadactyl you can turn on keypress passthrough 2011-11-12T16:00:58 *** roflmao2 has joined #aichallenge 2011-11-12T16:01:07 theres always a tradeoff damnit 2011-11-12T16:01:15 pentadactyl? 2011-11-12T16:01:25 vimperator fork by a few of the devs 2011-11-12T16:01:27 vim-like keyboard navigation for ff 2011-11-12T16:02:07 *** KP13 is now known as Kingpin13 2011-11-12T16:02:08 huh, somehow that made me think of a RTS-like client for ants 2011-11-12T16:02:28 does anyone here have experience with the visualizer? 2011-11-12T16:02:29 with keyboard navigation of course 2011-11-12T16:02:42 roflmao2: yes, I'm using it 2011-11-12T16:02:51 is it normal for it to say 2011-11-12T16:02:54 v setlinewidth 2 # unknown action 2011-11-12T16:03:05 no, the normal engine will say things like that 2011-11-12T16:03:06 you are outputting to stdout 2011-11-12T16:03:15 shouldnt do that 2011-11-12T16:03:18 and I want to output to? 2011-11-12T16:03:25 roklmao 2011-11-12T16:03:27 stdout is right for the visualizer 2011-11-12T16:03:36 yeah stdout should be right for the visualizer 2011-11-12T16:03:53 what is this visualizer anyway? 2011-11-12T16:03:58 *** mtmcannon has joined #aichallenge 2011-11-12T16:04:01 it visualizes 2011-11-12T16:04:07 "the visualizer" is something else i think 2011-11-12T16:04:30 *** roflmao1 has quit IRC (Ping timeout: 240 seconds) 2011-11-12T16:04:32 hey all. I'm trying to set up an unofficial game server so I can test bots faster with some of my friends 2011-11-12T16:04:40 *** roflmao2 is now known as roflmao 2011-11-12T16:04:49 mtmcannon, sorry to burst your bubble, but one already exists 2011-11-12T16:04:54 ants.fluxid.pl 2011-11-12T16:04:59 the compute nodes unzip and compile submissions, but don't play any games 2011-11-12T16:05:16 roflmao: I want one of my own :-( 2011-11-12T16:05:33 how often does that one play? 2011-11-12T16:05:42 i'll average about 60 games in a day? 2011-11-12T16:05:46 don't leave it running all the time 2011-11-12T16:05:58 bout a game per minute I'd say 2011-11-12T16:06:10 yeah I want to set one up for my coworkers... 2011-11-12T16:06:30 *** jmcarthur has quit IRC (Ping timeout: 240 seconds) 2011-11-12T16:06:30 game per minute for you, not for the whole server though 2011-11-12T16:07:08 but, if you want to set another up, that's cool :) 2011-11-12T16:07:17 can't help you unfortunately cause I have no idea how it works 2011-11-12T16:08:40 *** ALplus has joined #aichallenge 2011-11-12T16:09:16 *** ProperChaos has joined #aichallenge 2011-11-12T16:11:22 someone set up tcpants.com too 2011-11-12T16:13:56 thx 2011-11-12T16:14:40 I want that code :-( 2011-11-12T16:14:48 ? 2011-11-12T16:14:52 you can have it 2011-11-12T16:14:59 for setting up the tcp server? 2011-11-12T16:15:01 yes 2011-11-12T16:15:03 where? 2011-11-12T16:15:05 why not? 2011-11-12T16:15:12 you are encouraged to do that 2011-11-12T16:15:33 hooray! 2011-11-12T16:15:34 is is hard? 2011-11-12T16:15:40 is it hard* 2011-11-12T16:16:02 https://github.com/berak/ants-tcp 2011-11-12T16:16:05 has anyone already coded something to run single games with a given map etc over network? 2011-11-12T16:16:07 probably this is it 2011-11-12T16:16:31 mcstar: thanks! 2011-11-12T16:19:11 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2011-11-12T16:20:45 *** yahim91 has quit IRC (Ping timeout: 265 seconds) 2011-11-12T16:24:46 FINALLY got tcp working 2011-11-12T16:32:56 *** borist has joined #aichallenge 2011-11-12T16:33:21 *** jmcarthur has joined #aichallenge 2011-11-12T16:35:36 *** borist has quit IRC (Client Quit) 2011-11-12T16:37:09 for some reason tcpants is still pairing me with the testbot over and over 2011-11-12T16:37:22 I should probably stop it and let it reshuffle the pairings or something 2011-11-12T16:39:47 *** borist has joined #aichallenge 2011-11-12T16:42:02 *** borist has quit IRC (Client Quit) 2011-11-12T16:44:29 *** mtmcannon has quit IRC (Quit: Page closed) 2011-11-12T16:53:41 *** Harpyon has quit IRC (Quit: Computer has gone to sleep.) 2011-11-12T16:55:47 *** pairofdice has quit IRC (Quit: Just keep breathing) 2011-11-12T17:02:02 why is the attack radius squared? 2011-11-12T17:02:42 to be int 2011-11-12T17:03:10 ? 2011-11-12T17:03:38 roflamo: because no one in their sane mind computes distances, instead of squared distances 2011-11-12T17:04:27 i guess 2011-11-12T17:07:39 because of pythogoras 2011-11-12T17:07:43 that bastard 2011-11-12T17:15:48 at one time you gonna need to un-square it, but im not going to tell where... 2011-11-12T17:16:31 *** mikewintermute has joined #aichallenge 2011-11-12T17:20:30 grr! newbie at Java and MyBot is crashing: http://pastebin.com/w7Lrjuw9 2011-11-12T17:20:44 mikewintermute: try Matlab/Octave :P 2011-11-12T17:20:56 octave? 2011-11-12T17:21:02 dont pay attention 2011-11-12T17:21:13 lol 2011-11-12T17:21:15 =( 2011-11-12T17:21:28 mikewintermute: ive seen that error before 2011-11-12T17:21:50 yeah. gotta be something really basic 2011-11-12T17:22:03 http://myarch.com/classnotfound 2011-11-12T17:22:05 it has to do with some class path fuckup 2011-11-12T17:22:21 i dont know java, but someone had that here before 2011-11-12T17:22:27 aha! cool - will checkout 2011-11-12T17:23:18 mikewintermute: dont run it as a jar 2011-11-12T17:23:26 try with -cp and the class name 2011-11-12T17:24:23 like this? java -cp MyBot.class 2011-11-12T17:24:29 no 2011-11-12T17:24:34 java -cp MyBot 2011-11-12T17:24:46 but since MyBot.class is not coing to be tehre 2011-11-12T17:24:54 you need to specify the class path 2011-11-12T17:25:05 -cp path/to/damn/class/files 2011-11-12T17:25:25 Class files: JDK or just MyBot class files? 2011-11-12T17:25:27 i think somehow your starterbot got jarred up badly, not sure 2011-11-12T17:25:38 all the starterbots .class files 2011-11-12T17:25:49 where they are 2011-11-12T17:26:11 give it a try, should work, but dont blame me if it wont 2011-11-12T17:29:08 aha! answer was to use "java MyBot"! 2011-11-12T17:29:12 http://aichallenge.org/forums/viewtopic.php?f=25&t=1579&p=9736&hilit=exception#p9736 2011-11-12T17:29:28 ? 2011-11-12T17:29:29 (haven't the foggiest why!) 2011-11-12T17:29:40 if the classes are right there 2011-11-12T17:29:46 but if they arent, you need -cp 2011-11-12T17:29:54 mcstar: k - thanks. 2011-11-12T17:30:13 anyway, you should be able to run a jar 2011-11-12T17:30:23 but i really dont know why that failed 2011-11-12T17:30:46 me neither 2011-11-12T17:33:15 *** Regis has joined #aichallenge 2011-11-12T17:36:13 *** Redgis has quit IRC (Ping timeout: 240 seconds) 2011-11-12T17:40:29 *** willem has quit IRC (Quit: Lost terminal) 2011-11-12T17:43:17 *** Harpyon has joined #aichallenge 2011-11-12T17:43:21 *** jnwhiteh has quit IRC (*.net *.split) 2011-11-12T17:43:41 *** jnwhiteh has joined #aichallenge 2011-11-12T17:45:27 *** Zao has quit IRC (Ping timeout: 258 seconds) 2011-11-12T17:45:39 *** Zao has joined #aichallenge 2011-11-12T17:52:51 another newbie Java question: re tutorial code there is a line: "Ants ants = getAnts();". I assume Ants is a Java class but how does Java know where to find it? the only import I can see is "import java.io.IOException;" 2011-11-12T17:53:51 mikewintermute: the default package... or the package the file belongs to is automatically included as a "import" 2011-11-12T17:54:32 aha - you mean Bot.java? 2011-11-12T17:54:39 note also that a single file can have more than one class too 2011-11-12T17:54:43 ? 2011-11-12T17:55:22 so: public class MyBot extends Bot 2011-11-12T17:55:31 means import Bot.java, right? 2011-11-12T17:55:32 what are you talking about now? 2011-11-12T17:55:35 no 2011-11-12T17:55:44 its does not "import" like that 2011-11-12T17:55:44 *** Rinum has joined #aichallenge 2011-11-12T17:55:54 import is a namespace thing 2011-11-12T17:56:08 its not C headers 2011-11-12T17:56:23 amazing... xathis still maintains his lead 2011-11-12T18:00:48 *** apple8 has joined #aichallenge 2011-11-12T18:04:45 @rankings 2011-11-12T18:04:46 amstan: Top 10 players: xathis(89.8), GreenTea(86.2), delineate(84.9), MomoBot(84.9), sir_macelon(84.2), teapotahedron(84.1), GarySWest(83.8), ThisIsNotABug(83.6), Migi32(82.7), lordap(82.3) 2011-11-12T18:05:11 xathis hacked the matrix, thats why 2011-11-12T18:07:57 *** choas has quit IRC (Read error: Operation timed out) 2011-11-12T18:12:10 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-12T18:12:32 pguillory just resubmitted though so watch out :) 2011-11-12T18:14:17 it's interesting to see how everyone is following the "Big Bots" 2011-11-12T18:14:45 doh. feels as if the Tutorial is missing a step: http://pastebin.com/sCAHwTT6 2011-11-12T18:19:08 I really need an offensive strat (currently have none besides going after hills) http://ants.fluxid.pl/replay.31967 2011-11-12T18:20:17 anyone up to solving a math problem? 2011-11-12T18:20:23 *** delt0r_ has quit IRC (Ping timeout: 245 seconds) 2011-11-12T18:20:33 amstan: go for it 2011-11-12T18:20:35 what's the problem? 2011-11-12T18:20:46 is it round? 2011-11-12T18:20:54 i'm filling a blank(all dirt) map randomly with water 2011-11-12T18:20:58 mikewintermute: haha nice! wintermute... from neuromancer 2011-11-12T18:20:59 50% chance that it gets to be a water 2011-11-12T18:21:15 now.. placing a water is symmetric, so water might get placed to other areas too 2011-11-12T18:21:28 what percentage of water will i have at the end? 2011-11-12T18:21:51 Rinum: darn. discovered. 2011-11-12T18:21:59 *** ztfw has quit IRC (Remote host closed the connection) 2011-11-12T18:22:34 amstan: 50% chance each square is water? 2011-11-12T18:23:00 Rinum: yes, but i have symmetry going on, so placing water at 0,0 will also place it at 9,9(given the map is 10x10) 2011-11-12T18:23:33 isn't that around 75% chance, or am I solving it too quickly? 2011-11-12T18:23:39 idk 2011-11-12T18:23:47 hmm, i could go around this by making stuff to dirt 2011-11-12T18:23:48 amstan: are you looping through each square on the grid? 2011-11-12T18:23:50 yes 2011-11-12T18:24:26 so 50% 0,0 is water AND 50% 9,9 is water... and if either is water then both is water? 2011-11-12T18:24:55 yes 2011-11-12T18:25:04 50% 2011-11-12T18:25:24 it has to be > 50%, no? 2011-11-12T18:25:29 hmm yes 2011-11-12T18:26:04 you should just run through one of the symetric parts no? 2011-11-12T18:26:13 this way you get 50% water, simple 2011-11-12T18:26:17 well, I see the problem like this: you draw on the map randomly twice (50%) 2011-11-12T18:26:35 so you have 4 states in the end 2011-11-12T18:26:48 grwip: i don't know what my one symmetric area is 2011-11-12T18:26:55 oh ok 2011-11-12T18:27:04 1) none of the layers draw a pixel 2&3) 1 of the layers draw a pixel 4) both draw a pixel 2011-11-12T18:27:17 avdg: yes, i tried dividing by number of symmetric areas, but i will have much less than 50% then 2011-11-12T18:27:34 avdg: because it can't put water where there's already water 2011-11-12T18:27:42 yeah, what am I telling? 2011-11-12T18:27:51 (state 4) 2011-11-12T18:28:10 hmm? 2011-11-12T18:28:16 but maybe I'm just explaining it badly 2011-11-12T18:29:16 (or I don't get it) 2011-11-12T18:29:48 but you'll have a porblme doing it like that because you'll have much more water in the second part of your map 2011-11-12T18:30:10 no, I'm off, too late 2011-11-12T18:30:11 sorry 2011-11-12T18:30:16 amstan: as long as you don't run any probabilities on the symmetric squares it should come out to 50% 2011-11-12T18:30:27 grwip: putting water at 9,9 also puts water at 1,1 2011-11-12T18:30:50 not sure if that helps 2011-11-12T18:30:50 amstan: yep I got it, but a bit too late, 'think before typing' 2011-11-12T18:31:04 (hmm, I think I have to revert my question, so I get that 75% into a 50%) 2011-11-12T18:31:53 hmm, what about 33% 2011-11-12T18:32:06 amstan: if water is placed translationally symmetrically, than its always 50 percent 2011-11-12T18:32:19 *** Jak_o_Shadows has joined #aichallenge 2011-11-12T18:32:55 but of course this changes if you have other constraints 2011-11-12T18:32:55 if you fill the map 33%, add symmetric on top of it, it should result in around 50% and if I still don't get it, I may need a better example :p 2011-11-12T18:33:08 *** delt0r_ has joined #aichallenge 2011-11-12T18:33:28 I think you were right at the beginning, i.e.e you get 75% water 2011-11-12T18:34:06 ok.. would it be easier if for every square i turn it to dirt right before i decide if i should make it water again? 2011-11-12T18:34:10 will i get 50% then? 2011-11-12T18:34:42 amstan: what are you trying to do? 2011-11-12T18:34:47 yeah, i should, the state of the spot before will not affect the percentage probability of the output 2011-11-12T18:34:48 *** Rinum_ has joined #aichallenge 2011-11-12T18:34:55 it's 50% 2011-11-12T18:34:56 mcstar: fill the map randomly with 50% water 2011-11-12T18:35:00 I think amstan wants to generate a symmetric map with 50% water 2011-11-12T18:35:07 *** Rinum has quit IRC (Ping timeout: 265 seconds) 2011-11-12T18:35:18 50 percent of total area to water area 2011-11-12T18:35:23 amstan: an easy way with random walk, is just stop once its 50% 2011-11-12T18:35:27 if 2 squares are linked... the chance that they both are water ONLY depends on the 2nd square 2011-11-12T18:35:31 but you still have the connected part 2011-11-12T18:35:45 *** Antimony has joined #aichallenge 2011-11-12T18:35:48 delt0r_: yeah, i think i'll do random walk then the automata instead of this random business 2011-11-12T18:36:02 amstan: you better fill up a bravais cell with 50 percent random water anyway it happends, than tile this to fill the map 2011-11-12T18:36:12 therefore, either both squares are not water or they both are... so 50% 2011-11-12T18:37:12 btw, this is getting rather slow 2011-11-12T18:37:15 *** Rinum_ is now known as RInum 2011-11-12T18:37:17 how can i do this in pypy instead? 2011-11-12T18:37:31 use cython? 2011-11-12T18:37:32 oh.. nvm, i researched this, it's not in the ubuntu repo 2011-11-12T18:37:38 *** RInum is now known as Rinum 2011-11-12T18:37:39 RInum: nah, i don't want to restrict my code 2011-11-12T18:38:45 RInum: you flip two fair coins, if either one comes up heads you get a prize, you either get the prize or you don't what is the chance of winning a prize? (the answer isn't 50% ;) ) 2011-11-12T18:39:19 let's see how this random walk business works 2011-11-12T18:39:47 you have 75% to win the prize 2011-11-12T18:39:57 *** ProperChaos has quit IRC (Ping timeout: 265 seconds) 2011-11-12T18:40:04 yay apparently i'm doing well in ai class 2011-11-12T18:40:08 T-H H-T H-H against T-T 2011-11-12T18:40:26 and it's the same situation with your water on the map assuming each all squares are linked as pairs 2011-11-12T18:40:34 heh, I just don't the mathematical notations of randomness 2011-11-12T18:40:50 *don't get 2011-11-12T18:40:56 we dont either, dont worry 2011-11-12T18:41:37 avdg: you got it right when you enumerated the solution beofre 2011-11-12T18:41:49 janzert: yeah I was wrong... I should have said... disregard the 1st toin coss... only count the 2nd 2011-11-12T18:42:02 lol 2011-11-12T18:42:16 was this on purpose? 2011-11-12T18:42:43 grwip: yeah, but the question was how to get that 50% in the end starting from a full map and applying a symmetric thing on it 2011-11-12T18:43:04 so I then said 33%, but I'm still not sure 2011-11-12T18:43:06 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-12T18:43:26 In Ilk.java it says "Represents type of tile on the game map". Why's it called Ilk? 2011-11-12T18:43:29 (well, I'm thinking 33% + (33% / 2) 2011-11-12T18:43:48 I think 1/6 2011-11-12T18:44:09 hmm… "type" or "sort" 2011-11-12T18:44:19 I think we want to solve 3 * P(Water) = 1/2 2011-11-12T18:44:24 bizarre 2011-11-12T18:44:37 hmm 2011-11-12T18:44:38 hmm, 3 layers? 2011-11-12T18:45:03 because with the two linked tiles on the map we have WW WL LW against LL 2011-11-12T18:45:08 this random walk business up to 50% is hard, because there's no way for me to count how many squares are water except with an interator 2011-11-12T18:45:27 i can't do an accumulator because more than one water might be placed 2011-11-12T18:45:53 no that's not it... 2011-11-12T18:46:15 I'll stick with my 33% until I know more ;-) 2011-11-12T18:46:38 nvm, i can just divide 2011-11-12T18:47:57 hmm 2011-11-12T18:48:02 what about wolframalpha 2011-11-12T18:48:11 ok I guess it's 1/4 : with P(W) = p we have P(WW) = p^2, P(WL) = P(LW) = p(1-p) 2011-11-12T18:48:38 so we want to solve p^2 + 2p(1-p) = 1/2 2011-11-12T18:49:04 which is note 1/4 but I think it's what we want to solve 2011-11-12T18:49:28 http://www.wolframalpha.com/input/?i=2*P%28x%29+%3D+.5 <- I just typed this 2011-11-12T18:49:53 which is totally different from what I was thinking 2011-11-12T18:51:21 unless its calculating something else... 2011-11-12T18:51:46 it is calculating something else, here P is the zeta function 2011-11-12T18:52:10 ah 2011-11-12T18:52:14 http://www.wolframalpha.com/input/?i=permutations 2011-11-12T18:53:28 ok, I say P(w) = 29.28% 2011-11-12T18:54:21 which is the solution to p^2 + 2p(1-p) = 1/2 2011-11-12T18:54:48 this? http://www.wolframalpha.com/input/?i=+p%5E2+%2B+2p%281-p%29+%3D+1%2F2 2011-11-12T18:54:49 :p 2011-11-12T18:55:11 (I would try that blind) 2011-11-12T18:55:51 hmm, that first dot approximates that number you just told 2011-11-12T18:56:02 yes I confess I got it there 2011-11-12T18:56:16 @amstan: so you want to get 50% water in the end but water is placed at (x,y) and (w-x,h-y) ? 2011-11-12T18:56:17 pedrosorio: I'm sorry Dave, err pedrosorio; I cannot 'amstan:'. 2011-11-12T18:56:19 tried to solve by hand at first but my math is really off this evening 2011-11-12T18:56:26 pedrosorio: yes 2011-11-12T18:56:36 pedrosorio: but i got it now 2011-11-12T18:57:09 my primary concern now is how to make random walk maps where the bases stay connected without generating until my bfs works 2011-11-12T18:57:12 its still something interesting to see how we are trying to solve this 2011-11-12T18:57:43 hmm 2011-11-12T18:57:48 hmm, I was thinking at bfs until I read the last words 2011-11-12T18:58:08 well, i guess i could put the bfs inside the loop 2011-11-12T18:58:12 but that will be pretty slow 2011-11-12T18:58:38 what i really don't want through is have something like while(!bases connected): generate_new_map() 2011-11-12T18:58:38 hmm, I wonder how slow 2011-11-12T18:59:09 avdg: i'm looping 10 times over the whole map and it's very slow(a second) 2011-11-12T18:59:44 i still dont get what the problem is 2011-11-12T19:00:04 hmm, what about doing an a* to both sides every step? (quick idiot thinking) 2011-11-12T19:00:13 I also don't understand 2011-11-12T19:00:13 *both side of the placed tile 2011-11-12T19:00:15 what you are doing 2011-11-12T19:00:16 xD 2011-11-12T19:00:35 the way you generate a map is 2011-11-12T19:00:41 decide position of hills first 2011-11-12T19:00:43 and then? 2011-11-12T19:00:47 *** Antimony has joined #aichallenge 2011-11-12T19:00:56 add water and make sure the land stays connected 2011-11-12T19:01:10 all of it? 2011-11-12T19:01:22 you don't want islands ;-) 2011-11-12T19:01:28 yeah, probably 2011-11-12T19:01:32 connected in a mathematical sense? 2011-11-12T19:01:40 you can shrink them into a point? 2011-11-12T19:01:56 can you think about another kind of connected? 2011-11-12T19:02:00 yes 2011-11-12T19:02:11 with a hole in it 2011-11-12T19:02:31 and someone might still call that "connected" 2011-11-12T19:02:38 catch me up here: are we talking about generating maps? 2011-11-12T19:02:44 no 2011-11-12T19:02:50 only a couple of us 2011-11-12T19:02:59 as long you shoot holes in a plat without separating the plate 2011-11-12T19:03:05 *plate 2011-11-12T19:03:09 I would call the "with a hole in it" connected 2011-11-12T19:03:23 avdg: well that is not the simplest connected topology 2011-11-12T19:03:32 from wiki 2011-11-12T19:03:32 the simplest doesnt have whole in ot 2011-11-12T19:03:34 it 2011-11-12T19:03:35 "A topological space X is said to be disconnected if it is the union of two disjoint nonempty open sets. Otherwise, X is said to be connected. " 2011-11-12T19:04:01 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-11-12T19:04:31 i don't mind a few islands 2011-11-12T19:04:35 as long as there's not too many 2011-11-12T19:04:38 and too big 2011-11-12T19:04:45 and no hills on it 2011-11-12T19:04:57 amstan: you are being maticulus :D 2011-11-12T19:05:03 meticulus? 2011-11-12T19:05:03 probably 2011-11-12T19:05:07 hmm, a* to all hills? 2011-11-12T19:05:07 @mcstar: you probably meant simply connected 2011-11-12T19:05:08 pedrosorio: An error occured while trying to show the previous error. 2011-11-12T19:05:15 pedrosorio: yes 2011-11-12T19:05:19 thats the simples case 2011-11-12T19:05:22 t 2011-11-12T19:05:25 simplest? 2011-11-12T19:05:33 avdg: might take too long 2011-11-12T19:05:33 hmmm 2011-11-12T19:05:51 we call it by a different name 2011-11-12T19:05:53 maybe only call it when i'm closing an area 2011-11-12T19:06:00 I don't see how you could have a simply connected map 2011-11-12T19:06:04 in this game 2011-11-12T19:06:07 still faster than bfs 2011-11-12T19:06:16 avdg: i need it like 100 times faster 2011-11-12T19:06:28 avdg: actually, it might work 2011-11-12T19:06:35 you guys realize that a properly implemented BFS is linear in the size of the map 2011-11-12T19:06:42 it should be no worse than 'for all rows, for all cols, do something' 2011-11-12T19:06:48 BenJackson_: still too big 2011-11-12T19:06:52 ie a loop to about 10-20k 2011-11-12T19:06:52 n^2 i.e. 2011-11-12T19:07:01 40.000 is a lot :( 2011-11-12T19:07:03 n in the area of the map 2011-11-12T19:07:16 the problem with this map generation 2011-11-12T19:07:19 you want to think in vertices here 2011-11-12T19:07:23 thats n^2 2011-11-12T19:07:24 but I wonder how to find the closes tile quickly 2011-11-12T19:07:30 is that I haven't looked into all the kinds of symmetries you are using 2011-11-12T19:07:35 so I can't really say anything about it 2011-11-12T19:07:43 *** Antimony has quit IRC (Ping timeout: 248 seconds) 2011-11-12T19:07:51 on a modern CPU you could literally do 1 bfs *from every point on the map* in about 1 second 2011-11-12T19:07:56 it would be faster to find an earlier discovered tile 2011-11-12T19:08:24 BenJackson_: i think amstan wants to do that (needs) with every new water tile added 2011-11-12T19:08:28 aichallenge: Alexandru Stan amstan_mapgen * r8d65810 / (3 files): changed stuff to kwargs - http://git.io/gpbPew 2011-11-12T19:08:28 aichallenge: Alexandru Stan amstan_mapgen * ra6680ac / (ants/mapgen/mapgen.py ants/mapgen/util.py): modified random_upto - http://git.io/XtoaDQ 2011-11-12T19:08:28 aichallenge: Alexandru Stan amstan_mapgen * rd22a5c7 / ants/mapgen/cavemap.py : working on random walker instead of noise - http://git.io/NI87nA 2011-11-12T19:08:30 mcstar: n in the area of the map is the right way to think about it if you're going to apply graph algorithms to a grid problem 2011-11-12T19:08:50 this function takes a while: https://github.com/aichallenge/aichallenge/blob/d22a5c7ce882d7350c1cd2f0a6ae0f3da818d2c0/ants/mapgen/cavemap.py#L30 2011-11-12T19:08:54 *** Fandekasp has quit IRC (Ping timeout: 240 seconds) 2011-11-12T19:09:05 i'm doing it 10 times on 100x100 maps, and i'm just iterating 2011-11-12T19:09:12 BenJackson_: not in python :P 2011-11-12T19:09:19 *** dorisabayon has quit IRC (Ping timeout: 248 seconds) 2011-11-12T19:09:26 i don't want to do something similar every step of the random walk, it'll take ages 2011-11-12T19:09:57 pedrosorio: right, I said "properly implemented BFS", but that can be tricky if you lack (for example) a proper queue data structure 2011-11-12T19:10:11 if you simulate your queue with list copies then it IS n^2 in the area of the map 2011-11-12T19:11:10 *** Antimony has joined #aichallenge 2011-11-12T19:11:50 amstan: how do you use the random walk to generate the map? 2011-11-12T19:12:18 i random walk water until 50% of it is filled 2011-11-12T19:12:24 then i smooth it with some automata 2011-11-12T19:13:05 *** LouisMartin has joined #aichallenge 2011-11-12T19:13:54 hmm 2011-11-12T19:14:37 but your random walk can wander the whole map? 2011-11-12T19:14:40 so there is only 1 seed location 2011-11-12T19:14:43 pedrosorio: yes 2011-11-12T19:14:47 mcstar: yes 2011-11-12T19:15:12 so are you saying smooth() is slow? 2011-11-12T19:15:50 it shouldnt be, since you only apply an 8-point stencil on every piel 2011-11-12T19:15:53 pixel 2011-11-12T19:16:02 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-12T19:17:15 *** Antimony has joined #aichallenge 2011-11-12T19:17:46 for example, i can do that in my bot for a 100 times to become noticeable 2011-11-12T19:17:58 mcstar: yes, but this is python 2011-11-12T19:18:01 So your water will be connected. Couldn't you reverse it and start with all water and go around in a random walk deleting it? 2011-11-12T19:18:05 amstan: you could probably make the python map gen faster by doing it as an image 2011-11-12T19:18:18 and applying PIL functions to dialate/erode/etc images 2011-11-12T19:18:26 blur + threshold, eg 2011-11-12T19:18:38 *** datachomper has joined #aichallenge 2011-11-12T19:18:38 BenJackson_: PIL doesn't know about wrapped stuff 2011-11-12T19:18:39 More practically, numpy has a lot of tools for doing numeric manipulation in C 2011-11-12T19:18:57 amstan: tile the image 3x3 and then center cut it after 2011-11-12T19:19:16 nickjohnson: whouldn't adding land instead do the same thing 2011-11-12T19:19:36 amstan: If I understood correctly, you're worried about the land staying fully connected 2011-11-12T19:19:45 If you do a random walk of land instead of water, it is by definition 2011-11-12T19:19:46 g2g 2011-11-12T19:20:31 just shoot the whole map with random coordinates, remove the water if its still connected(has enough nbors) dont do it otherwise 2011-11-12T19:20:50 if the water==land stop 2011-11-12T19:22:39 ok 2011-11-12T19:22:42 problem 2011-11-12T19:22:47 c++ guys around? 2011-11-12T19:23:14 ok 2011-11-12T19:23:25 lets say i have many sources, and thus object files 2011-11-12T19:23:33 i only recompile the files i changed 2011-11-12T19:23:36 obviously 2011-11-12T19:23:44 and relink the binary 2011-11-12T19:23:56 now, after some changes i do, and relink 2011-11-12T19:24:03 sometimes my bot crashes 2011-11-12T19:24:15 if i run it under valgrind, theres no error 2011-11-12T19:24:16 you probably changed a header 2011-11-12T19:24:19 no crash 2011-11-12T19:24:25 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-11-12T19:24:32 if i do a full recompile, eversthings works 2011-11-12T19:24:33 some file you didn't recompile has the wrong shape for the structure 2011-11-12T19:24:34 *** TheLinker has joined #aichallenge 2011-11-12T19:24:42 you need dependencies in your makefile 2011-11-12T19:25:11 BenJackson_: but how come under valgrind it doesnt crash? 2011-11-12T19:25:13 the stock makefile makes *.o depend on *.h 2011-11-12T19:25:44 it'd be hard to tell you why exactly it doesn't work 2011-11-12T19:25:50 yep for me too 2011-11-12T19:25:54 the answer is really "don't do that" (don't mix .o's from mismatched .h's) 2011-11-12T19:25:59 and it is not like crashes right away, no 2011-11-12T19:26:06 a good while in the game 2011-11-12T19:26:12 k 2011-11-12T19:26:16 are you using the included makefile? 2011-11-12T19:26:24 ? 2011-11-12T19:26:28 where? 2011-11-12T19:26:31 the c++ starter has a makefile 2011-11-12T19:26:42 i have nothing to do with it 2011-11-12T19:26:52 you wrote yours in C++ from scratch you mean? 2011-11-12T19:27:01 doesnt everybody? 2011-11-12T19:27:15 http://aichallenge.org/starter_packages.php 2011-11-12T19:27:25 * mcstar would not have given free starter packages away 2011-11-12T19:27:34 BenJackson_: im here since march/may 2011-11-12T19:27:36 i know 2011-11-12T19:27:42 sure, I've rewritten a bunch of it 2011-11-12T19:27:48 but you should get the makefile it sounds like :) 2011-11-12T19:27:58 all you have to do is change the SOURCES= line to your list of .cc files 2011-11-12T19:28:05 actually ive seen it 2011-11-12T19:28:18 but still, its a weird error 2011-11-12T19:28:36 i would have expected something like that manifest eearlier after program loading 2011-11-12T19:28:42 that to* 2011-11-12T19:29:07 im going to make proper dependencies than 2011-11-12T19:29:35 *** antimatroid has quit IRC (Ping timeout: 248 seconds) 2011-11-12T19:32:02 BenJackson_: my name.cpp files depend on name.hh files 2011-11-12T19:32:32 probably more than one of your .cpp files depend on each .hh 2011-11-12T19:32:48 the starter makefile makes the simplifying assumption that all .cc depend on all .h 2011-11-12T19:32:50 but those are really unchanged 2011-11-12T19:32:56 aha 2011-11-12T19:33:13 well, thats too much dependency 2011-11-12T19:33:25 if you add a member to a class in the .h and some other .cpp (not the one that defines its methods, necessarily) declares one, any stale .o will use the wrong size 2011-11-12T19:33:32 and possibly wrong versions of inlined functions 2011-11-12T19:33:34 if i change a .h then everything recompiles 2011-11-12T19:33:45 sure 2011-11-12T19:33:52 how bit is your bot anyway? 2011-11-12T19:33:56 YES 2011-11-12T19:33:58 YES IT WORKS 2011-11-12T19:33:58 YES 2011-11-12T19:34:02 2000locs 2011-11-12T19:34:26 BenJackson_: how many lines is your battle code? 2011-11-12T19:34:31 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-12T19:34:46 just the combat or all of it? 2011-11-12T19:34:57 i meant combat by battle 2011-11-12T19:35:06 the kill-code 2011-11-12T19:35:12 * BenJackson_ just timed a parallel make at under 4 seconds 2011-11-12T19:35:21 276 lines 2011-11-12T19:35:42 7.9 secs for me 2011-11-12T19:35:51 not parallel 2011-11-12T19:35:53 another 40 I guess for a header which is fairly dedicated to combat 2011-11-12T19:36:04 mcstar: make -j 4 (or whatever) 2011-11-12T19:36:23 2 2011-11-12T19:36:32 see? not too many deps 2011-11-12T19:36:50 huh? 2011-11-12T19:36:54 what do you mean? 2011-11-12T19:37:10 2 = i onyl have 2 cores 2011-11-12T19:37:14 2 seconds doesn't seem like too high a price to pay to be sure your bot doesn't crash 2011-11-12T19:37:19 YOU DONT GET IT IT WORKS THATS THE POINT OMG 2011-11-12T19:37:24 :P 2011-11-12T19:37:29 (talking to myself here, sorry) 2011-11-12T19:37:36 roflmao: replay? 2011-11-12T19:37:38 k, manfred, now calm down 2011-11-12T19:37:58 mcstar: when doing make -j whatever usually it pays to pick a number greater than your cores 2011-11-12T19:38:29 ill see 2011-11-12T19:38:44 nah, 4.5 secs 2011-11-12T19:38:52 *** NoxiaZ^ has joined #aichallenge 2011-11-12T19:38:55 *** BenJackson_ is now known as BenJackson 2011-11-12T19:39:19 you know, i segmented my code soo nicely, because i was annoyed by a 5 sec compilation 2011-11-12T19:39:25 now on 2 cores i get 4.5 2011-11-12T19:39:28 im lame 2011-11-12T19:39:59 it all seems fast to me *shakes stick at kids on his lawn* 2011-11-12T19:40:28 are you that old grandpa? 2011-11-12T19:41:16 * BenJackson smacks his gums 2011-11-12T19:41:58 lol 2011-11-12T19:42:06 i hope i will learn english one day 2011-11-12T19:42:12 you know what's lame? 2011-11-12T19:42:20 to take 70ms to receive some input 2011-11-12T19:42:41 you measured? 2011-11-12T19:42:44 (octave - damn you octave!) 2011-11-12T19:42:46 *** abc12345 has joined #aichallenge 2011-11-12T19:42:48 what? 2011-11-12T19:42:53 pedrosorio: really? 2011-11-12T19:42:58 yep 2011-11-12T19:43:01 lol 2011-11-12T19:43:04 poor you 2011-11-12T19:43:06 why are you already speaking english? ;-) (my english isn't that good either, but sometimes I ask who cares as long the message came over clearly) 2011-11-12T19:43:15 there's one thing 2011-11-12T19:43:20 I'm measuring the time 2011-11-12T19:43:25 from the first line of input 2011-11-12T19:43:27 not from the "go" 2011-11-12T19:43:29 avdg: i was just wondering on the meaning "smacks his gum" 2011-11-12T19:43:35 im not sure what hea meant by that 2011-11-12T19:43:38 unlike the c++ starter package 2011-11-12T19:43:47 * avdg kills himself :p 2011-11-12T19:43:54 nononono 2011-11-12T19:43:57 dooont 2011-11-12T19:43:59 :D 2011-11-12T19:44:05 * avdg jumps 2011-11-12T19:44:12 * avdg on his bed 2011-11-12T19:44:21 stupid avdg tamagotchi 2011-11-12T19:44:27 mcstar: oh, just joking about old people without their dentures in 2011-11-12T19:44:35 dentures = false teeth 2011-11-12T19:44:46 BenJackson: does it mean: you are liking you gums? 2011-11-12T19:44:48 avdg lol 2011-11-12T19:44:54 r 2011-11-12T19:45:12 licking 2011-11-12T19:45:14 lol 2011-11-12T19:45:22 *** sigh has quit IRC (Remote host closed the connection) 2011-11-12T19:45:32 its late, my typos come out this time of night 2011-11-12T19:45:42 *** junky has quit IRC (Ping timeout: 265 seconds) 2011-11-12T19:46:31 anyway, that battle code is pretty long 2011-11-12T19:46:45 http://www.youtube.com/watch?v=evx3Tkf0IoE 2011-11-12T19:47:05 back to octave 2011-11-12T19:47:12 66ms to receive input 2011-11-12T19:47:23 40ms to send out orders to ants 2011-11-12T19:47:31 i think my game-tree search and evaluation is 150 lines 2011-11-12T19:47:37 *** abc12345 has quit IRC (Ping timeout: 265 seconds) 2011-11-12T19:47:46 20ms for everything else =) 2011-11-12T19:48:10 *** tsidie has joined #aichallenge 2011-11-12T19:48:34 so I played fluxid's bot on the aichallenge site 2011-11-12T19:48:52 BenJackson: great, i wont be able to sleep, like when i was 10 and wathched aliens 2011-11-12T19:49:02 I was thinking: wait, this is aichallenge, not fluxid.pl 2011-11-12T19:49:39 lol 2011-11-12T19:49:52 and you beat me ;P 2011-11-12T19:50:00 another surprise ;-) 2011-11-12T19:50:12 my bot is pretty good at mazes, less at random walks 2011-11-12T19:51:06 BenJackson: for your pleasure http://www.youtube.com/watch?v=yLLtv1ARKgU&NR=1 2011-11-12T19:51:32 what has been seen cannot be un-seen 2011-11-12T19:54:07 alllrighty time to put this new imperfect combat zapper to test 2011-11-12T19:54:22 he tries to gang up on single ants with two ants 2011-11-12T19:54:23 *** apple8 has quit IRC (Ping timeout: 265 seconds) 2011-11-12T19:55:13 Gentlemen, greetings. ಠ_ರೃ 2011-11-12T19:56:33 Any of you in C#? I have a pretty nice tip for you. 2011-11-12T19:56:49 "dont use it" 2011-11-12T19:57:27 *** sigh has joined #aichallenge 2011-11-12T19:57:28 stop stealing my jokes, mcstar. 2011-11-12T19:57:37 :) 2011-11-12T19:57:42 * avdg wants to say "dont use it yet" 2011-11-12T19:59:04 Why all the C# hate? :P 2011-11-12T19:59:23 srr, I just wanted to complete that sentence 2011-11-12T20:00:12 *** Draakon has quit IRC () 2011-11-12T20:00:14 *** cutoff has joined #aichallenge 2011-11-12T20:01:04 Does the game send the bots any information about the current scores or standings 2011-11-12T20:01:22 no, only at the end 2011-11-12T20:01:25 well too late. the Microsoft's little b*t*h inside of me is definitively hurting. -_- 2011-11-12T20:01:59 ok thank you 2011-11-12T20:02:33 http://aichallenge.org/visualizer.php?game=90336 2011-11-12T20:02:36 sweet 2011-11-12T20:03:09 :p 2011-11-12T20:03:48 i always end of watching cat videos when someone gives me a youtube link 2011-11-12T20:04:12 youtube.com *gives youtube link* 2011-11-12T20:04:17 Alright. I don't want to spam, but I'll do just one more time: Any C# coder in the house? Our code could run between 150% - 1000% faster without changing a single line of code, but I need you to vote on a motion, at the organizer request. 2011-11-12T20:04:42 lol 2011-11-12T20:04:47 * avdg wonders if mcstar would say "oh kitties" 2011-11-12T20:04:55 no 2011-11-12T20:04:58 he wouldnt 2011-11-12T20:05:06 he would say kitteh 2011-11-12T20:05:14 it would be funny if we only care about the not funny joke 2011-11-12T20:05:18 mcstar: don't take responsability, blame http://en.wikipedia.org/wiki/Toxoplasma_gondii 2011-11-12T20:05:25 k 2011-11-12T20:05:27 i know 2011-11-12T20:05:41 i have to check my dopamine levels anyway 2011-11-12T20:05:56 im sad like a poor little kitteh 2011-11-12T20:06:24 oh, you didn't saw the power of cuteness yet ;-) 2011-11-12T20:06:34 * avdg doesn't want to mention his sister 2011-11-12T20:06:47 LouisMartin: you probably will be able to get that optimization switch get through the bureucracy 2011-11-12T20:06:54 aaah it is so hard to iomplement certain evasive functionality without sacrificing production 2011-11-12T20:07:17 avdg: i dont like people 2011-11-12T20:07:27 ^^ he likes mcdonalds instead 2011-11-12T20:07:28 mcstar: Amstan told me he would switch on optimize=all if I get a few C# people to vote ... 2011-11-12T20:07:39 yeah i waas here 2011-11-12T20:07:48 mcstar: I understand that, but I still have to live with them 2011-11-12T20:08:13 LouisMartin: look up their nicks on the rankings and see if any of them is here, then address them by name 2011-11-12T20:08:32 rofl- coding AI bot while producing in your met lab is kinda cool if you ask me. It's totally worth loosing some of the product. 2011-11-12T20:08:48 loose: not tight 2011-11-12T20:08:55 lose: as in lost 2011-11-12T20:09:03 mcstar- good idea. 2011-11-12T20:09:14 heh awesome 2011-11-12T20:09:15 but i was referring to ant production 2011-11-12T20:09:15 food grabbing :P 2011-11-12T20:09:41 lol 2011-11-12T20:09:48 mcstar - how does I tell my french girlfriend that she lost her tighness then? I'm at loose for words! 2011-11-12T20:09:48 nice joke this meth lab 2011-11-12T20:10:11 rofl- oh, yea sorry :) 2011-11-12T20:10:52 subtle grammar hint is subtle. 2011-11-12T20:11:01 ? 2011-11-12T20:11:06 it was spelling, not grammar 2011-11-12T20:11:16 met lab -> meth lab. 2011-11-12T20:11:21 that too 2011-11-12T20:11:22 where are you from louis? :) 2011-11-12T20:11:35 it was trolling, not trying. 2011-11-12T20:11:56 roflmao: Montreal 2011-11-12T20:12:09 what about you, gentlemen? 2011-11-12T20:12:16 Ghent, Belgium 2011-11-12T20:12:16 im not 2011-11-12T20:12:19 mcstar: gentlemen or gentlemens? 2011-11-12T20:12:24 men 2011-11-12T20:12:27 but you knew that 2011-11-12T20:12:31 HA! sorry, it's just that when I first saw you and saw louis I was like ""hmm louis, that guy must be french" 2011-11-12T20:12:38 I'm from mexico 2011-11-12T20:12:40 mcstar: I was ;) 2011-11-12T20:13:00 i mean im not from montreal 2011-11-12T20:13:25 *** cutoff has quit IRC (Ping timeout: 260 seconds) 2011-11-12T20:13:37 roflmao: yea, well Montreal is mostly french speaking. But yea, that doesn't make us French :) 2011-11-12T20:14:05 ah yes sorry 2011-11-12T20:14:06 id probably live in canada 2011-11-12T20:14:09 nice people 2011-11-12T20:14:21 http://ants.fluxid.pl/replay.32092 — even though i epically lost, my combat code has improved, awesome 2011-11-12T20:14:22 mcstar: Detroit? Delaware? Washington D.C.? 2011-11-12T20:14:34 neither 2011-11-12T20:15:21 mcstar: compliment accepted. 2011-11-12T20:15:26 cats have fast reflexes 2011-11-12T20:17:13 i dont like animals, i just like watching cute cats, but i dont want to touch them 2011-11-12T20:17:27 mcstar: infectious increased dopamine turnover in the Substantia nigra. cats give me all of my powers. I live in the house of my cat. 2011-11-12T20:18:42 LouisMartin: say hi to your tapeworms, pinworms and other parasites for me, will you? 2011-11-12T20:18:57 All neuroscience geek jokes aside, please get tested for Toxoplasmosis. Could explain a few of your introspective questions perhaps. ;) 2011-11-12T20:19:18 aha 2011-11-12T20:19:24 that cat shiz? 2011-11-12T20:19:24 why would i have that? 2011-11-12T20:19:25 mcstar: they're reading you .. 2011-11-12T20:19:42 anyway, this cat craps into a real toilet 2011-11-12T20:19:45 very nice 2011-11-12T20:20:00 but, after the deed, he "scratches" the wc, like it was dirt 2011-11-12T20:20:17 to bury its feces that are under water 2011-11-12T20:20:44 neurotic cat 2011-11-12T20:20:51 Like Pedobear; I'm just kidding. 2011-11-12T20:21:06 LouisMartin: do you sleep with pedobear? 2011-11-12T20:21:06 haha nice mc 2011-11-12T20:21:12 here's some interesting bits: Correlations have been found between latent Toxoplasma infections and various characteristics:[30] Decreased novelty seeking behaviour[31] Slower reactions[32] Lower rule-consciousness and greater jealousy (in men)[31] Greater warmth, conscientiousness and moralistic behavior (in women)[31] 2011-11-12T20:21:33 WAIT WAT 2011-11-12T20:21:43 so I could be infected 2011-11-12T20:21:44 what erections? 2011-11-12T20:21:45 and not even notice? 2011-11-12T20:21:50 oh, reactions 2011-11-12T20:22:00 mcstar: well, I like to think he's the one sleeping with me.. but, no. 2011-11-12T20:22:08 *** Migi32 has quit IRC (Remote host closed the connection) 2011-11-12T20:22:25 rofl: yep. 2011-11-12T20:22:32 lol 2011-11-12T20:22:39 it is a part of you now 2011-11-12T20:22:46 shieeet 2011-11-12T20:23:02 unfortunately im mostly healthy 2011-11-12T20:23:20 some theorize that this parasite may even have some effects on some civilization since they modulate human behavior that much and infects as many... 2011-11-12T20:23:31 *** cutoff has joined #aichallenge 2011-11-12T20:23:38 lets kill all the cats then 2011-11-12T20:24:03 roflmao: spare you with the infection? 2011-11-12T20:24:10 yes' 2011-11-12T20:24:11 i say kill everything that can be infected 2011-11-12T20:24:46 omg, this cat is boxing with a snake 2011-11-12T20:25:18 Human LouisMartin = (Human) Toxoplasma_gondii; 2011-11-12T20:25:53 "unfortunately im mostly healthy" best thing I've read all day. 2011-11-12T20:25:56 that type coercion is not implemented 2011-11-12T20:25:56 :P 2011-11-12T20:27:47 damn. I suck at coding :( 2011-11-12T20:28:21 mcstar: why so depressed mcstar? 2011-11-12T20:28:34 am i? 2011-11-12T20:28:59 i just thought my bot will be better by now 2011-11-12T20:29:32 i wish i had sketch battle much earlier 2011-11-12T20:29:38 sketched* 2011-11-12T20:29:58 mcstar: There's new options out there. Heard of Agomelatine and Vybrid? All the antidepressant and emotions numbness without any of the erectile dysfunction! Now, that's good news for people with "blackdogs" 2011-11-12T20:31:00 mcstar: oh. feeling ok then? sorry about that then 2011-11-12T20:31:14 LouisMartin: are you a prescription-drug addict? 2011-11-12T20:31:45 you need to rethink your architecture to make place to a new battle system? 2011-11-12T20:32:47 nah, but I like RedBull and RockStar way, too, much. doesn't count? 2011-11-12T20:33:09 no, i just imlemented a game-tree search and it turned out not to be as productive as i hoped 2011-11-12T20:33:29 and i rewrite my path-handling too 2011-11-12T20:33:39 game-tree search? 2011-11-12T20:33:45 only to find that its too complicated, and there are unanticipated problems 2011-11-12T20:33:53 and a much simpler approach is better 2011-11-12T20:34:08 my bot beats all other bots it played with recently in food collection 2011-11-12T20:34:17 so i wont change that 2011-11-12T20:34:37 im just annoyed that i spent much time on something that didnt work out 2011-11-12T20:34:53 aichallenge: janzert epsilon * r1e35218 / ants/dist/starter_bots/python3/ants.py : Make the python3 starter use wall time not cpu time - http://git.io/YGyL2Q 2011-11-12T20:35:39 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-12T20:36:02 sure, but simpler approaches may not be the one winning the contest after a few succeed in making a smarter "Top-Bottom" AI. That's harder to get it right, but it will make the "little minimax dancers" look like fools when it does. 2011-11-12T20:36:53 mcstar: I hear you. I know that feeling and it sucks. 2011-11-12T20:37:13 LouisMartin: there are no minimax dancers 2011-11-12T20:37:16 there wont be any 2011-11-12T20:37:30 why? what do you mean? 2011-11-12T20:37:55 all those "dancing" bots do a 1-turn search 2011-11-12T20:38:01 that is hardly minimax 2011-11-12T20:38:11 the search space is so enermous, it wont work 2011-11-12T20:38:18 or only with very aggresive pruning 2011-11-12T20:38:26 and it would still be bad 2011-11-12T20:38:38 I disagree mcstar 2011-11-12T20:38:41 k 2011-11-12T20:38:49 you know the ironic aspect of all this? 2011-11-12T20:38:52 LouisMartin: ill let you live 2011-11-12T20:38:54 now 2011-11-12T20:39:08 we should really have lots of simple cpu's doing the processing of each ant 2011-11-12T20:39:11 the tile states are very limited. this isn't chess 2011-11-12T20:39:27 instead we have a single-core controlling all the ants :P 2011-11-12T20:39:37 LouisMartin: did you do some back-of-the-envelope calculation? 2011-11-12T20:39:43 or did you actually write it? 2011-11-12T20:39:53 one could include in his code a lot of situations and just do look-ups .. 2011-11-12T20:40:06 ? 2011-11-12T20:40:25 mcstar: "openings" 2011-11-12T20:40:25 lets talk about that when you have a working code 2011-11-12T20:40:36 you precalculate situations and save them in your code files 2011-11-12T20:40:49 pedrosorio: exactly 2011-11-12T20:40:58 until than, ill be watching this meter long centipede kill a snake 2011-11-12T20:41:15 what???????? xD 2011-11-12T20:41:19 link 2011-11-12T20:41:23 LouisMartin: i get what you want, but it simple is HUGE 2011-11-12T20:41:26 wont work 2011-11-12T20:42:19 you don't minimax the whole board all the time. you use it wisely and you search in the cache rather than calculate everything all the time. if you end up calculating a scenario you don't have in your lookup table, then at least cache it. 2011-11-12T20:42:25 pedrosorio: http://www.youtube.com/watch?v=ooFSFR2s7Ig&feature=related 2011-11-12T20:42:27 sorry 2011-11-12T20:42:33 its just 60cm 2011-11-12T20:42:48 *** Jak_o_Shadows has quit IRC (Read error: Connection reset by peer) 2011-11-12T20:43:01 mcstar: "but it simple is huge" .. what do you mean? 2011-11-12T20:43:16 simply 2011-11-12T20:44:06 mcstar: ah, ok, but why? it does work for chess software and pretty well. 2011-11-12T20:44:19 chess has a much smaller problem space 2011-11-12T20:44:19 *** Jak_o_Shadows has joined #aichallenge 2011-11-12T20:44:31 LouisMartin: really, just try it 2011-11-12T20:44:59 that's what pedrosorio talked about: opening book. 2011-11-12T20:44:59 i build up a 2vs1 player search tree in a reasonable time for 3 plies 2011-11-12T20:45:10 i cant* 2011-11-12T20:45:11 shit 2011-11-12T20:45:37 lets enumerate all the possibilities again 2011-11-12T20:45:54 how are you gonna handle symmetries? 2011-11-12T20:45:59 antimony: sure, but feasability depends on the (2d) space you process. if you try it on 100 x 100 then, yes, that's ridiculous. 2011-11-12T20:46:06 anyway, it's more like game theory than minimax due to simulataneous movement 2011-11-12T20:46:06 have you wrote any software like that before? 2011-11-12T20:46:16 chess is much more understood than Ants 2011-11-12T20:46:22 I've written simple Chess AIs before 2011-11-12T20:46:29 http://aichallenge.org/visualizer.php?game=90239&user=2772 2011-11-12T20:46:32 I still haven't managed to get any combat for my nats bot yet 2011-11-12T20:46:36 wow goldcadyy's bot looks awesome!! 2011-11-12T20:46:56 Antimony: i dont think thats a problem 2011-11-12T20:47:09 just imagine they are making its moves in succession 2011-11-12T20:47:32 but it might not be correct in a strict mathematical sense 2011-11-12T20:47:43 i dont know, am not a game theorist 2011-11-12T20:47:46 mcstar: you don't have to define any best move yourself. you write the code to find the best moves in a given context in a (very small) given spaces and then you include the data in your code file so as to use it as lookup tables... 2011-11-12T20:47:52 for now I'll be happy to be able to do a 1ply search assumign the enemy stays still 2011-11-12T20:48:06 Antimony: yo dont have to 2011-11-12T20:48:11 my code does that in no time 2011-11-12T20:48:20 i mean 1 ply is perfectly ok 2011-11-12T20:48:37 well I'll see how it is once I actually get something implemented 2011-11-12T20:48:39 well, of course you have to limit the number of participants 2011-11-12T20:48:45 that's my problem 2011-11-12T20:48:49 I don't limit participants 2011-11-12T20:48:56 you need to 2011-11-12T20:49:07 you cant do that on the whole map at once 2011-11-12T20:49:40 i tried a 10 player 1-ply search, took 10 seconds approximately 2011-11-12T20:49:49 *** Regis has quit IRC (Ping timeout: 240 seconds) 2011-11-12T20:49:52 fluxid, are you there? 2011-11-12T20:50:33 LouisMartin: a code can easily generate MUCH MUCH data 2011-11-12T20:50:42 i mean its like storing primes for factorization 2011-11-12T20:51:25 *** Jak_o_Shadows1 has joined #aichallenge 2011-11-12T20:51:25 *** Jak_o_Shadows has quit IRC (Read error: Connection reset by peer) 2011-11-12T20:51:53 how much space do you want to cover? 2011-11-12T20:52:16 goldcaddy is a sissy 2011-11-12T20:52:27 why doest it attack your ant 2011-11-12T20:52:34 I know but it looks awesome :P 2011-11-12T20:53:22 i probably need to upload my code with defense 2011-11-12T20:53:34 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-12T20:53:45 the code i have on the server doest know about its own hills 2011-11-12T20:54:04 in some strategies, you may want attack an enemy and let another one go 2011-11-12T20:54:16 LouisMartin: ? 2011-11-12T20:54:57 goldcaddy is awesome 2011-11-12T20:55:22 not sure if it's a bug or it's just too late for me to look at haskell code... but I had a look at the starter package and something looks strange 2011-11-12T20:56:00 Fluxid, hey. How are ya? We were speaking of battle AI and I tried to remember what you told me you had for that. didn't you told me about a 16x16 grid? was that for exploration (whole map) or for a minimax-like implementation of 16x16 pixels? 2011-11-12T20:56:21 http://www.youtube.com/watch?v=2qxBxn-YVvw&feature=related 2011-11-12T20:56:28 line 226-227 Ants.hs :ownerToTile Enemy2 = Enemy2Tile 2011-11-12T20:56:28 ownerToTile Enemy3 = Enemy2Tile 2011-11-12T20:56:31 this brown spider sucks soo much 2011-11-12T20:56:40 has flesh necrotizing venom 2011-11-12T20:57:19 LouisMartin: "minimax-like" tree with 16 leaves 2011-11-12T20:57:31 mcstar: looks like what we would call a brown recluse 2011-11-12T20:57:39 yes 2011-11-12T20:57:40 but i plan to rework it, just need to gather energy 2011-11-12T20:58:30 those and black widows are the only poisonous spiders around where I live \o/ 2011-11-12T20:58:41 you mean it eats your flesh 2011-11-12T20:58:46 but doesn't stop? 2011-11-12T20:59:08 fluxid: Okay. Is it actually minimax? 2011-11-12T20:59:17 roflmao: nah, they generally only bite humans when they feel threatened 2011-11-12T20:59:27 but 2011-11-12T20:59:29 once oyu get bit 2011-11-12T20:59:34 you have like an infection that just keeps on growing 2011-11-12T20:59:37 but the venom then kills a good chunk of flesh 2011-11-12T20:59:38 I've heard of once spider that once it bites you 2011-11-12T20:59:59 other spiders ( of the same species) will be attracted to you ( to bite you) 2011-11-12T21:00:16 LouisMartin: probably not, but it somehow works :D 2011-11-12T21:01:02 i'm bad with naming things, "it" does work but it's not the best 2011-11-12T21:01:03 Rinum: I've seen that used in a story but I've never heard of it being factual but no idea really 2011-11-12T21:01:03 Fluxid: haha! that's my kind of software engineering spirit :) 2011-11-12T21:01:15 roflmao: the venom spreads fast in the blood vessels, and in a couple of days that flesh simply dies 2011-11-12T21:01:19 rots away 2011-11-12T21:01:24 stops working 2011-11-12T21:01:26 awesome 2011-11-12T21:01:35 no, totally horrible 2011-11-12T21:01:39 and i have dejavu again 2011-11-12T21:01:44 havent we discussed this? 2011-11-12T21:01:47 you were attacked once? 2011-11-12T21:01:50 LouisMartin: i guess you code for a living 2011-11-12T21:01:56 no, all you ramble on about is your fascination for mcdonalds 2011-11-12T21:01:57 Fluxid: Alright. Well, it got you pretty far so it must do something good 2011-11-12T21:02:28 Fluxid- Yep. Doesn't most people here? 2011-11-12T21:02:52 LouisMartin: dunno, many are studying or even i hischool 2011-11-12T21:02:56 in* 2011-11-12T21:03:34 please stop talking about mcdonalds; I'm having glutamate monosodium withdrawal effects just reading about it 2011-11-12T21:03:42 http://www.youtube.com/watch?v=vT_NSNW8gTg 2011-11-12T21:03:44 i think Fluxid bactracks from final situations that are acceptable according to some criteria 2011-11-12T21:03:45 niiiiiiiiiiice 2011-11-12T21:04:09 Fluxid: I guess 2011-11-12T21:04:44 roflmao: indeed ^ ^ 2011-11-12T21:06:02 *** smiley1993 has quit IRC (Quit: Lost terminal) 2011-11-12T21:06:02 *** smiley1983 has quit IRC (Quit: Lost terminal) 2011-11-12T21:06:21 *** marijnfs_ has quit IRC (Quit: leaving) 2011-11-12T21:06:52 xRaider, xoposhiy, ifdef, coachbudka, Gadziferoth: If you didn't voted on this yet, please do -> http://aichallenge.org/forums/viewtopic.php?f=25&t=1872 2011-11-12T21:07:27 i dont want c# people to be in an advatage 2011-11-12T21:07:31 so i vote with no 2011-11-12T21:07:53 *** grwip has quit IRC (Quit: Leaving) 2011-11-12T21:07:54 *** Conorach has quit IRC (Ping timeout: 276 seconds) 2011-11-12T21:08:33 mcstar: don't worry, voters nickname will be checked and non C#'ers won't count ;) 2011-11-12T21:08:51 i dont see c# in your nick 2011-11-12T21:09:07 mine is almost csharp 2011-11-12T21:09:11 mcsharp 2011-11-12T21:09:14 LouisMartin: why won't you make a pull request with updated compiler settings? 2011-11-12T21:09:39 this was fun, nite ppl 2011-11-12T21:09:42 *** mcstar has quit IRC (Quit: WeeChat 0.3.6) 2011-11-12T21:09:56 ...and so the C* began ... 2011-11-12T21:10:27 Fluxid: Amstan asked me to find a few people in C# that would agree with the change 2011-11-12T21:11:47 *** sigh has quit IRC (Remote host closed the connection) 2011-11-12T21:12:27 *** orange8 has joined #aichallenge 2011-11-12T21:12:57 ah 2011-11-12T21:14:30 speed shouldn't be an issue in this contest, but if there is an opportunity, go for it 2011-11-12T21:14:33 Fluxid: I'm not allowed to ask you to vote on what you think is the right decision; that'd be wrong..... 2011-11-12T21:15:15 the real question is what are any downsides to enabling it? 2011-11-12T21:15:43 janzert: no downsides 2011-11-12T21:16:02 hmm, I would find that hard to believe ;) 2011-11-12T21:16:23 compilation time isn't different on my side 2011-11-12T21:16:37 how about less intelligible runtime error messages? 2011-11-12T21:16:41 there's downsides, but not in the contest settings 2011-11-12T21:16:44 more memory use? 2011-11-12T21:17:07 compile time? 2011-11-12T21:17:13 that'd be it (debugging), but we don't remote debug on your server unfortunately ;) 2011-11-12T21:17:27 compile time is the same for me 2011-11-12T21:17:39 not more memory use from what I see on my side 2011-11-12T21:18:12 I think I just hit my sweet spot 2011-11-12T21:19:56 is the server x86 or x64? 2011-11-12T21:20:06 LouisMartin: i'm not planning to vote, i don't even have an account on forums 2011-11-12T21:20:45 Fluxid: that was humor actually. may have been too subtle without voice tone.. 2011-11-12T21:23:33 avdg: you think been able to implement better (and usually more cpu voracious) algo can't make a difference in the contest? I think C/C++ can afford way more analysis than some other language.. why do you think it doesn't matter? 2011-11-12T21:23:38 *** antimatroid has joined #aichallenge 2011-11-12T21:23:51 LouisMartin: ah sorry 2011-11-12T21:24:06 http://ants.fluxid.pl/replay.32181 EPIC FAIL 2011-11-12T21:24:34 if someone sends a slower inefficient bot, should he win? but if something can be calculated faster, why not? 2011-11-12T21:25:20 its not really clarifying my thinking, just my opinion 2011-11-12T21:25:25 *** Anilm3 has joined #aichallenge 2011-11-12T21:26:18 roflmao: too scared! 2011-11-12T21:26:39 yeah no kidding 2011-11-12T21:26:42 roflmao: that match shows the value of timeout detection 2011-11-12T21:27:07 of when enemies time out? 2011-11-12T21:27:17 (and why would we punish someone because the server was executing the code slowly) 2011-11-12T21:27:20 yeah and stop moving. combat gets a lot easier :) 2011-11-12T21:27:28 very good point bugnuts 2011-11-12T21:27:38 because I have noticed that the same algo works well while enemies are moving 2011-11-12T21:27:43 but when they go stationary it freaks out 2011-11-12T21:28:05 *** Adam_ has joined #aichallenge 2011-11-12T21:28:12 Hello hello 2011-11-12T21:28:13 *** cutoff has quit IRC (Ping timeout: 252 seconds) 2011-11-12T21:28:20 hello hello hello 2011-11-12T21:28:21 *** Antimony has joined #aichallenge 2011-11-12T21:28:28 avdg: right, but I'm thinking between 2 equally competent and efficient programmers: the C++ one can afford much complex strategies/algo than some much slower language, and so is advantaged by the language. Which is fine, but better minimized when possible; as you agreed 2011-11-12T21:28:31 Quick question, how much time per turn do we have? 2011-11-12T21:29:02 @settings 2011-11-12T21:29:03 avdg: settings could be http://aichallenge.org/game_settings.php. 2011-11-12T21:29:07 500 ms atm 2011-11-12T21:29:12 my starter package says 1000ms 2011-11-12T21:29:17 500 is on th website. 2011-11-12T21:29:24 I guess it's 500 then :( 2011-11-12T21:29:27 damn. 2011-11-12T21:29:44 1s is your local settings ;-) 2011-11-12T21:30:12 is that changed through the command, or some py file? 2011-11-12T21:30:14 *** sigh has joined #aichallenge 2011-11-12T21:30:25 Adam_: edit test_bot 2011-11-12T21:30:26 check the files 2011-11-12T21:30:37 or edit play_game 2011-11-12T21:30:42 does the engine tell you when an enemy times out? 2011-11-12T21:30:43 Adam: in which language are you? plz tell me C# 2011-11-12T21:30:46 lol 2011-11-12T21:30:48 roflmao: no 2011-11-12T21:30:50 or do you have to deduce 2011-11-12T21:30:50 ok 2011-11-12T21:31:00 Java 2011-11-12T21:31:11 Adam: ok, nvm then 2011-11-12T21:31:14 lol 2011-11-12T21:31:24 why? 2011-11-12T21:31:50 What are the most compelling reason to detect timeout? 2011-11-12T21:32:09 Adam: because this http://aichallenge.org/forums/viewtopic.php?f=25&t=1872 2011-11-12T21:32:27 brb 2011-11-12T21:32:31 o 2011-11-12T21:33:55 *** djr_ has quit IRC (Ping timeout: 258 seconds) 2011-11-12T21:34:06 *** Adam_ has quit IRC (Quit: Page closed) 2011-11-12T21:40:56 Fluxid: btw I think your food rates don't match the current code 2011-11-12T21:41:16 I haven't watched a lot of games from the server lately but when I was they always seemed food-starved compared to the latest git and aichallenge.org 2011-11-12T21:44:17 *** treeform has joined #aichallenge 2011-11-12T21:45:18 is it possible to see past contest games or at least screenshots? 2011-11-12T21:45:29 Tron and Planetwars 2011-11-12T21:47:19 BenJackson: settings are the same as in http://aichallenge.org/game_settings.php 2011-11-12T21:47:26 just checked it 2011-11-12T21:47:43 *** amstan has joined #aichallenge 2011-11-12T21:47:43 *** ChanServ sets mode: +o amstan 2011-11-12T21:47:48 *** Anilm3 has quit IRC (Quit: Page closed) 2011-11-12T21:56:22 *** djr_ has joined #aichallenge 2011-11-12T21:56:32 *** TheLinker has quit IRC (Quit: Bye) 2011-11-12T21:56:34 *** Harpyon_ has joined #aichallenge 2011-11-12T21:57:27 *** anon has joined #aichallenge 2011-11-12T21:58:51 Does anyone know what Python libraries are installed on the tournament machine? For example is there a linear programming solver? 2011-11-12T22:00:31 anon: as in numpy? yes 2011-11-12T22:01:31 *** antimatroid1 has joined #aichallenge 2011-11-12T22:01:48 *** antimatroid has quit IRC (Read error: Connection reset by peer) 2011-11-12T22:02:20 anon: http://pastebin.com/sMYAf5UA 2011-11-12T22:03:31 *** Lerc has quit IRC (Quit: Lerc) 2011-11-12T22:05:17 amstan: thanks! 2011-11-12T22:09:00 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-12T22:09:03 *** roflmao has quit IRC (Quit: Leaving.) 2011-11-12T22:10:03 amstan: doesn't that just show what libraries are loaded by default, not what's available for import? 2011-11-12T22:10:16 that = "python -v" 2011-11-12T22:10:26 anon: you may be right, what should i do then? 2011-11-12T22:10:59 amstan: I dunno I'm a python newbie 2011-11-12T22:11:17 what do you mean by linear programming solver? 2011-11-12T22:11:33 *** Antimony has joined #aichallenge 2011-11-12T22:12:01 numpy, scipy 2011-11-12T22:12:15 amstan: is there any Tron & Planetwar website up? I'd like to see what it was like 2011-11-12T22:12:35 me too 2011-11-12T22:12:46 I only found *one* visualization of tron anywhere 2011-11-12T22:13:09 LouisMartin: tron's at csclub.uwaterloo.ca/contest 2011-11-12T22:13:21 LouisMartin: pw is temporarly inaccessible 2011-11-12T22:13:25 amstan: a linear programming solver solves linear optimization problems. See e.g. http://wiki.python.org/moin/NumericAndScientific/Libraries 2011-11-12T22:13:31 ok thanks 2011-11-12T22:13:36 LouisMartin: if you really want to get it point ai-contest.com to the ip of jpcameron.com in your hosts file 2011-11-12T22:13:45 anon: numpy and scipy are available 2011-11-12T22:14:05 anon: Worst-case, you can always bundle your library of choice with your bot. 2011-11-12T22:14:07 ok 2011-11-12T22:14:45 (Well, perhaps not if it requires additional binary extensions. But generally.) 2011-11-12T22:14:52 amstan: I've seen that site but it appears to be broken 2011-11-12T22:15:17 http://csclub.uwaterloo.ca/contest/visualizer.php?game_id=4160340 2011-11-12T22:15:26 I see controls but nothing else 2011-11-12T22:15:46 numpy and scipy are great to have, but unfortunately they don't seem to include an LP solver. 2011-11-12T22:16:29 BenJackson: i don't know myself actually 2011-11-12T22:16:38 Any LP solver advanced enough to do what I want will probably have portions written in C and hence not be includable with my bot 2011-11-12T22:17:01 jbroman: can you look into it? i don't have powers over that db@csclub, is the games table still there? 2011-11-12T22:17:03 Oh well. 2011-11-12T22:17:19 if you want a python graph library networkx is great 2011-11-12T22:17:38 amstan: Not sure what database that was in, or what the table was called. 2011-11-12T22:18:03 jbroman: probably contest, games table, there's a replay column 2011-11-12T22:19:00 anon: but Octave does have an LP solver :P 2011-11-12T22:19:17 amstan: the tron game viewer doesn't work for me. Tried it in Chrome and IE. The youtube video doesn't work neither. not my luck. 2011-11-12T22:19:47 amstan: Table contest.games doesn't exist. 2011-11-12T22:19:55 Table in contest: news, submissions, users 2011-11-12T22:19:56 jbroman: i was affraid of that, too bad 2011-11-12T22:20:06 someone probably got pissed because of its size 2011-11-12T22:20:10 and deleted it 2011-11-12T22:20:18 :( 2011-11-12T22:20:44 when I saw the broken tron site (and couldn't even find pw) I was afraid ants would go the same way 2011-11-12T22:21:14 pw's alive, but the there's no domain pointed at it 2011-11-12T22:21:15 *** Jak_o_Shadows has joined #aichallenge 2011-11-12T22:21:34 why not just pw.aichallenge.org? 2011-11-12T22:22:30 amstan: contest2.contest_games does exist, though. 2011-11-12T22:22:41 With 1,258,435 rows. 2011-11-12T22:22:46 pedrosorio: I don't want an LP solver badly enough to deal with the ickiness that is programming in MATLAB/Octave 2011-11-12T22:22:51 that's a lot of rows 2011-11-12T22:22:53 jbroman: well... whatever that means, lol 2011-11-12T22:22:58 jbroman: could be it 2011-11-12T22:23:02 Don't remember what you guys used contest vs contest2 for. 2011-11-12T22:23:07 yeah, no idea 2011-11-12T22:23:19 seems like too many rows to be just a test, so let's be optimistic 2011-11-12T22:23:24 anon: oooh =P 2011-11-12T22:23:50 BenJackson: I might try networkx library. Do you know if it is installed on the tournament machine? 2011-11-12T22:23:53 amstan: I see move_sequence populated, etc. 2011-11-12T22:23:56 So it looks complete. 2011-11-12T22:24:00 anon: probably not, I'm just recommending it in general 2011-11-12T22:24:00 jbroman: when you have a free day, could you help me migrate that db/data to the current server? 2011-11-12T22:24:10 whenever I want to try some grahp thing without actually implementing it myself I use networkx 2011-11-12T22:24:32 Mmm…maybe a bit. Don't know much about how you set it up the first time. 2011-11-12T22:24:36 *** Jak_o_Shadows1 has quit IRC (Ping timeout: 276 seconds) 2011-11-12T22:24:40 Do you not have access to that database on caffeine? 2011-11-12T22:24:45 select * from context_games | mail -s "database" amstan@... 2011-11-12T22:24:46 i could figure it out, nope 2011-11-12T22:25:01 Assuming j3camero doesn't mind, I would just give you access to it. 2011-11-12T22:25:04 (Don't see why he would.) 2011-11-12T22:25:25 BenJackson: Well presumably a mysqldump would be more useful than a printout like that. 2011-11-12T22:25:50 that was a joke... 2011-11-12T22:25:56 amstan: if we could get static crawls of each site we could stuff them into s3 and serve 'em from there and not have to worry about a server/database migration again 2011-11-12T22:26:11 janzert: Hasn't this come up at least a few times now? 2011-11-12T22:26:25 I seem to recall amstan claiming it was very difficult to do for some reason I don't remember. 2011-11-12T22:26:25 it had 2011-11-12T22:26:28 may have 2011-11-12T22:26:39 to be honest i just want to data for now, and not loose it 2011-11-12T22:26:46 then we can worry about optimizing and compression later 2011-11-12T22:27:43 *** Harpyon has quit IRC (Quit: Computer has gone to sleep.) 2011-11-12T22:27:44 *** Harpyon_ is now known as Harpyon 2011-11-12T22:31:08 amstan: /scratch/contest2.sql 2011-11-12T22:31:14 Just so you can have the data. 2011-11-12T22:31:31 You can remove it when you've got it. 2011-11-12T22:31:39 Actually, hang on. 2011-11-12T22:31:59 compressing it as it's 678M 2011-11-12T22:33:57 There, now grab /scratch/contest2.sql.bz2 (81M) 2011-11-12T22:33:57 only 81M? 2011-11-12T22:34:07 SQL compresses really well. 2011-11-12T22:34:17 Because repeated strings like "INSERT" get compressed to a few bits. 2011-11-12T22:34:46 yeah, decent 2011-11-12T22:35:01 So scp that somewhere you like and you have the database, at least. 2011-11-12T22:35:07 yeah 2011-11-12T22:35:40 well, i have it 2011-11-12T22:35:41 thanks 2011-11-12T22:36:41 *** Harpyon has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-11-12T22:38:11 jbroman: i forgot about the wonders of /scratch 2011-11-12T22:38:16 this crow: smarter than all of our ant bots: http://www.youtube.com/watch?v=ofjo26O0z_o 2011-11-12T22:40:06 Anyone know if they've finished porting scipy to python3? Supposedly they were working on it about a year ago, but I haven't found confirmation that it's done. 2011-11-12T22:41:07 anon: i was like you about 3 years ago, looking to learn python, starting with python3. but then I realized that nobody actually uses python3 yet, and it's still the same today 2011-11-12T22:41:30 anon: just do python2.7, it's more supported 2011-11-12T22:42:03 amstan: OK, I'll switch to python 2 2011-11-12T22:42:43 anon: if you really want to get a python library that we don't have, post on the forums(language support), we should get it inside a week if there's demand 2011-11-12T22:43:15 well.. might as well bring this stuff up 2011-11-12T22:43:33 yay 2011-11-12T22:43:35 jbroman: do you happen to have the src for tron too? 2011-11-12T22:44:08 or.. hmm 2011-11-12T22:44:16 i might have access to it, i should try it first before bothering 2011-11-12T22:44:39 Going to bed now. Bug me later if you really can't reach it. 2011-11-12T22:44:50 yeah, i have access to www, but not to parent, should be enough through 2011-11-12T22:44:53 though* 2011-11-12T22:44:56 jbroman: good night 2011-11-12T22:44:59 thanks again 2011-11-12T22:45:34 amstan: thanks a lot for your help! 2011-11-12T22:46:22 <== Leaving IRC client open, but may not pay attention anymore 2011-11-12T22:50:15 gn 2011-11-12T22:52:00 janzert: /var/www/archive/tron good? 2011-11-12T22:52:28 I suppose 2011-11-12T22:52:54 be sure to create a new database/database user for it though 2011-11-12T22:53:04 janzert: yep 2011-11-12T22:53:24 how large is the database? 2011-11-12T22:53:40 there is only 20GB total on that machine 2011-11-12T22:54:02 81mb bzip compressed, 678mb the sql file 2011-11-12T22:54:14 ok 2011-11-12T22:54:21 janzert: i was also considering to get pw here too 2011-11-12T22:54:30 didn't we have more? 2011-11-12T22:54:57 janzert: apparently /mnt is a 500GB partition 2011-11-12T22:54:58 that's all I ever set the ec2 servers up for 2011-11-12T22:55:09 but will go away if the machine shutsdown 2011-11-12T22:55:16 not good 2011-11-12T22:55:35 it was just meant to hold the current contest 2011-11-12T22:55:45 well, i didn't know this, when i was in the mailing list i had the impression that we had a ton of space on the current server 2011-11-12T22:56:16 *** pedrosorio has quit IRC (Quit: Page closed) 2011-11-12T22:57:22 well.. i'll do this as a demo to jeff and hopefully he'll go onboard 2011-11-12T22:57:28 janzert: please support me in my arguments 2011-11-12T22:57:43 yeah, ec2 is probably not the right way to do the archival server 2011-11-12T22:58:03 well, maybe a micro instance would work hmm 2011-11-12T22:58:33 for archival we want lots of disk space and should need pretty low cpu and bandwidth 2011-11-12T22:59:13 yes 2011-11-12T23:04:13 *** anon has quit IRC (Quit: Page closed) 2011-11-12T23:07:43 *** Rinum has quit IRC (Ping timeout: 265 seconds) 2011-11-12T23:13:12 *** foRei has quit IRC (Quit: Bye) 2011-11-12T23:13:19 *** NoxiaZ^ has quit IRC (Read error: Connection reset by peer) 2011-11-12T23:20:06 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-12T23:22:14 *** cutoff has joined #aichallenge 2011-11-12T23:22:55 *** ChrisH_ has joined #aichallenge 2011-11-12T23:25:58 wow I really don't grok the orientation code in ants.py at all 2011-11-12T23:26:03 ok, so far i have it almost done: http://tron.aichallenge.org/ 2011-11-12T23:26:10 i just need the tron maps folder 2011-11-12T23:28:56 *** orange8 has quit IRC (Quit: Page closed) 2011-11-12T23:28:59 amstan: I seem to recall the other tron page you linked above had a link to the "final map pack" 2011-11-12T23:29:19 BenJackson: where?? omg 2011-11-12T23:29:28 BenJackson: yeah.. the one from the other server's gone, but if there's a pack 2011-11-12T23:29:56 *** dvladim has joined #aichallenge 2011-11-12T23:30:51 <_flag> amstan: Is there anyway to search for a player on that site? 2011-11-12T23:30:57 _flag: nope 2011-11-12T23:31:07 <_flag> I remember I was disqualified from the final tournament because my bot timedout one game, so I can't find my entry 2011-11-12T23:31:15 _flag: i could look up your name manually 2011-11-12T23:31:18 player search wasn't added till planetwars 2011-11-12T23:31:23 jhttp://aichallenge.org/forums/viewtopic.php?f=3&t=1176 2011-11-12T23:31:31 <_flag> It was probably FlagCapper then too 2011-11-12T23:31:33 http://aichallenge.org/forums/viewtopic.php?f=3&t=1176 2011-11-12T23:31:49 it's a forum post with a link toa z ip of tournament maps 2011-11-12T23:32:02 BenJackson: that was planetwars 2011-11-12T23:32:24 jbroman: there seems to be a map folder in his svn checkout, i'm gonna try using those 2011-11-12T23:32:47 <_flag> Ah, I found it with my googling skills 2011-11-12T23:33:50 woot! http://tron.aichallenge.org/visualizer.php?game_id=3354720 2011-11-12T23:33:54 oh yeah! 2011-11-12T23:34:19 nice! 2011-11-12T23:34:58 amstan: Once you have it working, it would be nice if you could put in a redirect in csc/~contest 2011-11-12T23:35:17 Just so that there's one authoritative archive of it. 2011-11-12T23:35:25 jbroman: yep, 301? 2011-11-12T23:35:28 (well, one authoritative *accessible* one anyhow) 2011-11-12T23:35:33 amstan: Whatever. 301 is fine. 2011-11-12T23:35:59 janzert: the is a small bug in the hover text for game times on profile pages. The